Why this exists — in plain English, and for the kids
One honest list of every prediction this lab has ever frozen, in a single place: what was claimed, the day it was locked, what it was measured against, whether its deadline has passed, and how it scored. It even keeps a “failure museum” of dead ideas, each with the reason it died.
The whole point is to make the record impossible to fool itself with. If it quietly dropped its losers and only remembered its winners, it would be a lie. So everything stays here, forever. Over time this list is the single clearest answer to “is this thing actually getting better?”
This is the lab's big honesty notebook. Every single guess it ever wrote down lives here: what it guessed, when, and whether it turned out right or wrong.
And here's the important part: it keeps the wrong guesses TOO, right next to a note saying why they were wrong. Nobody gets to secretly hide their mistakes. That's how you can tell if the lab is really learning.
The track record, by what it actually proves
Verify any freeze against Bitcoin → /verifyMatured & graded vs an external benchmark: 11 — the real score. Frozen, awaiting maturity: 128 ledger rows — the countdown above counts atomic claim contracts, a finer unit (one row can hold many; the definitions manifest reconciles every headline number). Backtests (context, not proof): 5 rows. Self-assessed research (not a track record): 592 frozen claims across 8 crews. Retired with a stated cause of death: 346 rows. The graded number is young BY DESIGN — most windows have not opened. We publish it, unpadded, rather than dress backtests up as proof.
Predictions by lane — 1345 frozen · 778 graded · 337 open · 178 void · 3 lanes beating the null · as of 2026-09-09
778 graded predictions means 778 individual frozen calls in the per-lane call ledgers, across 8 lanes, frozen since 2026-07-29. That is not the same population as the 11 matured & graded ledger rows above (frozen 2026-06-28 to 2026-08-05): those are cross-lane research and strategy records where one row can summarise many claims, these are individual calls in the daily call ledgers. Neither bounds the other and they must never be summed — see About for the definitions manifest that reconciles every headline number on this site.
| Lane | Frozen since | Frozen | Open | Graded | Void | Brier | Null | Read |
|---|---|---|---|---|---|---|---|---|
| skywatch-rf v1 (retired) | 2026-08-28 | 36 | 0 | 0 | 36 | — | — | stuck — every matured call was VOID — unobservable, not a miss. A lane that cannot grade teaches nothing; the question needs re-asking, not more freezes. |
| skywatch-rf v2 | 2026-08-31 | 45 | 0 | 45 | 0 | 0.066 | 0.066 | at or behind the null |
| skywatch-rf (unattributed) | 2026-09-01 | 270 | 18 | 252 | 0 | 0.231 | 0.231 | at or behind the null |
| sdr-calls | 2026-08-27 | 52 | 11 | 33 | 0 | 0.252 | 0.258 | beats the null |
| earthwx | 2026-08-29 | 209 | 46 | 27 | 136 | 0.235 | 0.235 | at or behind the null |
| skywatch | 2026-07-29 | 397 | 22 | 373 | 0 | 0.176 | 0.208 | beats the null |
| market-weather | 2026-08-22 | 252 | 210 | 0 | 0 | — | — | no grades yet |
| lane-drumbeat | 2026-08-26 | 84 | 30 | 48 | 6 | 0.228 | 0.239 | beats the null |
The OKLM ledger — ladder null → raw → tokens → seq → cross-lane · as of 2026-09-08
| Lane · family | Rows | Forward | Null | Raw | Tokens | Rung |
|---|---|---|---|---|---|---|
| audibility | 187,818 | 121,101 | 0.146 | 0.123 | 0.124 | raw the vocabulary does NOT beat the ordinary reading |
| cytometry | 78,410 | 19,736 | 0.091 | 0.082 | 0.066 | raw accumulating — read on 2026-09-18, needs >= 5000 forward rows (have 19,736) |
| dialogue | 98 | 1 | 0.250 | 0.022 | 0.250 | null accumulating — 99 more graded claims before the read |
| rf-ablation | 6 | 6 | 0.218 | 0.250 | 0.250 | null accumulating — 294 more graded forward pairs before the read; nothing is claimed until 2026-12-01 |
| tts-pronunciation · pronunciation | 1,336 | 462 | 0.007 | 0.006 | 0.006 | raw accumulating — the read is pre-registered for 2026-11-06 and today is 2026-09-07; 462/300 forward rows so far |
| node-stream · fp-kind | 48 | 23 | 0.251 | 0.257 | 0.217 | null accumulating — 1977 more graded rows in the sealed forward window before the read |
| node-stream · psd-peaks | 4,803 | 4,533 | 0.250 | 0.127 | 0.097 | tokens the vocabulary earns its name here |
| node-stream · wspr-verify | 30,857 | 23,221 | 0.057 | 0.033 | 0.037 | retired RETIRED — degenerate target — the label is the site's own 3 % distance-tolerance rule between grid-centre and reported km (1.000 match for every path >= 200 km); a model of it learns a rounding rule, not radio. Retired 2026-09-01; the family's history is kept as an integrity check of the browser compute, not as a vocabulary result. |
Lab Ledger — the track record, separated by what it actually proves
Every prediction, sorted into four honest buckets so nothing is read as more than it is: MATURED & GRADED vs an external benchmark (the real score) · FROZEN, awaiting maturity (credible but immature) · BACKTESTS (context, not forward proof) · SELF-ASSESSED research (the crews' own reads — never counted as a track record) · plus RETIRED ideas kept with their cause of death. The flagship is Materials Risk; everything was frozen before its outcome and graded after. Verify any freeze independently at /verify.
Snapshot 2026-09-09 · 507 predictions tracked across 6 lobes
Matured & graded vs an external benchmark: 11 — the real score. Frozen, awaiting maturity: 128 ledger rows — the countdown above counts atomic claim contracts, a finer unit (one row can hold many; the definitions manifest reconciles every headline number). Backtests (context, not proof): 5 rows. Self-assessed research (not a track record): 592 frozen claims across 8 crews. Retired with a stated cause of death: 346 rows. The graded number is young BY DESIGN — most windows have not opened. We publish it, unpadded, rather than dress backtests up as proof.
No ledger changes since last cycle.
wealth· 12 live
The Graham long-horizon tier beats SPY across multiple 5-year windows.
The backfill made the dot-com bust (fired 0/32) and the GFC (fired 10/18) reachable with as-filed fundamentals -- the regimes the XBRL-era study could never see. SINGLE_WINDOW_ONLY: beat materially in exactly one window. Under the survival bar this is NOT a validated edge -- it is the multiple-testing warning made concrete. Returns are now SURVIVORSHIP-FREE for 12 delisted names: their real prices (Sharadar, death-truncated) are in the book (none cleared the gate, so the bias-free book matches here); costs charged (5bps/side on turnover); pre-XBRL parser error rate disclosed in the artifact.
The Safe basket matches SPY with less drawdown (risk-adjusted).
accumulating — not yet matured
The DCA basket beats a SPY-DCA floor over the holding window.
accumulating — not yet matured
The Moonshot basket's upside tail beats SPY.
accumulating — not yet matured
The $1B far-tail basket beats SPY (a lottery, sized as one).
accumulating — not yet matured
Four breadth-filtered quality edges freeze one book
accumulating — one snapshot is noise; the forward record decides, not the metaphor
The capacity sim retires thin-name edges as net-of-cost illusions
Inaugural run: swapping flat cost for Kyle/Amihud impact collapsed net illiquidity premium toward zero at meaningful AUM; three thin-name edges retired (net-surviving illiquidity premium, liquidity-gr
The distinctness test exposes de-correlated breadth as the same megacap bet
Inaugural: bare de-correlation failed distinctness at J=0.62 overlap; distinctness recovered only after re-ranking on gross-profit/assets — single construction, one snapshot.
Every hundredfold kill is an unowned choke point, not the firm's financials
Inaugural kill cycle: 16/16 kills at an unowned choke point, 0 on the firm's own financials — single cycle, re-judged each run.
NVDA first clears the gates in 2005; a standing re-screen catches flickering winners
WEAK: pooled capture ratio 1.56 in (1,2) — traits enrich less than claimed (9/59 passers vs 49/500 controls hit 10x)
The hundredfold survivability gates select for higher mortality, not survival
KILL CONDITION MET: passer mortality 1.78x >= control — the gates don't select survivors (4/59 passers vs 19/500 controls dead — small counts, re-judged each re-run)
Surfacing a circular held-name credit without removing it still reorders ranks
SAFE top-10: panel-vetted CPAY core 77.8 (highest core) ranks #2 behind NVR total 80.7/core 71.1; bookLift +9.4-9.7 still inside the scored value.
engine· 22 live
early tail head to head
SUPPORTED (interim): early pool 10x rate 1.23x the screen pool's (3/16 early vs 9/59 screen hit 10x)
The macro correction-risk read is calibrated (does elevated risk precede drawdown).
too few matured snapshots to judge calibration yet
Risk-off regimes precede more SPY drawdown than risk-on.
accumulating — need matured freezes in BOTH risk-on and risk-off to compare
Science-foresight picks beat SPY before the science meets the market.
accumulating — not yet matured
The CAPR Kalman-trend oversold-quality basket beats SPY.
accumulating — not yet matured
The AI Director's synthesized list beats SPY AND the engine's other lists.
accumulating — not yet matured
The mechanical cross-crew consensus top-20 beats SPY.
accumulating — not yet matured
The Hundredfold right-tail bench beats SPY over multi-year windows.
accumulating — not yet matured
Asset Growth / Capital Discipline (CMA) beats SPY over its window.
accumulating — not yet matured
Accruals Anomaly (Sloan) beats SPY over its window.
accumulating — not yet matured
Gross Profitability (Novy-Marx) beats SPY over its window.
accumulating — not yet matured
Piotroski F-Score (adapted) beats SPY over its window.
accumulating — not yet matured
Return Seasonality (same-month) beats SPY over its window.
accumulating — not yet matured
52-Week High Momentum (anchoring anomaly) beats SPY over its window.
accumulating — not yet matured
Time-Series Momentum (TSMOM) beats SPY over its window.
accumulating — not yet matured
Residual (Fundamental-Factor-Adjusted) Momentum beats SPY over its window.
accumulating — not yet matured
Cross-domain fusion hypotheses beat SPY.
accumulating — one snapshot is noise; the forward record decides, not the metaphor
Higher measured froth precedes weaker forward return vs SPY.
accumulating — froth is frozen each snapshot; grading opens at 90d
The crew-adjusted froth ranking predicts better than the mechanical one.
accumulating — the crew overlay freezes alongside the mechanical score
early mortality honesty
early mortality 0/16 vs screen-passer 4/59 — the stated cost of the cheap end, reported not judged
right tail enrichment
WEAK: pooled capture ratio 1.56 in (1,2) — traits enrich less than claimed (9/59 passers vs 49/500 controls hit 10x)
survivability gate works
KILL CONDITION MET: passer mortality 1.78x >= control — the gates don't select survivors (4/59 passers vs 19/500 controls dead — small counts, re-judged each re-run)
materials· 105 live
Magnet-material controls widen from separation-tech to finished NdFeB magnets.
accumulating — awaiting horizon
the lesson
No formal control step on finished magnets within the horizon; Nd/Pr oxide flat/down.
Permit throughput is dialed down enough to bite non-localized Western anode lines.
accumulating — awaiting horizon
the lesson
Western anode lead times do not extend >30%; permit throughput unchanged.
Antimony becomes the visible edge of bloc decoupling via defense-munitions supply.
accumulating — awaiting horizon
the lesson
Ex-China antimony falls back below 2x its pre-control level.
DRC re-uses export suspension as a price-defense lever, whipsawing battery makers.
accumulating — awaiting horizon
the lesson
No DRC suspension/quota action; cobalt stays at/near its cycle low.
Rare-earth trade regionalizes further after the US-specific ban.
accumulating — awaiting horizon
the lesson
Rare-earth intra-bloc trade share fails to rise >5 pts YoY.
REE-specific tightening shows up as REE miners outperforming the broad materials complex (proxy).
accumulating — awaiting horizon
the lesson
REE proxy fails to outperform the broad complex over the window (no REE-specific premium).
Lithium tightening shows up as lithium miners outperforming the broad complex (proxy).
accumulating — awaiting horizon
the lesson
Lithium proxy fails to outperform the broad complex (no Li-specific squeeze).
A hard REE squeeze lifts the REE miner proxy in absolute terms (proxy).
accumulating — awaiting horizon
the lesson
REE proxy fails to rise >20% in absolute terms (no hard squeeze).
As of 2026-07-11, PT Weda Bay Nickel (Eramet JV, Indonesia; halted since early June 2026 after exhausting its cut 12 Mwmt 2026 RKAB quota vs 42 Mwmt sold in 2025) stays effectively shut: Indonesia's ESDM does NOT grant WBN a supplementary 2026 quota of 10 Mwmt or more by 2026-08-31, Eramet's H1 2026 results cut rather than reaffirm prior Weda Bay full-year volume guidance, and ERA.PA underperforms the STOXX Europe Basic Resources index from 2026-07-11 to 2026-10-09. Call FAILS if ESDM grants >=10 Mwmt by 2026-08-31, or Eramet reaffirms prior Weda Bay volumes at H1, or ERA.PA outperforms the index over the window.
accumulating — awaiting horizon
As of 2026-07-11: the US Treasury/IRS does NOT extend the 45X FEOC 'impracticable-to-trace' graphite exemption beyond its 2026-12-31 expiry; Panasonic does NOT disclose a qualified non-FEOC graphite source covering >=50% of its North American anode needs before Q4 2026; Panasonic's energy segment flags anode/graphite materials cost or FEOC-compliance drag in its October 2026 or January 2027 guidance before LG Energy Solution or Samsung SDI flag equivalent costs; and PCRFY underperforms both 373220.KS (LGES) and 006400.KS (Samsung SDI) over the 180 days from 2026-07-11. Call FAILS if the exemption is extended by 2026-12-31, or >=50% non-FEOC coverage is disclosed before Q4 2026, or PCRFY outperforms both Korean peers.
accumulating — awaiting horizon
CMOC Group's (3993.HK) H1 2026 report, due by 2026-08-31, shows a cobalt-attributable inventory BUILD (not drawdown) with a working-capital or NRV/impairment disclosure tied to the DRC quota regime (117.5kt 2025 production vs 31.2kt 2026 export quota; ARECOMS 2026-06-29 forfeiture of unused H1 allocations to the state strategic quota); CMOC receives no supplementary quota and cobalt realized/exported volumes are more than 20% below production; 3993.HK underperforms Glencore over the 90 days from 2026-07-11. Call FAILS if the H1 report shows cobalt inventory drawdown, or a supplementary quota grant to CMOC, or realized volumes within 20% of production.
accumulating — awaiting horizon
Following MOFCOM's 2026-06-29 entity listing of Mitsubishi Electric units (banning any party from transferring China-origin dual-use rare-earth items to them, reaching Dy/Tb in sub-supplier magnets), Mitsubishi Electric's Defense & Space segment discloses materials-driven schedule slippage or cost provisions in its October 2026 or January 2027 results commentary, and 6503.T underperforms TOPIX over the 200 days from 2026-07-11 despite growing defense orders. Call FAILS if Defense & Space reports on-schedule deliveries with no materials-related caveat through 2027-01-31, or 6503.T performs in line with TOPIX while defense orders keep growing.
accumulating — awaiting horizon
United States Antimony (UAMY; sole-source $245M DLA antimony-ingot IDIQ, $57.3M ordered and ~$2.6M June 2026 invoicing as of 2026-07-11; expanded Thompson Falls MT furnaces commissioned Q2 2026): across H2 2026 incremental DLA orders total at least $20M (cumulative crossing ~$75M), Q3 2026 filings show Thompson Falls throughput UP vs Q2, and Bolivian/third-party antimony-flake feedstock deliveries normalize (no repeated delay disclosures in 8-Ks/10-Q). Call FAILS if incremental H2 2026 DLA orders are below $20M, or Q3 filings show throughput flat vs Q2, or UAMY underperforms XME over the 180 days from 2026-07-11.
accumulating — awaiting horizon
By 2027-04-30, Neo Performance Materials (NEO.TO; Silmet, Estonia — the EU's only operating rare-earth separation plant, first EU Dy/Tb line commissioned April 2026) appears on the first RESourceEU pilot-stockpile rare-earth or gallium tender award, or announces an EU-government-linked purchase/offtake for Silmet output. The sovereign-channel attachment is the claim; year-one volumes are expected immaterial. Call FAILS if the first RESourceEU REE/gallium tender is issued and awarded without Neo/Silmet participation, or no EU-sovereign purchase/offtake involving Neo exists by 2027-04-30, or NEO.TO underperforms REMX over the 270 days from 2026-07-11.
accumulating — awaiting horizon
By 2027-01-31, at least one NEW binding defense-linked Dy/Tb offtake or government-anchored contract naming Lynas Rare Earths (LYC.AX; only separated heavy-RE producer outside China, Kuantan Malaysia) is announced AFTER 2026-07-11 by a Japanese magnet maker (Shin-Etsu, TDK, or Proterial), JOGMEC/Japan MoD, or an EU sovereign channel — explicitly EXCLUDING the pre-existing March 2026 12-year JARE agreement and any restatement of it — and Lynas quarterly Dy+Tb oxide output stays above prior-quarter levels. Call FAILS if no such new contract by 2027-01-31, or Dy/Tb output stalls below prior quarter, or LYC.AX underperforms the MVIS Rare Earth index over the 180 days from 2026-07-11.
accumulating — awaiting horizon
By 2027-03-31, Panasonic Energy formally confirms validation of Novonix's (NVX) synthetic-graphite C-sample (delivered 2026-06-11, the first North American AAM C-sample), and neither contracted offtake (10kt over 4 years; minimum 32kt over 5 years) is cancelled or renegotiated down. Call FAILS if validation fails or remains unconfirmed by 2027-03-31, or either offtake is cut, or NVX underperforms an equal-weight graphite/anode peer basket (SYR.AX + POSCO Future M) over the 240 days from 2026-07-11.
accumulating — awaiting horizon
The DLA battery-grade lithium carbonate award (solicitation of 2026-07-02, up to ~16,000t/$300M ceiling, bids due 2026-07-17) is announced by 2026-12-31 naming Albemarle (ALB; Silver Peak NV, the only scaled operating US battery-grade Li2CO3 source) as supplier, and ALB then outperforms an equal-weight SQM + Pilbara Minerals basket over the following 30 trading days. Call FAILS if the award goes to a non-Albemarle supplier (an equally informative print implying allied-origin material was acceptable), or no award is posted by 2026-12-31, or an ALB award is not followed by the 30-trading-day relative outperformance.
accumulating — awaiting horizon
Arafura Rare Earths (ARU.AX; Nolans FID 2026-05-21 backed by the first — non-binding — Export Finance Australia Strategic Reserve Letter of Support for up to 500tpa NdPr): (1) financial close/first debt drawdown completed by 2026-12-31 (ASX announcement); (2) the EFA LOS converts into binding offtake/contract terms; (3) at least one additional sovereign-adjacent offtake (Korean or EU buyer) announced by 2027-03-31. Call FAILS if financial close is not completed by 2026-12-31, or the LOS lapses without conversion to binding terms, or ARU.AX underperforms LYC.AX by more than 20% over the 270 days from 2026-07-11.
accumulating — awaiting horizon
By 2027-01-09, at least one of LG Energy Solution, SK On, or Samsung SDI explicitly QUANTIFIES a US FEOC-graphite compliance cost in guidance or earnings materials (a won/kWh, margin-basis-point, or absolute-cost figure — not a qualitative mention), and/or discloses above-market take-or-pay terms (premium vs a Chinese anode reference price) in a new non-China graphite/anode contract SIGNED AFTER 2026-07-11 — while Korea's HS 2504+3801 import volumes from China stay within -10% YoY (a cost event, not a shortage event). Call FAILS if no such quantification or new-contract disclosure appears by 2027-01-09, or if equal-weight 006400.KS + 373220.KS OUTPERFORM EWY by >10% over the window with no such news.
accumulating — awaiting horizon
By 2026-11-08, a named German Tier-1 or OEM (e.g., ZF, Schaeffler, Bosch, or an auto/wind OEM) makes a NEW disclosure after 2026-07-11 of a magnet-related production impact or allocation warning that the company itself attributes to Chinese export-license timing, allocation, or transshipment enforcement (MOFCOM Announcement No. 26 of 2026-06-24, effective 2026-07-01) — a price/cost complaint alone does not count — while spot NdPr stays roughly flat. Call FAILS if by 2026-11-08 no such attributed disclosure exists AND Germany's HS 8505.11 monthly import volume from China (Eurostat) is down less than 5% YoY.
accumulating — awaiting horizon
By 2027-01-09, MP Materials or USA Rare Earth disclose in Q2-Q4 2026 filings or earnings calls an equipment- or extractant-chemistry-sourcing constraint linked to MOFCOM's 2026-06-22 entity listing (which bans any party from transferring China-origin dual-use goods to them), or a slip of at least one quarter in a magnet/heavy-rare-earth commissioning milestone; and MP lags Lynas (LYC.AX, ADR-adjusted) by >=10% over the 180 days from 2026-07-11. Call FAILS if MP's disclosed commissioning timelines are unchanged across its Q2 and Q3 2026 filings AND MP does not lag LYC.AX by >=10% over the window.
accumulating — awaiting horizon
By 2026-12-31, Export Finance Australia's A$1.2bn Critical Minerals Strategic Reserve announces a SECOND project-level commitment (letter of support, offtake, or floor contract) — a NEW instrument announced after 2026-07-11; the first was Arafura Nolans on 2026-05-21, and pre-existing Critical Minerals Facility loans (e.g., Iluka Eneabba's) or restatements of them do NOT count — with contractual floor pricing visible in the ASX/government announcement. Call FAILS if no such second commitment by 2026-12-31, or if equal-weight LYC.AX + ARU.AX + ILU.AX vs ASX 200 Materials is below +0% over the 170 days from 2026-07-11.
accumulating — awaiting horizon
(1) Malaysia's HS 2846 export value rises >=25% YoY in the 2026 UN Comtrade vintage with volume growth LESS THAN HALF the value growth (premium capture, not tonnage); and (2) a NEW downstream heavy-rare-earth commitment into Malaysia is announced after 2026-07-11 that is NEITHER Lynas's already-announced October 2025 5,000tpa HRE facility NOR the existing JARE offtake (e.g., a Japanese/European magnet-alloy or metallization JV, or a third-party HRE separation project). Interim check: Lynas's October 2026 quarterly shows growing Dy+Tb oxide output at a disclosed Western premium — flat/declining HRE output or China-benchmark-parity pricing fails the call early. Call FAILS on HS 2846 value growth under 25% YoY or the volume condition breaking.
accumulating — awaiting horizon
In H2-2026 trade vintages (graded via UN Comtrade delta): Indonesia's HS 7202 (ferroalloy) export unit value relative to the LME nickel monthly average rises >=10% vs the H1-2026 baseline while export volumes flatten — the single-desk effect of state-owned DSI being sole exporter of >=8% Ni ferroalloys since 2026-06-01 under PP 24/2026, combined with the 260 Mt wmt 2026 RKAB cap and own-smelter-only supplementary quotas. Call FAILS if the HS 7202 unit-value-to-LME ratio is not up >=10% vs H1-2026, or if the DSI export monopoly is rolled back or waived for major producers before 2026-12-31 (dated primary-source event).
accumulating — awaiting horizon
(1) DRC's 10% strategic-mineral royalty (May 2026 decree, up from 3.5%) is applied to Zijin/Manono's first lithium export cargoes (~Q4 2026) with NO publicly negotiated Zijin carve-out or rebate before first export; and (2) at least one disclosed ARECOMS state-strategic-quota cobalt sale/deployment (from the ~9,600t of forfeited H1 2026 allocations under the 2026-06-29 order) occurs by 2027-01-31. Secondary confirmation: 2026 Comtrade DRC cobalt export unit value up >=15% YoY while volume stays within +/-5% of the 96,600t cap pro-rata. Call FAILS if Zijin publicly secures a royalty carve-out before first export, or no ARECOMS state-quota sale is disclosed by 2027-01-31, or Fastmarkets standard-grade cobalt falls below ~$10/lb and holds through 2026-12-31.
accumulating — awaiting horizon
By 2027-01-11: (1) the DLA award on its 2026-07-02 battery-grade lithium carbonate solicitation (~16,000t, $300M ceiling, bids due 2026-07-17) is announced at a disclosed or inferable unit price >=10% ABOVE Fastmarkets US battery-grade Li2CO3 spot at award date — the price-floor precedent, not the origin, is the claim; and (2) a DLA/NDS solicitation SPECIFICALLY for antimony or natural graphite appears on sam.gov — generic other battery-material solicitations do not count. Call FAILS if any two of: no lithium award by 2027-01-11; award priced at/below spot; no antimony/natural-graphite-specific solicitation by 2027-01-11.
accumulating — awaiting horizon
Indonesia's ESDM supplementary RKAB decision for PT Weda Bay Nickel (Eramet, ERA.PA; 2026 base quota cut to 12 Mwmt vs ~42 Mwmt guided; ore production halted June 2026), as published or reported by Mysteel/SMM/Eramet by 2026-09-30, restores WBN to LESS THAN 35 Mwmt for 2026. This call is WRONG if WBN is restored to >=35 Mwmt by 2026-09-30. Secondary leg: if the quota stays cut, ERA.PA total return underperforms the STOXX Europe 600 Basic Resources index by >=8% between 2026-07-13 and 2026-10-11.
accumulating — awaiting horizon
USA Rare Earth (NASDAQ: USAR) will NOT confirm its Stillwater, OK plant's 600 t/yr sintered NdFeB run-rate on schedule: in Q3 2026 through Q4 2026 disclosures (reporting through 2027-01-31), USAR discloses either a ramp delay past the end-2026 600 t/yr target or a feedstock-driven gross-margin deterioration vs prior guidance, caused by the 2026-06-22 MOFCOM entity listing barring any party worldwide from transferring China-origin rare-earth dual-use items to USAR (a 600 t/yr finished-magnet run-rate requires roughly 700-900 t/yr of alloy input absent heavy swarf recycling, and the non-China oxide-to-metal-to-alloy chain at that spec does not exist in 2026). This call is WRONG if USAR confirms the 600 t/yr run-rate on schedule by 2027-01-31 reporting with identified non-China feedstock and gross margin within prior guidance. Secondary leg: USAR underperforms MP Materials (MP) by >=15% over 2026-07-13 to 2026-11-10.
accumulating — awaiting horizon
CMOC Group (3993.HK), holding a 31,200 t 2026 DRC cobalt export allocation (~27% of its 2024 output) under ARECOMS's regime whose 2026-06-29 order forfeits unused quota to the state with no carry-forward, will disclose a visible financial mark within two quarters: in its H1 2026 report (~Aug 2026) or Q3 2026 disclosures, at least one of (a) a cobalt production-plan cut at the TFM/KFM circuits, (b) a quantified trapped cobalt inventory/working-capital disclosure exceeding $1B equivalent at current prices, or (c) a provision tied to quota forfeiture/expropriation risk. This call is WRONG if none of (a)-(c) appears in H1/Q3 2026 reporting AND ARECOMS grants CMOC a 2027 allocation >=45kt or any carry-forward by 2026-12-31. Price check (corroboration): 3993.HK vs Hang Seng Materials, 2026-07-13 to 2026-11-10.
accumulating — awaiting horizon
Mitsubishi Electric's (6503.T) H1 FY2026 results (late October 2026) will show defense & space segment revenue converting from backlog BELOW plan, or will contain materials-related delay/provision language (Dy/Tb magnet or gallium input constraints stemming from China's 2026-06-29 MOFCOM listing of Mitsubishi Electric units and the 06-24 anti-transshipment whistleblower regime). This call is WRONG if defense & space converts backlog at or above plan with no materials-related delay/provision language in the late-Oct 2026 disclosure. Secondary leg: 6503.T underperforms a Japan defense basket (7011.T, 7013.T) by >=8% over 2026-07-13 to 2026-11-10.
accumulating — awaiting horizon
By 2027-01-31, a named commercial linkage involving Neo Performance Materials (NEO.TO) SPECIFICALLY (not Solvay La Rochelle, Europe's other separator) is disclosed: an EU/member-state stockpile-linked supply, tolling, or storage-processing agreement involving Neo, OR a Japanese magnet customer contracting Neo's Narva NdFeB plant capacity (Phase 1 = 2,000 t) beyond Neo's existing book. Context: Neo's Silmet (Estonia) is the first producer of separated Dy/Tb in Europe (April 2026) and owns Europe's only integrated oxide-to-magnet chain. This call is WRONG if by 2027-01-31 no such agreement is disclosed AND Narva has announced no Phase 1 customer commitments beyond its existing book. Relative check (corroboration): NEO.TO vs MP over the window.
accumulating — awaiting horizon
A NEW disclosure made after 2026-07-13 and by 2026-11-30, by a Japanese automaker or magnet maker (e.g., Proterial, Shin-Etsu 4063.T, TDK 6762.T, Toyota, Honda), explicitly cites heavy rare-earth (Dy/Tb) or high-temperature magnet-grade availability as constraining EV/hybrid traction-motor output or costs - the auto chain, not defense, absorbing the shortage (China Dy/Tb oxide exports to Japan at zero since Nov 2025; defense demand holds priority claim on Lynas Malaysia's 1,500 tpa circuit, the only non-China heavy stream; Shin-Etsu's Fukui refinery produces nothing before ~2028). Pre-existing generic magnet-disruption disclosures (Nissan, Suzuki 2025-26) do NOT count. FAILED if by 2026-11-30 (a) no such new heavy-REE-specific disclosure exists AND (b) China customs show Dy or Tb oxide exports to Japan resumed >5 t/month for 2 consecutive months. Corroboration only: Japan auto names vs TOPIX spread -4% or worse.
accumulating — awaiting horizon
Magnet-driven supply constraints hit Indian e-2W makers (Bajaj, TVS, Ather) into the Oct-Nov 2026 festival season, because MOFCOM Announcement No. 26 (effective 2026-07-01, whistleblower bounties) criminalizes the third-country re-export channel while India's licensed MOFCOM channel is only a handful of foreign-owned Tier-1s (Continental India, Hitachi unit, Jay Ushin) under end-user certificates - too narrow for OEM-scale volume. FAILED if by 2026-12-31 VAHAN e-2W registrations for Oct+Nov 2026 grow >10% over Oct+Nov 2025 AND no Indian OEM publicly cites magnet supply as a constraint. Trade check: India HS 8505 (permanent magnets) monthly imports from China - no decline = no squeeze.
accumulating — awaiting horizon
By 2026-12-31, MP Materials' (MP) SEC filings (8-K/10-Q/10-K) show EITHER (a) a disclosed timing change to 10X/Fort Worth magnet commissioning or magnet capacity guidance, OR (b) a new heavy-REE supply agreement or risk-factor language that explicitly references sourcing constraints, premium pricing, or China-origin supply-chain restrictions - evidencing that MOFCOM's 2026-06-22 entity listing (which bars ANY party, including non-Chinese intermediaries, from transferring China-origin dual-use REE items to MP and USA Rare Earth) actually bit, contra the 'symbolic non-event' consensus. A routine offtake announcement with no constraint/premium/restriction language does NOT count as confirmation. FAILED if neither (a) nor (b) appears by 2026-12-31 with capacity guidance unchanged. Graded on the disclosure event, not price.
accumulating — awaiting horizon
The DLA National Defense Stockpile lithium carbonate award (solicitation 2026-07-02, ~36M lb / ~16,000 t battery-grade >=99.5% over FY2026-30, $300M ceiling, bids due 2026-07-17) is announced by 2026-12-31 AND the award price (derivable from obligated value and quantity on sam.gov/FPDS) is ABOVE prevailing China-linked spot lithium carbonate indices at award date - creating the first public US federal reference price/floor for domestically-favored carbonate. FAILED if no award is announced by 2026-12-31, OR the award price is at/below prevailing China-linked spot indices. Differentiated sub-signal (not load-bearing for the grade): a non-incumbent winner (e.g., Standard Lithium, Piedmont chain) evidences policy intent to seed new supply.
accumulating — awaiting horizon
SHORT-HORIZON DRUMBEAT: China customs monthly export data for July 2026 and August 2026 (published by ~2026-09-25) will show dysprosium and terbium oxide exports to Japan each remaining below 1 t/month - the zero-flow regime in place since November 2025 persisting through and past the 2026-06-29 entity-list escalation, feeding the Japan auto-magnet at-risk thesis. WRONG if either month's data shows Dy or Tb oxide exports to Japan at or above 1 t (a resumption signal that would also start the 2-consecutive-month falsifier clock on the main Japan call).
accumulating — awaiting horizon
CMOC Group (3993.HK), holder of a 31,200 t 2026 DRC cobalt export quota (~27% of its 2025 output of 117,549 t) and subject to ARECOMS's 2026-06-29 forfeiture order, reports in its H1 2026 interim report (due ~Aug 2026) cobalt sales volumes down MORE than 30% y/y (claimed: >50%) against roughly flat mine output, with a disclosed cobalt inventory build or carrying-value adjustment; AND 3993.HK does not outperform GLNCY by >10% over the 90 days from 2026-07-16. Falsified if volumes fall <30% y/y, or no inventory build/discounting is disclosed, or CMOC outperforms GLNCY by >10%.
accumulating — awaiting horizon
Nickel Industries (NIC.AX): Indonesia's 2026-07-09 implementing decree (MoF 32/MK/BC/2026 under Permendag 17/2026) recaptures NPI under HS Ex. 7202.60.00 with mandatory surveyor reports/export authorization. Claim: NIC's September-2026 quarterly report (due Oct 2026) shows the friction fingerprint — realized NPI price discount to the SMM Indonesia NPI index widening vs the June 2026 quarter plus working-capital stretch — and mandatory DSI (Danantara) channeling of NPI from 2027-01-01 is confirmed with NO written exemption for NIC's RKEF lines. Falsified early if NIC or its IMIP partners disclose a granted DSI/NPI exemption or the Trade Ministry publishes a written NPI carve-out before September 2026; cross-check NIC.AX vs VALE 90-day relative from 2026-07-16.
accumulating — awaiting horizon
AXT Inc (AXTI), whose GaAs/InP/Ge substrates are made in Beijing by subsidiary Tongmei under per-shipment MOFCOM export permits (GaAs permits already 'came in light' in Q1 2026): at the Q2 2026 earnings print (~early Aug 2026), Q3 revenue guidance midpoint comes in BELOW street on permit timing, against consensus extrapolating the Q1 beat; AXTI underperforms SOXX over the 60 days following the print. Falsified if the Q3 guide midpoint is at/above street and management describes permits as normalizing.
accumulating — awaiting horizon
Lynas Rare Earths (LYC.AX): the September-2026 quarter (FY27 Q1) report shows a visible average-selling-price uplift vs the June 2026 quarter, driven by the Dy/Tb mix shift from the 1,500 tpa Malaysian heavy-RE circuit's first full-rate quarters (first ex-China separated Dy May 2025, Tb June 2025) — an uplift NdPr-only sell-side models understate. Single-leg falsifier: September-quarter ASP flat vs June quarter = wrong. Cross-check LYC.AX vs MP 90-day relative from 2026-07-16. Pre-registered adversarial pair with impact-forge-lynas-token (which claims Dy+Tb volumes are token).
accumulating — awaiting horizon
China stainless/nickel role-reversal: at the 2027-01 China-customs/Comtrade vintage, (a) China's NPI/FeNi import volume from Indonesia for H2 2026 is down >=8% y/y (Indonesia's 260 Mt RKAB hold vs 379 Mt approved 2025, plus the 7/09 FeNi/NPI export controls), AND (b) transmission evidence exists — SHFE stainless-to-NPI margin proxy compressed vs the H1-2026 average or documented Chinese stainless mill run-cuts attributed to feed cost — AND (c) Philippine laterite CIF-China realizations are up >=15% y/y. Falsified if imports are not down >=8%, if no transmission evidence beyond the already-printed Jan-Apr -8% trend, or if the Philippines' share of China's nickel-ore import tonnage falls below 85%.
accumulating — awaiting horizon
South Korea graphite gap-year: a China graphite permit slowdown lands in H2 2026, before POSCO Future M's domestic spherical capacity (Saemangeum, W396bn) operates in 2027 — by 2026-12-31, at least one month of Korea customs graphite imports from China (HS 2504 + HS 3801 combined) prints <= -15% y/y, OR POSCO Future M (003670.KS) underperforms KOSPI by >=10% from 2026-07-16. Falsified only if BOTH hold: no single month <= -15% y/y AND no >=10% underperformance by 2026-12-31.
accumulating — awaiting horizon
Philippines two-buyer ore premium: by 2027-01-31 (PSA/Comtrade 2026 vintages + PSE prices), Philippine nickel-ore export tonnage for H2 2026 is up >=20% y/y (Indonesia roughly doubling purchases from 15.8 Mt toward ~30 Mt while the Philippines supplies >90% of China's nickel-ore imports), OR Nickel Asia (NIKL.PS) outperforms the PSEi by >=10% from 2026-07-16. Falsified only if BOTH fail.
accumulating — awaiting horizon
Argentina becomes the marginal ex-China carbonate supplier of record: at the 2027-Q1 Comtrade/INDEC vintage, Argentina's 2026 lithium carbonate export tonnage is up >=35% y/y AND Argentina's y/y export-tonnage increment EXCEEDS Chile's y/y increment (drivers: RIGI 30-year fiscal stability with 16 projects/~$30bn committed by May 2026; Rincon first carbonate March 2026 + $2.5bn 60kt expansion; Fenix and Sal de Vida first production Q2 2026; vs Chile CEOL 50% state-take friction and DRC's lithium royalty tripled to 10%). Falsified if either leg fails.
accumulating — awaiting horizon
US stockpile award composition: the DLA National Defense Stockpile lithium award (solicitation SP8000-26-R-0021; up to 16,167 t battery-grade carbonate, $300M ceiling, front-loaded 3,657 t year one, bids closed 2026-07-17) posts on sam.gov as a SPLIT award across multiple suppliers AND/OR accepts allied-origin (Chile/Argentina-fed) material, naming at least one non-obvious mid-cap — revealing US domestic battery-grade incapacity (Albemarle Silver Peak ~5kt/yr is essentially the whole domestic base). Leg 1 falsified if by 2026-10-17 the posted award is a single sole-source purely-domestic-origin award, or if no award posts by 2026-12-31. Leg 2: a further DLA critical-minerals stockpile solicitation (cobalt, antimony, or graphite) posts on sam.gov by 2027-01-16; falsified if none.
accumulating — awaiting horizon
Share-gated transmission: re-running the engine's pooled 30-event supply-event study stratified by frozen chokepoint share at event date (deterministic, all inputs already in the PIT vault; grade by 2026-09-15), the >=0.90-share stratum (rare-earth separation 0.90, anode-grade graphite 0.92, gallium refining 0.94) shows pooled 20-trading-day proxy drift >= +4%, exceeds the <0.80-share stratum (lithium 0.65, cobalt 0.76) by >= 3 percentage points, and passes a stratified placebo test at p <= 0.15. Killed if the >=0.90 stratum's drift is < +3%, or the spread vs the <0.80 stratum is < 3 pp, or placebo p > 0.15.
accumulating — awaiting horizon
US anode-graphite pull-forward before the 2026-11-27 expiry of China's suspension of the stricter-review layer on US-bound graphite (Announcement No. 46 of 2024): (1) US imports of Chinese natural + artificial graphite (HS 2504 + HS 3801, anode-relevant lines) for Aug+Sep 2026 average >=25% above the Mar-Jul 2026 monthly baseline (baseline frozen in the PIT vault before August data exists; graded when September 2026 Census/Comtrade data publish ~2026-11-05); (2) >=1 new binding US anode offtake/qualification announcement (Novonix Chattanooga, Anovion, or Syrah Vidalia) by 2026-10-31. DEAD if import volumes are <10% above baseline AND no new offtake lands.
accumulating — awaiting horizon
Gallium suspension-cliff hoarding: as China's suspension (Announcement No. 74; runs 2025-11-09 through 2026-11-27) of its US-directed gallium export ban enters its final ~90 days, deadline-conditioned restocking lifts the ex-China (US/Europe) gallium spot price by >=25% vs the 2026-07-15 PIT-vault mark by 2026-10-15 — no new China gallium control/renewal event required. DEAD if by 2026-10-15 the ex-China gallium spot price is up <25% from the 2026-07-15 mark AND the supply-event ledger records no new China gallium control/renewal event.
accumulating — awaiting horizon
Ex-China heavy-RE separation is token because HEAVY FEEDSTOCK binds (Mt Weld is light-RE dominant; the heavy-concentrate channel remains Myanmar-to-China, KIA-taxed 35,000 yuan/t): Lynas's June-2026-quarter activities report (due ~late July 2026) discloses combined Dy+Tb oxide output UNDER 15 tonnes, AND the chokepoint ledger's China rare-earth SEPARATION share stays >=0.85 through 2026-11-12. KILLED if the quarterly discloses >=15 t combined Dy+Tb for the June quarter, OR the China separation share is revised below 0.85 on cited ex-China volume evidence. Pre-registered adversarial pair with impact-co-lyc (Lynas ASP-uplift call) — one of the two loses.
accumulating — awaiting horizon
Cobalt tightness is a stockpile illusion: standard-grade cobalt metal (or hydroxide payable equivalent) FALLS >=10% from its 2026-07-15 PIT-vault frozen mark by 2026-10-13, as DRC quota shipments plus ~12+ months of accumulated hydroxide (production continued through the 2025 suspension; the ban was imposed FOR oversupply per ARECOMS Decision 001) clear into destocking Chinese refiners (76% chokepoint share). KILLED if the price has NOT fallen >=10% by 2026-10-13, OR if a new ARECOMS quota CUT or re-suspension is recorded before then. Interim checkpoint 2026-08-15: ARECOMS quota utilization — fully subscribed quotas plus rising price = early PLAUSIBLE-WRONG. Frozen explicitly as the adversarial counterparty to the board's cobalt-tightness calls (impact-co-3993hk, impact-co-glncy).
accumulating — awaiting horizon
Standing CONDITIONAL, armed only on a qualifying event: if the supply-event ledger records a new China rare-earth licensing/denial/export-control event (denial wave, processing-time extension, or magnet-license tightening under the Apr-2025 MOFCOM Announcement No. 18 regime) dated on or before 2026-11-12, then the PIT-vault rare-earth proxy basket 20-trading-day response is >= +8% — roughly 3x the ~2-3% consensus/pooled-model response, because at 90% separation share a denial strands demand rather than rerouting it. KILLED if a qualifying event occurs and the 20-day response is < +4% or the placebo window is not clean. EXPIRES UNSCORED (no credit claimed) if no qualifying event occurs by 2026-11-12.
accumulating — awaiting horizon
The Codelco-SQM lithium joint venture will NOT have completed Chilean Contraloria toma de razon (the legal registration precondition for JV effectiveness and for SQM's 2031-2060 Atacama quota extension) by 2026-10-15 — because CGR Oficio No. 78132 (2026-04-22) rejected registration and Codelco chose in July 2026 to litigate rather than resubmit, putting registration on a judicial clock that runs months. KILL if, on or before 2026-10-15, the Contraloria completes toma de razon of the JV/lithium-project decree (checkable in the CGR public registry / Diario Oficial) OR either company declares the JV legally effective. Single mechanical leg, no escape hatch. Secondary recognition read, recorded as evidence but with no power to save or kill: SQM vs ALB cumulative spread from the PIT price vault, freeze 2026-07-19 to 2026-10-15.
accumulating — awaiting horizon
DLA solicitation SP800026R0021 (up to ~16,167t battery-grade lithium carbonate for the National Defense Stockpile, $300M ceiling over five years, front-loaded ~3,657t year one, LPTA method, bids closed 2026-07-17) results in an award announced on sam.gov by 2026-10-31 to a domestic/allied producer (not a trader/intermediary with FEOC-ambiguous material), and the named awardee outperforms an equal-weight ALB/SQM/LAC lithium producer basket by 10+ percentage points in the 20 trading sessions after the award notice. KILLED if no award by 2026-10-31, or the award goes to a trader/intermediary with non-domestic FEOC-ambiguous material, or the awardee fails the +10ppt relative test.
accumulating — awaiting horizon
MP Materials (NYSE: MP) heavy-feedstock gap: after MOFCOM's 2026-06-22 entity-listing barred any party from transferring China-origin dual-use items to MP, its mid-2026-commissioned heavy separation circuit (200 t/y Dy+Tb nameplate off ~3,000 t/y feedstock; Mountain Pass bastnaesite is heavy-lean) will show a constraint. KILLED if by 2026-11-15 (Q3-2026 reporting) MP confirms on-schedule separated Dy+Tb oxide production AND names at least one secured third-party heavy-rare-earth feedstock source, with no disclosed heavy-feedstock constraint in the Q3 10-Q.
accumulating — awaiting horizon
US domestic magnet ramp (MP Materials / USA Rare Earth channel) underperforms the decoupling narrative after MOFCOM's 2026-06-22 entity-listing cut both firms off from China-origin Dy/Tb feedstock, separation consumables and metal-making equipment: MP's Q3-2026 report (~November 2026) shows a magnet/NdPr segment output shortfall vs prior guidance OR discloses incremental Dy/Tb sourcing cost or timeline caveats in 10-Q text. KILLED if segment output meets or beats prior guidance AND no such caveat appears. Weakening condition (recorded): a disclosed non-China Dy/Tb supply agreement at reasonable terms before 2026-10-31.
accumulating — awaiting horizon
The November-2025 suspension (Announcement 74, through 2026-11-27) of China's US-directed antimony ban delivered headline relief but not flows: China-origin US imports of antimony (HTS 2825.80 antimony oxides plus HTS 8110 antimony metal, country of origin China, US Census Bureau data) remain below 33% of the 2023 pre-control China-origin monthly average in EVERY month July, August, and September 2026 — because the discretionary MOFCOM license layer underneath the suspended prohibition never went away. KILLED if any single one of those three months prints at or above 33% of the 2023 monthly baseline for the same HTS set. Grade on 2026-11-15, when all three monthly prints are published (September data ~2026-11-05). One threshold, no dead zone, no event-based escape.
accumulating — awaiting horizon
Export Finance Australia / the responsible ministers will announce the Critical Minerals Strategic Reserve's FIRST binding transaction — an offtake, floor-price contract, or purchase agreement naming a producer, with rare earths among the covered commodities — on or before 2026-11-16 (enabling Act passed 2026-04-01; published schedule says operational H2-2026; FY2026-27 appropriations flowing from 2026-07-01). MOUs, shortlists, consultations, and non-binding Letters of Support (including the 2026-05-21 Arafura LoS) do NOT count. KILL if no such binding instrument is announced by 2026-11-16, OR if the first executed instrument contains no rare-earth component. Frozen watch basket for recognition read (recorded, non-saving): LYC.AX, ILU.AX, ARU.AX from the PIT vault.
accumulating — awaiting horizon
Nickel Industries (ASX: NIC) is squeezed from both ends by Indonesia's PP 24/2026 state export monopoly (DSI sole channel with pricing authority for ferroalloys >=8% Ni, phased from 2026-06-01, mandatory DSI platform 2027-01-01) plus the 2026 RKAB base ore quota held at 260-270 Mt vs 379 Mt approved for 2025. KILLED if NIC's December-half 2026 reporting shows NPI margin per tonne within 10% of the June-half 2026 level with no disclosed DSI-related shipment delays, OR if Indonesia formally exempts standard NPI from the DSI channel before 2026-09-01.
accumulating — awaiting horizon
Ford (NYSE: F) suffers the differential magnet halt: a Ford production stoppage, shift cut, or force-majeure attributed to magnet/rare-earth supply occurs by 2026-12-31 under China's enforcement-driven squeeze (2026-06-24 MOFCOM Announcement No. 26 whistleblower/transshipment regime), while GM — which pre-contracted domestic magnet supply via MP Fort Worth and Noveon Magnetics — does not halt for the same cause. KILLED if no such Ford stoppage occurs by 2026-12-31 while at least one Chinese enforcement action against transshipment is publicly reported (mechanism fires but Ford is unhurt). If no enforcement action is reported at all, the mechanism never fired and the call is VOID rather than wrong.
accumulating — awaiting horizon
POSCO Future M (KRX: 003670) hits the 18-month feedstock gap: its Sejong natural-graphite anode line runs on spherical graphite 100% imported from China under China's permit regime, while its ex-China chain (Saemangeum ~$321M plant; Tanzania Mahenge 60kt/y) only arrives ~2027-2028. KILLED if by 2027-01-31 POSCO Future M has disclosed no graphite feedstock shortage or Chinese permit delay affecting anode output, AND natural-graphite anode volumes grow year over year.
accumulating — awaiting horizon
Olin's (NYSE: OLN) Winchester segment shows the primer-grade antimony trisulfide squeeze (no qualified US commercial source; MOFCOM Announcement No. 26 of 2026-06-24 closes third-country grey channels): KILLED if Winchester segment adjusted EBITDA margin in H2-2026 is within 100bp of H1-2026 AND the FY2026 10-K (Feb 2027) contains no antimony-related supply-risk disclosure beyond boilerplate. Survives only on 100bp+ H2 margin compression or a specific antimony supply-risk disclosure.
accumulating — awaiting horizon
Nickel Industries Limited (ASX: NIC), whose production base is entirely Indonesian RKEF/NPI plus HPAL, is harmed by Indonesia's PP 24/2026 single-gate export regime (signed 2026-05-20, effective 2026-06-01; private export rights ending 2026-09-01 after President Prabowo's 2026-07-20 acceleration from 2027-01-01), which makes state-owned PT Danantara Sumberdaya Indonesia (DSI) the sole exporter of record for ferroalloys. TWO-LEG CLAIM as of 2026-07-24: (1) the official DSI regulated-commodity list published on or before 2026-12-31 includes nickel pig iron / ferronickel (HS 7202.60) either outright or with a stated phase-in date — i.e. the FINI/SMM 'NPI is permanently excluded' grey-zone comfort is wrong; AND (2) the harm appears as realised-price erosion rather than volume: NIC's H2 2026 realised NPI price, disclosed in its H2/FY2026 results on or about 2027-02-28, prints at a WIDER discount to the SMM 8-12% Indonesian NPI index than its 2025 average discount band. GRADED WRONG if the published list excludes HS 7202.60 outright with no phase-in date, OR if the H2 2026 realised price stays within the 2025 historical discount band. Grade 2027-02-28.
accumulating — awaiting horizon
VIGO Photonics S.A. (WSE: VGO; market cap ~PLN 437m at 2026-07-20) was named on China's MOFCOM Announcement No. 30 of 2026-07-24, which lists 14 EU entities and bars ANY party anywhere from re-transferring China-origin dual-use items to them. Vigo's entire product line is HgCdTe and InAsSb epitaxial infrared detectors whose consumables are high-purity tellurium, cadmium, indium and gallium plus CdZnTe/GaAs substrate wafers. The claim is NOT that non-China supply is absent (5N Plus in Canada and Coherent/II-VI produce 6N Te and CdZnTe); it is that Vigo's MBE recipes are lot- and vendor-qualified into fielded missile-seeker and gas-analyser bills of material, so substitution is a 9-18 month requalification and the harm lands as inventory build and gross-margin compression, not a revenue cliff. CLAIM: by its FY2026 report (about March 2027) Vigo shows EITHER gross margin more than 200bp below its trailing-four-quarter average as at 2026-07-24, OR a disclosed feedstock provision, supply-chain charge or substrate requalification programme. GRADED WRONG if BOTH the Q3 2026 report (November 2026) and the FY2026 report show gross margin within 200bp of that trailing average with no such disclosure; ALSO WRONG if Vigo announces a COMPLETED non-China substrate or tellurium qualification, or obtains a MOFCOM exceptional-case licence, before 2026-12-31. Vigo's July 2026 statement to Reuters that the listing 'will not have a critical impact on operations' concerns the order book and does not grade this call.
accumulating — awaiting horizon
Olin Corporation (NYSE: OLN) will show its Winchester segment squeezed by ANTIMONY, the one ammunition input with no exchange contract and therefore no hedge, concentrated in the fixed-price military/small-caliber mix (Lake City plus the 2025 small-caliber and AMMO Inc. asset acquisitions) that the Q1 2026 commercial price increases structurally cannot reach. Baseline facts frozen at 2026-07-24: Rotterdam antimony was about US$51,800/t on 2026-07-21; China's export ban to the US was SUSPENDED in November 2025 and replaced by a licensing regime running to 2026-11-27; management named copper, brass and propellant — not antimony — on the Q4 2025 call, where Winchester segment earnings fell from US$42m to US$0.6m and Olin posted an US$85.7m net loss. GRADED WRONG if EITHER (a) Winchester segment EBITDA margin, as disclosed in the Q3 2026 release (about late October 2026) or the Q4/FY2026 release (about late January 2027), returns to or above its 2024 full-year average while the Fastmarkets Rotterdam antimony free-market price is still above US$40,000/t, OR (b) Olin discloses before 2027-02-09 a fixed-price antimony supply agreement, an antimony escalation clause on military contracts, or a qualified non-antimonial primer or core substitution. GRADED RIGHT only if Winchester segment EBITDA margin stays below the 2024 full-year average in BOTH Q3 and Q4 2026 AND disclosed military/small-caliber volume mix is flat-to-up over the same period, so a pure commercial-demand miss cannot be scored as confirmation. Grade 2027-02-09.
accumulating — awaiting horizon
Sinomine Resource Group (SZSE: 002738) will have its Bikita, Zimbabwe lithium concentrate stranded by Zimbabwe's 2027-01-01 concentrate export ban, which Mines Minister Polite Kambamura refused to defer on 2026-07-17 ('We are still sticking with January 1'). The binding fact is that BOTH compliance paths are closed on the same date and neither is political: Zimbabwe's only operating lithium sulphate plant, Zhejiang Huayou Cobalt's at Arcadia, cannot process third-party feed, and Bikita's own sulphate plant is incomplete and not expected operational before the deadline. GRADED WRONG if Zimbabwe publishes a Statutory Instrument, gazette notice or ministerial exemption before 2026-12-31 deferring, phasing or carving out lithium concentrate exports, OR if Sinomine announces a binding tolling agreement with the Huayou sulphate plant, or commissions its own Zimbabwean sulphate capacity, before 2027-03-31. GRADED RIGHT if the ban takes effect on 2027-01-01 without deferral AND Sinomine has neither a tolling agreement nor commissioned in-country conversion at 2027-03-31. Grade 2027-03-31.
accumulating — awaiting horizon
Neo Performance Materials (TSX: NEO) benefits because its Narva plant (opened September 2025) is the only sintered NdFeB magnet facility inside the EU, Phase 1 nameplate 2,000 t/yr ramping toward 5,000 t/yr, fed by its Silmet (Sillamae, Estonia) heavy-rare-earth solvent-extraction line commissioned April 2026 — a SMALL-SCALE line at modest nameplate that proves Europe can separate terbium and dysprosium, not that it can supply them at volume. The demand event is MOFCOM Announcement No. 30 of 2026-07-24, which listed 14 EU entities including magnet-dependent Lafert and TATRA, legally forcing named EU consumers off China-origin magnets with no transitional supply. Disclosed offsetting exposure: Neo retains Magnequench manufacturing inside China. THREE-LEG CLAIM, all must hold to be RIGHT: (1) Neo announces at least one NEW European magnet supply agreement or LOI with an EU industrial or defence customer by 2027-03-31; (2) Narva demonstrates a 2,000 t/yr Phase-1 run-rate in Neo's Q4 2026 or Q1 2027 disclosure; and (3) Neo's magnetics segment revenue for FY2026 grows year over year. GRADED WRONG if any leg fails. Grade on disclosed contracts, disclosed run-rate and segment revenue only — never on the NdPr price. Grade 2027-03-31.
accumulating — awaiting horizon
Compass Diversified (NYSE: CODI) is mispriced as a distressed holdco (Lugano fraud CODI itself called 'pervasive, complex and isolated to Lugano', FY2022-24 restated, distributions suspended, US$50.4m of management fees clawed back) while owning Arnold Magnetic Technologies, the largest Western producer of samarium-cobalt magnets, with non-Chinese samarium oxide contracted from Solvay's La Rochelle line via Less Common Metals (2025 agreement) and a March 2026 mutual distribution agreement with USA Rare Earth. The demand driver is EO 14415, signed 2026-07-20, under which from 2027-01-01 the Secretary of War ceases issuing 10 U.S.C. 4872(c)(1) nonavailability waivers ABSENT an exhaustive mitigation plan — waivers are conditioned and tightened, NOT abolished — with SmCo restrictions extending across the chain from ore or recycled feedstock to finished magnet. GRADED WRONG if ANY of: (a) Arnold's adjusted EBITDA growth in CODI's FY2026 10-K is below +20% year over year measured LIKE-FOR-LIKE, excluding the US facility-relocation and Arnold CEO severance add-backs and disclosing the Thailand capacity-ramp contribution that drove part of the ~100% Q1 2026 growth; (b) CODI sells or spins Arnold at an EV/EBITDA multiple at or below the Sterno transaction multiple (US$292.5m EV comparable), i.e. no strategic critical-materials premium; or (c) Arnold discloses it cannot secure samarium feedstock for 2027 deliveries. Size the re-rating against CODI's FY2026 guidance of US$95-105m total Industrial adjusted EBITDA, of which Arnold is a portion. Grade 2027-05-20.
accumulating — awaiting horizon
Finland's exposure to Russian-fed nickel refining is NOT closed by the EU's 21st sanctions package (Council Regulation (EU) 1848/2026, in force 2026-07-23), which prohibits imports of Russian-origin metal ORES including nickel ores (HS 2604) and bans exports of nickel powders and alloys, but does NOT touch nickel matte. Norilsk Nickel Harjavalta Oy refines up to about 65,000 t/yr of Class-1 nickel in Finland with roughly 94% of feedstock arriving as converter matte under CN 7501 — the third consecutive package to spare that line, after the 20th package (April 2026) banned CN 7503-7505 while explicitly leaving CN 7501 and CN 7502 untouched, and the same 21st-package annex banned Russian COPPER up to matte (CN 7401). PRIMARY CLAIM, graded 2026-12-31: Finnish Customs (Tulli ULJAS) Russian-origin CN 7501 nickel-matte import volume for Q3 2026 lands within +/-25% of Q3 2025 AND Nornickel Harjavalta announces no production suspension before 2026-12-31 — i.e. the flow is live and anyone marking it as sanctioned-out is mismarking. GRADED WRONG if Harjavalta suspends output before 2026-12-31, or Q3 2026 Russian-origin CN 7501 imports fall more than 40% year on year. SECONDARY escalation leg, graded separately at 2027-06-30: an EU regulation published in the Official Journal on or before 2027-06-30 extends import prohibitions to CN 7501 or CN 7502.
accumulating — awaiting horizon
South Korea's binding critical-materials constraint is anode-grade graphite, not NdFeB magnets, and it is DEEPENING: China supplied 90.34% of Korean graphite imports January-May 2026 versus 84.32% in 2021, against a China chokepoint share of about 92% of anode-grade graphite and a permit regime (from 2026-06-22) that throttles throughput with no formal ban and therefore no shock date. Anode is roughly 10-15% of cell cost, and Korea must cut Chinese content for US tax-credit eligibility exactly as its Chinese share climbs. CLAIM graded on two supply-chain legs only: (1) China's share of Korean natural-graphite (HS 2504) plus artificial-graphite (HS 3801.10) imports stays at or above 88% on EVERY rolling three-month period ending on or before 2027-03-31 per KITA/Korea Customs; AND (2) at least two of LG Energy Solution (373220.KS), Samsung SDI (006400.KS) and SK Innovation (096770.KS) cite anode or graphite sourcing as a QUANTIFIED margin or utilisation factor across their Q3 2026 through Q1 2027 results — generic risk-factor boilerplate does not count. GRADED WRONG if the China share falls below 88% for any qualifying rolling three-month period, or if fewer than two issuers make a quantified citation. No index-relative leg: EWY is priced off memory semiconductors and cannot grade an anode-cost claim. Grade 2027-05-10.
accumulating — awaiting horizon
On MOFCOM Announcement No. 30 of 2026-07-24 (14 EU entities), the measurable input-denial casualty is the chokepoint-dependent SME, not the defence prime: Rheinmetall's CEO has publicly put Chinese content at about 1% of the company, while VIGO Photonics (WSE: VGO, 2024 revenue about PLN 78m, loss-making, single-site, no second substrate source) builds HgCdTe and InAs/GaSb infrared detectors from Chinese-refined gallium, indium, tellurium, selenium and antimony. CLAIM, graded 2027-03-31: BOTH (a) Vigo Photonics discloses, in an ESPI current report or in its H1 or FY2026 financial statements before 2027-03-31, a component, substrate or order constraint attributed to export controls; AND (b) Rheinmetall AG (RHM.DE) issues NO supply-driven guidance cut over the same window. GRADED WRONG if no such Vigo disclosure appears by 2027-03-31, or if Rheinmetall issues a supply-driven guidance cut while Vigo reports no constraint (meaning the prime, not the SME, was the real target). No equity-relative leg: VGO-versus-WIG over 90 trading days is noise on a micro-cap and does not grade the mechanism. DISCLOSED CORRELATION: this call shares its subject entity and thesisFamily with impact-co-vgo-wa-feedstock-2026, which grades margin and provisions; this one grades the SME-versus-prime divergence.
accumulating — awaiting horizon
In the 2026 tightening cycle China is the IMPORT-DEPENDENT party in stainless, and the first quantifiable damage lands on Chinese import costs in September-November 2026. Indonesia's PP 24/2026 single gate was pulled forward on 2026-07-20 to 2026-09-01 (from 2027-01-01), making state-owned PT Danantara Sumberdaya Indonesia the transaction gate for ferroalloys with a mandate to set price benchmarks and retain export proceeds; ESDM held the 2026 RKAB ore quota at 260 Mt wmt versus 379 Mt in 2025, and MoF Decree 32/MK/BC/2026 with Trade Reg 17/2026 imposed Surveyor Reports and BUMN-channelled authorisation on HS ex-7202.60. Indonesian nickel exports to China ran about US$7.86bn in 2025. CLAIM: China GACC imports of HS 7202.60 from Indonesia for September-November 2026 show EITHER a CIF unit value at least 10% above the June-August 2026 average, OR a volume decline of at least 15% year on year. GRADED WRONG if unit value is flat-to-down AND volume lands within +/-10% year on year (DSI proves an administrative pass-through, not a toll), or if the 2026-09-01 single-gate date slips again by published Indonesian regulation. Chinese stainless mill margins are explicitly NOT a grading leg — they move on property demand and mill oversupply. Grade 2026-12-31.
accumulating — awaiting horizon
Zimbabwe's lithium-concentrate export ban, confirmed for 2027-01-01 by Mines Minister Polite Kambamura on 2026-07-17 ('We are still sticking with January 1'), functions in the graded window as a CHINA-CONSOLIDATION event rather than a sovereignty event, because the only operating in-country converter is Zhejiang Huayou Cobalt's Arcadia sulphate plant (about 400 ktpa concentrate, no third-party capacity) and the next two — Sinomine's Bikita and Kamativi — are also Chinese-linked and not expected operational before the deadline. CLAIM graded 2027-06-30: BOTH (a) no non-Chinese-owned lithium sulphate or carbonate plant has commissioned in Zimbabwe by 2027-06-30; AND (b) at least one Zimbabwean lithium asset has changed hands to, or signed exclusive toll-conversion with, a Chinese-controlled converter (Huayou, Sinomine or Tsingshan), evidenced by a published transaction, ZIMRA notice or ministerial announcement. GRADED WRONG if a non-Chinese converter commissions in Zimbabwe in the window, OR the ban is deferred by published Statutory Instrument before 2027-01-01, OR no such transaction or toll agreement occurs by 2027-06-30. 'Does the ban hold' is deliberately NOT the graded question — it is a high-probability, low-information claim and is carried only as a lead-time trigger.
accumulating — awaiting horizon
Japan's measurable 2026 damage from China's Japan-specific control campaign lands on the DEFENCE AND PRECISION entities Beijing actually named on MOFCOM Announcements No. 27 and No. 28 of 2026-06-29 (~40 Japanese entities, including the National Institute for Defense Studies, Mitsubishi Electric and Mitsubishi Heavy units, Mitsui E&S, Mitsui Bussan Aerospace, Terra Drone, ACSL and Mitsubishi Nuclear Fuel) — not on the auto chain, which has engineered dysprosium out since 2010 via grain-boundary diffusion and Dy-lean NdFeB and holds JOGMEC stock. Context frozen at 2026-07-24: Chinese customs recorded zero Dy, Tb, Y and Ga shipments to Japan in June 2026, with Tb/Dy at zero since November 2025. CLAIM: at least TWO of Mitsubishi Electric (6503.T), Mitsubishi Heavy Industries (7011.T), Mitsui E&S (7003.T) and ACSL (6232.T) cite the 2026-06-29 MOFCOM listings or Chinese dual-use export controls as a supply-chain, cost or order factor in H1 FY26 results published on or before 2026-11-30. GRADED WRONG if fewer than two do. No EWJ-versus-EEM leg and no JMTBA machine-tool leg — an unrelated macro selloff must not decide a supply-chain claim, and JMTBA publishes no clean aero/defence monthly cut. The acknowledged risk to this call is that Japanese issuers disclose supply constraints sparingly, which is why conviction is 0.5.
accumulating — awaiting horizon
Indonesia captures RENT, not merely tonnage, from the DRC's cobalt quota regime, because Indonesian cobalt is a by-product of HPAL/MHP nickel output on already-sunk capex and is therefore near price-INELASTIC: it arrives regardless of what the DRC does to defend price. The DRC's ARECOMS order of 2026-06-29 forfeits unused H1 2026 export quotas (no carry-forward, reassigned to the state strategic quota) inside an annual cap near 96,600t, while Indonesian output is forecast to grow about 21.2% in 2026 to about 59,800t from about 14.9% of world production in 2025 versus the DRC's about 72%. CLAIM: BOTH (a) Indonesian cobalt-in-MHP/hydroxide export VALUE per contained tonne (Indonesia BPS trade statistics) for FY2026 rises year on year while DRC cobalt export volume falls year on year; AND (b) USGS Mineral Commodity Summaries 2027 puts Indonesia's share of world cobalt mine production at 16% or more. GRADED WRONG if Indonesian unit value is flat-to-down despite falling DRC volume (rent capture falsified even if the tonnage forecast lands), OR if USGS MCS 2027 puts Indonesia's share below 16%, OR if the DRC lifts quotas and 2026 DRC exports rise year on year. A tonnage-only threshold was deliberately rejected: it would re-grade an already-published forecast. Grade 2027-03-01.
accumulating — awaiting horizon
The under-modelled Australian critical-minerals asset is Bayer-process alumina liquor, not a mine: the 2026-07-14 Final Investment Decision by Australia, Japan, the United States and Alcoa on a gallium plant co-located at Alcoa's Wagerup alumina refinery in Western Australia (reported at up to about 10% of global gallium demand, about US$307m total, sovereign partners funding the SPV and taking offtake in proportion to funding) establishes a retrofit template at roughly US$3m per annual tonne of capacity, against a gallium refining chokepoint of about 94% China. The claim graded is REPLICATION, not Wagerup itself. CLAIM: before 2027-12-31, BOTH (a) a SECOND gallium or germanium extraction unit co-located at an alumina or zinc refinery in Australia, Canada, Japan or the EU reaches FID with nameplate of at least 25 t/yr AND has either a signed government funding commitment or published offtake — feasibility studies, MOUs and letters of intent explicitly do NOT qualify; AND (b) the Wagerup project reports construction start or long-lead equipment orders. GRADED WRONG if no qualifying second unit reaches FID by 2027-12-31, or if the Wagerup FID is deferred, descoped below 100 t/yr, or its tri-government offtake structure is amended before that date. Note frozen at 2026-07-24: no global primary-gallium tonnage denominator was verified, so no world-share percentage is claimed.
accumulating — awaiting horizon
Chile's state-JV lithium execution SPEEDS UP despite the nationalisation label, because Contraloria General de la Republica Resolucion No. 14/2026, published in the Diario Oficial on 2026-07-09, amended Resolucion No. 36/2024 to let corporate and joint-venture agreements of Codelco and Enami be adopted WITHOUT prior toma de razon, substituting ex-post legality review (controles de reemplazo) for the ex-ante bottleneck; Codelco withdrew its related action after publication. The claim is about speed, not deregulation — control is substituted, not eliminated. CLAIM: at least TWO Codelco or Enami lithium corporate agreements, joint ventures or CEOLs are executed and published between 2026-07-09 and 2027-06-30 under the ex-post regime without a Contraloria objection. GRADED WRONG if fewer than two are executed and published in that window, or if the Contraloria reverses or re-tightens Resolucion 14/2026 before 2027-06-30. The SEA admissibility of the Salar Futuro EIA is deliberately excluded as a grading leg — admissibility is a near-automatic formality and evidences nothing about velocity. Grade 2027-06-30.
accumulating — awaiting horizon
SHORT-HORIZON TRIGGER CHECK. Indonesia's PP 24/2026 single-gate export regime takes effect for ferroalloys on 2026-09-01 as accelerated by President Prabowo on 2026-07-20 (from the original 2027-01-01), without further deferral: by 2026-09-15, published Indonesian instruments (State Gazette regulation, Ministry of Trade or ESDM implementing rule, or an official DSI/Danantara notice) confirm that PT Danantara Sumberdaya Indonesia is the mandatory exporter of record or transaction gate for ferronickel/NPI exports for shipments from 2026-09-01, ending private exporters' own export rights for that commodity group. GRADED WRONG if a published Indonesian instrument dated on or before 2026-09-15 defers the ferroalloy start date beyond 2026-09-01, or if no implementing instrument or official notice confirming the 2026-09-01 start exists by 2026-09-15. This is the load-bearing lead condition for both impact-co-nic-ax-dsi-npi-2026 and impact-cty-china-npi-toll-2026 and is frozen separately so the trigger itself is scored.
accumulating — awaiting horizon
SHORT-HORIZON, ABOVE-BASE-RATE EVENT COUNT (forge candidate hyp-a7701b9e, adopted with a POST-SNAPSHOT baseline). During the Announcement 72/74 suspension window toward the US (running to 2026-11-27), China's restrictive pressure on gallium, germanium, antimony, graphite and rare earths reroutes to non-US targets rather than relaxing. CLAIM: by 2026-10-17 the cited supply-event ledger records at least TWO NEW qualifying China-origin restrictive events with event dates from 2026-07-25 to 2026-10-17 inclusive. MOFCOM Announcement No. 30 of 2026-07-24 (14 EU entities) is EXPLICITLY EXCLUDED from the count, because it had already fired at freeze and a falsifier that is part-satisfied at freeze is not a prediction. Qualifying means: China-origin act (MOFCOM, GACC or MIIT); material in {gallium, germanium, antimony, graphite, rare earths}; restrictive signal type (export control, targeted ban or entity listing); and the cited text names a non-US target jurisdiction or entity. US-targeted events, renewals of existing US measures and non-China events do NOT count. The ledger's own measured rate for this event class is about 0.3/month (one visible 2026 instance before the snapshot: the 2026-04-24 MOFCOM listing of seven European defence entities), so the base-rate expectation over the window is roughly one. GRADED WRONG if fewer than two qualifying events with dates in that window are recorded by 2026-10-17.
accumulating — awaiting horizon
REGIME-PERSISTENCE CALL, ANTI-CONSENSUS vs the captured naked prior (kraken worldline origin — see origin note). Between 2026-07-29 and 2026-09-26 inclusive, US BIS export-control rulemaking does NOT flip back to expansion dominance: fewer than 3 Federal Register RULE documents (final or interim-final) classified EXPAND, OR EASE >= EXPAND over the window. MECHANICAL CLASSIFICATION (frozen): EXPAND iff the document's own summary/preamble states ANY of (a) adds >=1 item/ECCN to the CCL or imposes new controls on items; (b) adds >=1 entity/address to the Entity List, MEU list, or any EAR restriction list; (c) extends license requirements to new destinations, end-uses, or end-user classes; (d) removes/narrows/revokes a license exception or authorization (incl. VEU revocations) — regardless of co-occurring easing provisions (any-expansion-action-wins tie-break; residual ambiguity resolves toward EXPAND, i.e. AGAINST this claim). EASE iff ONLY easing actions (removal/narrowing of controls or listings, streamlining, favorable treatment, suspension of a prior expansion, relaxed review policy) and none of (a)-(d). All other rules (conforming, corrections, admin/process/fee, tariff-process) count neither. PREREGISTERED BASE TABLE (all 16 trailing-13-month BIS RULEs classified at freeze under this rule): EXPAND = 2026-07-23 suppressor CCL controls; 2025-10-09 Entity List +29; 2025-09-30 affiliates end-user IFR; 2025-09-16 Entity List +32; 2025-09-02 VEU revocation. EASE = 2026-07-14 UAE favorable treatment; 2026-04-09 IC-designer extension; 2026-01-21 drone streamlining; 2026-01-15 advanced-computing review relaxation; 2025-11-12 Entity List removals; 2025-11-12 suspension of affiliates expansion; 2025-09-02 Syria relaxation. NEITHER = 2026-02-04 Cambodia conforming; 2025-12-05 CCL technical; 2025-09-30 firearms-requirements revision; 2025-09-17 Section-232 process. Only the Sep-Oct 2025 burst would have failed this claim; every rolling 60-day window since 2025-11-01 passes it. GRADED WRONG iff EXPAND >= 3 AND EXPAND > EASE. Panel note recorded at freeze: the disjunctive pass condition is lenient — only a Sep-2025-scale burst fails it; that regime persistence IS the claim, and the naked prior still leans the other way (0.4). [Origin: kraken worldline candidate kraken-fc:934f519621 'Trade Tensions Escalate' (KRAKEN-PROTOTYPE working memory, support 0.79 from 22 locally-extracted claims) surfaced the US control-tempo question; primary-source grounding measured the OPPOSITE recent tempo and the two-skeptic panel flipped the direction — two prior forms killed in-panel 2026-07-28, reasons kept in the forge ledger (hyp-ef0a1e07 lineage). The frozen edge: the kraken world's US-side escalation is NOT confirmed on this horizon.]
accumulating — awaiting horizon
LG Energy Solution (KRX 373220) operating margin EXCLUDING US AMPC/45X credits (disclosed separately in LGES quarterly earnings releases) declines quarter-over-quarter in BOTH Q3-2026 and Q4-2026, WHILE CATL (SZSE 300750) gross margin in those quarters holds within 1 percentage point of its Q2-2026 level — the pass-through indexation-gap prediction: LGES's metal-linked contracts index LME/spot references that miss MHP payables, cobalt sulfate premia and spherical graphite FOB, and its US plants uniquely eat the anode-duty leg. Both members required; graded at Q4-2026 earnings (~Jan-Feb 2027).
accumulating — awaiting horizon
Conditional two-branch, frozen 2026-08-05. Trigger (mechanical): MOFCOM's suspended October-2025 second-wave rare-earth/magnet export-control package, suspension expiring 2026-11-10, is re-imposed/frictioned vs extended. Branch A (re-imposition or documented license friction): BorgWarner (NYSE: BWA) — the major Western e-motor supplier with NO disclosed captive or contracted ex-China magnet source — cuts ePropulsion margin guidance or calls out magnet cost/allocation in its Q4-2026 results/FY-2027 guide (~Feb 2027), while magnet-secured peers (GM/MP chain) make no equivalent callout → right. Branch B (suspension extended or made permanent by 2026-11-10): BWA makes no magnet-related disclosure through the Q4-2026 print → thesis expires wrong.
accumulating — awaiting horizon
On MOFCOM Announcement No. 30 (2026-07-24, 14 EU entities listed), graded by 2027-01-31: (leg 1) Rheinmetall (RHM.DE) makes NO disclosure of a China-materials-driven program delay or cost provision — its listing is symbolic, exposure runs through tiers; AND (leg 2) the midstream bite is visible: Eurostat China-to-Italy NdFeB magnet imports (HS 8505.11) fall at least 10% YoY in H2-2026, OR Sumitomo Heavy (6302.T) discloses a Lafert supply disruption or magnet requalification program. Wrong if Rheinmetall discloses a material China-materials impact, or if Italian magnet flows hold (<10% decline) with no Lafert-chain disclosure.
accumulating — awaiting horizon
Two-branch, branch selected mechanically by the published DSI commodity list at the 2026-09-01 go-live of Indonesia's single-gate export system (PP 24/2026). Branch A (NPI/ferronickel HS ex 7202.60 included at launch): China customs import unit values of Indonesian ferronickel/NPI (HS 7202.60, GACC monthly) widen to a premium of at least 5% versus LME-implied contained-nickel value relative to the H1-2026 average baseline, within 90 days of 2026-09-01; falsified if the premium stays within ±3% of the H1-2026 baseline through 2026-12-31. Branch B (FeNi-only launch scope): no material premium expected before the Jan-2027 phase 2 — grading defers one quarter and re-tests on Q1-2027 data. If the DSI gate formally slips past 2026-10-01, grading defers one quarter (pre-registered).
accumulating — awaiting horizon
EU ores-vs-matte seam, graded by 2027-03-31: EITHER (member 1) Finland's monthly imports of Russian-origin nickel mattes (CN/HS 7501, Eurostat Comext) run at least 15% below the 2025 monthly average for two or more months BEFORE any formal ban extension — bank/insurer de-risking after the 21st package's Russian nickel-ORES ban (Reg. 1848/2026, in force 2026-07-23) that legally spares Nornickel Harjavalta's ~94%-Russian matte feed; OR (member 2) EU 22nd-sanctions-package text published in the Official Journal extends nickel prohibitions beyond ores to mattes/unwrought (CN 7501/7502). Falsified if HS 7501 flows stay within ±15% of 2025 volumes AND no extension text appears by the date.
accumulating — awaiting horizon
Waiver-cliff papering rush plus scrap-rule follow-through, two members: (member 1) at least 3 DISTINCT new non-Chinese magnet supply or qualification agreements disclosed by US defense primes or their tier-1s (8-K/press releases naming NdFeB magnet supply or qualification) between 2026-08-05 and 2027-01-15, ahead of EO 14415's (2026-07-20) cessation of 10 U.S.C. 4872 waivers from 2027-01-01 — fewer than 3 falsifies the timing claim; AND (member 2) a Commerce implementing rule under the 2026-07-30 DPA Section 101 recoverable-materials determination published in the Federal Register with export-licensing coverage of black mass by 2027-03-31 — absence falsifies the margin-transfer leg. Frozen for the record: DLA cancelled its $300M/16,167t lithium carbonate solicitation on 2026-08-03; that leg is dropped and flips to evidence of the domestic-conversion-vs-price tension, re-tested on any re-issue. MP-vs-SPX is supportive color, not a gate.
accumulating — awaiting horizon
LME 3-month nickel records at least one daily official close at or above 1.12x its 2026-07-17 official close (frozen from the PIT price vault at freeze) on or before 2026-10-30 — the claim that consensus structural-surplus complacency ignores Indonesia's PP 24/2026 statutory quota lever plus the EU removal of Russian-origin nickel from the deliverable Western pool. Single leg, no OR-escape: only the price level grades it; killed if no qualifying close occurs by 2026-10-30.
accumulating — awaiting horizon
By 2026-11-16, at least one NEW restrictive antimony supply event — export duty, licensing regime, quota, suspension, or state-marketing monopoly — is announced by a NON-China jurisdiction (candidate set led by Tajikistan, plus Myanmar, Turkey, Bolivia, Russia), evidenced by a citable dated primary source (official gazette or ministry announcement) and entered on the cited supply-event ledger under its normal citation standard. Chinese events explicitly do NOT count (a MOFCOM renewal or expiry-window action cannot rescue this), price moves do not count, and there is no OR-leg. The ledger's measured base rate for this event class is zero; killed if no qualifying event by the date.
accumulating — awaiting horizon
By 2026-09-15, the officially published commodity list in force for Indonesia's DSI (PT Danantara Sumberdaya Indonesia) single-gate export system at its 2026-09-01 go-live INCLUDES nickel pig iron / ferronickel under HS ex 7202.60 among the mandatorily routed export commodities (per the implementing ministerial regulation or official Danantara/government announcement). Predicted outcome: inclusion — MoF Decree 32/MK/BC/2026 already routes HS ex 7202.60 through surveyor/BUMN-Ekspor channels. Graded NO if the launch scope is FeNi-only or excludes NPI. This call is the mechanical branch-selector for the frozen NIC.AX, China-nickel, and Indonesia rent-capture calls.
accumulating — awaiting horizon
Sinomine Resource Group (SZSE: 002738) hits a hard stop on Bikita lithium concentrate exports when Zimbabwe's concentrate export ban starts 2027-01-01 because its Bikita lithium sulphate plant is not commissioned. WRONG if ANY of: (1) Sinomine announces (SZSE filing or company release) commissioning/first production of a Bikita lithium sulphate or hydroxide plant dated on or before 2026-12-31; (2) the Government of Zimbabwe grants Bikita Minerals a specific concentrate export exemption or deferral published by 2027-01-31; (3) 002738.SZ total return exceeds the CSI Lithium Battery theme index from 2026-08-24 to 2027-01-31.
accumulating — awaiting horizon
The BIS DPAS black-mass allocation order (Federal Register 2026-16078, effective 2026-08-27) acts as a feed subsidy to operating US hydromet recyclers, Cirba Solutions (private, Lancaster OH) first. WRONG if ANY of: (1) US Census exports of HS 8549.11+8549.19 for Sep-Dec 2026 average within 20% of the Jan-Jun 2026 monthly average in USD value; (2) the Fastmarkets or Benchmark Mineral Intelligence US NCM black-mass payables assessment for Dec 2026 is not at least 10 percentage points of payables below the corresponding Asia (Korea CIF) NCM black-mass payables assessment; (3) Cirba Solutions publicly announces capacity idling, plant closure, or a financing distress event by 2027-01-31.
accumulating — awaiting horizon
Iluka Resources (ASX: ILU) secures a HEAVY-rare-earth-specific sovereign instrument for Eneabba ahead of the EO 14415 US defense magnet waiver cut-off (2027-01-01). WRONG if BOTH: (1) between 2026-08-24 and 2027-02-28 Iluka issues no ASX announcement of a Dy and/or Tb offtake, price floor, or US DoD/DFC or Japanese JOGMEC commitment that names Eneabba heavy rare-earth product and discloses a volume or a floor price (NdPr-only deals and generic Australian critical-minerals-reserve headlines do not count); AND (2) ILU.AX total return trails LYC.AX total return from 2026-08-24 close to 2027-02-28 close.
accumulating — awaiting horizon
The BIS DPAS black-mass rule (effective 2026-08-27) transfers pricing power to US recyclers. WRONG if EITHER: (1) the Fastmarkets US NMC black-mass payables assessment (% of contained Ni+Co+Li value) averaged over Nov-2026, Dec-2026 and Jan-2027 is less than 5 percentage points below its Jan-Jun 2026 average; OR (2) US Census exports of HS 8549.11+8549.19 to all destinations for Sep-2026..Jan-2027 combined fall by less than 70% in USD value versus Sep-2025..Jan-2026 combined.
accumulating — awaiting horizon
The ARECOMS DRC cobalt export quota (~96,600 t for 2026, roughly 45% of pre-ban run-rate) binds physically on the shipping-lag clock: China GACC customs imports of DRC-origin cobalt intermediates (HS 8105.20, cobalt hydroxide/mattes and intermediates) for August 2026 plus September 2026 combined print at or below 70% of August 2024 plus September 2024 combined tonnage. WRONG (killed) if the GACC September-2026 release (available by 2026-10-25) shows Aug+Sep 2026 DRC-origin tonnage above 70% of the Aug+Sep 2024 baseline.
accumulating — awaiting horizon
China's MOFCOM 2026-04-24 entity listing of Hensoldt AG bars Chinese exporters from shipping controlled germanium and gallium IR-optics inputs to it, and the squeeze concentrates in the Optronics segment: Hensoldt's 9M-2026 report (expected ~2026-11-13) prints Optronics segment adjusted EBIT margin at least 100 basis points below the Optronics adjusted EBIT margin in the 9M-2025 report. WRONG (killed) if the 9M-2026 report, graded by 2026-11-20, shows Optronics adjusted EBIT margin flat or up year-on-year, or down by less than 100bp. A Hensoldt disclosure naming Ge/Ga sourcing is confirmatory color only and cannot rescue a failed margin gate.
accumulating — awaiting horizon
First full-month tell of the BIS DPAS black-mass rule (effective 2026-08-27): US Census exports of HS 8549.11 plus HS 8549.19 to all destinations for September 2026, as published in the October/November 2026 Census release (~2026-11-05), are more than 70% below September 2025 in USD value. WRONG if the September 2026 combined value is at or above 30% of the September 2025 combined value.
accumulating — awaiting horizon
Iluka Resources (ASX: ILU) secures a HEAVY-rare-earth-specific sovereign instrument for Eneabba ahead of the EO 14415 US defense magnet waiver cut-off (2027-01-01). WRONG if BOTH: (1) between 2026-08-24 and 2027-02-28 Iluka issues no ASX announcement of a Dy and/or Tb offtake, price floor, or US DoD/DFC or Japanese JOGMEC commitment that names Eneabba heavy rare-earth product and discloses a volume or a floor price (NdPr-only deals and generic Australian critical-minerals-reserve headlines do not count); AND (2) ILU.AX total return trails LYC.AX total return from 2026-08-24 close to 2027-02-28 close.
accumulating — awaiting horizon
SHORT-HORIZON DRUMBEAT: Eramet's (ERA.PA) H1 2026 results, expected ~end-July 2026 (graded by 2026-08-12), will disclose that PT Weda Bay Nickel ore production remains halted or operating only within the unchanged 12 Mwmt 2026 quota, and will NOT confirm an approved ESDM supplementary quota restoring WBN to >=35 Mwmt as of the release date (the ESDM application window is 1-31 July; approvals were expected July-September, so a confirmed >=35 Mwmt restoration inside July would fire against us early). WRONG if the H1 release or an accompanying announcement confirms an approved restoration to >=35 Mwmt.
hit
SHORT-HORIZON DRUMBEAT: CMOC Group's (3993.HK) H1 2026 interim report (HKEX deadline 2026-08-31) will disclose H1 2026 DRC cobalt production materially exceeding cobalt sales/export volumes - a visible in-country inventory buildup consistent with its 31,200 t 2026 export allocation (~27% of 2024 output) under the ARECOMS quota regime. WRONG if the H1 2026 interim report shows cobalt sales volumes at or above cobalt production (no buildup), or discloses no cobalt production/sales volume split at all.
hit
China's Announcement No. 74 (2025) suspension of the US-bound gallium export ban expires 2026-11-27; that dated cliff triggers order front-running while MOFCOM permits still flow. AXT Inc (AXTI) — the one US-listed compound-substrate maker producing INSIDE China (Tongmei) and shipping gallium substrates only under Chinese export permits — is the cleanest XBRL-visible conduit, so the pull-forward surfaces in its Q2-2026 results (report ~2026-08-06/10) before it shows anywhere else. Consensus read the suspension as 'gallium risk resolved' and prices no expiry premium; this call claims the deadline, not the relief, is the operative fact. [Origin: Hypothesis Forge inaugural pass 2026-07-15, mechanism-hunter lens, survived both skeptics, director rank #1.]
miss
Glencore (GLEN/GLNCY): the H1 2026 production report (~2026-07-30) confirms Glencore shipped its full DRC cobalt quota allocation (KCC 16,100 t + Mutanda 6,700 t 2026 quotas; 2025 carryover exported by April 2026), and the H1 2026 results (~2026-08-06) guide marketing-division EBITDA toward the TOP of its range with cobalt explicitly cited; the long-GLEN/short-3993.HK pair is positive over the 60 days from 2026-07-16. Falsified if marketing EBITDA guidance is unchanged with no cobalt attribution, or the production report shows Glencore failing to ship its DRC quota.
miss
SHORT-HORIZON DRUMBEAT checkpoint (first gate of impact-forge-lynas-token, split out for a 30-day grade): Lynas Rare Earths' (LYC.AX) June-2026-quarter activities report, due ~late July 2026, discloses combined dysprosium + terbium oxide production of UNDER 15 tonnes for the June 2026 quarter. KILLED if the report discloses >=15 t combined Dy+Tb oxide for the quarter. Grades from a single dated public disclosure already on the trigger-monitor calendar.
miss
MP Materials' (NYSE: MP) Q2-2026 Form 10-Q or accompanying earnings release (expected ~2026-08-06) will disclose a realized price per REO metric ton (or, if that metric is absent, realized NdPr price per kg) at least 20% above the corresponding Q1-2026 disclosed figure — because the April-2026 China control cluster (MOFCOM entity-listings 2026-04-24, MIIT tightening 2026-04-29) bifurcated the ex-China price from the China domestic NdPr quote consensus marks MP against. KILLED if the disclosed Q2 figure is less than 20% above Q1-2026. Grade within 5 trading days of filing, deadline 2026-08-20. VOID only if MP discontinues both realized-price disclosures. The filing number alone decides; no market-price leg.
miss
MP Materials' (NYSE: MP) Magnetics segment gross margin (XBRL segment revenue minus segment cost of sales) in the Q2-2026 10-Q (due ~2026-08-10) prints below Q1-2026 — the claim is at least 200bp of compression, because the grain-boundary-diffusion magnet grades MP ships to GM require dysprosium/terbium that Mountain Pass does not produce, and ex-China Dy/Tb input costs were driven up by China's Announcement No. 18 licensing (2025-04-04) plus Kachin/Myanmar heavy-RE disruptions. Single mechanical gate: KILLED if the Q2-2026 10-Q (grade by 2026-08-31) shows Magnetics segment gross margin flat or up versus Q1-2026. No disclosure leg, no OR-leg. Internally consistent pair with impact-co-mp-q2-realized-price (Materials realized price up, Magnetics COGS squeezed, same filing).
miss
The 2026-07-13 to 2026-07-19 nickel froth surge in the engine's radar (0.596 -> 0.808 in radar_history.jsonl) occurred with zero new cited nickel event since Indonesia's PP 24/2026 (2026-05-20) — an event-orphaned spike. Claim: the froth composite is picking up pre-citable narrative of PP 24/2026 implementing rules, and a new citable Indonesian nickel restrictive event — signalType in {export-control, permit-regime, quota, targeted-ban}, region=Indonesia, event date after 2026-07-13 — will enter the cited supply-event ledger under its normal citation standard by 2026-09-02. KILLED if by 2026-09-02 no such event exists in the ledger; a price move without a citable event does not count; no OR-leg, one hard date. A kill grades the froth composite as a lagging echo on nickel — itself a usable calibration result.
miss
AND-gate, graded from segment tables by 2026-09-05: Albemarle's (NYSE: ALB) Specialties segment net sales for Q2-2026 (10-Q ~2026-08-06) decline year-over-year vs Q2-2025, AND ICL Group's (NYSE: ICL) Industrial Products segment sales for Q2-2026 (report ~2026-08-12) decline year-over-year vs Q2-2025 — because record ex-China antimony-trioxide prices (post MOFCOM Announcement No. 46, 2024-12-03) are pushing compounders off Br+ATO flame-retardant systems entirely, destroying bromine demand two links downstream of the antimony headline. If EITHER segment grows YoY, the hypothesis is killed. Attribution commentary is mechanism color only, never the gate.
miss
robotics· 6 live
Picks-and-shovels thesis: the actuation supplier basket (reducers/motors/bearings — the largest humanoid BOM slice) outperforms the humanoid-OEM-adjacent froth proxy over 180 days.
accumulating — awaiting horizon
the lesson
Actuation basket fails to beat the OEM froth proxy by 5pts over 180d (the brand, not the bottleneck, captured the move).
The broad robotics-SUPPLIER complex outperforms the humanoid-OEM froth proxy over 180 days — value accrues to the bottleneck, not the brand.
accumulating — awaiting horizon
the lesson
The broad supplier complex fails to beat the OEM froth proxy over 180d.
Teradyne (robotics + semiconductor-test integration) grows revenue >6% YoY by horizon — real deployment shows up as supplier revenue, not demo videos.
accumulating — awaiting horizon
the lesson
Teradyne revenue YoY comes in <=6% at the next 10-K.
Ouster (LiDAR perception — a robotics-eyes pure-play) grows revenue >25% YoY by horizon.
accumulating — awaiting horizon
the lesson
Ouster revenue YoY comes in <=25% at the next 10-K.
Despite the humanoid IPO wave, the sector's humanoid truth mass (share of programs at L3+ / paid pilot or better) stays <0.25 at +365d — deployment reality lags the narrative.
accumulating — awaiting horizon
the lesson
Humanoid truth mass rises to >=0.25 within 365d (real paid deployment arrived faster than the null expected).
At least 2 cited safety/compliance events (incident, recall, regulatory action, or teleoperation reveal) enter the ledger within 365d as fenceless humanoid operation scales.
accumulating — awaiting horizon
the lesson
Fewer than 2 cited safety/compliance events land within 365d.
quantum· 6 live
Contrarian picks-and-shovels: the quantum supply-chain basket outperforms the pure-play froth proxy over 180 days — froth (pure-plays running hot on ~$1-130M revenue) reverts toward the shovels.
accumulating — awaiting horizon
the lesson
The pure-play froth proxy keeps beating the supply basket over 180d (hype persists / accelerates).
The post-quantum-crypto layer (the concrete-money side of 'quantum') outperforms the pure-play froth proxy over 180 days.
accumulating — awaiting horizon
the lesson
PQC-security names fail to beat the pure-play froth over 180d.
IonQ revenue YoY decelerates below 60% at the next 10-K — triple-digit growth off a tiny base (FY latest ~$130M, +200% YoY) does not sustain.
accumulating — awaiting horizon
the lesson
IonQ sustains >=60% revenue YoY at the next 10-K.
No fault-tolerant USEFUL quantum advantage (Q5) is demonstrated within 365d — the honest zero under the 'quantum breaks encryption / revolutionizes X' hype. Below-threshold QEC (Q3) exists; useful logical advantage does not.
accumulating — awaiting horizon
the lesson
A credible fault-tolerant USEFUL quantum advantage (Q5) is demonstrated within 365d.
At least 3 public roadmap milestones slip past their promised date within 365d — the industry slip-rate no one else publishes.
accumulating — awaiting horizon
the lesson
Fewer than 3 roadmap milestones slip within 365d (the industry hits its promises on time).
At least 1 quantum-advantage claim is credibly contested/debunked within 365d — the field's signature failure mode continues.
accumulating — awaiting horizon
the lesson
No advantage claim is credibly contested/debunked within 365d.
telescope· 10 live
Bee-research predictions frozen by hash; self-assessed, awaiting external outcomes.
68 frozen claims, accumulating 2026-07-03 → 2026-08-02 · 10 self-assessed · 0 externally graded
Octopus-research predictions frozen by hash; self-assessed, awaiting external outcomes.
38 frozen claims, accumulating 2026-07-03 → 2026-07-19 · 2 self-assessed · 0 externally graded
Telescope-research predictions frozen by hash; self-assessed, awaiting external outcomes.
62 frozen claims, accumulating 2026-07-03 → 2026-08-02 · 10 self-assessed · 0 externally graded
Quantum_Research-research predictions frozen by hash; self-assessed, awaiting external outcomes.
104 frozen claims, accumulating 2026-07-06 → 2026-08-01 · 0 self-assessed · 0 externally graded
Materials_Research-research predictions frozen by hash; self-assessed, awaiting external outcomes.
125 frozen claims, accumulating 2026-07-06 → 2026-07-24 · 0 self-assessed · 0 externally graded
Robotics_Research-research predictions frozen by hash; self-assessed, awaiting external outcomes.
92 frozen claims, accumulating 2026-07-06 → 2026-07-31 · 0 self-assessed · 0 externally graded
Sandbox_Research-research predictions frozen by hash; self-assessed, awaiting external outcomes.
29 frozen claims, accumulating 2026-07-07 → 2026-07-07 · 0 self-assessed · 0 externally graded
Ants-research predictions frozen by hash; self-assessed, awaiting external outcomes.
74 frozen claims, accumulating 2026-07-07 → 2026-07-30 · 0 self-assessed · 0 externally graded
Eye-first per-operator calibration beats a fixed frequency peg
Model-only: Arnold-tongue fixed-drive R=0.49 vs IAF-drive R=1.00 (d=4.0), fixed-drive coupling-vs-IAF r=-0.955; no real per-operator data yet.
Bee-derived bistable-breadth regime detector with a no-brake tail
Frozen bee model only: bistable ~4→(7-15 separatrix)→~37 stings, governor caps ~37 vs ~67; wealth analog not yet tested out-of-sample against a linear baseline.
Paper-only research + personal R&D. Not investment advice.
Time-to-recognition — was the engine early?
materials · protocol v1 · 2026-08-26Nothing here is a “lead” yet. This is the backward pass: for each frozen edge, did consensus-grade coverage of the specific second-order claim already exist at freeze? An edge that clears that bar is differentiated at freeze — a necessary condition, not proof of value. It becomes a proven-lead only when consensus actually arrives later (lead-days = recognition − freeze). Those accrue over 12–24 months.
Consensus-grade (frozen v1) = the same entity+mechanism claim in ≥1 tier-1 source OR ≥2 independent tier-2 sources. Conservative bias: already-consensus when uncertain — the grader’s job is to refute the engine’s novelty. All 32 already-consensus rows carry the exact quote that sank them. Rules were frozen (RECOGNITION_PROTOCOL.md v1) before this data.
The export-layer cash drag arrives (09-01) before any ore-tightness margin benefit shows, and it hits the RKEF/NPI slice, not the MHP stakes (HNC, Excelsior). Claim the observable: NIC's Q3 2026 quarterly (due ~late October) discloses (a) RKEF NPI sales tonnes down >5% vs Q2 while production is flat (shipment deferral), or (b) an explicit DSI/BUMN fee, reference-price or documentary-delay item, or (c) receivables/DSO above 60 days in H2 accounts. Danantara's stated 'no profit margin' is part of
Grader confidence: 0.60 · engine conviction 0.42
Searched for the specific second-order claim: NIC's RKEF/NPI cash drag from the DSI export layer arriving 09-01 before ore-quota margin benefit, hitting NPI not MHP. Findings: (1) Tier-1 Reuters (May 22) reported NPI EXEMPTED from DSI centralisation, and SMM (Jul 9) reported NPI later folded into LS/export-authorization controls with BUMN-only export from Jan 2027 — both event coverage, no NIC lin
Listed-entity denial is qualitatively different from price tightness: denied buyers must requalify a magnet source and Narva is the obvious one for EU motor makers. Name the counterparty class: by 2027-02-28 Neo discloses at least one NEW European magnet supply agreement or qualification with a MOFCOM-listed entity or an EU defense/motor OEM (Lafert-class PM motor lines are the prime candidate), or announces a Narva Phase-2 expansion decision explicitly citing EU demand. Eurostat HS 850511 China
Earliest coverage: None found matching entity AND mechanism. Nearest misses: Bloomberg 2025-09-19 "Europe's EV Makers Line Up at Estonia's New Rare-Earth Magnet Plant" (Narva as general EU alternative for EV makers — supply-diversification mechanism, not listed-entity denial); SMM analysis 2026-07-29 names Lafert as PM-motor maker exposed to NdFeB but names no alternative supplier; euinsider/eutoday (non-tier) say Rheinmetall "can redesign procurement and qualify new suppliers, but doing so takes time" without naming Neo/Narva.
Grader confidence: 0.70 · engine conviction 0.50
The listing of 14 EU entities (MOFCOM Announcement No. 30, 2026-07-24) including Rheinmetall and Lafert S.p.A. is Tier-1 consensus as an event (Bloomberg, Reuters, S&P Global, OSW, SCMP). But the engine's specific claim — that denial to listed entities forces magnet-source requalification and that Neo's Narva plant is the resulting counterparty (new agreement/qualification with a listed entity or
The claim is a HEAVY-rare-earth-specific sovereign instrument, not generic support: by 2027-02-28 Iluka announces a Dy/Tb (not NdPr) offtake, price floor or DoD/DFC/Japanese (JOGMEC) commitment naming Eneabba heavy-RE product, with volume or floor price disclosed. Generic Australian critical-minerals-reserve headlines or NdPr-only deals do NOT count.
Earliest coverage: None for the sovereign heavy-RE claim. Nearest adjacent (does not match mechanism): Iluka Resources ASX notice 'Rare earths offtake agreement', 22 June 2026 (2026-06-22)
“Beyond being Iluka's first, the agreement is significant in that it encompasses the full suite of light and heavy magnet rare earth oxides and contains minimum prices agreed between commercial parties that are independent of those backed by governments.”
Grader confidence: 0.60 · engine conviction 0.45
Searches for an Iluka/Eneabba heavy-rare-earth (Dy/Tb) sovereign instrument (US DoD, DFC, JOGMEC, Australian Critical Minerals Strategic Reserve offtake/price floor) returned nothing naming Eneabba Dy/Tb product from Bloomberg/Reuters/FT/WSJ/government documents, nor from two independent trade-press/think-tank sources. Coverage that exists is (1) the Australian A$1.65bn EFA loan (financing, not an
The damage is name-based, not volume-based: Japan MOF HS 8505.11 import tonnage from China for Sep-2026..Jan-2027 stays within -10% yoy (non-listed buyers keep importing, so a volume reading looks falsely calm) AND Mitsubishi Electric's FY2026 Q3 (Jan-Feb 2027) Defense & Space segment commentary carries an explicit rare-earth/magnet sourcing caveat. Country-ETF leg dropped as macro noise.
Grader confidence: 0.70 · engine conviction 0.40
Underlying event (China Jan-6-2026 dual-use ban on Japan; Feb-24 entity list of 20 firms incl. MHI/KHI/IHI/NEC subsidiaries + 20-firm watchlist incl. TDK/Subaru/Mitsubishi Materials; June expansion to ~80 firms) is heavily Tier-1 covered (Reuters via marketscreener/UPI, CNBC 2026-06-29, SCMP, CSIS 2026-01-13, Benchmark 2026-01-13, S&P Global 2026-01-27). The name-based/entity-list MECHANISM is the
The country-level observable is Malaysian HS 2846 exports to Japan+US, a clean single-plant series. The mix shift toward Dy/Tb at 10-20x NdPr per kg should show as unit value, not tonnage: Malaysia DOSM/Comtrade HS 2846 exports to Japan+US for Jul-Dec 2026 show USD/kg unit value up >30% yoy WHILE tonnage is within +/-10% yoy. Lynas reported record prices in Q4 FY26, so the falsifier must isolate mix from price via the tonnage leg.
Earliest coverage: Nearest (mechanism only, not the claim): Lynas Q4 FY26 quarterly report as reported by Australian Mining, "Lynas revenue hits four-year high on record rare earth prices" — states record ASP A$98.2/kg was "driven by improved NdPr pricing, a higher mix of heavy rare earth sales and rising premiums over the market index." This is a company-level mix/price statement, not the Malaysian HS 2846 unit-value-vs-tonnage observable. (2026-07 (Q4 FY26 quarterly, late July 2026; exact day not confirmed))
Grader confidence: 0.60 · engine conviction 0.50
Searches for the specific claim (Malaysian HS 2846 exports to Japan/US, unit value vs tonnage, as a Dy/Tb mix-shift diagnostic) returned nothing from Bloomberg/Reuters/FT/WSJ, government documents, or trade press. What IS consensus at freeze: (1) Lynas separates Dy/Tb in Malaysia (Fastmarkets, Mining.com, Reuters-syndicated 2025); (2) Lynas's own Q4 FY26 report and trade-press coverage attribute t
The same policy that lifts the metal transfers margin AWAY from Indonesia-domiciled foreign-listed producers: quota-tight ore raises NIC's feed cost, and DSI intermediation (fees, FX/proceeds retention, payment lag) claws the export leg. The open NPI-scope question resolves ~2026-09-01; if NPI is included, NIC underperforms the metal. Consensus prices the metal, not the intermediation wedge.
Earliest coverage: none establishing consensus (nearest adjacent: mining.com.au "Septembergate", trading-house margin capture — fails entity test) (2026-05-30)
Grader confidence: 0.70 · engine conviction 0.60
The underlying events were saturated consensus by freeze (2026-08-05): the RKAB quota cut (Argus, Mysteel, Benchmark), the DSI single-gate export regime with fees/proceeds retention and its Sep-1 acceleration (ICIS 07-29, Benchmark, SMM), and even the open NPI-scope question with a ~September inter-ministerial resolution (SMM 06-10; FINI clarity request). But the protocol requires the SAME second-
The NOVONIX slip (disclosed in NVX FY2026 filings) converts Panasonic's US anode plan from 'covered by 2026' to 'uncovered through H2-2027' exactly as the final duties landed. The margin leak concentrates in North America Energy and should surface as cost commentary/guidance trim within two quarters — before consensus models catch it.
Earliest coverage: None found for the specific claim. Closest non-qualifying: Bloomberg newsletter "Tesla and Panasonic Lose a Costly Battle Over Battery Sourcing" (first-order duty exposure, predates the slip) (2025-07-23)
Grader confidence: 0.68 · engine conviction 0.55
The specific second-order claim requires the SAME entity (Panasonic North America Energy) AND SAME mechanism (NOVONIX qualification slip converting the US anode plan from covered-by-2026 to uncovered-through-H2-2027 precisely as final duties landed, producing a concentrated margin leak due to surface as cost commentary/guidance trim within two quarters). Nine targeted searches, including a Tier-1
LGES's metal-linked pass-through clauses index LME/spot references that do not capture MHP payable percentages, cobalt sulfate premia, or spherical graphite FOB — so H2-2026 physical-premium tightening plus the US anode duty leg compresses LGES ex-AMPC operating margin QoQ in both Q3 and Q4 2026 while CATL gross margin holds within 1pt, an amount demand weakness alone cannot explain.
Earliest coverage: none found matching entity AND mechanism; nearest-miss = S&P Global Platts subscriber note launching daily MHP CIF North Asia cobalt payable (2026-06-02), which evidences the LME-vs-payables pricing gap as a trade-press topic but makes no LGES margin claim (2026-06-02)
Grader confidence: 0.60 · engine conviction 0.50
The frozen protocol requires the SAME specific second-order claim (same entity AND same mechanism) in a tier-1 source or two independent tier-2 sources at/before 2026-08-05. Seven targeted searches found only: (a) heavy consensus coverage of the underlying headline events — the ~220% total US graphite anode duty (final determination 2026-02-11, LGES assigned 93.5% AD), DRC/producer cobalt export q
The 2026-11-10 expiry is public and priced as a macro headline; the differentiated claim is entity-level: BWA is the major Western e-motor supplier with NO disclosed captive or contracted ex-China magnet source (unlike GM/MP and offtake-holding peers), so any re-imposition or license friction reprices BWA's magnet input to the DoD-anchored Western level (~40-80% over China spot) rather than merely delaying it — visible first as magnet cost/allocation language in BWA's Q4-2026 guide while magnet-
Grader confidence: 0.60 · engine conviction 0.45
The underlying event (2026-11-10 expiry of Beijing's suspension of the October-2025 controls, and autos-at-risk generally) was thoroughly consensus by freeze — Bloomberg, CSIS, IEA, trade press all covered it, and the DoD $110/kg NdPr floor / ~63-100% Western-vs-China premium was public. But the protocol excludes underlying-event coverage. The engine's specific second-order claim is entity-level:
UAMY's cash flows are now delivery-cadence driven, not spot-driven — the $57.3M booked against the $245M ceiling converts to revenue as the expanded furnaces ramp, so quarterly DLA delivery revenue grows ≥50% QoQ in Q3-2026 and again in Q4-2026 regardless of the falling antimony tape. Almost nobody models the invoice run-rate.
Earliest coverage: None found meeting the tier definition. Closest non-qualifying antecedent: UAMY's own Q1-2026 guidance ($125M FY26, ~$75M from DLA deliveries) as carried by Yahoo Finance/AccessNewswire — company self-disclosure via non-tier channels.
Grader confidence: 0.60 · engine conviction 0.60
Applying protocol v1 literally: no Tier-1 source and no pair of independent Tier-2 sources stated the specific second-order claim (UAMY cash flows delivery-cadence-driven against the $245M DLA ceiling, quarterly DLA revenue growing >=50% QoQ in Q3/Q4-2026 regardless of the falling antimony tape) at or before the 2026-08-05 freeze. Targeted searches at Reuters/Bloomberg and at Fastmarkets/Argus/S&P
EO 14415 does not create an escape-hatch-free prohibition — waivers survive under formal mitigation plans with exhaustive unavailability evidence — but that is precisely the mechanism: primes must document that compliant sources were unavailable, forcing qualification engagement with Stillwater during H2-2026. Claim: at least one defense-prime or DoW-linked supply/qualification agreement disclosed before 2027-01-31 and USAR outperforms MP over the window. Watch item: China listed USAR on its own
Grader confidence: 0.70 · engine conviction 0.55
The underlying event (EO 14415) and its waiver-mitigation mechanism were consensus by late July 2026 (Fastmarkets, Breaking Defense, law-firm alerts), and the generic reading "good for US rare earths / USAR among beneficiaries" was also consensus (government stake Feb 2026, Cantor Overweight May 2026, Motley Fool MP-vs-USAR pieces). But the engine's SPECIFIC second-order claim — that waiver surviv
The mapping is the edge: the listed entities (Lafert's IE4/IE5 PM motor lines above all) need qualified ex-China sintered magnets in months, and Narva is the only EU source with PPAP-stage capability in 2026. Claim: Neo discloses at least one NEW European magnet customer win or a Narva expansion decision within 8 months, pulled forward by the July listing — turning Narva from 'option' to 'sold out Phase 1A'.
Earliest coverage: None found making the specific claim. Closest second-order piece: Shanghai Metals Market (SMM) analysis of the EU watchlist's rare-earth supply-chain impact — mentions Lafert's magnet exposure but names no alternative supplier and never mentions Neo or Narva. (2026-07-29)
Grader confidence: 0.70 · engine conviction 0.55
The underlying event (MOFCOM Announcement No. 30, 2026-07-24, listing 14 EU entities incl. Lafert S.p.A. and Rheinmetall AG) was widely covered at/before freeze, but per protocol underlying-event coverage does not count. The engine's specific second-order claim requires the mapping: listed entities (Lafert's IE4/IE5 PM motor lines) need qualified ex-China sintered magnets in months, Narva is the o
This is a RE-squeeze prediction, not a description of the resolved 2025 episode (Chetak output recovered): under the 2026 MOFCOM escalation template, India's dual-EUC queue re-lengthens first among major importers — observable as approval turnaround >60 days (OEM-reported) and HS 8505.11 imports from China down >30% YoY in at least one month of H2-2026, BEFORE any VAHAN registration dip; at least one listed OEM (BAJAJ-AUTO.NS, OLAELEC.NS, TVS) discloses a magnet-driven cut or launches a reduced-
Grader confidence: 0.70 · engine conviction 0.60
The engine's specific second-order claim is a conditional H2-2026 re-squeeze prediction with India-first sequencing (dual-EUC turnaround >60 days and HS 8505.11 imports down >30% YoY before any VAHAN registration dip, listed-OEM disclosure by 2026-12-31) under the 2026 MOFCOM escalation template. The resolved 2025 episode and India's Rs 7,280-crore magnet scheme were saturated consensus, but the e
The list targets the MIDSTREAM, not the primes: Lafert and Sindlhauser are magnet/motor nodes for European industrial drives, and III-V Lab + Vigo are a GALLIUM/InP photonics chokehold — the first EU action linking the REE list to the gallium regime. Rheinmetall's inclusion is largely symbolic (exposure runs through tiers). Prediction: bite appears as lead-time extension in EU industrial-motor order books and share shift to unlisted EU motor makers within two quarters, while Rheinmetall disclose
Earliest coverage: Geopolitechs (Substack newsletter by Peng Zhang) — closest prior articulation of the mechanism, but non-consensus-grade by rule; earliest countable partial match: SMM (Shanghai Metals Market) analysis, 2026-07-29 (2026-07-24)
“Gallium is the most obvious pressure point: China supplies more than 90 per cent of global primary gallium production and has required export licences for gallium-related items since 2023.”
Grader confidence: 0.60 · engine conviction 0.55
Underlying event: MOFCOM Announcement No. 30 (2026-07-24) listing 14 EU entities, retaliating for the EU's 21st sanctions package — saturation event coverage everywhere, which per protocol does not count. The engine's specific composite (midstream-not-primes; Lafert/Sindlhauser magnet-motor nodes; III-V Lab+Vigo gallium/InP chokehold tied to the gallium regime; Rheinmetall symbolic; bite via indus
The NEW piece is the US side: consensus reads the DPA scrap determination as a US-recycler subsidy story, not a KOREA import-shock story. If implementing rules restrict black-mass exports, Korean cathode makers lose contracted discounted US feed at the same moment graphite permits squeeze the anode — a simultaneous two-input margin compression nobody models as one event.
Earliest coverage: (none consensus-grade; closest adjacent: Discovery Alert newsletter noting US shippers to "South Korean and Japanese processors" affected — single non-tier source, US-shipper framing, no cathode-margin or graphite coupling)
Grader confidence: 0.72 · engine conviction 0.50
The underlying event (Presidential Determination under DPA §101 on recoverable critical minerals, signed 2026-07-31, Federal Register 2026-08-04) was widely covered by the freeze date of 2026-08-05, but every tier-1 and government-primary source frames it exactly as the engine says consensus did: a keep-scrap-home / US-recycler / anti-China story. Reuters (2026-07-30, full text checked) names no f
The unpriced detail is the LEGAL BOUNDARY: 'ores' does not capture matte — the package as written spares Harjavalta — but the campaign infrastructure to close exactly that gap (Global Witness, NICKEL22 initiative) now has a statutory hook to extend to mattes in the 22nd package. Finnish imports of Russian matte start falling BEFORE any formal extension as banks/insurers de-risk. Finland's refining node — not Russia — carries the operational tail risk; BASF's EU precursor economics are the second
Earliest coverage: Nearest-miss (non-qualifying): Global Witness, "Sanctions gap lets Russian-mined nickel flow to Western markets" — NGO campaign piece documenting the CN 7501 matte gap feeding Harjavalta and naming BASF among buyers of the refined nickel; fails the tier test (not tier-1, not listed tier-2, single advocacy source). NICKEL22 (nikkeli.eu) is the campaign itself, likewise non-qualifying. (2026 (pre-freeze; exact date unverifiable — page returned HTTP 403, content recovered via search snippets))
Grader confidence: 0.68 · engine conviction 0.45
The engine's edge is a compound second-order claim: (i) the 21st package's "ores" line legally spares matte-fed Harjavalta, (ii) the NICKEL22/Global Witness campaign now has a statutory hook to extend to mattes in the 22nd package, (iii) Finnish matte imports fall BEFORE formal extension via bank/insurer de-risking, and (iv) BASF's EU precursor economics are the second-order casualty. The loophole
The beneficiary is the Indonesian STATE, not Indonesian-listed miners — that inversion is the edge. Quota discipline squeezes miners' volumes while DSI captures export margin, fees, and FX retention: OPEC-style price administration with rent accruing to the sovereign, visible in realized export unit values, not IDX mining equities. Prediction: post-Sept-1, Indonesia's FeNi/NPI realized export unit value (Comtrade, HS 7202.60) rises vs LME-implied content by ≥5% while listed Indonesian nickel min
Grader confidence: 0.60 · engine conviction 0.60
The underlying event (DSI single-gate export control over ferroalloys incl. HS 7202.60 FeNi, Sept-1 sole-exporter rollout, quota cut to 260-270Mt, 100%/12mo FX retention, PP 24/2026 "reasonable margin" clause) was saturated with pre-freeze coverage — but the protocol excludes event coverage. The engine's specific second-order claim is the distributional INVERSION: rent accrues to the sovereign via
The DLA cancellation is flipped into evidence of the entry's own tension (domestic-conversion preference vs price), to be re-tested on any re-issue. Remaining claim, with teeth: (a) ≥3 DISTINCT new non-Chinese magnet supply or qualification agreements disclosed by defense primes/tier-1s between 2026-08-05 and 2027-01-15 — the pre-cliff papering rush; fewer than 3 falsifies the timing claim; (b) Commerce implementing rule under the scrap determination published in the Federal Register with export
Earliest coverage: Reuters (syndicated), 'Trump orders restrictions on export of critical minerals scrap, White House officials say' — NEAREST MISS ONLY: underlying event, fails entity+mechanism test (2026-07-30)
Grader confidence: 0.70 · engine conviction 0.60
The underlying events were consensus at freeze: the 2027 NDAA magnet ban (Breaking Defense 2025-12, Federal News Network 2026-07) and the 2026-07-30 Presidential Determination empowering Commerce to license black-mass exports (White House primary doc; Reuters syndication; Forbes). The protocol excludes underlying-event coverage. The engine's specific second-order claims were not found in any tier-
The under-modelled input is antimony, and the reason is structural: Olin can pass through and hedge copper and lead against COMEX/LME, but antimony has no exchange contract and no hedge, so it is the one input with unhedged, uncapped cost — and the licensing suspension expires 2026-11-27, inside the grading window. Winchester's growth is in the military/small-caliber mix (Lake City plus the 2025 small-caliber and AMMO Inc. asset acquisitions) — fixed-price government work with weaker escalation
Earliest coverage: NEAR-MISS, NON-QUALIFYING: Apaton Finance GmbH sponsored commentary by Nico Popp, "Rüstungsboom auf tönernen Füßen: Profiteur Antimony Resources, Risiken bei RTX und Olin" (kapitalerhoehungen.de; English mirror news.financial; republished finanznachrichten.de) (2026-03-20)
Grader confidence: 0.58 · engine conviction 0.55
SEPARATING THE UNDERLYING EVENT FROM THE CLAIM. Three components of the edge are individually public and do NOT count: (i) China's suspension of the antimony export ban expiring 2026-11-27 with the military-end-user carve-out intact is covered at tier-1/tier-2 (Fastmarkets, mining-technology, CSIS, Pillsbury) — underlying event; (ii) antimony has no LME/COMEX contract and is priced off Fastmarkets
Magnets are not Korea's binding constraint; anode-grade graphite is, and unlike the magnet number (roughly flat for 15 years) it is actively DEEPENING. The lever is a permit regime, not a ban — so there is no shock date to trade and the impact surfaces as gross-margin and utilisation lines in Korean cell-maker quarterlies rather than as a headline. The squeeze is compounding: Korea must cut Chinese graphite content to keep US credits at exactly the moment its Chinese share is climbing. Anchor fo
Earliest coverage: NONE consensus-grade. Nearest misses (each fails entity+mechanism match or tier test): Seoul Economic Daily, "Korea's Graphite Reliance on China Deepens..." (2026-06-24) — deepening FACT only, no permit-regime/margin/magnet-comparison, and a general business daily not on either tier list; The Diplomat, "South Korea Has Diversified Some Critical Minerals. The Hardest Dependencies Remain." (2026-06-16) — names graphite among hard dependencies but emphasises INDIUM, does not compare with magnets, no permit regime, no cell-maker economics, and is not Tier-1 or enumerated Tier-2; Korea Herald, "Korean battery-makers under cost pressure as US targets Chinese graphite" — different lever (US 93.5% anti-dumping tariff, a shock-dated event), not the Chinese permit regime. (n/a — no consensus-grade coverage identified at or before 2026-07-24)
Grader confidence: 0.58 · engine conviction 0.66
I searched to REFUTE, targeting each load-bearing element separately and then the composite. Every element of the engine's claim exists somewhere in the public record in isolated form: the deepening share (Seoul Economic Daily 2026-06-24, which the engine itself concedes; The Diplomat 2026-06-16 with a longer 2015-2025 series), the export-permit/licence regime (CSIS "China's New Graphite Restricti
In stainless the counterparty risk runs the other way, and it is about to be measurable. Chinese mills (Tsingshan, Delong, Baosteel Desheng) are structurally short Indonesian NPI/FeNi and from 2026-09-01 face a SINGLE state seller with an explicit mandate to set price benchmarks and retain export proceeds — a monopsony buyer turned into a monopoly seller against them. So the first quantifiable damage of the 2026 tightening cycle lands on Chinese import COSTS in Sept-Nov 2026, not on Western magn
Grader confidence: 0.66 · engine conviction 0.55
Adversarial attempt to refute novelty. The UNDERLYING EVENT is thoroughly covered at/before freeze: DSI (PT Danantara Sumberdaya Indonesia) becomes sole exporter of record for coal, palm oil and ferroalloys incl. ferronickel from 1 Sept 2026 (announced 20 May 2026; PP No.24/2026; PMK/MoF Decree No. 32 and MoT Reg 17/2026), with phase-2 full state purchase-and-resale from Jan 2027. Trade press (SMM
The measurable harm is NOT in autos. Japan's auto chain has been engineering Dy out since 2010 (grain-boundary diffusion, Dy-lean NdFeB) and sits on JOGMEC stock; the entities Beijing actually named are defence research and precision dual-use units — National Institute for Defense Studies, Mitsubishi Electric and Mitsubishi Heavy units, Mitsui E&S, Mitsui Bussan Aerospace, Terra Drone, ACSL, Mitsubishi Nuclear Fuel. So Japan's 2026-27 damage shows up in the FILINGS OF THE NAMED ENTITIES, not in
Earliest coverage: NONE meets the consensus bar. Closest near-miss (single Tier-2, does not satisfy branch b alone): Benchmark Mineral Intelligence — Neha Mukherjee, Research Manager, 'What are the implications of China's latest rare earth export restrictions on Japan?' (~2026-01 (Benchmark near-miss; exact date unconfirmed — article body returned HTTP 403, date established only via syndicated quotation))
Grader confidence: 0.60 · engine conviction 0.50
CLAIM UNDER TEST (two halves): (A) the harm is located in the specifically named Japanese defence/dual-use entities (NIDS, Mitsubishi Electric/Heavy units, Mitsui E&S, Mitsui Bussan Aerospace, Terra Drone, ACSL, Mitsubishi Nuclear Fuel), measurable in their issuer disclosure; and (B) the harm is NOT in autos, because the Japanese auto chain engineered Dy out since 2010 (grain-boundary diffusion, D
The material Indonesian win in this cycle is COBALT, not nickel — and the reason is structural. Indonesian cobalt is a by-product of HPAL/MHP nickel output, so its supply curve is set by nickel capex already sunk and is near price-INELASTIC: it keeps arriving regardless of what the DRC does to defend price. That makes the DRC quota regime a mechanical transfer of both price AND share to Indonesia — the DRC surrenders volume to hold price, and Indonesia sells the surrendered volume INTO the price
Grader confidence: 0.62 · engine conviction 0.57
ADVERSARIAL TEST APPLIED — I searched to REFUTE, and the refutation attempt failed on the mechanism, not on the entity. WHAT THE PROTOCOL REQUIRES: entity match (Indonesia) AND mechanism match (price-INELASTIC by-product supply set by sunk nickel HPAL/MHP capex => DRC quota is a mechanical transfer of price AND share/RENT to Indonesia, with offsetting irrelevant). Entity match is ambient and easy
The restoration will be structurally partial: the own-smelter test caps WBN's supplementary grant far below the ~30 Mwmt gap, so Weda Bay contribution stays near zero into Q4 and Eramet is forced into a second FY26 guidance cut at the Q3 update — against a consensus that assumes a July fix.
Earliest coverage: none qualifying (nearest miss: SunSirs, single non-roster source, different mechanism, published on freeze date)
Grader confidence: 0.72 · engine conviction 0.55
The components of the edge were separately public but never assembled into the specific second-order claim by any consensus-grade source at/before freeze. (i) The generic ESDM rule — supplementary quota 'must be tied to the applicant's own smelting capacity within Indonesia' — appeared in a Mysteel FLASH (2026-06-30), a single non-roster trade source that never applies the rule to WBN, never sizes
The July decree closes the NPI classification loophole but phases in: the Q3-2026 fingerprint is surveyor/export-authorization friction — NIC's Sep-quarter realized NPI price discount to the SMM Indonesia NPI index widens and working capital stretches — while mandatory DSI channeling of NPI from Jan 1 2027 (absent a written exemption for NIC's RKEF lines) is confirmed a quarter before consensus models it. Kill early if NIC or its IMIP partners disclose a granted DSI exemption.
Earliest coverage: Nearest-miss only (event coverage, fails entity+mechanism test): SMM flash 'Indonesia government officially releases new export controls on FeNi and NPI' (2026-07-09)
Grader confidence: 0.70 · engine conviction 0.50
The engine's stated prior consensus is corroborated: Reuters via Airlangga (2026-05-22) reported NPI excluded from DSI centralization, and SMM's 2026-06-10 analysis called NPI an unresolved 'grey zone' with supplemental guidance unpublished. The underlying July event is public by freeze: SMM (2026-07-09) reported MoF Decree 32/MK/BC/2026 imposing controls on HS Ex.7202.60.00 'including certain NPI
Both USAR and MP were entity-listed June 22 — the pair asymmetry is that MP has DoD-anchored economics (price floor, offtake) and an operating separation/magnet chain, while USAR's entire value is a Phase 1a ramp still exposed to China-origin equipment spares, process validation, and thin ex-China Dy/Tb feedstock. Specific claim: 600 tpa exit-2026 run-rate slips or capex rises at the Q3/Q4 prints, and USAR underperforms MP over 120 days despite identical headline 'beneficiary' status.
Grader confidence: 0.70 · engine conviction 0.50
The engine's claim is a second-order PAIR-ASYMMETRY thesis: despite identical headline 'beneficiary' status after the June 22 entity listing, USAR's 600 tpa exit-2026 Phase 1a run-rate slips or capex rises at Q3/Q4 prints (mechanisms: China-origin equipment spares, process validation, thin ex-China Dy/Tb feedstock) and USAR underperforms MP over 120 days, MP being protected by DoD price-floor/offt
Direction, not volatility: as REE/gallium bargaining escalates through Q3, permit throughput for a US-owned exporter degrades again — GaAs permits already 'came in light' in Q1 per the company. Specific claim: Q2 (early-Aug print) or Q3 guidance disappoints on permit timing, against a consensus extrapolating the Q1 beat.
Earliest coverage: none found for the specific claim (closest near-miss: Investing.com 'AXT stock falls after lowering Q2 revenue guidance on export challenges' — but dated 2025-07-10 and about fiscal Q2 2025, a prior-year event within the acknowledged permit-volatility baseline, not the frozen Q2/Q3 2026 forward claim)
Grader confidence: 0.75 · engine conviction 0.45
Entity resolved to AXT Inc (AXTI)/Tongmei. The engine's edge is the DIRECTIONAL forward claim: Q2 2026 (July 30 print) or Q3 2026 guidance disappoints on permit timing as REE/gallium bargaining escalates, against consensus extrapolating the Q1 beat. Pre-freeze public record confirms the engine's stated consensus baseline (permit volatility known: Q4 2025 miss, Q1 2026 beat, CEO noting GaAs permits
The ramp math: June's ~$2.6M invoice month scaling to a 400-500 t/month smelter implies a Q3/Q4 revenue run-rate ABOVE the cadence implied by the $125M FY guide — specific claim: monthly DLA invoicing exceeds $6M by September and guidance is raised at/before the Q3 print. The binding constraint to watch is ore FEEDSTOCK, not demand — a feedstock shortfall is how this fails.
Earliest coverage: United States Antimony Corp press release / Form 8-K (underlying event only — first DLA shipments, ~$2.6M June invoices; NOT the second-order run-rate/guidance-raise claim) (2026-07-01)
Grader confidence: 0.72 · engine conviction 0.50
The underlying facts were fully public at freeze via issuer disclosures (July 1 PR/8-K: ~$2.6M June DLA invoices, $57.3M cumulative orders; 500+ t/mo Thompson Falls target; $125M FY26 guide), matching the engine's stated consensus baseline. However, no consensus-grade source made the engine's specific second-order claim: that the ramp math implies monthly DLA invoicing exceeding $6M by September a
A permit-regime squeeze inside the window forces cell makers to lock in qualified ex-China supply early. Specific claim: Panasonic formal validation lands by end-2026 (earlier than the skeptic case) AND/OR a second binding US cell-maker offtake is signed — this is a qualification-conversion call, not a graphite-price call. (Note: this entry received no skeptic verdict; retained on original terms, conviction unchanged, and it pairs with the adopted forge call on the 2026-11-27 graphite suspension
Earliest coverage: none found establishing consensus (closest non-qualifying challenger: Novonix's own 2026-06-11 GlobeNewswire/6-K release saying validation sits with Panasonic 'over the coming months' — a company PR, not a listed tier source, and no end-2026 commitment)
Grader confidence: 0.70 · engine conviction 0.45
The engine's edge is a qualification-conversion timing call: Panasonic formal validation by end-2026 (vs the skeptic anchor of the H2 2027 mass-production slip) and/or a NEW second binding US cell-maker offtake forced by the 2026-11-27 permit-suspension cliff. At freeze, the underlying events were saturated in coverage (C-sample delivery 2026-06-11; H2 2027 guidance reaffirmed; Stellantis terminat
A renewed closure lands in the Aug-Oct 2026 festival build while India still lacks any bilateral end-user channel. Tells, in order: DGCI&S HS 8505.11 imports from China roll over (-30% YoY), then VAHAN e-2W registrations go negative YoY while ICE 2W stay positive, then share rotation TOWARD OEMs with ferrite/light-RE motor programs (Bajaj's redesigned Chetak motors are the proven fallback). Concentrate the market expression in pure-play Ather (ATHER.NS) rather than diversified Bajaj/TVS, where E
Earliest coverage: None found for the specific claim. Closest non-matching coverage: Reuters-syndicated/Business Standard/Forbes India/Al Jazeera on the 2025 squeeze and Bajaj's light-RE Chetak redesign (underlying event, 2025 festival window); CSIS/IEA/trade press on the Nov 10, 2026 reimposition of China's 0.1% rule (different timing and mechanism, not India e-2W/Ather specific).
Grader confidence: 0.60 · engine conviction 0.55
The underlying events are fully consensus: the April-2025 China HRE curbs, the 2025 festival-season production cuts at Bajaj/Ather/TVS, Bajaj's light-rare-earth Chetak motor redesign as the proven workaround, the reopening of flows, mid-2026 friction (China demanding written no-US-diversion guarantees from India), and the scheduled Nov 10, 2026 snapback of the suspended 0.1%-rule controls. But the
The timing gap is the claim: a permit slowdown in H2 2026 lands exactly in the window where Korea's substitution capacity is announced but NOT commissioned (2027). Transmission is anode-maker margin compression plus FEOC-compliance premium at the cell makers (paying up for scarce non-China units), NOT cell-line halts — consensus frames halt risk, we claim margin-shape. Korea, not Japan, shows the first customs-visible import dip because Japan's anode mix is domestic-synthetic-heavy.
Earliest coverage: None found matching the specific claim. Closest adjacent coverage: Fastmarkets "Why underinvestment in ex-China anode supply chains could make graphite exception to new FEOC rules" (FEOC context only, no premium/margin transmission mechanism) (2024-04-22)
Grader confidence: 0.70 · engine conviction 0.50
The underlying event is heavily consensus (China permit regime, Nov-2026 suspension expiry, Korea ~90-93% China dependence, POSCO capacity announced for 2027-2028: Fastmarkets, CSIS, KED Global, HSF Kramer, The Diplomat June 2026). But the engine's SPECIFIC second-order claim is a composite with three differentiators, and none was found in Tier-1 or 2x-independent-Tier-2 coverage at/before freeze:
Stripped of the already-public stoppage fact, the residual claims: (1) guidance divergence — at least one German Tier-1 cuts guidance citing 'material availability' while no German OEM cuts production guidance in the same window, because OEMs triage scarce magnets to flagship EV lines; (2) the unlisted Mittelstand motor/sensor makers, with no political channel to Beijing, take the deepest hit (watch insolvency/short-time-work filings, not just listed names); (3) the reversal trigger is an EU-Chi
“""”
Grader confidence: 0.60 · engine conviction 0.50
The engine correctly conceded the consensus baseline (VDA halt warnings, CLEPA stoppages, rare-earth risk to German autos — all wire-covered since June 2025 and renewed in the 2026 episode where Berlin was excluded from first-wave licence batches). Auditing only the residual: (1) the guidance-divergence prediction with the flagship-EV triage mechanism appears nowhere; the closest published allocat
This is a volume-PLUS-premium story decoupled from LME nickel: with two rival buyers bidding for the same laterite, Philippine ore CIF-China realizations rise even in an LME surplus, and Nickel Asia (NIKL.PS) / Global Ferronickel (FNI.PS) capture margin regardless of the metal price. Consensus's bearish-LME frame causes it to miss an ore-price rally that never touches the exchange.
Earliest coverage: none consensus-grade; nearest miss = Bloomberg News (via Mining.com syndication), "Philippine nickel sales to Indonesia to jump after Jakarta curbs" — volume-only, no premium/margin/decoupling content (2026-02-13)
Grader confidence: 0.62 · engine conviction 0.65
The engine's composite second-order claim — Philippine ore CIF-China realizations rising even in an LME surplus because two rival buyers (China and Indonesia) bid for the same laterite, with NIKL.PS/FNI.PS capturing margin decoupled from the exchange price — was not found in any tier-1 source or in two independent enumerated tier-2 sources at/before the 2026-07-16 freeze. The pieces existed scatte
The Reserve has switched from policy to DEAL-FLOW, and the sequence is the tell: Canberra clears Chinese holders from the register FIRST (Northern Minerals, twice), THEN funds. Specific claim: a second project-level Reserve offtake/support instrument lands within ~6 months, most likely in heavy REE (Northern Minerals' Browns Range feed into Eneabba, or Iluka directly). The re-rating catalyst is the announcement, not REE prices — which are DOWN 26% over 90d in our vault, so consensus sees no reas
Earliest coverage: NONE qualifying. Closest non-qualifying precursor: Northern Minerals' own executive chair (Adam Handley) linking register clean-up to funding, carried only by non-tier outlets (Small Caps / TradingView syndication / Kalkine) and referring to EXIM-EFA debt, not a Reserve offtake instrument. (n/a (closest non-qualifying item is undated on page, post-2026-05-17 disposal order))
Grader confidence: 0.70 · engine conviction 0.60
The engine's edge is a forward, entity+mechanism-specific prediction: a SECOND project-level Critical Minerals Strategic Reserve offtake/support instrument within ~6 months of 2026-07-16, most likely heavy REE (Northern Minerals' Browns Range feeding Eneabba, or Iluka directly), with Canberra's register-clearing of Chinese holders as the sequencing tell. Pre-freeze public record establishes only t
The gradable claim consensus doesn't make: Argentina's INCREMENTAL export tonnage overtakes Chile's increment in the 2026 Comtrade vintage — i.e., Argentina becomes the marginal ex-China carbonate supplier of record — and Western strategic stockpiles (the DLA solicitation closing 7/17 and RESourceEU) start accepting Argentine-origin carbonate as the FEOC-safe unit, embedding a persistent origin premium for RIGI-protected supply over Chilean state-encumbered supply.
Earliest coverage: none establishing consensus; closest-call = Bloomberg (via Yahoo Finance), DLA $300M lithium tender article — underlying event only, no Argentina/FEOC/premium mention (2026-07-07)
Grader confidence: 0.72 · engine conviction 0.60
The underlying events are consensus (DLA tender covered by Bloomberg 2026-07-07; Argentina's 66% growth and Rincon/RIGI widely covered; JP Morgan forecast of Argentina overtaking Chile in TOTAL production by 2028). But the engine's specific compound second-order claim fails to appear anywhere at consensus grade before 2026-07-16: (a) no source frames the comparison as incremental 2026 export tonna
The award COMPOSITION will reveal domestic incapacity, and that's the tradable detail: US battery-grade carbonate capacity that can deliver front-loaded volume is essentially Albemarle Silver Peak (~5kt/yr) — so we claim the award (expected within ~90 days of bid close) is SPLIT across multiple suppliers and/or accepts allied-origin (Chile/Argentina-fed) material, naming at least one non-obvious mid-cap winner. Expression fix per skeptic: this is NOT an ALB-beta trade — Albemarle wins only in th
Earliest coverage: none found — closest coverage is announcement-level Bloomberg wire (via Mining Weekly, 2026-07-08) and SMM Analysis (2026-07-03, single source, 'expected to favor North American producers'), neither making the split-award/domestic-incapacity claim
Grader confidence: 0.70 · engine conviction 0.50
The underlying event (DLA's first-ever lithium NDS solicitation, posted 2026-07-02) was consensus by 2026-07-08 via Bloomberg/E&E/trade press, but the protocol excludes event coverage. The engine's specific second-order claim — that the award composition will reveal domestic incapacity (front-loadable US battery-grade carbonate ≈ Albemarle Silver Peak ~5kt/yr), forcing a split across multiple supp
The 06-30 rule change structurally biases the decision AGAINST WBN's application: quota must attach to the applicant's own smelting capacity, and WBN is predominantly an ore seller. We claim partial restoration at best (below ~25 Mwmt), a halt extending into Q4, and a second guidance cut - while consensus models full restoration because the smelters 'need the ore' (they can instead buy from smelter-owning miners, which is exactly what the rule engineers).
Earliest coverage: none found matching entity+mechanism (nearest miss: Mysteel FLASH 'ESDM supplements the RKAB nickel ore quota approval rules', 2026-06-30 — rule only, no WBN mention)
Grader confidence: 0.70 · engine conviction 0.60
The underlying events are fully consensus (quota cut to 12 Mwmt, May/June exhaustion and halt, Eramet's July revision request targeting ~42 Mwmt with July-September approval, IWIP ~30 Mwmt deficit) — covered by Mining Weekly, Bloomberg Technoz, Petromindo, Argus, S&P Global, SMM, Mysteel. The 06-30 rule change itself is also public: Mysteel's 06-30 FLASH describes the supplementary-quota criteria
The ramp arithmetic, corrected: sintered NdFeB machining/pressing yields mean alloy INPUT exceeds finished-magnet output (~700-900 t/yr of alloy at a 600 t/yr finished run-rate absent heavy swarf recycling). The non-China oxide-to-metal-to-alloy chain at that spec and volume does not exist in 2026, and the 06-22 listing legally forecloses the quiet China-origin backfill that pre-revenue magnet makers have used. We claim a disclosed ramp delay past the end-2026 600 t/yr target, or a feedstock-dri
Earliest coverage: None found for the specific claim. Closest miss: Mining.com (2026-06-22) noting the blacklist "tightens constraints on Chinese-origin processing equipment, magnet technology and robotics" with "practical impact on the two companies remains unclear" - equipment/tech receive-side, not the alloy-feedstock mass-balance/ramp-delay mechanism, and Mining.com is not on the frozen Tier-2 list.
Grader confidence: 0.70 · engine conviction 0.50
The underlying event (MOFCOM adding MP Materials and USA Rare Earth to the dual-use export-control list on 2026-06-22) was saturated Tier-1 news, but the frozen rule requires the SAME second-order claim: USAR-specific alloy INPUT arithmetic (~700-900 t/yr alloy for a 600 t/yr finished run-rate), non-existence of a non-China oxide-to-metal-to-alloy chain at spec/volume, the listing legally foreclos
The 06-29 forfeiture rule flips the option value: stockpiled DRC cobalt is not a warehouse asset awaiting quota relief, it is escalating expropriation surface - and ARECOMS's expansion to lithium/REE shows the regime is broadening, not sunsetting. Sharpened per skeptic: with cobalt prices RISING under the quota, an IFRS NRV write-down is the least likely form - we predict instead a disclosed cobalt production-plan cut at TFM/KFM circuits, a quantified trapped-inventory/working-capital disclosure
Grader confidence: 0.68 · engine conviction 0.50
The forfeiture order itself (Bloomberg/Reuters/Mining Weekly 06-29/06-30) and the 07-03 Reuters glitch story (CMOC could lose almost all Q2 quota; ~20kt/$1.1B industry-wide missed shipments) are underlying-event/first-order coverage with a different mechanism. Tier-source consensus at freeze in fact ran OPPOSITE to the engine's edge: CMOC maintained 100-120kt cobalt guidance, framed as a copper gr
The routing-around is legally foreclosed this time: the transshipment prohibition plus whistleblower bounty makes third-country laundering of China-origin gallium and Dy/Tb prosecutable at the Chinese exporter, and non-China capacity at spec cannot cover 2026 delivery schedules. We claim defense & space delivery slippage or cost provisions become visible in the H1 FY2026 (Oct) disclosure - materials-driven, named or thinly veiled - while the market models only demand, not input legality.
Earliest coverage: None found for the specific claim. Closest near-miss (event-level only, does not qualify): Bloomberg via The Japan Times, "China's mineral squeeze testing Japan's military buildup" (2026-06-24)
Grader confidence: 0.70 · engine conviction 0.45
The underlying events are fully consensus (Bloomberg 06-23, CNN 01-06, CNBC 06-29, Reuters 07-06; MOFCOM Announcement No. 26 whistleblower/transshipment coverage) but protocol excludes event coverage. Full-text verification of the three closest candidate sources shows none contain the second-order claim: (1) Bloomberg/Japan Times 06-24 piece has no defense-contractor delivery/financial-impact sent
The eligible-supplier set for FRONT-LOADED, spec-compliant, preference-compliant delivery is close to a set of one: developers cannot deliver in FY2026, Chinese-origin material is excluded, and Argentina lacks a US FTA. We claim ALB wins a meaningful share of the award (announcement plausible within ~60-90 days of the 07-17 bid date) and that this is the first purchase of a standing US floor-buyer regime for lithium - a repeatable revenue-and-sentiment channel the market is assigning to the wron
Earliest coverage: None found matching entity+mechanism. Closest non-matching: SMM Analysis (2026-07-03, non-listed trade press) — tender "expected to favor North American producers" (category, no entity); INN (2026-07-09, non-listed) quoting RK Equity on a generic strategic-reserve-steps-in-as-buyer concept (no ALB, no award-share claim).
Grader confidence: 0.70 · engine conviction 0.55
The headline event (DLA $300M/5yr battery-grade lithium carbonate solicitation, published 07-02, bids due 07-17) was fully consensus by 07-08 via Bloomberg and trade press — exactly as the engine's own consensus statement says. But the edge is the second-order claim: the eligible-supplier set for front-loaded, spec- and preference-compliant delivery collapses to roughly Albemarle, ALB wins a meani
Two claims consensus does not hold: (1) SEQUENCING - the engine's arrival-order sim puts antimony in the next tightening wave with measurable lead time, implying a second demand/price impulse into UAMY's exact ramp window (H2 2026), not after it; (2) the bottleneck is SMELTING, not mining - every Western antimony ore or concentrate strategy (including Perpetua's Stibnite, unfinanced smelter-wise) terminates at UAMY's furnaces, making it the toll-collector of the entire North American decoupling
Grader confidence: 0.70 · engine conviction 0.55
The underlying facts (UAMY = only permitted NA antimony smelters; DLA $245M IDIQ; $27M DPA award; Perpetua-USAC collaboration since 2021) are heavily covered, but the engine itself concedes those as priced consensus. The audited edge is two second-order claims. Claim 1 (arrival-order sequencing producing a SECOND demand/price impulse timed into UAMY's exact H2 2026 ramp window): no tier-1/tier-2 c
Timing collision: IF China turns the permit dial in H2 2026 (the sim's arrival-order claim, graphite radar already 'active'), qualification - not price - becomes the scarce asset, and Novonix is the only US name whose anode material is already inside automotive qualification pipelines. We claim the tell will be COMMERCIAL, not spot-price: an expanded/accelerated offtake, customer prepayment, or DOE loan closing on improved terms within 120 days of any China graphite permit event. The sim's ~6-st
Earliest coverage: none found matching entity AND mechanism (closest non-matching: Benchmark Source event coverage of China graphite controls and Novonix offtakes; Crux Investor 2025-11-25 covers the Nov-2026 permit cliff but never mentions Novonix)
Grader confidence: 0.70 · engine conviction 0.45
The engine correctly conceded the consensus baseline (offtakes, DOE support, slipping Panasonic timeline are all widely covered). Its edge is a conditional second-order structure: China permit re-tightening in H2 2026 (a real, covered policy cliff - the enhanced-controls suspension expires 2026-11-27) making QUALIFICATION rather than price the scarce asset, with Novonix as the only US name inside
Tightened per skeptic (uniqueness premise corrected): Europe has a set of TWO separators, but usable-form stockpiling and magnet offtake both favor the integrated chain over Solvay's oxide-only line. We claim a named commercial linkage - a RESourceEU/member-state purchase, tolling, or storage-processing agreement, or a Japanese magnet customer contracting Narva capacity - involving NEO specifically (not Solvay) is disclosed within two to three quarters, converting Neo from 'niche processor' to '
Grader confidence: 0.70 · engine conviction 0.40
The specific second-order claim is a NAMED commercial linkage — a RESourceEU/member-state purchase, tolling, or storage-processing agreement with Neo, or a Japanese magnet customer contracting Narva capacity — repricing Neo as the EU's designated separation-and-magnet asset. Targeted searches found no such disclosure or Tier-1/2x-Tier-2 assertion at/before the 2026-07-13 freeze. What exists is (i)
The non-obvious detail: contracted non-China volumes (Syrah, Westwater ~34 kt over 2027-31, Samsung SDI 10 kt/yr) cover under ~10% of Korean anode demand before 2027, so the binding constraint through 2027 is Chinese permit cadence. Specific claim: a Korea-directed graphite permit slowdown (approval lag >30 days or monthly import volume drop >20% YoY) appears in H2 2026 as the stockpile regime absorbs supply, forcing at least one Korean cell maker to flag anode sourcing in guidance.
Earliest coverage: None found matching the specific claim. Closest adjacent: Fastmarkets "China grants graphite export approval to South Korea but concerns remain" (underlying 2023 event, opposite direction) and Seoul Economic Daily 2026-06-24 dependence piece (verified via fetch to lack the permit-slowdown, volume-math, and guidance-flag claims).
Grader confidence: 0.70 · engine conviction 0.55
Nine targeted searches plus a direct fetch of the closest consensus candidate found no consensus-grade coverage of the engine's specific claim at/before the 2026-07-13 freeze. Existing coverage falls into four non-matching buckets: (a) Korea-China dependence stats (the engine's stated consensus baseline, confirmed to lack the specific claims); (b) the 2023-24 permit episode, where Tier-2 sources (
Tightened per skeptic: India is a loser before it is a winner - but not because it has zero licenses. The existing MOFCOM channel is a handful of foreign-owned Tier-1s (Continental India, Hitachi unit, Jay Ushin) under end-user certificates, structurally too narrow for OEM-scale e-2W volume, which leaned on third-country re-export routes that Announcement No. 26 (whistleblower bounties) specifically criminalizes. Specific claim: magnet-driven supply constraints hit e-2W makers (Bajaj, TVS, Ather
Grader confidence: 0.60 · engine conviction 0.50
The components of the engine's claim are all individually public at freeze: Announcement No. 26 whistleblower bounties incl. third-country routing (Bloomberg 2026-06-24, SCMP, Morgan Lewis 2026-07, geopolitechs); the narrow licensed channel of foreign-owned Tier-1s under end-user certificates - Continental India, Hitachi, Jay Ushin, DE Diamond (Business Standard 2025-10-30); and the 2025 productio
Inversion, tightened per skeptic to condition on the mechanism firing: IF a RESourceEU stockpile tender covering rare earths or gallium is published by 2026-12-31, THEN state purchases compete with industrial restocking against zero new supply before ~2029 (German Tier-1s, unlike Japanese peers with JOGMEC, carry no national buffer) and the EU-vs-China premium on NdPr oxide AND gallium widens vs the 2026-06-30 baseline within 2 quarters of tender publication. If no tender is published by 2026-12
Earliest coverage: Bloomberg, "EU-Backed Group Warns of Expensive Critical Metals Stockpiling" (nearest adjacent Tier-1 coverage; does NOT match the edge's mechanism) (2026-05-28)
“Although stockpiling and floor pricing are definitely not the final solution — it can become very expensive for taxpayers — it might be a transformative measure that you have to undertake”
Grader confidence: 0.60 · engine conviction 0.50
The underlying event (EU shortlisting rare earths/gallium/tungsten for the first RESourceEU joint stockpile, May 2026) was saturated Tier-1 news (Reuters, Bloomberg) — excluded per protocol. The engine's specific second-order claim is the INVERSION: a published stockpile tender makes the state a competing buyer against industrial restocking in zero-new-supply markets, WIDENING the EU-vs-China prem
The structure is a deliberate price-maker play, and it works for the COUNTRY even if it hurts individual producers: quota tied to own-smelting + a single export desk gives the state simultaneous volume and price levers over ~half of world-class nickel units. Specific claims: (1) NPI gets CONFIRMED inside DSI scope (the grey zone resolves toward more control, not less), and (2) Indonesian ferroalloy export unit values firm in 2026H2 vintages vs H1 as the desk asserts pricing - improving terms of
Earliest coverage: Closest near-miss (fails match): Reuters wire, "Indonesia to bring commodity exports under centralised control, president says" — event coverage with skeptical framing, not the engine's second-order claim (2026-05-20)
Grader confidence: 0.60 · engine conviction 0.55
The edge's two enumerated falsifiable claims were affirmatively contrarian to covered consensus at freeze, not merely uncovered: (1) NPI-in-DSI was universally described as an UNRESOLVED grey zone with contradictory ministerial signals (Airlangga publicly suggesting NPI exempted per detik.com; FINI formally requesting clarification per SMM; IDNFinancials asking if NPI is excluded) — no consensus-g
The materiality is the PRICE-SETTING mechanism, not the volume: a front-loaded, domestically-favoring federal award creates a PUBLIC reference price above China-linked spot for US-produced carbonate - a financeable benchmark for marginal North American projects (the same trick the DoD's MP price floor did for NdPr). Specific claims: award lands by end of FY2026 window (Sep-Oct 2026); winner-and-price disclosure shows a premium to spot; a NON-incumbent winner (Standard Lithium/Arkansas, Piedmont
Earliest coverage: Shanghai Metals Market (SMM) analysis — nearest partial mechanism match only, NOT consensus-grade (2026-07-03)
“Should the final contract prices exceed prevailing Asian market prices, the procurement could effectively create a policy premium for qualified suppliers.”
Grader confidence: 0.70 · engine conviction 0.55
The underlying event (DLA solicitation, ~16 kt battery-grade lithium carbonate, $1M-$300M over 5 years, bids due 07-17) was widely covered from 07-02/07-08, including Tier-1 Bloomberg (07-08), but consistently as a stockpile/China-derisking story — matching the engine's stated consensus read. The engine's specific second-order claim (award functions as a public reference price above China-linked s
Tightened per skeptic: the forfeited quota is a NEW sovereign asset, not just a supply cut - ARECOMS can allocate the reassigned strategic tonnage bilaterally, with the buyer expected from the US framework-deal or Gulf channel. The EU Russian-cobalt wind-down (Reg 2026/506, LME 07-25) is demoted from forcing event to background: Russian cobalt was already absent from EU warranting for over a year, so it adds sanction-hygiene tailwind, not a new demand shock. Specific claim: within ~2 quarters th
Earliest coverage: NONE establishing consensus. Nearest miss: Reuters column (Andy Home), "Congo pivots westward under cover of cobalt controls" — westward pivot via private deals (Virtus/Chemaf, EGC-Trafigura-EVelution, Lobito), not a state-directed bilateral sale of strategic-quota tonnage. (2026-06-24)
Grader confidence: 0.68 · engine conviction 0.50
The engine's edge has two load-bearing parts: (a) the forfeited quota is a NEW sovereign asset ARECOMS can allocate bilaterally, and (b) the prediction that within ~2 quarters the DRC announces a government-directed sale/allocation of strategic-quota cobalt to a non-China counterparty at quota-referenced prices. Consensus-grade coverage at freeze treats the June 30 forfeiture exactly as the engine
The supplementary window's own-smelter test means WBN gets no quota or a token one: expect a second Weda Bay guidance cut at Eramet's H1 results (late July) or after ESDM's August approvals, with full-year Weda Bay contribution near zero rather than the assumed H2 restart. Consensus is treating a legal-structure problem as a paperwork delay.
Earliest coverage: Mysteel FLASH: "ESDM supplements the RKAB nickel ore quota approval rules" (nearest-miss; establishes mechanism only, NOT the specific WBN claim) (2026-06-30)
Grader confidence: 0.68 · engine conviction 0.60
The underlying facts are fully consensus-grade at freeze: the 11 Feb quota cut to 12 Mwmt and the ~4 Jun production halt are covered by tier-1 sources (Reuters via Mining.com/TradingView, Mining Weekly 2026-06-04) and multiple tier-2 outlets (SMM, Mysteel, Petromindo). The supplementary-window own-smelter RULE itself became public via Mysteel FLASH on 30 Jun 2026: "The additional quota must be tie
Anchored on 45X only (the 30D credit died Sept 2025): the bite is 45X material-assistance compliance on Panasonic's US-made cells plus 25%-tariffed feedstock. Claim: Panasonic's energy segment flags anode/materials cost or FEOC-compliance drag in Oct 2026 or Jan 2027 guidance BEFORE Korean peers (LGES/Samsung SDI, who diversified anode sourcing earlier), because Panasonic's sole qualified non-FEOC route (Novonix) cannot ship at scale until 2H2027 — a 12+ month uncovered window no US cell peer sh
Earliest coverage: (no source for the specific claim) — closest underlying-event coverage: The Globe and Mail / TipRanks, "NOVONIX Delays Panasonic Anode Mass Production to 2027" (underlying event only, not the comparative claim) (2026-06-11)
Grader confidence: 0.60 · engine conviction 0.55
The engine itself concedes the component facts are consensus (China graphite dominance, EV softness, Kansas repurposing). The claim to check is the differentiated synthesis: a comparative, timing-specific prediction that Panasonic discloses anode/FEOC-compliance drag in Oct 2026/Jan 2027 guidance BEFORE Korean peers because Novonix cannot scale until 2H2027, opening a 12+ month uncovered window no
Staked on the H1 2026 accounts due by 2026-08-31: cobalt-attributable inventory rises by a quantifiable multiple of H1 quota shipments, with a working-capital or NRV/impairment charge disclosed — and any post-forfeiture reallocation goes to the ARECOMS state reserve rather than back to CMOC, capping CMOC's cobalt realization even if Fastmarkets cobalt prices keep rising. Screens modeling cobalt upside at spot x production are overstating CMOC revenue by the quota gap.
Earliest coverage: thedailybreakdown.com (newsletter) — nearest miss; documents stranded inventory but NOT the accounting/impairment/revenue-overstatement claim. No consensus-grade source found for the specific claim. (2026-06 (approx early June 2026))
Grader confidence: 0.58 · engine conviction 0.50
Targeted searches confirm the REGIME/EVENT layer is fully consensus-grade at freeze: forfeiture to ARECOMS strategic reserve with no carryforward (SMM/metal.com, InvestingNews, discoveryalert), CMOC's ~4:1 production-to-export ratio and ~78,800t stranded (thedailybreakdown newsletter), and 'cobalt upside politically constrained regardless of spot' (ad-hoc-news aggregator). The engine explicitly co
The transfer-prohibition clause reaches the magnet sub-suppliers, not just direct purchases — the choke is invisible at the prime level until deliveries slip. Claim: materials-driven schedule slippage or cost provisions appear in the Defense & Space segment commentary within two quarters (Oct 2026 or Jan 2027 results), during Japan's defense-backlog boom when the segment is supposed to be the growth story.
Grader confidence: 0.66 · engine conviction 0.45
The entity-listing event and the generic 'China controls reach sub-tier magnet suppliers of Japanese defense firms' narrative are both well-covered by Tier-1/Tier-2 sources at freeze, but neither constitutes the engine's specific second-order claim. The differentiated element is the falsifiable, entity-specific financial-statement forecast: that Mitsubishi Electric's Defense & Space SEGMENT commen
The scale-up claim runs through feedstock, not furnaces: with expanded Thompson Falls capacity commissioned, quarterly DLA invoicing steps up several-fold from the ~$2.6M June baseline ONLY if Bolivian/third-party antimony-flake deliveries normalize (UAMY's June 8-K disclosed Q2 diesel-shortage feedstock delays). Claim: H2 2026 DLA order/delivery cadence (published in 8-Ks) confirms the ramp — cumulative orders crossing $75M — before it appears in consensus revenue estimates, with feedstock rece
Earliest coverage: United States Antimony July 1, 2026 press release / 8-K (company self-disclosure, not independent consensus); nearest independent tier-2: Simply Wall St, July 3, 2026 (furnace-centric framing, does NOT match the feedstock-gated leading-indicator claim) (2026-07-01)
Grader confidence: 0.60 · engine conviction 0.55
The engine concedes the consensus layer: UAMY as a known antimony pure-play that already re-rated on the $245M DLA contract-ceiling headline, priced skeptically due to tiny historic throughput. That IS broadly covered (MINING.COM, investing.com, multiple trade outlets). The edge to check is the second-order analytical/predictive claim: the ramp is gated by FEEDSTOCK (Bolivian/third-party flake) no
The claim is about WINNING the first EU tender/offtake slot, not volumes: when RESourceEU issues its first REE/gallium purchase tenders, Silmet is effectively the sole EU-origin separated-REE vendor that qualifies — presence on the first award or an EU-government-linked purchase by 2027-04-30 is the repricing event; volumes stay immaterial to revenue in year one, and first Narva EU-OEM qualification is the confirming leg.
Grader confidence: 0.68 · engine conviction 0.55
The specific second-order claim is not merely 'Neo is Europe's only separated-REE producer' (that IS consensus: tier-1 BNN Bloomberg CEO video, CNBC, plus tier-2 Fastmarkets/Benchmark) — it is that Silmet WINS the first RESourceEU tender/offtake slot as the sole qualifying EU-origin vendor, and that being on the first award (or an EU-gov-linked purchase) by 2027-04-30 is the stock repricing cataly
The claim: within two quarters (by 2027-01-31) at least one NEW binding defense-linked Dy/Tb offtake or government-anchored contract naming Lynas — announced after 2026-07-11 by a Japanese magnet maker (Shin-Etsu/TDK/Proterial), JOGMEC/Japan MoD, or an EU sovereign channel, excluding the pre-existing March 2026 JARE agreement — as a direct consequence of the 6/29 sub-supplier transfer ban, on a timeline consensus has not committed to, with quarterly Dy+Tb oxide output above prior quarter as the
Earliest coverage: CSIS, "China's Rare Earth Campaign Against Japan" (and S&P Global, "Rare earth supply bottlenecks set to persist in 2026") — cover the underlying ban and Lynas-as-benchmark, but NOT the specific new-contract claim (2026-01-27)
Grader confidence: 0.70 · engine conviction 0.60
The engine's stated consensus is accurate and heavily corroborated (Reuters, S&P Global, CSIS, Benchmark, Metal Tech News): Lynas is the canonical non-China / only-HREE-producer-outside-China play, and the March 2026 JARE 12-year deal (floor price + 50% of HRE output to Japan) is fully consensus; Sojitz already imports Dy/Tb from Lynas. The edge, however, is a forward, entity+mechanism-specific cl
The binary is closer than the 2027 production date suggests: Panasonic's validation decision lands within ~2 quarters, and a pass converts conditional offtakes into binding demand and unlocks strategic/government financing at exactly the moment the FEOC clock forces cell-makers to book non-China anode supply. The qualification EVENT, not production, is the repricing catalyst.
Earliest coverage: None consensus-grade for the specific claim. Closest partial coverage: NOVONIX C-sample press release (GlobeNewswire, 2026-06-11) and Fastmarkets FEOC-graphite analysis — each covers only an underlying ingredient, not the synthesized qualification-as-repricing-catalyst thesis. (2026-06-11)
Grader confidence: 0.70 · engine conviction 0.50
See above.
The origin clause is the explicit hinge: IF the solicitation's NDS domestic-source preference operates as a bar (checkable in sam.gov documents before 7/17), ALB/Silver Peak is close to the only compliant existing producer for the front-loaded FY2026 deliveries, and an award naming ALB at a premium to Fastmarkets CIF-NA spot is the tradeable signal (relative re-rate vs SQM+PLS basket over 30 trading days). IF allied-origin material is acceptable, expect a trader/South-American-supplied award and
Grader confidence: 0.71 · engine conviction 0.50
Searched the solicitation against ALB/Silver Peak, 'compliant producer', 'domestic vs allied source', 'likely winner', 'award/trade/re-rate', and fetched the two nearest candidate pieces (Mining Weekly = Tier-1; TheDeepDive = newsletter). The underlying $300M DLA/NDS solicitation is heavily Tier-1 covered, and 'Silver Peak = only US producer' plus 'NDS subject to domestic/allied-source preferences
A COST event, not a shortage event — graded on quantification, not offtake counts: by 2027-01-09 at least one of LGES/SK On/Samsung SDI explicitly QUANTIFIES a FEOC-graphite compliance cost in guidance or earnings materials (a won/kWh, margin-bp, or absolute-cost figure — not a qualitative mention), and/or discloses above-market take-or-pay terms (premium vs Chinese anode reference price) in a new non-China contract signed after 2026-07-11 — while Korean graphite import volumes from China look n
Earliest coverage: The Korea Herald — "Korean battery-makers under cost pressure as US targets Chinese graphite" (closest coverage found; establishes the general cost-pressure narrative but explicitly does NOT quantify any company FEOC-graphite compliance cost or disclose premium take-or-pay terms) (2025-07-18)
Grader confidence: 0.77 · engine conviction 0.60
The frozen protocol requires matching BOTH entity AND mechanism, not the underlying event. The underlying event (Chinese-graphite tariff/FEOC cost pressure) and base-rate behavior (non-China anode offtake announcements) are indeed consensus-grade and well-covered — but the engine explicitly excluded those and graded its edge on QUANTIFICATION: a specific compliance-cost number in a Korean cell mak
Contrarian detail: the listing bites the rebuilders' EXPANSION TIMELINES, not their current output. Graded primarily on a disclosed equipment/chemistry-sourcing constraint or a >=1-quarter commissioning-milestone slip for magnet/heavy-RE capacity in MP or USA Rare Earth Q2-Q4 2026 filings/calls; relative leg: listed US rebuilders lag non-listed Lynas (LYC.AX, ADR-adjusted) by >=10% over 6 months. Country-level read: US magnet independence arrives later than the narrative says, keeping the DoD pr
Grader confidence: 0.70 · engine conviction 0.45
The freeze-date consensus was unambiguously the symbolic/bullish read the engine described: tier-1 outlets (Bloomberg 2026-06-22 'China Places Two US Rare Earths Producers on Export Control List'; WaPo 2026-06-22; Al Jazeera 2026-06-22) framed the listing as 'mostly symbolic, as both companies say they have largely cut off supplies of equipment and materials from China,' and MP held a Buy consensu
The reserve is a serial FINANCING MACHINE, not a one-off. Specific claim: a SECOND project-level Reserve commitment (most likely heavy-RE — Northern Minerals' Browns Range or an Iluka Eneabba-linked deal — or an Arafura follow-on) lands by 2026-12-31, as a NEW Strategic Reserve instrument (letter of support, offtake, or floor contract) announced after 2026-07-11 — pre-existing Critical Minerals Facility loans (e.g., Iluka Eneabba's) or restatements of them do NOT count — with Western NdPr floor-
Grader confidence: 0.72 · engine conviction 0.65
The engine's stated consensus (Australia = critical-minerals winner with government support; reserve is 'repeatable financing architecture') IS broadly acknowledged in trade press and think-tanks at freeze. But the sharp edge is narrower and forward-looking: a NEW Strategic Reserve instrument for a SECOND specific project (Browns Range / Iluka Eneabba deal / Arafura follow-on), announced after 202
Sinomine disclosed plans for a Bikita lithium sulphate plant in 2025; the claim is that it is NOT commissioned by 2027-01-01, that Zimbabwe monthly concentrate exports (MMCZ/ZimStat) for Oct-Dec 2026 run >=30% above the Q3 2026 monthly average (front-loading), followed by a Q1 2027 print >=50% below the Q4 average. Secondary: Huayou (603799.SS) announces third-party tolling at Arcadia or a government relaxation of the third-party rule by 2027-03-31.
Earliest coverage: Reuters, "Zimbabwe's sole lithium salt plant unable to process third-party metal, official says" (syndicated via mining.com, TradingView, China Global South) (2026-07-18)
“Sinomine Resource Group's Bikita Minerals and Kamativi Mining Company, the Zimbabwean subsidiary of China's Yahua Group, are in the process of building plants which are unlikely to be ready by January 2027.”
Grader confidence: 0.85 · engine conviction 0.45
The edge's primary leg — same entity (Sinomine/Bikita) and same mechanism (sulphate plant NOT commissioned by 2027-01-01, i.e. behind the concentrate ban) — was stated flatly by Reuters on 2026-07-17/18, a Tier-1 source, well before the 2026-08-24 freeze. Corroborating Tier-1 coverage the day after freeze (Reuters 2026-08-25, syndicated by Mining Weekly/CNBC Africa/Kitco: "Sinomine says it is buil
The DPAS order is a FEED subsidy in kind captured first by the operator with a running hydromet line, not by the press-favored names. Make the price observable the claim: by 2027-01-31, US HS 8549 exports for Sept-Dec 2026 fall >=40% vs the H1 2026 monthly average AND a domestic black-mass price (Fastmarkets/Benchmark US NCM black-mass payables) trades at a widening discount (>=10 pts of payables) to Asian import parity. A Cirba or Ascend feed contract with a named US cell plant or DoD by 2027-0
Earliest coverage: Federal Register / BIS temporary final rule "DPAS Directive Allocation Order and Additional Requirements for Recoverable Critical Minerals and Materials" (tier-1 primary regulator document); earliest tier-2 mechanism headline: Benchmark Mineral Intelligence, "US black mass export restrictions could bolster domestic refiners" (2026-08-06 (Federal Register); 2026-08-07 (Benchmark))
“"US black mass export restrictions could bolster domestic refiners" (Benchmark Mineral Intelligence headline, 2026-08-07). Rule purpose as summarized verbatim by Bergeson & Campbell from the FR text: "The action is intended to reduce foreign dependencies on critical minerals by incentivizing recovery from finished goods and scrap and providing feedstock for domestic processors." (Federal Register direct fetch was blocked by an unblock redirect; quote taken from the law-firm summary of the rule.)”
Grader confidence: 0.60 · engine conviction 0.55
The underlying event (BIS 100% domestic-sales requirement for black mass, Schedule B 8549.13/.14/.19, effective 2026-08-27) was covered by Bloomberg (08-05), Reuters (08-06), Argus (08-04) and the Federal Register (08-06) — event coverage alone does not count. The specific second-order mechanism — the order functions as an in-kind feed subsidy accruing to whoever operates domestic hydromet refinin
The non-obvious hit is the black-mass rule, not graphite, and the casualty is recycler utilization, not LGES margins. Tightened bar: US Census exports of HS 8549.11/8549.19 to Korea for Oct-2026..Jan-2027 fall >70% yoy in value (near-total shutoff, not a trim) AND at least one of SungEel HiTech (365340), Ecopro or POSCO HY Clean Metal discloses a feedstock-driven utilization cut or non-US feed replacement contract by 2027-02-28.
Earliest coverage: Bloomberg News, "US Bans Export of Tungsten, Battery Waste to Secure Supplies" (Martin Ritchie, Annie Lee), syndicated on Mining.com; quoting Lee Allen, senior strategic markets analyst, Fastmarkets (2026-08-05)
“"US export restrictions could significantly tighten black mass availability for Asian recyclers," particularly for importers of nickel, cobalt and manganese materials in South Korea and Southeast Asia.”
Grader confidence: 0.80 · engine conviction 0.50
The engine's edge has two parts: (1) the second-order hit is the BIS black-mass allocation rule (not graphite) and (2) the casualty is Korean recyclers' feedstock/utilization (not LGES margins). Bloomberg's 2026-08-05 story on the rule (Tier-1) already carried, via Fastmarkets' analyst, the exact entity+mechanism: the US black-mass rule tightens feedstock availability for South Korean recyclers/im
The first-order winner is the US RECYCLER, not the miner: a captive, price-disadvantaged seller base means black-mass payables fall and US recyclers (Redwood, Ascend, Cirba, Li-Cycle Rochester) get feed below world price. Claim: US NMC black-mass payables (Fastmarkets/Benchmark) for Nov-2026..Jan-2027 average >=5pp below the 2026-H1 mean while US Census HS 8549.1x exports (all destinations) fall >70% yoy. Miners' equity is priced; recycler feed spreads are not.
Earliest coverage: Bloomberg, "US Bans Export of Tungsten, Battery Waste to Secure Supplies" (syndicated at MINING.COM and Bloomberg Law) (2026-08-05)
“Asian buyers — particularly South Korean and Southeast Asian importers of nickel, cobalt, and manganese materials — were consistently willing to pay more for US feedstock than domestic processors could offer. [NOTE: surfaced by a search-engine summary of the Bloomberg/MINING.COM syndication alongside the sentence "Massachusetts-based startup Ascend Elements Inc. filed for bankruptcy in April, while Li-Cycle Holding Corp. sought bankruptcy protection last year, with key creditor Glencore Plc taking over some of its assets."; the paywalled full text (bloomberg.com, mining.com, bloomberglaw.com all 403/paywalled) could NOT be fetched to verify verbatim. Verified Tier-2 corroboration: Benchmark, 2026-08-07, "A ban on the export of black mass should redirect feedstock towards US recyclers" (headline: "US black mass export restrictions could bolster domestic refiners").]”
Grader confidence: 0.55 · engine conviction 0.55
Underlying event: BIS temporary final rule (published 2026-08-06, Federal Register; Presidential DPA determination 2026-07-30) requiring 100% domestic allocation of black-mass (Schedule B 8549.13/14/19) sales from 2026-08-27 — heavily covered by Bloomberg, Argus, Benchmark, law firms, trade press, all before the 2026-08-24 freeze. The engine's SPECIFIC second-order claim is: US recyclers (Redwood/
Transmission is ENTITY-specific, not country-wide. Country-level customs will NOT show a Japan-wide magnet shortage (Lynas/JOGMEC buffers hold the civilian chain), but named defense-adjacent small caps — ACSL and Terra Drone, with no alternative Dy/Tb-magnet channel — take program slippage first. The pain shows in specific filings and defense-program schedules, not aggregate trade flows.
Earliest coverage: China Ministry of Commerce (MOFCOM) Announcements No. 27/No. 28 adding ACSL and Terra Drone to the dual-use export watch list (reported same day by CNBC and Geopolitechs) (2026-06-29)
“the drone sector (e.g., ACSL and Terra Drone) are likely to face persistent uncertainty in contract execution, deliveries, and overseas maintenance and support”
Grader confidence: 0.65 · engine conviction 0.60
The engine's stated consensus baseline (country-level 2010-era dependence story) was stale by freeze. On 2026-06-29 — five weeks before the 2026-08-05 freeze — MOFCOM Announcements No. 27/28 (a primary government document, tier-1 branch a) named ACSL and Terra Drone specifically on China's dual-use export watch list and stated exports for Japanese military users/applications or defense-strengtheni
The generalization is the edge: the CEEC/ARECOMS apparatus is being converted from a cobalt tool into a template for ALL strategic minerals, lithium first — Manono's certified first cargo makes DRC a lithium exporter under state-gate rules from day one, unlike every retrofit regime. Prediction: a formal DRC lithium export-quota or levy framework within two quarters, and the reclaimed strategic quota becomes a state revenue line via premium resale — quota administration as a profit center.
Earliest coverage: Bloomberg — "Congo Triples Royalty Rate on Lithium With New Strategic Minerals" (reinforced by the DRC decree itself, a Tier-1 primary government document, and June 1 trade coverage explicitly framing the generalization) (2026-05-31)
“Democratic Republic of Congo's lithium miners could soon be required to triple royalty payments after the government approved the inclusion of several metals to its list of strategic minerals.”
Grader confidence: 0.72 · engine conviction 0.55
The engine's edge is the generalization: CEEC/ARECOMS converted from a cobalt tool into a template for ALL strategic minerals, lithium first, with Manono state-gated from day one and a lithium quota/levy framework predicted within two quarters. But that generalization was formally ENACTED before freeze: a late-May 2026 government decree (Tier-1 primary document) added lithium plus tantalum, niobiu
The gradeable detail is the commodity list, not the policy. The industry is still waiting for the official list; the reported timetable is that NPI enters 'gradually' (Bisnis: bertahap) with a hard backstop of 2026-12-31 and full enforcement from January 2027. So the market is pricing a permanent exclusion against a regulation whose own timetable implies inclusion within five months. And the cost is not a tariff — it is a state intermediary inserted between producer and offtaker, taking margin,
“"The regulation imposes stricter export controls covering ferronickel (FeNi) ingots and lumps (≥8% Ni), Sponge FeNi and Nugget FeNi (≥4% Ni), and low-grade FeNi products (2% ≤ Ni < 4%, Fe ≥75%), including certain Nickel Pig Iron (NPI) products." — plus, from the same 2026-07-09 release: "From 1 January 2027, exports will generally be limited to state-owned export enterprises" (BUMN Ekspor). Corroborating mechanism quotes from SMM's 2026-06-10 analysis of PP No.24/2026: Art. 3(4) grants DSI authority to "charge a margin at a reasonable level in accordance with prevailing regulations," and Art. 3(2) states the selling price of strategic commodities is "determined by the BUMN Ekspor." Corroborating headline, Bisnis 2026-05-22: "Ekspor Nickel Pig Iron Bertahap Masuk Skema Danantara Sumberdaya Indonesia (DSI)."”
Grader confidence: 0.82 · engine conviction 0.55
The edge decomposes into three components; two of them — the gradeable ones — were already public at consensus grade before the 2026-07-24 freeze, and only the third (a measurement choice) appears uncovered. (1) "The market prices permanent NPI exclusion against a regulation whose own timetable implies inclusion." This is NOT differentiated. Bisnis published it as a headline on 2026-05-22 — "Eksp
The harm is inbound feedstock, not outbound sales, and it appears as an inventory build and gross-margin compression 2-4 quarters out rather than a revenue cliff. Stating the counter-source explicitly, per skeptic: non-China 6N tellurium and CdZnTe DO exist (5N Plus in Canada, Coherent/II-VI), so the claim is NOT that substitutes are absent. It is that Vigo's MBE epitaxy recipes are qualified to a named substrate vendor and lot, and requalifying a detector already designed into a missile-seeker
Earliest coverage: VIGO Photonics S.A. — raport bieżący (current report) Nr 35/2026 to KNF/ESPI, "Analiza skutków wpisania VIGO Photonics na listę podmiotów objętych zakazem eksportu towarów typu dual-use z Chin" (carried verbatim by Bankier.pl) (2026-07-24 (14:03 CET, same day as freeze))
“"Jedynym produktem typu dual use pozyskiwanym z Chin w ramach tego segmentu były podłoża do produkcji warstw półprzewodnikowych" — and, on the alternative supplier: "ale z uwagi na jego ograniczone moce produkcyjne zapotrzebowanie Emitenta może nie zostać w pełni zaspokojone"; with "ok. połowa przychodów tego segmentu generowana była w oparciu o podłoża dostarczane dotychczas z rynku chińskiego". (Translation: the only dual-use product sourced from China in this segment was substrates for semiconductor-layer production; the European alternative's limited capacity may not fully satisfy the Issuer's demand; roughly half of the segment's revenue was generated on substrates previously supplied from the Chinese market.)”
Grader confidence: 0.90 · engine conviction 0.60
The engine's edge is that the harm from the 2026-07-24 MOFCOM listing runs through INBOUND Chinese semiconductor feedstock/substrate supply rather than outbound sales, and that management's "no critical impact" line addresses the order book only because "no one has asked them about wafer origin." That framing is refuted on the freeze date itself. At 14:03 CET on 2026-07-24 — the freeze date — Vigo
Deferral does not rescue Sinomine, because BOTH compliance paths are closed on the same date and neither is political. The only working in-country conversion asset belongs to a direct competitor (Huayou) and is technically unable to accept third-party feed; Bikita's own sulphate plant is confirmed incomplete and not expected operational before the deadline. So the outcomes are (a) the ban holds and Bikita's concentrate is stranded, or (b) Sinomine funds and commissions its own plant, which canno
Earliest coverage: Reuters (wire ID newsml_L8N43J0NP), syndicated via MINING.COM, Zimbabwe Situation, NewZimbabwe, ZimLive, Miningmx, China-Global South Project (2026-07-17)
“"We don't have the capacity to process other minerals from outside. Our concentrator plant produces around 400,000 tons per annum, so we don't have space for other players." [mine manager Mthokozisi Goliath, Prospect Lithium Zimbabwe] ... "Sinomine Resource Group's Bikita Minerals and Kamativi Mining Company...are in the process of building plants which are unlikely to be ready by January 2027." ... "The January 2027 deadline is still on. We cannot talk of extending right now." [Mines Minister Polite Kambamura]”
Grader confidence: 0.93 · engine conviction 0.50
The engine claimed its edge was that consensus was pricing political softening while the true binding constraint was physical: (i) Huayou's plant technically cannot accept third-party feed, (ii) Bikita's own sulphate plant is incomplete and cannot be operational by the deadline, so (iii) deferral does not rescue Sinomine because both compliance paths close on the same date. Every load-bearing leg
The claim is QUALIFICATION POSITION, not tonnage — and the skeptic's correction is carried: Silmet's Dy/Tb line is small-scale and running at a modest nameplate, so it establishes that Europe CAN separate heavies, not that it can supply them at volume. What is unpriced is the dated coincidence: Narva is the only sintered NdFeB plant inside the EU, it can point to non-Chinese in-house heavy feed, and China blacklisted a set of EU magnet consumers this week — so a listed EU buyer needing a complia
Earliest coverage: Reuters (INSIGHT: "Rare earth magnet users jolted into paying premium prices for ex-China supply"), syndicated via MINING.COM (2025-07-01)
“"Suleman's company, Neo Performance Materials, launched output of permanent magnets at its Estonia plant in May."”
Grader confidence: 0.82 · engine conviction 0.60
ADVERSARIAL FINDING: the engine's stated consensus baseline is a strawman, and the claim's mechanism was already consensus-grade well before freeze. 1. The asserted baseline ("legacy Magnequench bonded-powder business... long-promised, slow-ramping Estonian project... Europe's answer assumed years away") is contradicted by the public record at freeze: Narva began magnet output May 2025, was forma
The market is pricing an accounting scandal CODI itself described as 'isolated to Lugano', and ignoring that the same entity owns the West's scarcest defence-magnet asset heading into a statutory tightening. SmCo is the harder half: unlike NdFeB there is essentially no US greenfield SmCo build, and Arnold has already locked samarium oxide from Solvay's La Rochelle line via LCM metallisation — the verified chain from French separation chemistry to a US magnet maker sits inside this holdco. Qualit
Earliest coverage: Fastmarkets — "Arnold Magnetic enters samarium oxide supply agreement with Solvay, Less Common Metals" (2025-12-10)
“Samarium is a heavy rare earth element used in samarium-cobalt (SmCo) permanent magnets, which are primarily used for defense applications. [Entity link established in the same article's opening paragraph: "New York-based Arnold Magnetic Technologies, a subsidiary of US conglomerate Compass Diversified..."]”
Grader confidence: 0.88 · engine conviction 0.60
The edge's specific second-order claim is: CODI (entity) is mispriced as a distressed holdco because the market ignores that it owns Arnold, the West's scarcest defence-magnet (SmCo) asset, with a verified non-China chain Solvay La Rochelle -> LCM metallisation -> Arnold, into a statutory tightening. Both halves were already public and matched at/before freeze. MECHANISM + ENTITY, Tier-2 x2 (diff
The ban is on ORES; Harjavalta's feed is converter MATTE under CN 7501. The 20th package (April 2026) already banned CN 7503-7505 while deliberately leaving CN 7501 matte and CN 7502 unwrought untouched — 2026-07-23 makes this the THIRD consecutive package to spare exactly the tariff line that keeps Harjavalta alive, and the same annex bans Russian COPPER all the way up to matte (CN 7401), which sharpens the asymmetry into a deliberate carve-out rather than an oversight. Two consequences nobody
Earliest coverage: S&P Global (Platts) — "20th sanctions package to close EU market to more Russian metals," by Katya Bouckley (republished by EUROMETAL); earlier corroborating tier-2: Global Witness report, September 2025 (2026-02-10)
“Nickel matte accounted for $653.4 million of the total, with the entire amount imported by one EU country: Finland. It is home to Harjavalta refinery, which receives most of its feedstock — nickel matte – from its parent company, Russian nickel, copper and platinum group metals producer Nornickel.”
Grader confidence: 0.80 · engine conviction 0.72
The engine's second-order claim is: the EU nickel ban does not touch the tariff line that feeds Nornickel's Harjavalta refinery (nickel matte), so Finland's Russian-nickel exposure is a live, unrealised flow rather than a closed one. I verified the underlying fact pattern independently — the 21st package (2026-07-23) restricts "nickel ores" on the import side (sofiaglobe factbox of the Council rel
Rheinmetall is the decoy — its CEO has publicly put Chinese content at about 1% of the company, so that listing is symbolic and will not move a P&L. The entity that actually loses production inputs is a loss-making photonics SME. The generalisable claim is the SELECTION RULE: Beijing's list is calibrated at chokepoint-dependent SMEs, not at primes, because primes hold buffers and SMEs do not — so the first MEASURABLE casualty of 2026-07-24 is a Polish ESPI current report, not a German guidance c
Earliest coverage: Reuters (wire item, distributed via Refinitiv/TradingView and syndicated same day; corroborated by VIGO Photonics ESPI current report carried by Bankier.pl 2026-07-24 14:26) (2026-07-24)
“"Vigo Photonics Says Being Placed On China's Dual-Use Goods Export Ban List Will Not Have Critical Impact On Operations" (Reuters, 2026-07-24); corroborating primary disclosure, VIGO Photonics ESPI current report, 2026-07-24 14:26: "w ocenie zarzadu decyzja Ministerstwa Handlu Chin nie bedzie miala krytycznego znaczenia dla VIGO Photonics" ("in the board's assessment the decision of China's Ministry of Commerce will not be of critical significance for VIGO Photonics"), which itself specifies the InP-substrate mechanism, that roughly half of semiconductor-materials segment revenue depended on Chinese InP substrates, alternative Japanese and European suppliers, and over one year of substrate inventory.”
Grader confidence: 0.86 · engine conviction 0.58
Both halves of the engine's edge were in the public record at or before the freeze date, so the claim does not clear the frozen bar. (1) The "Rheinmetall is symbolic" leg rests on a fact the engine itself concedes is public: CEO Armin Papperger's on-the-record statement to Politico that Chinese components are about 1% of the company. Same-day coverage of the 2026-07-24 listing already carried that
The ban's first-order effect in the graded window is the OPPOSITE of its stated purpose: it does not create Zimbabwean capacity by 2027-01-01, it hands Zimbabwe's lithium to the two Chinese owners (Huayou at Arcadia, Sinomine at Bikita) who are the only parties with, or credibly building, in-country conversion. A policy sold as sovereignty is a CHINA-CONSOLIDATION event — independent and junior concentrate is stranded or sold into those same Chinese plants on distressed domestic terms. So the cl
Earliest coverage: Reuters (syndicated via MINING.COM): "Zimbabwe's sole lithium salt plant unable to process third-party metal, official says" (2026-07-17)
“"We don't have the capacity to process other minerals from outside. Our concentrator plant produces around 400,000 tons per annum, so we don't have space for other players" — Mthokozisi Goliath, mine manager, Prospect Lithium Zimbabwe (owned by China's Zhejiang Huayou Cobalt), Zimbabwe's sole lithium sulphate plant, reported ahead of the January 2027 concentrate export ban; the same report notes Sinomine's Bikita Minerals and Kamativi are building plants "unlikely to be operational before the January 2027 deadline."”
Grader confidence: 0.78 · engine conviction 0.55
The engine's edge has two load-bearing components: (i) the ban does not create Zimbabwean-owned conversion capacity by 2027-01-01, and (ii) it therefore routes the country's lithium to the two Chinese owners (Huayou/Arcadia, Sinomine/Bikita), stranding independent and junior concentrate or forcing it into those Chinese plants on domestic terms. Both components were in Tier-1 print before the freez
The under-modelled Australian asset is not a mine — it is BAYER-PROCESS ALUMINA LIQUOR. The Wagerup FID establishes the template that an existing alumina refinery can be retrofitted into sovereign gallium capacity with tri-government offtake at roughly US$3m per annual tonne of capacity (derived from the reported ~US$307m total against 100 t/yr). That converts every large allied alumina refinery — Wagerup, Pinjarra, Kwinana, Worsley, Gove and their Canadian/EU/Japanese analogues — into a live ca
Earliest coverage: METLEN Energy & Metals / European Commission — CRMA Strategic Project designation for gallium production at the existing Aluminium of Greece alumina refinery (corroborated by EuropaWire, same date) (2025-03-26)
“METLEN's gallium production investment "has been officially recognized as a Strategic Project by the European Commission under the Critical Raw Materials Act (CRMA)" — the project encompassing "the production of 50 MT of gallium" at the existing Aluminium of Greece bauxite/alumina complex. Corroborating tier-2, CSIS, "Beyond Rare Earths: China's Growing Threat to Gallium Supply Chains," 2025-07-17: "The Department of Energy's Office of Industrial Efficiency and the Department of Commerce should jointly create a pilot-scale retrofit fund—modeled on the 48C Advanced Energy Project Credit—specifically targeted at adding gallium recovery capabilities at existing facilities like South32's Worsley Alumina in Western Australia and Nyrstar's Clarksville Smelter," and "Industry analysis suggests that building side-stream gallium recovery plants at these sites could be cost-effective and yield up to 40 t annually."”
Grader confidence: 0.92 · engine conviction 0.52
The edge's gradeable second-order claim is REPLICATION: that existing allied alumina refineries can be retrofitted off Bayer-process liquor into sovereign gallium capacity with government capital/offtake, making the allied refinery base (Wagerup, Pinjarra, Kwinana, Worsley, Gove and Canadian/EU/Japanese analogues) the under-modelled strategic asset. Entity (existing allied alumina refineries) and
The 2026-07-09 Contraloria change cuts the OPPOSITE way from the nationalisation narrative — it accelerates state-JV execution by removing prior Comptroller approval, which had been the single largest procedural delay on Codelco/Enami lithium contracting. The same institution everyone cites as Chilean sclerosis just made the state's own dealmaking faster. Stated precisely: control is substituted ex-post, not eliminated, so the frozen claim is about SPEED of state-JV execution, not deregulation.
“"[La resolución busca] compatibilizar sus facultades de fiscalización con la necesidad de que las empresas estatales puedan actuar con mayor rapidez en operaciones estratégicas, especialmente en sectores altamente competitivos como el cobre y el litio" — and, on the mechanism itself: "reemplaza en ciertos casos los controles preventivos por mecanismos de fiscalización posterior".”
Grader confidence: 0.88 · engine conviction 0.50
The engine's edge is that the 2026-07-09 Contraloría change cuts opposite to the nationalisation-sclerosis narrative because it removes prior Comptroller approval (toma de razón) on Codelco/Enami corporate operations, substituting ex-post control, and therefore accelerates state-JV lithium execution — the engine explicitly restates the claim as being about SPEED of state-JV execution, not deregula
The differentiated part is print-level: the H1 2026 interim (Aug) shows cobalt sales volumes down >50% y/y against flat mine output, discloses an inventory build/carrying adjustment, and 3993.HK underperforms GLNCY over the 60-day window — generalist consensus still treats CMOC as a copper story with price-bullish cobalt quotas.
Earliest coverage: Bloomberg News — "China's CMOC Bets on Copper Growth, Maintains Cobalt Target" (Tier-1); print-level specificity earliest in SMM/news.metal.com "CMOC Q1 2026: Cobalt Production Up Slightly, Sales Drop Sharply Year-on-Year" (2026-04-24) (2026-01-16)
“CMOC kept cobalt guidance at 100,000 tons to 120,000 tons this year following a record output of 117,549 tons in 2025 ... CMOC — the world's largest cobalt miner — is expected to be able to export 31,200 tons of the material in 2026”
Grader confidence: 0.80 · engine conviction 0.60
The engine claimed its differentiation was "print-level": the Aug 2026 H1 interim revealing cobalt sales down >50% y/y vs flat output plus an inventory build, with 3993.HK underperforming GLNCY. But the print-level fact was already printed and covered before freeze: CMOC's Q1 2026 disclosure (published early May per CMOC; production/sales data reported by SMM on 2026-04-24) showed sales -91.79% y/
The dated, checkable specifics: the production report (~July 30) confirms full DRC allocation shipped, and H1 2026 results (~Aug 6) guide marketing-division EBITDA toward the top of range with cobalt explicitly cited — expressed as the long-GLEN/short-3993.HK pair that captures the quota asymmetry generic 'cobalt exposure' screens miss.
Earliest coverage: Bloomberg News (via Mining.com) — 'CMOC bets on copper growth, maintains cobalt target' (CMOC leg); Reuters (via KFGO) — 'Glencore first-quarter copper output jumps 19%, marketing unit set to exceed target' (Glencore leg, the piece that completes consensus) (2026-04-30)
“its marketing division remains on track to exceed the top end of its annual earnings guidance”
Grader confidence: 0.75 · engine conviction 0.50
The edge's three components each collapse to pre-freeze consensus under the frozen protocol. (1) 'Full DRC allocation shipped' at the ~July 30 production report: Glencore's own Q1 2026 report (2026-04-30) guided that finished cobalt inventories were 'sufficient to fully deliver into near-term quota levels' and exports would 'normalise... in line with its remaining 2026 quotas' — carried by Reuters
Primary claim: the Sept-quarter (FY27 Q1) report shows a visible ASP uplift from the Dy/Tb mix shift — the heavy-RE circuit's first full-rate quarters — which NdPr-only sell-side models understate. An EU/member-state stockpile tender naming separated NdPr/Dy/Tb (RESourceEU publications, Rotterdam storage news) is the lead-time trigger and upside accelerant, not a co-equal claim.
Earliest coverage: UBS sell-side note on Lynas (as reported in FNArena Weekly Ratings, Targets, Forecast Changes 09-05-25) (2025-05-09)
“the imminent addition of dysprosium (Dy) and terbium (Tb) capacity adds incremental value”
Grader confidence: 0.60 · engine conviction 0.50
The engine's primary claim has two parts: (i) mechanism — Lynas's blended ASP rises from the Dy/Tb mix shift as the heavy-RE circuit reaches full rate; (ii) differentiation premise — sell-side models are NdPr-only and so understate it. Part (ii) is directly refuted by a Tier-1 source: UBS (bulge-bracket) in May 2025 explicitly stated the Dy/Tb capacity 'adds incremental value' to Lynas and raised
Role reversal: in nickel, the choked party is CHINA — but the falsifier demands transmission, not trend continuation: the squeeze must reach Chinese stainless mill margins (SHFE stainless-to-NPI margin compression vs H1-2026 average, or documented feed-cost run-cuts) BEFORE any LME nickel rally, with China's ore fallback >90% single-sourced from a Philippines now also selling ~30 Mt/yr to Indonesia and coordinating via the IndoPhil corridor.
Earliest coverage: Reuters, "Chinese firms warn Indonesia's nickel quotas, tax hikes threaten investment" (republished by Kitco off-the-wire); FT embassy-letter story (letter dated 21 April 2026) corroborates (2026-05-13)
“Nickel ore mining quotas have been sharply reduced this year, with cuts for large mines exceeding 70% and total reductions reaching 30 million metric tons ... the changes had raised costs and could undermine existing projects and future investment [Chinese firms operating "smelters, stainless steel plants and battery-material projects"]”
Grader confidence: 0.72 · engine conviction 0.60
The engine's headline edge — role reversal, China as the CHOKED party in nickel — was already in Tier-1 print ~2 months before the 2026-07-16 freeze: Reuters (2026-05-13) reported Chinese firms (including their stainless steel plants) warning that Indonesia's quota cuts raised costs and threatened operations, and the FT obtained the Chinese embassy's 21-April letter describing an "unsustainable ec
Two tiers: 20 entities face outright export prohibition; 20 (incl. Mitsui E&S 7003.T) face no-general-license stricter review. Pain concentrates in the listed drone/defense small-mids (ACSL 6232.T, Terra Drone 278A.T, Mitsui E&S) via supplier-tier contamination — their whole Tier-1/2 chain becomes unable to source China-origin dual-use inputs — while non-listed civilian magnet buyers keep receiving licenses. Expression: equal-weight listed basket vs TOPIX; first disclosed sourcing delay or deliv
“Overseas organizations and individuals are prohibited from transferring or providing dual-use items originating from China to these entities. ... These measures do not affect normal economic and trade exchanges between China and Japan. Japanese entities that operate in good faith and comply with the law have no cause for concern.”
Grader confidence: 0.85 · engine conviction 0.55
The freeze (2026-07-16) came 17 days after MOFCOM Announcements No. 27/28 (2026-06-29), which literally ARE the engine's first element: 20 entities barred outright, 20 (incl. Mitsui E&S) stripped of general licenses and subject to non-timeline-bound stricter review. That is a primary regulator document, echoed same-day by CNBC, the AP wire, Al Jazeera, and Global Times. The claimed second-order me
Two-tier-market claim graded on instruments that isolate the certified-origin premium: primary = ex-China Dy/Tb oxide assessments (Argus/SMM ex-works Malaysia or ex-China quotes) vs China domestic, requiring a sustained >=15-20% premium; secondary = Comtrade HS 2846 Malaysia-vs-China export unit-value spread, acknowledged as mix-confounded (Malaysia's basket is NdPr-heavy and product weights shift as Dy/Tb ramps). The regulatory-inversion leg (MITI permits/incentives for the new HRE facility by
Earliest coverage: Argus Media press release, 'Argus expands critical rare earth pricing to US' (its European ex-China rare earth assessments expanded July 2025) (2025-12-11)
“"The fortnightly US assessments include key rare earth magnet raw materials — neodymium-praseodymium oxide (NdPr), praseodymium oxide, neodymium oxide, dysprosium oxide, and terbium oxide." ... "the Chinese export controls have dramatically widened premiums in non-Chinese markets."”
Grader confidence: 0.90 · engine conviction 0.55
The engine's specific second-order claim — a two-tier ex-China heavy rare earth market in which certified/non-China-origin Dy and Tb oxide (Lynas Malaysia being the only ex-China producer) trades at a sustained premium (>=15-20%) to China domestic prices, measured via Argus/SMM ex-China assessments — was consensus-grade well before the 2026-07-16 freeze. Same entity: Lynas Malaysia / ex-China Dy-T
Tightened per skeptic: auto, not defense, absorbs the Dy/Tb shortage - defense-adjacent demand (incl. the DoD agreement anchored to Lynas LAMP) takes priority claim on the only non-China heavy-REE stream, and Shin-Etsu's Fukui refinery, though it does target Dy/Tb, produces nothing before ~2028 so it is irrelevant to this squeeze. Specific claim: a NEW disclosure after 2026-07-13 by a Japanese automaker or magnet maker (Proterial, Shin-Etsu 4063.T, TDK 6762.T) explicitly citing heavy-REE (Dy/Tb)
Earliest coverage: Reuters (syndicated via Kitco off-the-wire), "China squeezes Japan over rare earths in repeat of 2010 showdown" (2026-05-22)
“Major Japanese magnet maker Shin-Etsu has stopped accepting new orders for dysprosium-containing magnets, according to a Western customer who spoke on condition of anonymity.”
Grader confidence: 0.62 · engine conviction 0.60
The operationalized falsifier — a Japanese magnet maker or automaker citing Dy/Tb or high-temp magnet-grade availability as constraining output/costs — was already in tier-1 coverage before freeze: Reuters reported on 2026-05-22 that Shin-Etsu (one of the engine's three named entities, 4063.T) had stopped accepting new orders for dysprosium-containing magnets, i.e., Dy-specific (not generic-magnet
Tightened per skeptic: the third-party reach is the detail - MP's magnet ramp (Fort Worth + 10X) still needs Dy/Tb it does not mine in meaningful volume, and the only ex-China source (Lynas, 1,500 tpa) is oversubscribed with Japanese priority claims. Specific claim: by 2026-12-31 MP's filings show EITHER (a) a disclosed timing change to 10X/Fort Worth magnet commissioning or capacity guidance, OR (b) a new heavy-REE supply agreement or risk-factor language explicitly referencing sourcing constra
Earliest coverage: Reuters — "West scrambles to fill heavy rare earth gap as China rivalry deepens" (Eric Onstad; syndicated on MINING.COM) (2025-11-19)
“MP's Mountain Pass mine in California contains only traces of those two elements, which are used in small quantities in the magnets but are vital.”
Grader confidence: 0.85 · engine conviction 0.50
The engine's edge is that MP's magnet ramp (Fort Worth/10X) still depends on Dy/Tb it does not mine in meaningful volume and that ex-China heavy-REE supply is constrained (Lynas 1,500 tpa Japan-committed), contra the 'listing is symbolic' read after China's 2026-06-22 export-control listing of MP. But this specific entity+mechanism was Tier-1 consensus months before freeze: Reuters 2025-11-19 repo
Tightened per skeptic into a two-part claim, both legs required for a full pass: (1) a second project-level Reserve offtake/LoS is announced by 2027-01-31 (half-credit alone, as this is semi-consensus rollout); AND (2) the differentiated leg - Japanese counterparties (JOGMEC, Sojitz, or a Japanese trading house/magnet maker) co-anchor a heavy-REE-relevant Australian project (Browns Range, Eneabba, or similar) via offtake or equity BEFORE any Korean or EU-institutional co-anchor appears, reflecti
Earliest coverage: JOGMEC news release: "Securing Supply of Heavy Rare Earths to Japan with Additional Investment to Lynas" (primary Japanese government document) (2023-03-07)
“JARE has decided an additional AUD 200million worth investment in the equity of Lynas ... [in return Lynas will] supply up to 65% of the HRE (dysprosium and terbium) [to Japan, because] HRE supply sources are currently concentrated to China.”
Grader confidence: 0.85 · engine conviction 0.55
The edge's differentiated leg (2) — Japanese counterparties (JOGMEC, Sojitz, or a Japanese trading house) co-anchoring a heavy-REE-relevant Australian project via offtake/equity, driven by Japan's Dy/Tb insecurity, before any Korean/EU-institutional co-anchor — was already a realized, publicly documented pattern well before the 2026-07-13 freeze, not a non-consensus forward call. Tier-1 branch (a)
Tightened per skeptic (pre-freeze announcements excluded): Kuantan becomes the price-setting venue for Western Dy/Tb (Lynas premium-to-China ASPs sustained in quarterly disclosures) AND the agglomeration CONTINUES beyond what is already public - at least one NEW downstream REE commitment in Malaysia announced after 2026-07-13 by a party other than Lynas or the existing Lynas-JS Link JV (e.g., a Japanese magnet/metal maker or a second Korean entrant), by 2027-01-31. Plus a visible jump in Malaysi
Earliest coverage: Bloomberg — "Lynas Grows Only Heavy Rare Earths Hub Beyond China" (Tier-1) (2025-10-28)
“Lynas Grows Only Heavy Rare Earths Hub Beyond China”
Grader confidence: 0.72 · engine conviction 0.55
The edge's second-order mechanism has three legs and each was consensus-grade before 2026-07-13. (1) Kuantan as the ex-China heavy-REE strategic node: Bloomberg's 2025-10-28 headline itself makes the node claim (entity: Lynas Kuantan; mechanism: sole ex-China heavies hub, expanding). (2) Sustained premium-to-China Dy/Tb pricing: Benchmark Mineral Intelligence explicitly forecast the ex-China heavi
The LOS is non-binding, and whether it converts IS the bet. Three dated claims: (1) financial close/first debt drawdown by 2026-12-31 (ASX announcement); (2) conversion of the EFA strategic-reserve LOS into binding offtake/contract terms; (3) at least one additional sovereign-adjacent offtake (Korean or EU buyer, with the EU's May 2026 REE stockpile shortlist as the demand channel) by end-Q1 2027 — all ahead of the market's 'wake me at first production' schedule.
Earliest coverage: Arafura ASX announcement — "Final Investment Decision to Deliver Australia's First..." (primary regulatory filing), corroborated by Australian Mining / discoveryalert.com.au coverage (2026-05-21)
“The non-binding Letter of Support received from EFA, in conjunction with its four previous binding offtake arrangements announced to the market, represent 93% of Arafura's binding offtake target. ... Arafura remains in discussions with the Australian Government and the EFA to secure final approvals and due diligence for the potential support outlined in the LoS.”
Grader confidence: 0.60 · engine conviction 0.50
The audited second-order claim is "the EFA LOS is non-binding and whether it converts IS the bet" (entity=Arafura, mechanism=non-binding LOS conversion + financial close as the operative near-term catalyst). At/before the 2026-07-11 freeze, this exact mechanism was consensus-grade: the non-binding nature of the EFA Critical Minerals Strategic Reserve LOS is disclosed in Arafura's primary ASX FID f
The new bite is transshipment enforcement: the quiet SE-Asia/Turkey rerouting that kept EU Tier-1s supplied since the 2025 controls becomes reportable and punishable inside China. We predict a NEW (post-2026-07-11) disclosed magnet-allocation warning or production-pause from a named German Tier-1 or OEM within ~120 days, surfacing in Q3 guidance (Oct-Nov), attributed by the company to export-license timing, allocation, or transshipment enforcement — NOT merely price/cost — while spot NdPr barely
Earliest coverage: MOFCOM Announcement No. 26 of 2026 (as reported by Morgan Lewis / China Briefing / Benchmark Mineral Intelligence), reinforced by IEA commentary and German FM Wadephul's confirmation of Germany's license-batch exclusion (2026-06-24)
“On June 24, 2026, MOFCOM published Announcement No. 26 of 2026 ... The scope of reportable conduct includes routing exports through third countries to circumvent controls, and transshipment scrutiny applies when Chinese-origin magnet content moves through a third country.”
Grader confidence: 0.58 · engine conviction 0.55
The engine concedes the base story (Europe's magnet dependence, 2025 controls, RESourceEU shortlist) is front-page, and locates its edge in transshipment enforcement surfacing as a German Tier-1 licensing/lead-time shock decoupled from spot NdPr price. Both load-bearing analytical claims were consensus-grade at/before the 2026-07-11 freeze: (1) the transshipment-enforcement mechanism — the engine'
Country-level flow claim, purged of already-announced events: (1) Malaysia's HS 2846 export VALUE rises >=25% YoY in the 2026 Comtrade vintage with volume growth under HALF the value growth — premium capture, not tonnage; (2) a NEW downstream commitment into Malaysia announced after 2026-07-11 that is NEITHER the Oct-2025 5,000 tpa Lynas facility NOR the existing JARE offtake — e.g., a Japanese/European magnet-alloy or metallization JV sited in Malaysia, or a third-party HRE separation project.
Earliest coverage: Columbia Emerging Markets Review — "Malaysia Climbing the Rare-Earth Value Chain" (2025-12-20)
“Downstream magnet manufacturing captures significantly more value than midstream refining, enabling Malaysia to secure a larger share of profits for its economy.”
Grader confidence: 0.72 · engine conviction 0.60
The engine's edge is framed as two forward-looking, deliberately event-purged predictions: (1) an HS 2846 export value/volume divergence in the not-yet-released 2026 Comtrade vintage, and (2) a new post-freeze non-Lynas/non-JARE downstream commitment into Malaysia. Neither the raw Comtrade data nor a literal post-2026-07-11 announcement can be "covered" at freeze by construction. But the protocol
The uncovered detail is the SINGLE DESK: DSI's export monopoly makes Indonesia an OPEC-of-NPI able to set realization prices. Specific claim: Indonesia's HS 7202 export unit value RISES relative to the LME nickel average by >=10% vs the H1-2026 baseline in H2-2026 customs/Comtrade vintages while volumes flatten — margin captured by the state and by integrated miners-with-smelters (the quota rules structurally favor them over pure ore sellers, e.g., Vale Indonesia INCO.JK with own downstream vs p
Earliest coverage: mining.com.au — "Indonesia's Septembergate: How the new export regime impacts global nickel and coal markets" (corroborated by Reuters via Indonesia Business Post, and Bloomberg 24 Jun 2026) (2026-05-30)
“The model was buying at domestic prices and selling at international benchmarks, pocketing the logistics and blending spread. DSI is designed to capture exactly that spread.”
Grader confidence: 0.72 · engine conviction 0.55
The engine frames its edge as an uncovered detail: the DSI SINGLE DESK making Indonesia an OPEC-of-NPI that sets realization prices, with margin captured by the state and integrated miners-with-smelters over pure ore sellers. Both legs of this specific second-order mechanism were already in consensus-grade coverage at/before the 2026-07-11 freeze. LEG 1 (single desk = margin/realization-price sett
The quota-forfeiture mechanism quietly hands the state a TRADABLE stockpile: ARECOMS can meter cobalt sales like OPEC meters oil. Primary graded claims: (1) the 10% royalty is applied to Manono's first lithium export cargoes (~Q4-2026) with NO publicly negotiated Zijin carve-out or rebate before first export — fiscal capture sticks even against China's biggest miner; (2) at least one disclosed ARECOMS state-strategic-quota cobalt sale/deployment by 2027-01-31. Secondary confirmation: 2026 Comtra
Earliest coverage: Semafor — 'DR Congo's cobalt quotas force miners to stockpile, pivot, or pause' (corroborated by Reuters/Bloomberg-sourced reporting on the ARECOMS strategic-reserve decree and S&P Global Oct-2025 / IEA ARECOMS Decision 004/2025) (2026-05-04)
“Through the strategic reserve, administered by ARECOMS, Kinshasa is positioning itself as a swing producer — much like OPEC in the oil markets.”
Grader confidence: 0.82 · engine conviction 0.55
The specific second-order claim to check is 'quota-forfeiture hands the state a TRADABLE stockpile; ARECOMS can meter cobalt like OPEC.' This exact framing — same entity (ARECOMS), same mechanism (forfeited quotas -> government-controlled strategic reserve the state can market/deploy to swing supply like OPEC) — was already mainstream, not differentiated, at freeze. Semafor (2026-05-04) states 'Th
Two narrowed claims: (1) the DLA lithium award is announced by 2027-01-11 at a disclosed/inferable unit price >=10% ABOVE Fastmarkets US battery-grade Li2CO3 spot at award date — the price-floor PRECEDENT is the claim, not the origin; (2) serialization narrowed to the differentiated call: a DLA/NDS solicitation specifically for ANTIMONY (the dual-use bellwether on our radar) or natural graphite appears on sam.gov by 2027-01-11 — generic 'any battery material' does not count. This is a precedent
Earliest coverage: sam.gov — DLA Strategic Materials Natural Graphite Flakes solicitation (SP8000-25-R-0013), primary government procurement document (2025-11-10)
“The DLA Strategic Materials is seeking bids for an Indefinite Delivery and Indefinite Quantity (IDIQ) contract (SP8000-25-R-0013) to procure up to 49,433,284 kilograms of Natural Graphite Flake over five years for the National Defense Stockpile (NDS), with a ceiling of $65,000,000.”
Grader confidence: 0.68 · engine conviction 0.50
The edge bundles two claims. (1) The lithium price-floor claim (DLA award priced >=10% above Fastmarkets US battery-grade Li2CO3 spot, as a precedent) IS genuinely not-yet-consensus at freeze: the award has not occurred (bids due 2026-07-17), and coverage of the tender's 5-year FIXED-price structure (Mining Weekly, SMM, TheDeepDive, Bloomberg 2026-07-08) is directional at most ('could command stro
Time-to-recognition — was the engine early?
robotics · protocol v1 · 2026-08-26Nothing here is a “lead” yet. This is the backward pass: for each frozen edge, did consensus-grade coverage of the specific second-order claim already exist at freeze? An edge that clears that bar is differentiated at freeze — a necessary condition, not proof of value. It becomes a proven-lead only when consensus actually arrives later (lead-days = recognition − freeze). Those accrue over 12–24 months.
Consensus-grade (frozen v1) = the same entity+mechanism claim in ≥1 tier-1 source OR ≥2 independent tier-2 sources. Conservative bias: already-consensus when uncertain — the grader’s job is to refute the engine’s novelty. All 24 already-consensus rows carry the exact quote that sank them. Rules were frozen (RECOGNITION_PROTOCOL.md v1) before this data.
By 2026-12-31 neither BMW nor Catalyst Brands discloses payment terms, unit counts, or contract value for Figure 03, and no third named site is disclosed with terms -- Figure stays L2 while Agility holds L4. The transfer-block and valuation legs are dropped (the 07-08 notice followed a $100M employee tender at the $39B Series C and reads as excess secondary demand, not a mark under pressure).
Grader confidence: 0.60 · engine conviction 0.50
Underlying facts are publicly reported but only in outlets outside the frozen tier lists: (1) BMW deal financial terms/duration undisclosed since Jan 2024 (TechCrunch, PR Newswire, Robot Report, Fortune/Yahoo 2024 piece questioning Adcock's BMW claims); (2) Catalyst Brands deal (May 26 2026) — Sourcing Journal/WWD and Humanoids Daily note 'the companies have not disclosed the scale or structure of
Conversion CADENCE accelerates: at least 6 systems flip to operational in Q4 FY2026 (56 -> >=62, above the trailing ~4-5 per quarter) AND a second named non-Walmart reorder appears. The edge is the cadence number, not the metric.
Grader confidence: 0.60 · engine conviction 0.55
The freeze date is 2026-08-24. The most recent public data point before the freeze is the 2026-08-05 Q3 FY2026 print: 4 systems went operational (52->56), 11 new starts, 77 in deployment, and Q4 guidance was given only in revenue ($760-780M) and adjusted EBITDA terms. A fetch of the full Q3 call transcript found no management statement on how many systems would flip to operational in Q4 and no exp
By FY2026 results (2027-03-31): a state bank / SOE-leasing or policy-financing structure is NAMED in connection with Walker S2 deployments, and cumulative Walker orders are disclosed at >=2x the last figure on record at freeze (the stale >800M-yuan L3 citation must be refreshed first; if a higher tally has already been disclosed, the 2x bar applies to that figure).
Grader confidence: 0.60 · engine conviction 0.45
The edge has two prongs. Prong (a), the mechanism claim (a NAMED state bank / SOE-leasing / policy-financing structure attached to Walker S2 deployments), is what would have to be consensus-grade. Targeted English and Chinese searches (UBTech Walker S2 leasing / state bank / policy bank / 融资租赁 / 国资租赁 / 银行授信 租赁 交付) returned no source of any tier tying a named state-bank or leasing vehicle to Walker
US pilot STARTS on legged bodies fall for 2-3 quarters after 2026-07-28 because the ~$16k-$90k Unitree/UBTech/Fourier price tier has no FCC-authorized substitute; integrators re-scope pilots onto non-covered form factors (floor scrubbers/AMRs) -- the LaGuardia pilot already ran 3 platforms of which only 1 was a quadruped.
Earliest coverage: Nearest-neighbour coverage (NOT consensus-grade under protocol): The Robot Report, 'Experts react to FCC limits on U.S. imports of new humanoid and mobile robots' (analyst Georg Stieler quote) — covers the price-tier/no-substitute ingredient, not the pilot-start/re-scope mechanism. (2026-07-29)
“"In the near term, the measure could slow U.S. physical AI innovation by cutting startups and researchers off from future low-cost Chinese platforms before comparable Western alternatives exist."”
Grader confidence: 0.55 · engine conviction 0.55
Splitting the edge into its parts: (1) 'no FCC-authorized substitute at the Unitree/UBTech/Fourier price tier' was widely aired before freeze by The Robot Report (07-29, Stieler/Scriven quotes), MIT Technology Review (08-03: Unitree quadruped ~$4,600 vs Boston Dynamics ~$278,000; '90% of recent robotics research papers from US universities relied on robots from Unitree'; 'could slow down the indus
Harmonic Drive's FY3/27 'double-to-triple humanoid orders' guidance is a single-program bet on one undisclosed North American customer at ~¥1.2B/quarter; the concentration, not the moat, is the graded object -- customer-identity-agnostic.
Earliest coverage: Harmonic Drive Systems FY3/27 Q1 earnings briefing (2026-08-07; transcript posted 2026-08-17), as summarised by BigGo Finance — NON-consensus-grade under protocol; closest coverage of the facts, not of the concentration claim (2026-08-17)
“"For AI robots, the company secured orders of approximately ¥1.2 billion from a North American customer for the second consecutive quarter." ... "While there is currently 'one large customer engaged in meaningful volume production,' Kamijo expressed the view that 'there is ample possibility that one or two more customers producing at the scale of 10 or 100 units will emerge over the next one to two years.'"”
Grader confidence: 0.55 · engine conviction 0.45
The engine's specific claim = HDS FY3/27 'double-to-triple' humanoid guidance is effectively a single-program bet on one undisclosed NA customer (~¥1.2B/qtr), with concentration (not moat) as the graded mechanism. Evidence found: (1) Investing.com 2025-11-19 (non-tier) reported the double/triple FY3/27 indication and quoted Macquarie's Wendy Pan — on the ¥10bn target being aggressive, NOT on conce
China's MIIT lifecycle-ID platform becomes the first COUNTABLE national deployment ledger: coded units rise from 28,000 (2026-05-28) past 60,000 by 2027-06-30, and the procurement pull is state/SOE tenders under 揭榜挂帅 (the Walker S2 pattern), not consumer sales. A second Chinese program reaches L3+ with disclosed terms, forcing the stale UBTech L3 row to be re-cited.
Earliest coverage: Xinhua, "China rolls out 'digital ID' system to regulate booming humanoid robot sector" (underlying event only; not a match for the engine's mechanism) (2026-05-28)
Grader confidence: 0.60 · engine conviction 0.60
Underlying facts were widely reported before freeze: the MIIT-standardization-committee lifecycle platform (29-digit code, "no code, no market access", 100+ firms, 200+ models, 28,000+ units) via Xinhua/CGTN 2026-05-28 and SCMP; the MIIT+SASAC 2026-06-09 notice (primary regulator document) mandating central SOEs to file >=10-20 real scenarios and targeting 10,000 humanoids in commercial use by end
Two filing-graded specifics: (1) Symbotic's backlog is diversifying -- the FY2026 10-K Walmart share of backlog prints below the FY2025 level; (2) Agility is the only L4 humanoid and now the only one with a protected US authorization regime, so its first post-SPAC annual report shows >= $300M committed orders and a cumulative Digit delivery count consistent with the 10k/yr plan. The Harmonic link is tested, not asserted: only if Agility's S-4/10-K names Harmonic Drive does the ¥1.2B/qtr line bec
Grader confidence: 0.60 · engine conviction 0.55
Refutation attempt by component. (1) Symbotic backlog diversification: FY2025 10-K (2025-11-24) says backlog is 'vast majority' Walmart+GreenBox with no percentage split; Zacks/Nasdaq/Seeking Alpha (not Tier-1/2) discuss 84% revenue concentration and diversification strategy; Barclays/KeyBanc target revisions cite new verticals but no source states the specific test that FY2026 10-K Walmart backlo
If (and only if) PRC-origin AMRs are covered and MiR secures a US authorization path, MiR is the direct share-taker in new US warehouse AMR authorizations -- invisible at the Teradyne consolidated level but visible in Denmark HS 8479.50 exports to the US (>= +15% YoY for 2026H2) and a zero count of new FCC authorizations for PRC AMRs per quarter. The $110M Teradyne segment leg is context only.
Grader confidence: 0.60 · engine conviction 0.40
The underlying event (FCC added foreign-produced advanced robotic devices, explicitly including warehouse AMRs, to the Covered List on 2026-07-28) is heavily covered: FCC fact sheet/FAQs, Sidley, K&L Gates, Wiley, White & Case, Morgan Lewis, IEEE Spectrum, CNBC, NPR, Fox, Robot Report, SDCExec/Interact Analysis, Manufacturing Dive, Supply Chain Dive. Note this partially undercuts the engine's stat
The impact channel is the ROBOTICS-lidar line (118,282 units in Q1, +137.8% YoY) — that segment, not China ADAS, is where US/global autonomy exposure sits, and design-outs happen 2-4 quarters before revenue shows. Statutory clock corrected per skeptic verification: the near-term forcing functions are the already-effective 2026-06-30 federal procurement restriction and S.4429's 2027-01-01 SOFTWARE prohibition — the hardware ban starts 2030-01-01, which is exactly why consensus underweights the ur
Grader confidence: 0.62 · engine conviction 0.55
The underlying events were widely covered by freeze: DoD military-entity listing, CNBC's 2026-07-07 piece naming Hesai's US AV/trucking customers (Zoox, Waabi, Kodiak, Nuro, Aurora), and S.4429's 2026-07-22 committee approval with its Jan-2027 software / 2030 hardware phasing (Commercial Carrier Journal, TheTrucker, CDLLife, law-firm advisories, bill text). But under the frozen rule, event coverag
The codification event is the 2026 Ulsan wage settlement (typically Aug-Sep): the union's consent demand has been on public record since January 2026, so a clause is not marginal wage-round noise — probability of codification raised accordingly. Pre-registered clause definition: any text requiring prior labor-management agreement/consent before deployment of new robot/automation technology on production lines. A clause pushes the Atlas rollout to the US Metaplant lane only — roughly halving the
Earliest coverage: none found for the specific second-order claim; nearest near-miss: gcn.com "the single labor clause... changes the math" (blog, non-tier, no quantified mechanism) (2026-08-02)
Grader confidence: 0.60 · engine conviction 0.55
The underlying event was internationally saturated by freeze (strikes, January consent demand, BD acquisition, Georgia-first rollout) — the engine's own consensus description "barely covered outside Korea" is too generous to itself on that flank. But the protocol excludes event coverage and requires the SAME second-order claim. Nine targeted searches and two full-article fetches found no Tier-1 so
Order QUALITY, not order size. The graded tell is the FY2026 interim report (announcement due by 2026-09-30 under HKEX rules): if trade receivables and contract assets grow faster than recognized robot revenue, the RMB800M is announcement-backlog, not cash. Forward supply trigger: ADDITIONAL conversion/placement filings on HKEX (noting the ~87% H-share ratio leaves little domestic stock left to convert). And the Beijing subsidy is capped at RMB1,500 per item — roughly 1-2% of a robot's price — f
Earliest coverage: NONE consensus-grade. Nearest non-qualifying antecedent: Sohu-hosted Chinese finance-media analysis of UBTech FY2025 report, '优必选2025财报：机器人业务爆发难掩财务隐忧 应收账款接近全年营收 坏账准备计提比例近三成' (receivables near full-year revenue, ~30% bad-debt provisioning) — same entity, adjacent mechanism, but not a Tier-1/Tier-2 source and does not make the FY2026-interim/RMB800M-Walker-backlog claim (~2026-04 (post-FY2025 annual results; nearest non-qualifying antecedent only))
Grader confidence: 0.60 · engine conviction 0.50
The engine's edge is a composite second-order claim: (1) the FY2026 interim report (due by 2026-09-30) as a graded tell — receivables+contract assets growing faster than recognized robot revenue would mark the RMB800M Walker S2 orderbook as announcement-backlog not cash; (2) additional HKEX conversion/placement filings as a forward supply trigger given the ~87% H-share ratio; (3) the Beijing RMB1,
The differentiated, pre-registered claim is revenue-mix PERSISTENCE: the first post-listing disclosure periods show revenue still concentrated in research/education-type buyers with no paid-deployment revenue class emerging, forcing the growth narrative onto Chinese consumer demand whose marginal driver is a RMB1,500-capped subsidy. Grading re-anchored per skeptic to disclosures that must exist: top-5 customer identities/types and revenue-by-product/customer categories, with a pre-registered cla
Grader confidence: 0.62 · engine conviction 0.50
The edge is about Unitree Robotics' STAR Market IPO (subscriptions 2026-08-10, i.e., after freeze). The static revenue mix (73.6% research/education, ~9% industrial, commercial mostly display use) is prospectus-disclosed and universally covered pre-freeze (Caixin, SCMP, The Robot Report, ChinaTalk, 36kr, Forbes contributor, Bloomberg sector primer per paraphrase) — but that is the underlying discl
The root-cause MIX is the signal: design and process-control failures clustering on the two NEWEST platforms (dV5, SP) is the classic ramp-strain signature. Tightened against base rates per skeptic: the predicted NEW recall must be on dV5 or SP specifically AND carry a design or process-control root cause in the FDA record (labeling/customer-use causes excluded), or be the FDA-inspection branch (483/warning letter touching manufacturing), within two quarters — plus litigation tail from the stapl
Earliest coverage: NEAREST MISS ONLY (not consensus-grade): Simply Wall St via Yahoo Finance, 'ISRG Is Down 6.3% After Class II Recall... Has The Bull Case Changed?' — single non-tier publisher, generic quality-scrutiny framing, no root-cause-mix recurrence prediction (2026-07-16)
Grader confidence: 0.65 · engine conviction 0.40
The underlying events are saturated with coverage: each 2026 recall (E-brake pin, SureForm stapler Class I with a death, dV5 foot pedal Class II, SP drapes, X/Xi screws), the -6.3% recall-day move, the Q2 blowout ($2.89B, +16% procedures) and the post-earnings selloff, plus analyst target cuts (JPMorgan $550-to-$450, explicitly motivated by procedure-growth slowdown and valuation, not recall recur
Two independent accelerants the street hasn't connected: the segment inflected on its own demand AND its Chinese competition was just legally removed from the US institutional market. Named assumption stated per skeptic: AMR coverage under the FCC definition is an inference (vacuums/mowers confirmed in scope, wheeled AMRs very likely), and the bar applies only to NEW authorizations after 2026-07-28 — already-authorized units keep selling; also Unitree's vacated US footprint was quadruped inspect
Earliest coverage: None qualifying. Closest non-qualifying: The Robot Report, "Experts react to FCC limits on U.S. imports of new humanoid and mobile robots" (2026-07-29) — mentions Teradyne only in a list of manufacturers with expanded US presence, without the beneficiary/mechanism claim. (2026-07-29)
Grader confidence: 0.70 · engine conviction 0.60
The two underlying events were each separately consensus before the 2026-08-04 freeze: the FCC covered-list order (2026-07-28, primary regulator document plus PBS/CBS/NBC/CNN/Forbes coverage) and Teradyne Robotics' demand inflection (first-ever $100M quarter, +33% YoY, US mix rising to 32%, reported 2026-07-28 and covered by trade press). But the protocol explicitly excludes underlying-event cover
Statutory clock fixed per skeptic: the near-term forcing functions are the already-effective federal procurement restriction and S.4429's 2027-01-01 SOFTWARE prohibition — not the 2030-01-01 hardware ban — and the FCC covered list does not itself cover lidar components (context, not mechanism). The requalification pipeline forms NOW, 2-4 quarters before statutes bite, because autonomy programs can't design 2027 stacks on hardware facing the software prohibition — and Ouster is the default domest
Earliest coverage: Closest non-qualifying near-miss: BingX analyst-note summary (citing Northland/Oppenheimer/Roth/Cantor) — first-order "Section 164 one-player market for OUST" claim; does NOT qualify under either consensus branch and lacks the software-clock/requalification mechanism (2026-07 (approx.; pre-freeze analyst-note summaries))
Grader confidence: 0.60 · engine conviction 0.55
The underlying events are heavily covered (Reuters-adjacent trade press on S.4429's July 2026 Senate Commerce advance; NDAA Section 164 effective 2026-06-30; CNBC 2026-07-07 on Hesai cyber risk), but the engine's SPECIFIC second-order claim — that the already-effective federal procurement restriction plus S.4429's 2027-01-01 SOFTWARE prohibition (not the 2030 hardware ban) forces a commercial requ
The engine's core divergence read is NEGATIVE (-0.222): deployed reality now exceeds the cooled narrative. The market sold the entire category while the one graded-real program compounded evidence (L3→L4 with citations). Tightened per skeptic: the PIPE guarantees minimum proceeds but does not mechanically suppress redemptions — so the graded claim is pre-registered as redemptions <50%, with minimum-cash-condition satisfaction as the fallback partial, so a high-redemption-but-closes outcome grade
Grader confidence: 0.60 · engine conviction 0.50
Adversarial decomposition of the edge: (1) the scarcity/'first public pure-play with real deployment revenue' component is NOT novel in substance — it is the deal's own marketing pitch, present verbatim in the 2026-06-24 press release and Form 425 ('the only U.S. publicly listed pure-play humanoid company with proven, active commercial deployments') and echoed across Forbes, Yahoo/24-7WallSt, Seek
BOM attribution fixed per skeptic: magnets are a low-single-digit-percent, no-substitute input INSIDE the 40-60% actuation slice — allocation, not cost, is the bind. The pinch lands on Germany's automation/robot-component sector in a datable window: the European qualification bridge (Neo Narva, PPAP H1-2026, Phase 1 = 2,000 t/yr, ~15% of EU demand) does not deliver qualified volume until 2027 — leaving German drive/motor makers a naked 2-3 quarter H2-2026 window in which magnet allocation, not o
Earliest coverage: Nearest NON-qualifying mechanism-adjacent coverage: Quality Digest, "Supply Chain Crunch Could Be on Horizon for Automation OEMs" (tier-3 trade press; generic automation-OEM rare-earth access concern, no German sector, no H2-2026 allocation window, no Narva qualification bridge) (2025-11-26)
Grader confidence: 0.68 · engine conviction 0.60
The audit tried to refute novelty and failed on the specific second-order claim. Pre-freeze tier-1 coverage establishes only the engine's stated consensus (German AUTOS exposed to Chinese rare earths, e.g., Reuters' VDA exclusive) — protocol explicitly excludes underlying-event coverage. For the claimed entity (German automation/robot-component drive/motor makers), all pre-freeze consensus-grade c
The door-opens narrative fails on both of Korea's own dependencies: Korea cannot scale the alternative it is supposed to be while its robot motors are ~89% Chinese-magnet-dependent, and the Atlas rollout start likely slips past H1 2027 because the union consent clause (a demand on public record since January 2026) gets folded into the 2026 wage settlement. Korean robot-parts suppliers priced for the Atlas ramp get hurt first. Truth-board discipline: zero Atlas units are at L3+ today.
Earliest coverage: None found at consensus grade. Closest partial match: Asia News Network (Korea Herald syndication), "US moves to curb China robots, opens door for Hyundai" — contains an expert caveat that Korean-made humanoid robot motors likely rely heavily on Chinese rare earth magnets, but is a single non-tier-listed source inside a bullish frame with no timeline-slip or supplier-hurt mechanism. (n/a (no consensus-grade source found; closest partial match circa mid-2026, exact date unverifiable — page returned HTTP 403))
Grader confidence: 0.68 · engine conviction 0.55
Adversarial audit tried to refute novelty and failed on the specific composite claim. The INGREDIENTS were all public before freeze: (1) union consent clause on record since 2026-01-22 (Seoul Economic Daily; Just Auto) — which the engine itself discloses; (2) July 2026 strike/wage-talk collision widely covered (Korea Times 06-25, TechTimes, Forbes, eWeek); (3) Korea's ~88-90% Chinese-magnet depend
The bite is not US unit sales — it is the severed DISTRIBUTION channel: Unitree's global mindshare was built through Western universities and developers, and the FCC bar on new-model authorizations means the next generation (post-2026 models) never enters that ecosystem, pushing Western labs to allied platforms exactly as Unitree's prospectus must defend an overseas growth story. Gradeable: Unitree's first post-IPO annual report shows overseas revenue share DOWN vs 2025. Parallel: US OEMs design
“"If Washington severs this symbiotic relationship, it will almost certainly make it harder for American researchers to maintain their lead in the software side of embodied AI." (ChinaTalk, non-consensus-grade); nearest Tier-2: "In the near term, the measure could slow U.S. physical AI innovation by cutting startups and researchers off from future low-cost Chinese platforms before comparable Western alternatives exist." (The Robot Report, 2026-07-29)”
Grader confidence: 0.60 · engine conviction 0.55
The underlying events (FCC Covered List action 07-28, Unitree IPO timing, prospectus US-risk disclosure) were saturation-covered by freeze, including Tier-1 Reuters — but Reuters' framing matched exactly the first-order consensus the engine pre-registered as consensus (US sales exposure percentage, future models blocked from the US market). The engine's second-order claim — the bite is the severed
Hold them together: the trade is INTRA-Japan dispersion. China's 2026-01-06 controls specifically targeting Japan hurt its magnet/motor chain (Proterial, TDK, Shin-Etsu — feedstock-starved, May magnet imports fell sharply) while leaving reducers untouched. The qualification moat binds via US policy: NDAA Sec.163-style procurement rules make a Chinese reducer single-source unqualifiable for every non-Chinese humanoid program — Agility, Apptronik, Figure, Atlas volumes default to Japanese reducers
Earliest coverage: none found for the joint claim (closest non-qualifying: nextfinancial Substack 'The Joint Problem'; note.com Japan humanoid stock list; brianartex.github.io humanoid deep-dive May 2026 — all single blogs/newsletters, not consensus-grade)
Grader confidence: 0.65 · engine conviction 0.65
Adversarial search found abundant pre-freeze consensus coverage of every INGREDIENT: (1) Japan magnet makers squeezed — Benchmark Mineral Intelligence, Reuters (Shin-Etsu refinery, 2026-06-11), UPI 2026-06-22 (magnet exports to Japan fell), Bloomberg-syndicated 'Corporate Japan's rare-earth warnings' 2026-07-06 all name Proterial/TDK/Shin-Etsu as hurt by the 2026-01-06 controls; (2) reducers-benef
The sharp version is a ROTATION claim with a date: humanoid disillusionment doesn't shrink automation budgets, it redirects them to deployable form factors within 1-2 quarters, and Denmark's cluster is the redirect's landing zone (labor-scarce, high-wage Europe converts pilots to paid fastest; UR/MiR sell L4/L5-equivalent reality today). Gradeable near-term: Teradyne Robotics posts a SECOND consecutive $100M+ quarter in Q3 2026 (reported late Oct) — consensus models still treat Q2 as a one-off A
Grader confidence: 0.72 · engine conviction 0.60
The underlying event (Teradyne Robotics' first $100M quarter, reported 2026-07-28..31) was consensus-covered, but its consensus attribution was the AI-datacenter buildout — the very framing the engine positions against. The engine's specific second-order claim has three components absent from all qualifying coverage at/before the 2026-08-04 freeze: (1) humanoid disillusionment REDIRECTING automati
Stronger than originally written, per skeptic verification: Bosch is Humanoid (UK)'s contract MANUFACTURER (production capacity for 100,000 units) and Schaeffler its anchor customer with a 1,000-robot order — so Europe's first non-Chinese humanoid magnet chain runs through Estonia, and its qualification is datable. Sharpest gradeable form: a disclosed Narva-to-robotics qualification (PPAP or supply designation for a named automation/humanoid program, e.g., HMND 01) by 2027-06-30 — an outcome zer
Earliest coverage: None found making the specific claim. Closest tier-1 coverage of the entity is Bloomberg's EV-framed Narva piece, which confirms rather than refutes the engine's stated consensus framing (2025-09-19)
Grader confidence: 0.78 · engine conviction 0.55
The edge's components are individually consensus (Bosch as Humanoid's contract manufacturer with 100k-unit capacity and Schaeffler's 1,000+-robot order were covered by Forbes/eWeek/trade press in May-July 2026; the Narva plant and its H1-2026 PPAP timeline were covered by Bloomberg/Fastmarkets/Argus), but the protocol grades the SPECIFIC second-order claim: that the Narva plant constitutes the mag
A sole-source qualification moat whose expiry is published by the government. OSHA effectively requires NRTL-listed equipment for US workplace deployment, and for the SCIEE-robot class one NRTL currently holds the standard. Two non-obvious consequences: revenue scales with MODEL PROLIFERATION rather than unit volume, so every model and hardware revision is a separate paid test cycle that pays even as the ramp disappoints (only 2 of 9 humanoid programs are at L3+); and the moat's end is observabl
Grader confidence: 0.72 · engine conviction 0.40
ADVERSARIAL POSTURE: I tried to refute novelty and could not, on the mechanism as specified. WHAT THE ENGINE'S CLAIM ACTUALLY IS. Three limbs beyond the headline fact: (i) a sole-source NRTL qualification moat for the UL 3300 / SCIEE-robot class; (ii) revenue scaling with MODEL PROLIFERATION rather than unit volume, so every model and hardware revision is a separate paid test cycle that pays even
The moat and the hostage are the same asset. Japan is the only economy under a country-wide dual-use ban on the seven controlled rare earths and NdFeB magnets (MOFCOM 2026-01-06) rather than firm-specific measures — the unqualified 'only country-wide ban ever' is wrong, since MOFCOM Announcement No. 46 of 2024 did exactly that to the United States on gallium/germanium/antimony. The specific predicted signature is a MARGIN miss with BACKLOG INTACT — rising inventory days plus flat-to-declining gr
Earliest coverage: No consensus-grade source found for the specific claim. Nearest non-qualifying candidates: S&P Global, 'Robot rise could be slowed by metals shortage: analysts' (2026-02-19, single Tier-2, wrong entity/exposure); simplywall.st, 'Harmonic Drive Systems (TSE:6324) Margin Compression Challenges Bullish Growth Narratives' (2026-05-15, non-enumerated retail algo-blog, no rare-earth mechanism attribution); WSJ, 'China Deprives Japan of Rare-Earths Supply, Escalating Dispute' (2026-01-08, Tier-1 but underlying event only).
Grader confidence: 0.58 · engine conviction 0.58
The engine's second-order claim decomposes into three limbs, and I searched each independently. Limb 1 (Japan's actuation moat is simultaneously its hostage under a Japan-specific country-wide dual-use ban): no source found inverting the picks-and-shovels-winner framing for Japanese reducer makers. Limb 2 (the falsifiable financial signature: margin miss with backlog INTACT — rising inventory days
Taiwan's robotics win is neither silicon nor assembly margin — it is that Foxconn is buying EQUITY IN ITS OWN FUTURE CUSTOMERS, the specific mechanism by which a contract manufacturer pulls the humanoid BOM's largest slice (actuation, ~40-60%) toward its own supply chain and away from the Japanese incumbents who own it today. THE TRADE-FLOW LEG IS NOW PRIMARY and the S-4 leg is DROPPED as a confirm: a lead $200M PIPE investor is near-certain to be named in a mandatory registration statement, so
Grader confidence: 0.66 · engine conviction 0.45
I tried to refute the edge and could not. What IS public and well-covered at/before the 2026-07-25 freeze is the EVENT and two generic background facts, none of which match entity+mechanism: (a) Foxconn led a $200M PIPE in Agility Robotics' Churchill Capital XI SPAC (announced ~2026-06-24, covered through July by Embodied Global 07-07, roboticsandautomationnews 07-07, theroboticsmedia, manufacturi
A technology-mix mismatch INSIDE the consensus-favorite layer: the humanoid BOM's actuation dollars (40-60% of BOM) skew to strain-wave/planetary sockets, not Nabtesco's RV geometry sized for industrial-arm base joints. The market is paying a humanoid multiple for capacity being doubled into the wrong socket — the picks-and-shovels trade is right about the layer and wrong about this name.
Earliest coverage: Next Financial (Substack newsletter) — "The Joint Problem: Who Owns the Most Expensive Part of a Humanoid Robot — and Who's Taking It" (NOT consensus-grade under protocol; recorded as earliest same-claim coverage found) (2026-06-02)
Grader confidence: 0.70 · engine conviction 0.35
The edge's second-order claim (entity: Nabtesco; mechanism: humanoid BOM actuation dollars skew to strain-wave/planetary sockets, not Nabtesco's RV geometry, so the humanoid-driven rerating and RV capacity doubling target the wrong socket) was NOT consensus-grade at freeze. Tier-1 sweep: no Bloomberg/Reuters/FT/WSJ article makes this claim; Goldman's humanoid supply-chain coverage prefers harmonic
The category selloff is indiscriminate, but the evidence is not: Agility is the one humanoid whose backlog is milestone-contracted and third-party-verifiable through SEC filings, and whose commercial launch has a specific, checkable gate — NVIDIA Halos/TUV functional-safety certification. Pre-close CCXI trades near trust value, so the market is charging almost nothing for the only US-listed L4 humanoid pure-play while punishing it for Figure's and Tesla's sins.
Grader confidence: 0.60 · engine conviction 0.40
The underlying events were fully public before the 2026-07-17 freeze: the $2.5B CCXI/Agility de-SPAC (announced June 24/July 5, 2026; company press release, Forbes contributor, TechCrunch, AP syndication), the $300M+ milestone-contracted order book with its SEC footnote, and the NVIDIA Halos/TUV Rheinland functional-safety inspection of Digit (NVIDIA newsroom, June 23, 2026). But the engine's SPEC
The edge rests entirely on the monopoly-window clock, not the ban headline: defense-drone high-res 3D lidar RFQs currently have exactly one compliant vendor, competitor qualification onto Blue UAS historically takes quarters-to-years, and the funded capacity means the window converts to BOOKINGS before revenue — so the differentiated, dated claim is sequential defense/robotics bookings inflecting +20% QoQ within two quarters while the Blue UAS list still shows zero high-res 3D competitors. Both
Earliest coverage: None found matching entity AND mechanism at consensus grade. Closest non-qualifying: mid-tier analyst commentary ("one-player market for OUST", "re-rating as bookings accelerate") relayed via aggregators around the 2026-07-15 Oppenheimer PT raise — Oppenheimer/Roth/MKM are not bulge-bracket and the relaying outlets are not tier-1/tier-2.
Grader confidence: 0.60 · engine conviction 0.40
The underlying event (Ouster as first/only high-res 3D lidar on Blue UAS; Hesai locked out by NDAA sec.164 effective 2026-06-30; the stock run) is saturated coverage, exactly as the engine's stated consensus concedes. But the edge's second-order claim — the monopoly-window CLOCK converting to sequential defense/robotics bookings inflecting +20% QoQ within two quarters, before revenue, while the Bl
The gradeable specifics: (a) the $300M committed-order figure is milestone-CONTINGENT and dominated by a single three-year ~1,000-robot contract from one undisclosed customer — the edge is that this anchor customer is identified in the S-4/proxy and first milestone conversions appear as RECOGNIZED revenue within 12 months of listing, making Agility the first humanoid P&L auditable via XBRL (the engine's revenue-truth lens) while grading customer-concentration risk honestly; (b) Halos/TUV certifi
Earliest coverage: TechTimes, "Agility Robotics Nears Public Debut: Safety Cert, S-4, and a 34x Valuation" (closest NEAR-MISS only — single non-tier outlet reportedly framing third-party safety certification as gating the $300M backlog; does not satisfy either consensus branch) (2026-07-11)
Grader confidence: 0.62 · engine conviction 0.60
The edge decomposes into background facts plus distinguishing forward claims. Background (milestone-contingent $300M dominated by one three-year ~1,000-robot contract from an undisclosed customer) was consensus-visible at freeze — it is footnote (1) of the SEC-filed investor deck and was echoed by Forbes (6/24), TechCrunch (6/24), GeekWire (~7/7), TechTimes (7/4, 7/11) — but the protocol excludes
The specific claim consensus hasn't made: the PIPE is the tell — Foxconn becomes the disclosed CONTRACT MANUFACTURER for a Western humanoid OEM (Agility the most likely) within 12 months, replicating the iPhone playbook (anchor investment → assembly contract → BOM aggregation) in humanoids. That makes Taiwan the assembly rent-collector of Western humanoid scaling regardless of which OEM brand wins — and Foxconn's robotics BU starts breaking out humanoid-related revenue, the XBRL-gradeable confir
Grader confidence: 0.70 · engine conviction 0.60
The anchor event is fully public before freeze: the June 24, 2026 Businesswire/SEC-filed announcement of the $2.5B Agility-Churchill XI SPAC names Foxconn as lead of the ~$200M PIPE, and the S-4 draft was submitted July 14, 2026. But no Tier-1 outlet was found covering the deal at all (Reuters/Bloomberg/WSJ/FT site-restricted search returned nothing on the SPAC or a Foxconn manufacturing role), an
Consensus stops at 'Japan has the moat OR China is eating it.' The edge is the SPLIT: Chinese share gains will stay confined to China-domestic OEM programs (Optimus's cost-driven sourcing is the exception, and Optimus is L1 with Musk himself calling the ramp 'extremely slow'), while every WESTERN program that actually converts to paid (Agility L4, Atlas captive, Apptronik) qualifies Japanese reducers/bearings because safety certification (Halos/IEC 61508) freezes the qualified BOM — qualificatio
Earliest coverage: Investing.com coverage of Harmonic Drive Systems earnings briefing (closest adjacent coverage — underlying event only, does NOT establish consensus on the split mechanism) (2025-11-19)
Grader confidence: 0.60 · engine conviction 0.55
The consensus frame the engine cites (Japan moat / China cost erosion / Shuanghuan-Optimus) is abundantly covered, but the specific second-order SPLIT — Chinese gains confined to China-domestic OEMs while Western paid-converting programs qualify Japanese reducers because safety certification (Halos/IEC 61508) freezes the qualified BOM, qualification not price binding — appears in NO Tier-1 or list
Re-anchored per skeptic: the near-term supply leg rests on real new paper (Unitree's RMB 4.2B STAR IPO plus the count of follow-on embodied-AI CSRC registrations), not the immaterial UBTech unlock. The two genuinely differentiated legs stand: (a) MIIT's 'no code, no market access' platform functioning as forced consolidation toward the three MIIT-committee incumbents (Unitree, UBTech, Fourier) — at least 10 of the 100+ coded makers exit, merge, or fail to re-register models within 12 months; (b)
Earliest coverage: none (no consensus-grade source found for the specific second-order claims)
Grader confidence: 0.70 · engine conviction 0.55
The headline event (MIIT-linked Humanoid Full Lifecycle Management Service Platform, 29-digit IDs, 100+ firms, 28,000 units, 'no code no market access') was saturated Tier-adjacent coverage from 2026-05-27 onward (Xinhua/CGTN/SCMP/trade press), but protocol excludes underlying-event coverage. Leg (a) — the coding platform functioning as FORCED CONSOLIDATION toward the three MIIT-committee incumben
The quantified equity outcome, not the flow-down idea: Hesai's Americas revenue share declines vs prior-year quarter in BOTH of the next two quarterly reports, zero new named US/Americas robotics or infrastructure design wins are disclosed through 2027-04-11, and HSAI underperforms OUST by >15% over the window — while the market grades only the (irrelevant) direct DoD line.
Grader confidence: 0.80 · engine conviction 0.50
The edge is explicitly the quantified equity outcome, not the flow-down mechanism (engine concedes flow-down is trade-press consensus). Applying the frozen protocol literally: consensus requires the SAME specific second-order claim (same entity AND mechanism) in >=1 tier-1 or >=2 independent tier-2 sources at/before freeze (2026-07-12). Four targeted searches surfaced the adjacent landscape but no
Severity-gated, not count-gated (ISRG's base rate of routine recalls is high): elevated probability that by 2027-01-11 ISRG suffers (a) a NEW Class I recall on any da Vinci product or ANY new recall specifically on dV5/IS5000, or (b) a disclosed FDA Form 483/warning letter tied to the quality system, or (c) a dV5 placement-guidance walk-down in H2-2026 prints. Routine new Class 2/3 recalls on legacy platforms do NOT count as a win.
Earliest coverage: Amy J. Reed Collaborative (Northeastern Law) FDA petition for Class I recall + warning letter on da Vinci 5 (nearest-miss, NON-consensus) (2026-02-04)
Grader confidence: 0.58 · engine conviction 0.40
The engine's edge is SEVERITY-GATED and forward: elevated probability by 2027-01-11 of (a) a NEW Class I recall on any da Vinci product or ANY new recall on dV5/IS5000, (b) a disclosed FDA 483/warning letter tied to the quality system, or (c) a dV5 placement-guidance walk-down in H2-2026 prints, explicitly EXCLUDING routine legacy Class 2/3 recalls. I tested each specific mechanism against the fro
UBTech is the only L3 program on the Truth Board whose evidence has aged past the 180-day staleness threshold while its PR stays loud on a different, pre-delivery product line. The graded claim: the H1-2026 interim (Aug 2026) shows trade receivables + contract assets growing >1.3x faster than recognized humanoid revenue, with NO new dated primary-source Walker S2 delivery/reorder citation in the window, and 9880.HK underperforming Hang Seng Tech.
Earliest coverage: Humanoid.guide — "Ubtech humanoid revenue jumps 2200% in 2025" (closest generic receivables-risk coverage; does NOT match the specific mechanism) (2026-03)
Grader confidence: 0.78 · engine conviction 0.40
Searches confirm the general bull story (revenue ~+2200%, orders >RMB800M, MIIT/mass-production) and a GENERIC receivables-risk narrative (AR ~RMB1.3-1.8B, +40% YoY, govt-client payment delays, credit impairment ~RMB151M flagged as a risk factor) are both widely and independently covered before freeze. But the engine's edge is a distinct, forensic, forward-graded prediction about a report that doe
The specific sequencing gap: NO materially disclosed DoD/prime lidar award (>$20M) lands within 2 quarters because FY2027 procurement was budgeted before Section 164 compliance flowed down, so Q3/Q4 prints show bookings growth WITHOUT the federal kicker the multiple assumes — and the 2026-07-02 raise at $55.22 marks the local top. Asymmetric with our HSAI short: Hesai loses design-ins now, Ouster gets paid later — both legs datable.
Earliest coverage: No consensus-grade source matches the specific claim. Closest (insufficient — generic valuation caution, different mechanism): 24/7 Wall St / Cantor Fitzgerald cautious $33 target commentary in OUST forecast coverage (2026-07-03)
Grader confidence: 0.74 · engine conviction 0.35
The claim has three load-bearing components, none in consensus-grade form at/before the 2026-07-12 freeze: (1) procurement-sequencing MECHANISM — no material DoD/prime lidar award >$20M within 2 quarters because FY2027 procurement was budgeted before Section 164 compliance flowdown, so Q3/Q4 bookings grow WITHOUT the federal kicker the multiple assumes; (2) the datable call that the 2026-07-02 $20
The utilization-plus-pricing-power signal, graded through disclosed proxies (Chinese Q3 reports don't disclose utilization or ASPs): if the moat binds, the Q3-2026 print shows revenue roughly doubling with the capacity ramp WITHOUT year-over-year gross-margin compression — pricing power at the choke point through a 2x ramp, while consensus models assume price deflation. The digital-ID regime raises switching costs against unqualified rivals.
Earliest coverage: No consensus-grade source matches the specific entity+mechanism claim; the earliest CONTRARY (deflation) consensus datapoint is Morgan Stanley's China humanoid coverage carrying a ~15% 2026 ASP-decline / price-war assumption (2026-06-24)
Grader confidence: 0.82 · engine conviction 0.50
The engine's edge is a specific, contrarian, forward-gradeable second-order claim: that Leaderdrive's Q3-2026 print will show revenue ~doubling on the capacity ramp WITHOUT year-over-year gross-margin compression (pricing power at the choke point through a 2x ramp), while consensus models assume price deflation; plus that the digital-ID regime raises switching costs against unqualified rivals. Adv
The flow-down mechanism (not the headline ban) is the transmission — over the next 2-3 quarters, US commercial integrators with ANY federal exposure preemptively de-design Hesai (~43% global ADAS long-range lidar share, 10M-unit robotics backlog) to keep contract eligibility, so Hesai's disclosed non-China revenue share declines within two 6-K cycles even though nothing legally forces commercial customers out. Second: Chinese humanoid OEMs' compute stack gets pinned to domestic chips, showing fi
Earliest coverage: sUAS News — "Section 164 Is Here: The Complete Guide to NDAA-Compliant LiDAR in 2026" (closest, but NON-consensus and non-matching) (2026-07-01)
Grader confidence: 0.68 · engine conviction 0.60
The engine's edge is a specific, measurable second-order transmission claim on two axes: (1) COMMERCIAL Hesai customers with federal exposure preemptively de-designing to preserve their own contract eligibility (nothing legally forces them), visible as declining disclosed non-China revenue within two 6-K cycles; and (2) Chinese humanoid OEM compute-stack pinning to domestic chips surfacing FIRST a
The 2026H2-2027H1 pain lands on the INTEGRATOR layer, not the sensor layer — expect at least one disclosed program delay or cost overrun attributed to sensor requalification at a US defense-robotics prime/sub by mid-2027, and expect Ouster ASPs to RISE (scarcity pricing) rather than volumes to spike, capping how much of Hesai's 1.6M-unit/yr volume actually transfers. The trade is short the integration gap, not long the obvious winner.
Earliest coverage: AI CERTs News — 'Congress Moves on LiDAR: National Security vs Supply Chains' (closest coverage found, but non-consensus-grade: single newsletter/blog publisher, and does not match the specific entity+mechanism) (2026-07 (pre-freeze; generic supply-chain framing, no matching second-order claim))
Grader confidence: 0.70 · engine conviction 0.50
The consensus at freeze is exactly what the engine said it was: the compliant-lidar-winner trade (Hesai ban helps Ouster, one-player federal market, priced in after the raise and the ~270% run). The engine's edge is a contrarian SECOND-ORDER structural claim with two distinct, entity+mechanism-specific legs: (1) the pain migrates to the INTEGRATOR/defense-prime layer, manifesting as a disclosed pr
The conversion happens specifically in the Southeast right-to-work corridor (SC-GA-TX-TN), because union geography — not technology — is the binding conversion variable (KMWU blocking Korean plants; German works councils at home). Predict: the NEXT humanoid L2-to-L3+ promotion on the truth board is a US Southeast site, before any EU or Korean site, by mid-2027. Secondary: MP's Independence output gets pre-committed before 10X commissions — a disclosed non-GM, robotics-sector magnet offtake in 20
Grader confidence: 0.60 · engine conviction 0.70
Applied protocol literally. The engine concedes 'reshoring' and 'picks-and-shovels' are broad consensus and carves its edge as the specific mechanism (union geography = THE binding conversion variable), the specific locus (SC-GA-TX-TN right-to-work corridor), and a falsifiable directional prediction (next L2->L3 humanoid promotion is US Southeast before EU/Korea by mid-2027), plus a secondary MP M
The PIPE converts into a disclosed manufacturing agreement: Foxconn (or a Foxconn plant) becomes a named manufacturing partner for Digit or another Western humanoid by mid-2027, making Taiwan the default ex-China assembly geography for Western humanoids. The qualification moat is Foxconn's NVIDIA co-development (Isaac GR00T at Houston): whoever industrializes humanoid assembly first becomes hard to displace, exactly like phone assembly. Clean directional kill: Agility naming a non-Foxconn volume
Earliest coverage: BizTech Weekly (closest near-miss; does NOT establish consensus for the specific claim) (2026-06-25)
Grader confidence: 0.70 · engine conviction 0.60
Freeze date 2026-07-12. Engine concedes as consensus: Foxconn diversifying into EVs/robots, the PIPE being reported, and 'Taiwan supply chain wins from AI.' Its differentiated edge is the CONVERSION of the financial PIPE into a disclosed/named manufacturing agreement for Digit by mid-2027, plus Taiwan becoming the default ex-China assembly geography for Western humanoids, moated by Foxconn's NVIDI
Staked on the uncovered leg only: predict a SECOND ASEAN government social-insurance/rehab-robotics procurement (PERKESO expansion or a neighboring agency, with disclosed counterparty and value) by 2027-07-31 — establishing 'government social-security agencies' as a new paid-deployment customer class the Western truth board doesn't yet track, with the verified 2026-06-16/17 PERKESO centre as the template event.
Earliest coverage: Closest adjacent (does NOT match mechanism): The Star — 'Govt working to expand use of PERKESO rehab centres beyond just contributors'; underlying-event coverage earliest at Malay Mail 2026-06-16. No consensus source matches the specific customer-class / second-procurement claim. (2026-06-29)
Grader confidence: 0.80 · engine conviction 0.45
The engine deliberately stakes its edge only on the uncovered leg, not the widely-reported PERKESO event. Targeted searches confirm the underlying event is heavily covered (tier-1/tier-2: Malay Mail, Yicai Global, BERNAMA, The Star) but ALL of that coverage matches entity-plus-event, not the engine's specific mechanism. No source at/before the 2026-07-12 freeze (a) treats 'government social-securi
The transfer clamp and the litigation posture are connected symptoms of mark pressure: companies restrict secondary trading when prices below the last round would become visible (Forge quotes near ~$174 give a live metric). Graded claim: by 2027-02-28 Figure's BMW engagement STILL lacks disclosed commercial terms (stays ≤L2 on the truth board) and no priced up-round above $39B closes. Who's hurt: SPV/secondary buyers at or above the last mark, and crossover funds carrying it at $39B.
Earliest coverage: TechCrunch — "Figure AI sent cease-and-desist letters to secondary markets brokers" (Marina Temkin) (2025-04-29)
“Existing shareholders were trying to sell their stock at a price that was below the new hoped-for $39.5 billion valuation, these brokers said.”
Grader confidence: 0.70 · engine conviction 0.55
The edge's analytic core — transfer clamp as a symptom of mark pressure (companies restrict secondary trading when below-round prices would become visible) — was published for this exact entity and mechanism by TechCrunch on 2025-04-29, 15 months pre-freeze, and independently by Humanoids Daily in April 2025. The companion components were also published pre-freeze: BMW commercial terms undisclosed
Everyone's picks-and-shovels trade is Harmonic Drive and the Japanese reducer names, which already re-rated (actuation basket +10% rel90d). Schaeffler is the same BOM slice through a name NOT in any robot basket — and it gets paid by every humanoid program including the ones that fail, while its deployer side converts its own labor cost if the robots work. Conditioned per skeptic: Humanoid (UK) is a well-funded but pre-scale startup ($152M raise, wheeled-platform pivot), so the end-2026 German g
Earliest coverage: Reuters (syndicated via Yahoo Finance, BusinessDay, TimesLive): "Schaeffler sees humanoid robotics orders in three-digit million euros by 2030" (2026-05-05)
“Schaeffler is "collaborating with around 45 humanoid robotics players globally", "currently has five customer contracts in the segment", targets an "order book in the hundreds of millions of euros by 2030", and estimates "around 50% of humanoid robots' materials bill represents an addressable market".”
Grader confidence: 0.90 · engine conviction 0.55
The engine framed Schaeffler as an overlooked humanoid picks-and-shovels name "NOT in any robot basket" with robotics "absent from the narrative". Refuted on both premises before the 2026-08-04 freeze: (1) Tier-1 Reuters coverage on 2026-05-05 carried the CEO's own disclosure of ~45 humanoid engagements, five customer contracts, a hundreds-of-millions-euro order-book target and the ~50%-of-BOM add
The Ulsan union fight (at-risk board) does NOT hit Mobis symmetrically: actuator qualification, tooling, and development revenue attach to the program regardless of WHERE robots deploy, and initial volumes route to the US Metaplant lane the union can't block. Mobis is the 'gets paid either way' leg of the Atlas program — the structural position Schaeffler holds for European humanoids, at a Korea-discount price. Graded claim (framework-to-commitment conversion): Mobis discloses a robotics-actuato
Earliest coverage: Hyundai Mobis official press release via PR Newswire (CES 2026 announcement) (2026-01-07)
“Hyundai Mobis, which is responsible for humanoid components, is expected to operate the actuator plant. ... In Korea, the group's union has strongly opposed deploying robots at production sites, citing employment shocks. (Seoul Economic Daily, 2026-05-20; the Metaplant routing: "Hyundai Motor Group plans to expand Atlas deployment starting with Hyundai Motor Group Metaplant America (HMGMA) in 2028")”
Grader confidence: 0.95 · engine conviction 0.45
The edge fails at two levels. (1) The graded claim itself — Mobis discloses a robotics-actuator order, dedicated capex, or robotics business line by 2027-06-30 — was already fulfilled ~7 months before the 2026-08-04 freeze: Mobis announced the Boston Dynamics Atlas actuator supply agreement at CES 2026 (2026-01-07, its first official robotics customer), and the Mobis-operated US actuator plant (35
Two gradeable claims, tightened per skeptic: (1) Malaysia's heavy-RE compound export subline (HS 284690, not the whole 2846 chapter — avoiding an NdPr price artifact grading true for the wrong reason) inflects in H2-2026 vintages as the existing Dy/Tb/Sm circuits ramp — a trade-flow delta the engine's Comtrade grading scores directly; (2) magnet-stage capacity follows the oxide: within 12 months a magnet or alloy maker announces Malaysian capacity to co-locate with the only ex-China heavy feed —
Earliest coverage: Reuters (wire, newsml FWN4380XL; binding deal reported 2026-07-06, following the Lynas-JS Link MOU of July 2025), corroborated verbatim by TechNode Global 2026-07-07 (2026-07-06)
“JS Link will establish a magnet factory near Lynas' existing advanced materials plant in Kuantan ... Lynas will also supply rare earth materials to JS Link's existing magnet factory in Yesan, South Korea, and the planned Malaysian factory at commercial prices under an exclusive supply arrangement until January 2038.”
Grader confidence: 0.93 · engine conviction 0.60
Sub-claim 2 is not merely consensus — the predicted event already occurred before freeze: Lynas and JS Link signed an MOU in July 2025 and announced the binding co-location deal (A$50M equity, 3,000 tpa NdFeB plant sited near the Lynas Kuantan plant, exclusive feed supply to Jan 2038) on 2026-07-06/07, four weeks before the 2026-08-04 freeze, with Reuters wire coverage plus SCMP, Mining.com, Minin
An issuer only reaches for upstream-SPV void clauses when the secondary tape is printing marks it does not want established — the 2026-07-08 notice IS the price signal, not paperwork. July 2026 printed two clearing prices for humanoid evidence that did not exist at the last round: Agility, the only humanoid at L4 with SEC-filed committed orders of $300M+, listing at $2.5B pre-money; and Hyundai buying 100% of Boston Dynamics, an L4 with a decade of field fleets, at ~$3.3B on 2026-07-16. Figure c
Earliest coverage: Silicon Canals (completing source for the 2-Tier-2 rule; first of the pair was TechCrunch, 2026-07-05) (2026-07-07)
“Agility Robotics is going public at $2.5B while Figure AI sits at $39B on less proven ground — and the gap between those two numbers is the entire humanoid robotics thesis in one line”
Grader confidence: 0.76 · engine conviction 0.50
ADVERSARIAL FINDING — the engine's stated consensus baseline was already publicly false before freeze. The engine asserts consensus held that "Figure is the humanoid category leader and the most valuable private program at a ~$39B mark." TechCrunch published the opposite framing on 2026-07-05, twenty days BEFORE the 2026-07-25 freeze: it describes Figure as having "self-reported that it closed on
The mitigation the market is crediting cannot work, and this is checkable against statutory text rather than opinion: Section 164 names Hesai Technology as the covered ENTITY, so a Thai plant changes country of origin and nothing about the statutory bar, while FAR flow-down contaminates any US integrator's federal-nexus program. The granular, dated consequence is upstream of Hesai's own P&L: US robotics, AMR and drone integrators carrying federal or Build America Buy America revenue must re-qual
Earliest coverage: FY2025 NDAA Section 164 statutory text (primary government document), as quoted verbatim in Covington's Inside Government Contracts alert "FY2025 NDAA: Congressional Efforts to Bolster U.S. Resilience Against Chinese Tech and Influence" (2025-02-14)
“"Under section 164, effective June 30, 2026, DoD is prohibited from operating, procuring, or using certain LiDAR (Light Detection and Ranging) technology that is manufactured, developed, or reliant on software, network connectivity, or data storage associated with China, Russia, Iran, or North Korea." — and, in the same alert, the provision "expressly prohibits LiDAR developed by Hesai Technology, headquartered in Shanghai, China," while "DoD's use of any system that incorporates or interfaces with such LiDAR technology is also prohibited."”
Grader confidence: 0.55 · engine conviction 0.40
Adversarial read, applying the frozen rule literally. The engine's edge rests on two load-bearing mechanism claims: (1) Section 164 attaches to Hesai as a named ENTITY/developer, so relocating manufacture to Thailand cannot cure the statutory bar; (2) the bar contaminates downstream US integrators with a federal nexus, forcing sensor re-qualification. Both are stated, not merely implied, in the st
The backlog's second pillar and the selling shareholder are the same counterparty — the specific thing nobody is underwriting. The SoftBank-controlled GreenBox JV must FUND its warehouse buildout from the same sponsor that has sold Symbotic equity at successively LOWER prices ($55.00 in Dec-2025, then ~$50.415 on 2026-05-27, cutting the aggregate stake to ~31.3%) and has now exited robotics entirely via the 2026-07-16 Boston Dynamics sale. So the test is not 'is the backlog big' but 'is GreenBox
“"GreenBox has yet to sign any unrelated customers more than two years after its formation, describing the arrangement as 'the type of circular transaction that investors are increasingly scrutinizing.'" (same note: "Over half of Symbotic's $22.5 billion backlog—approximately $11.6 billion—consists of sales to this JV, which is 35% owned by Symbotic and 65% by SoftBank." and "System shipments to GreenBox will likely generate limited initial cash flow".)”
Grader confidence: 0.85 · engine conviction 0.35
REFUTATION SUCCEEDS. The engine's stated consensus ("$22.7B backlog de-risks years of revenue; SoftBank's selling says nothing about the business") is a straw man that did not hold at freeze. Eight months before the 2026-07-25 freeze, Goldman Sachs — a bulge-bracket sell-side note, explicitly Tier-1 under protocol branch (a) — downgraded SYM to Sell with ~45% downside on exactly the mechanism the
RE-CENTRED FROM 'CHINA' TO UNITREE, because the engine's own Truth Board already carries UBTech Walker S2 at L3 with ~RMB 800M of disclosed orders at named industrial customers — so 'the Truth Board refuses to count China' is false as written. The defensible claim is narrower: Unitree's post-IPO disclosures will show research, education and government as the MAJORITY buyer mix while ASP per humanoid unit falls QoQ for two consecutive quarters — units up, revenue per unit DOWN. The mechanism is s
Earliest coverage: The Robot Report — "Unitree IPO shows a real hardware business, but the humanoid case is still early" (Deepak Jayaraj), reporting Unitree's own IPO prospectus (2026-03-25)
“73.6% of humanoid revenue came from research and education in the first nine months of 2025. Another 17.4% came from commercial consumption uses such as demonstrations and display environments. Only 9.01% came from industrial applications. ... the average selling price fell from about $85,000 (¥593,400) in 2023 to about $25,000 (¥167,600) in the first nine months of 2025.”
Grader confidence: 0.84 · engine conviction 0.52
REFUTATION SUCCEEDS on the load-bearing, observable part of the edge. The engine's defensible claim is stated as: Unitree's post-IPO disclosures will show research, education and government as the MAJORITY buyer mix while ASP per humanoid unit falls — "units up, revenue per unit DOWN." Both legs were already public and consensus-grade four months before the 2026-07-25 freeze, sourced to Unitree's
The sector now has a public clearing price for humanoid evidence: Agility — the ONLY program with disclosed committed revenue ($300M, SEC-filed) and an L4 rating — is listing at $2.5B pre-money. Figure carries ~15x that mark on evidence two rungs lower with zero disclosed payment. The countersuit converts the safety-culture question from rumor into checkable court-docket events with dates.
Earliest coverage: Silicon Canals (earliest statement of the specific mechanism); branch-(b) two-publisher threshold crossed 2026-07-11 by TechTimes ("Agility Robotics Nears Public Debut: Safety Cert, S-4, and a 34x Valuation") (2026-07-07)
“Agility Robotics is going public at $2.5B while Figure AI sits at $39B on less proven ground — and the gap between those two numbers is the entire humanoid robotics thesis in one line”
Grader confidence: 0.70 · engine conviction 0.50
The engine's edge decomposes into (i) Agility's $2.5B SPAC listing with $300M SEC-filed committed orders as the sector's public evidence-anchored price, (ii) Figure carrying ~15x that mark on weaker evidence with no disclosed payment, and (iii) the countersuit making the safety question checkable in court. Component (i) was saturation-covered from the 2026-06-24 announcement including in primary S
Precision-fixed: the RMB 800M figure is cumulative Walker-SERIES orders since early 2025 per the 2025-11-17 PR, and that same PR claimed a first batch of 'several hundred' delivered S2 units — so the sharpened test is not 'no deliveries claimed' but 'no recognized revenue disclosed against the claimed deliveries.' The interim report must show Walker-attributable recognized revenue consistent with several-hundred delivered units plus the RMB 250M and RMB 159M named contracts; delivery claims with
Earliest coverage: UBTech Robotics FY2025 Annual Results Announcement (filed/published via HKEXnews disclosure platform) (2026-03-31)
“revenue from full-size embodied intelligent humanoid robot products and services grew rapidly by approximately 2,203.7%, from RMB35.6 million for the year ended December 31, 2024, to RMB820.6 million for the year ended December 31, 2025, becoming UBTECH's largest source of revenue ... achieved revenue of approximately RMB820 million ... and achieved sales volume of 1,079 units”
Grader confidence: 0.78 · engine conviction 0.45
The edge's specific second-order claim is a forward test on UBTech (entity) with mechanism = reconciliation of claimed Walker deliveries against recognized revenue ("delivery claims without matching revenue recognition is the falsifiable tell," to be checked in "the interim report"). That exact reconciliation was already published before the 2026-07-17 freeze: UBTech's FY2025 annual results announ
Anchored to what is externally verifiable: a Class I recall (Z-1855-2026, 1 death/4 serious injuries, March 2026 stop-use) plus at least four additional 2026 FDA recall entries across the da Vinci fleet including the flagship dV5 — an accelerating cadence the market ignored while grading the record placement quarter. The claim that root causes are migrating to device design rests on the engine's own ledger reads of the FDA cause fields; stated as ledger-sourced, with the openFDA trigger (next re
Earliest coverage: Simply Wall St (syndicated on Yahoo Finance): "Intuitive Surgical Recalls Put da Vinci Revenue Reliability Under Scrutiny" (earliest recall-pattern-to-revenue-risk framing; Tier-1-grade market coverage of the recall repricing followed 2026-05-05/11) (2026-04-24)
“The recalled instruments "sit at the heart of the recurring instruments and accessories business that made up about 86% of Q1 2026 revenue." [and, establishing the market had priced the recall risk by 2026-05-11:] "Intuitive Surgical shares dropped 6.67% to $420.06, hitting a 52-week low after an FDA Class I recall and reduced 2026 procedure-growth forecasts."”
Grader confidence: 0.60 · engine conviction 0.25
The engine's differentiation premise — that the market treated 2026 recalls as routine, fully absorbed field corrections and priced an uninterrupted dV5 cycle — is contradicted by the pre-freeze public record: ISRG fell ~6.6% to a 52-week low on 2026-05-11 explicitly on the Class I classification (1 death/4 serious injuries), stayed at 52-week lows into mid-July (down ~28-30% YTD), and investor me
The only Western industrial with a production-ready, in-house-manufactured 60-250 Nm integrated humanoid actuator AND first-party humanoid deployment economics (hundreds of units planned in its own plants, plus paying-customer status on Digit), trading entirely on auto-cycle multiples outside the actuation basket that rerated 19.5%. First disclosed actuator order/revenue is the trigger, not the premise.
Earliest coverage: Reuters — "GRAPHIC: How Schaeffler's bet on humanoids is beating auto sector blues" (syndicated copy on Sahm Capital) (2026-02-12)
“Schaeffler has emerged as a rare humanoid-robotics play in European stock markets, shifting the narrative around this high-precision auto parts supplier. ... Schaeffler trades at a 94% premium to the broader auto sector, per LSEG data, underscoring investors' willingness to pay up for its humanoid-robot ambitions.”
Grader confidence: 0.95 · engine conviction 0.45
The edge fails on both legs. (1) The specific second-order claim (Schaeffler = the Western industrial humanoid-actuator play with in-house actuators plus first-party deployment economics) was Tier-1 consensus by 2026-02-12: Reuters explicitly framed Schaeffler as "a rare humanoid-robotics play in European stock markets," cited its plan to integrate several hundred humanoids into its own production
Captive ownership lets Atlas SKIP the paid-pilot chasm that traps every other Western humanoid at L2 — internal deployment needs no disclosed commercial terms. Hyundai's published roadmap says parts-sequencing DEPLOYMENT starts 2028, so the non-consensus claims are: (a) quantified fleet metrics (units/hours at HMGMA) disclosed by CES 2027 — ahead of the public 2028 schedule, exactly because captive deployment needs no commercial terms; and (b) a named Korean actuator/bearing/reducer supplier pub
Earliest coverage: Hyundai Mobis / Boston Dynamics strategic collaboration announcement at CES 2026 (PRNewswire + bostondynamics.com); earliest independent news coverage: The Korea Herald (2026-01-07)
“Hyundai Mobis will provide actuators for Atlas — the humanoid robot developed by Boston Dynamics, Hyundai Motor Group's US-based robotics unit — when the model enters mass production.”
Grader confidence: 0.85 · engine conviction 0.65
The edge fails on both prongs. Prong (b) — 'a named Korean actuator/bearing/reducer supplier publicly designed into the Atlas BOM within 12 months, the naming event nobody has made' — is factually stale at freeze: Hyundai Mobis was publicly announced as the actuator supplier for next-gen Atlas at CES 2026 (Jan 7, 2026), roughly six months before the 2026-07-17 freeze, with the exact entity (Hyunda
The specific non-obvious claim: the German AUTOMATION-MACHINERY layer (not the automakers) is the impact point, because German auto OEMs' humanoid/physical-AI qualification decisions are being made in China (Walker S2 at FAW-VW) and the US (Figure at BMW Spartanburg) — zero German integrators appear anywhere on the 9-program truth board. We claim a German OEM announces a EUROPEAN-plant humanoid pilot with a non-German (Chinese or US) vendor before any German-vendor humanoid pilot exists, cementi
Earliest coverage: Reuters (syndicated via Investing.com): "Mercedes-Benz takes stake in robotics maker Apptronik, tests robots in factories" (2025-03-18)
“The carmaker is testing use of the robots at its Digital Factory Campus in Marienfelde, Berlin, and in Kecskemet, Hungary”
Grader confidence: 0.88 · engine conviction 0.60
The edge's centerpiece forward call — "a German OEM announces a EUROPEAN-plant humanoid pilot with a non-German (Chinese or US) vendor before any German-vendor humanoid pilot exists" — was not a prediction at freeze (2026-07-17): it had already happened, twice, with tier-1 coverage. (1) Reuters 2025-03-18: Mercedes-Benz (German OEM) testing US-based Apptronik's Apollo at Berlin-Marienfelde (Europe
Consensus assigns the win to Australia (the miner's flag). The edge: the PROCESSING rent physically accrues in MALAYSIA — Kuantan is where every Western magnet-maker's ex-China Dy/Tb chain must touch, giving Malaysia chokepoint-adjacent leverage nobody prices. Gradeable claims: (a) Lynas heavy-RE output scales past ~15 t/quarter within 12 months (vs 9 t in the first quarter of production); (b) a magnet-making or actuator-component investment is announced co-located in Malaysia within 18 months a
“JS Link will establish a magnet factory in Kuantan, Malaysia, with an operating capacity of 3,000 tonnes per annum of neodymium-iron-boron (NdFeB) permanent sintered magnets.”
Grader confidence: 0.85 · engine conviction 0.60
The edge fails the novelty test on its own gradeable claims. Claim (b) — a magnet-making investment announced co-located in Malaysia as buyers move to the Dy/Tb source — was a completed, tier-1-covered event BEFORE the 2026-07-17 freeze: the Lynas–JS Link Kuantan magnet-factory deal was signed 2026-07-07 (Reuters wire; also SCMP, Nikkei, NST, FMT), following a July 2025 MOU. Claim (c) — Malaysia e
Single-digit utilization of the humanoid lines through 2026 and a humanoid-segment revenue shortfall visible in the Q3-2026/FY-2026 reports, with rare-earth magnet input costs biting first — the specific downstream-supplier casualty of the sector's most-hyped L1 program. The Truth Board's L1 grade plus the CEO's own 2026-07-01 anti-raise is a direct, dated contradiction of the volume assumptions in Tuopu's capex.
Earliest coverage: Sina Finance / 野马财经 (syndicated to Eastmoney): "拓普集团的'进与退'：295亿营收新高却遭利润首降，机器人百万台产能面临'证伪'考验" (2026-04-09)
“意味着公司为机器人业务超前建设的产能利用率仅约0.11%，几乎全部闲置（This means the utilization rate of the company's forward-built production capacity for robotics was only approximately 0.11%, almost entirely idle）”
Grader confidence: 0.87 · engine conviction 0.45
The engine's edge is the company-level, dated second-order consequence: single-digit utilization of Tuopu's humanoid lines, a humanoid-segment revenue shortfall, and a direct contradiction of the volume assumptions baked into Tuopu's capex/capacity story. That exact second-order claim (same entity Tuopu, same mechanism: idle humanoid-executor lines + segment revenue shortfall refuting the million-
BMW commercial terms will STILL not be disclosed within the window (keeping Figure at L2 while Agility banks L4 revenue), and the countersuit path means safety-record discovery risk lands before any safety certification does — pressuring the next raise/secondary marks. Secondary-market prints (Forge/EquityZen) are the falsifiable tape for a private name.
Earliest coverage: TechMarketBriefs — Figure AI pre-IPO analysis ($39B Valuation, Risks & Bull Case) (2026-04-28)
“The case is unresolved and the dollar exposure is small, but the optics are not: a humanoid robotics company being sued for cutting safety corners to close a $39 billion round, in a sector that depends on regulators and customers trusting the safety story, is a structural overhang on any IPO marketing.”
Grader confidence: 0.62 · engine conviction 0.45
The engine's edge is a compound claim about Figure AI. Its load-bearing components were already consensus-grade at freeze: (i) BMW commercial terms undisclosed / Figure stuck at pilot while Agility books commercial revenue is the dominant framing in multiple independent 2026 trade sources (Silicon Canals 2026-07-07 'Agility going public at $2.5B while Figure sits at $39B on less proven ground'; TM
The LINEAR leg, which is still open (the April-2026 Hexagon rotary deal does NOT count — it pre-exists the freeze): the roller-screw bottleneck binds BEFORE the harmonic/rotary one as western humanoid programs scale, and Ewellix is one of ~3 qualified western sources. By 2027-07-11 Schaeffler discloses a NEW humanoid roller-screw/linear-actuation supply agreement or design win beyond Hexagon (new counterparty or explicitly linear scope), or cumulative Ewellix roller-screw capacity capex >= EUR 1
Earliest coverage: GlobalBankingAndFinance, 'Schaeffler's bet on humanoids beating auto-sector blues' (carrying UBS/BofA sell-side notes); corroborated by Schaeffler 'Humanoids at Schaeffler' IR presentation (2026-02-05) and MarketScreener 'Schaeffler: Powering Humanoid Motion' (2026-01-12) (2026-02-12)
“Schaeffler, which reports on March 3, trades at a 94% premium to the broader auto sector, per LSEG data, underscoring investors' willingness to pay up for its humanoid-robot ambitions.”
Grader confidence: 0.62 · engine conviction 0.55
The edge's stated consensus premise ('Schaeffler priced with zero robotics premium') is demonstrably false at freeze: Schaeffler traded at a 94% premium to the auto sector for its humanoid ambitions, with bulge-bracket sell-side (UBS, BofA) explicitly valuing the humanoid business (tier-1 branch a). Matching entity (Schaeffler/Ewellix) AND mechanism (roller-screw/linear actuation for humanoids): S
The listed-equity arbitrage is datable: the $325M put print FIXES a public mark ($3.4B) at 1/10th of the private-market comp for a demonstrably superior evidence position (a real L4 + Atlas with a captive deployment site). Catalysts inside the window: RMAC opening, first dated Atlas production-work citation (L1->L2/L3), Stretch reorder flow. Hyundai is the cheapest truth-heavy robotics exposure on any exchange. Caveat carried: BD at $3.4B is small vs Hyundai's cap, so the re-rate requires the na
Earliest coverage: UPI — "KB Securities boosts Hyundai Motor target on Boston Dynamics robots" (2026-01-21)
“KB Securities kept its investment rating at buy and set a target price of 800,000 won, up 158% from its previous target. Analyst Kang Sung-jin said Boston Dynamics could be a turning point for Hyundai Motor's productivity efforts and argued the company has built a structural foundation to move beyond a traditional automaker model.”
Grader confidence: 0.70 · engine conviction 0.50
The engine's edge decomposes into (i) a core investable second-order claim — Hyundai's listed equity is the cheapest/underpriced way to get Boston Dynamics/humanoid exposure via a valuation gap vs private pure-plays — and (ii) a distinctive numeric framing that the $325M put print FIXES a low public mark ($3.4B) at ~1/10th of private comps, plus catalysts (RMAC, Atlas L1->L2/L3, Stretch reorders).
The sequencing claim (disbelief premise dropped per the tape): third-party functional-safety certification (NVIDIA Halos/TUV-style) for Digit is announced by 2027-04-11 AND the final S-4/proxy confirms >= $250M committed orders — converting milestone-gated backlog into recognized RaaS revenue before any competing humanoid program books its first disclosed paid external dollar, with Foxconn de-risking the 10k/yr ramp. Redemption/dilution mechanics are the real risk; the triggers watch the proxy t
Earliest coverage: SEC Form 425/8-K — Churchill Capital Corp XI / Agility Robotics business-combination filings (deal announcement), corroborated by TechTimes "Agility Robotics Nears Public Debut: Safety Cert, S-4, and a 34x Valuation" (2026-06-25)
“The contractual milestone that gates the bulk of that backlog is Digit v5 passing the third-party industrial safety certification required before it can operate cooperatively alongside human workers without physical barriers.”
Grader confidence: 0.68 · engine conviction 0.45
The edge's specific second-order claim decomposes into: (1) third-party functional-safety cert converting milestone-gated Digit v5 backlog into recognized RaaS revenue; (2) the S-4/proxy confirming >=$250M committed orders; (3) Foxconn de-risking the 10k/yr ramp; (4) redemption/dilution as the real risk; (5) Agility reaching recognized external RaaS revenue ahead of competitors. Adversarial search
Safety-certification regimes make F/T sensing MANDATORY content, so certified-sensor content per humanoid rises with every Truth-Board L2->L3+ transition — and ATI/Varo is the only western humanoid-specific six-axis F/T product from an incumbent with existing robot-OEM qualifications. We claim a disclosed humanoid design win or humanoid-driven ATI growth callout within 12 months, giving NOVT the sensing-layer re-rate lidar names already received — from a base where the market embeds zero. Displa
Earliest coverage: Morgan Stanley — "The Humanoid 100: Mapping the Humanoid Robot Value Chain" (Novanta/NOVT listed in the humanoid value chain) (2025-02-04)
“The Robot Report (2025-07-28) features "Nick Damiano from Novanta Robotics and Automation discussing safety in humanoid robotics and component-level safety ratings," noting Novanta offers "drives, encoders, motors, and force torque sensors" for humanoids. (Morgan Stanley Humanoid 100 PDF confirmed via multiple independent search results to list Novanta/NOVT as of 2/4/2025; verbatim PDF fetch timed out.)”
Grader confidence: 0.83 · engine conviction 0.45
The edge's premise — that NOVT is seen purely as a medtech/photonics compounder with 'no humanoid narrative' and the 'market embeds zero' humanoid F/T sensing — is refuted by consensus-grade sources predating the 2026-07-12 freeze. Entity match (Novanta/NOVT/ATI) and mechanism match (humanoid force/torque sensing exposure as an investment beneficiary) both appear: (1) Morgan Stanley's Humanoid 100
Japanese reducer makers capture only the ex-China slice of humanoid actuation while priced for the global slice, AND lose incremental Western design slots. Restated honestly: Chinese reducers now sit inside qualified actuator assemblies at Optimus (Shuanghuan via Tuopu/Sanhua) and are named in Figure's chain (Zhongda Leader) — the qualification moat binds per-program, and once passed it does not un-pass (domestic-Chinese harmonic share ~15% in 2022 to ~38% by early 2024 and rising). Expect Harmo
Earliest coverage: Jefferies (bulge-bracket sell-side downgrade note), reported by Investing.com — 'Jefferies downgrades Harmonic Drive Systems stock on competitive concerns' (2025-10-15)
“Jefferies cited "growing concerns about the company's competitive position against local players in China, particularly in the humanoid and collaborative robot sectors," and cautioned that "current valuation levels suggest the market is overestimating the company's fundamental strengths" (new price target JPY2,400, >20% downside).”
Grader confidence: 0.80 · engine conviction 0.55
The engine's second-order claim is that Japanese reducer makers (specifically Harmonic Drive Systems, 6324) are priced for global humanoid-actuation exposure but will lose the China slice and incremental Western design slots to already-qualified Chinese reducers, so their premium valuation is unjustified and Greater-China orders disappoint. The engine frames consensus as the uniform opposite ('Jap
Two things a Bloomberg reader has not internalized: (1) the digital-ID platform is the first machine-readable deployment CENSUS anywhere — predict its officially disclosed coded-unit count exceeds 60,000 by 2026-12-31 and 100,000 by 2027-06-30 (from the 28,000 baseline), and the engine should ingest it as a graded truth-mass series; (2) the winners are not the OEM brands (whose float is unlocking into the mandate — UBTech full circulation, Unitree IPO) but the domestic BOM layer the mandate forc
Earliest coverage: Morgan Stanley China humanoid shipment forecast (via SCMP/CNBC), reinforced by KraneShares "Humanoid Robotics in 2026: Pilot to Platform" picks-and-shovels supplier thesis (2026-06-24)
“Morgan Stanley raises China humanoid robot shipment forecast to 50,000 units”
Grader confidence: 0.68 · engine conviction 0.65
The edge bundles two second-order claims both billed as "not internalized by a Bloomberg reader." (Prong 2 — BOM/component suppliers outperform the humanoid OEM brands) is demonstrably ALREADY-CONSENSUS at freeze: multiple independent pre-2026-07-12 sources — KraneShares, Rare Earth Exchanges, Pandaily (2026-06), Tianxia Gongchang (2026-07-07) — carry the "picks-and-shovels / component suppliers m
Humanoid ASSEMBLY (not just components) begins commercial operation in Bac Ninh on roughly the filed schedule — a checkable, dated, primary-source event most robotics investors have never heard of because it's buried in a Vietnamese provincial licensing filing. If the Nov 2026 milestone hits within a quarter's slip, Vietnam leapfrogs India/Mexico as the designated low-cost node of the Western-aligned humanoid supply chain, with follow-on robot-line licenses in 2027.
Earliest coverage: TechNode Global — "Foxconn to manufacture humanoid robots in Vietnam" (2026-04-21)
“manufacture humanoid robots in its factory in Vietnam's northern province of Bac Ninh ... enter trial production in September 2026, and start official production from November 2026”
Grader confidence: 0.87 · engine conviction 0.50
The engine's edge is a checkable, dated, primary-source event: Foxconn/Fushan humanoid production at Bac Ninh beginning commercial operation on the filed Nov 2026 schedule, sourced from an April 2026 Vietnamese provincial (VSIP Bac Ninh) environmental permit renewal filing. The engine asserts this is 'buried' and unknown to most robotics investors. That novelty premise is refuted at freeze. TWO in
Time-to-recognition — was the engine early?
quantum · protocol v1 · 2026-08-26Nothing here is a “lead” yet. This is the backward pass: for each frozen edge, did consensus-grade coverage of the specific second-order claim already exist at freeze? An edge that clears that bar is differentiated at freeze — a necessary condition, not proof of value. It becomes a proven-lead only when consensus actually arrives later (lead-days = recognition − freeze). Those accrue over 12–24 months.
Consensus-grade (frozen v1) = the same entity+mechanism claim in ≥1 tier-1 source OR ≥2 independent tier-2 sources. Conservative bias: already-consensus when uncertain — the grader’s job is to refute the engine’s novelty. All 12 already-consensus rows carry the exact quote that sank them. Rules were frozen (RECOGNITION_PROTOCOL.md v1) before this data.
Rigetti does not publish a dated, third-party-checkable >=99.5% median 2Q on a 108Q system by 2026-12-31 (second consecutive slip on the same metric in the Roadmap ledger). Independently of cash need, a new S-3/424B/ATM sales-agreement filing appears in EDGAR within 120 days of the slip date, and RGTI trails px.quantum_pureplay_proxy by >15% over that window. The fidelity leg is graded on its own; the raise leg is the second, separately graded claim.
Grader confidence: 0.70 · engine conviction 0.55
The underlying facts are widely covered: Rigetti's Jan-2026 push of Cepheus-1-108Q GA to end-Q1 2026 (Investing.com, StreetInsider, Quantum Insider, company 8-K), GA at 99.1% median 2Q fidelity in April 2026, and a reaffirmed target of ~99.5% "later in 2026 / by year-end" (Q2-2026 results 2026-08-06; Investing.com earnings preview 2026-08-06 calls 99.5% by year-end "the harder and more relevant pr
Post-close 8-K discloses redemptions above 70% AND pro-forma cash to Pasqal below $250M (half the no-redemption headline). PSQL closes below the $10 reference 60 trading days after listing while the neutral-atom narrative is intact (no QuEra debunk). Pasqal is NOT moved to Q3+ on the Board by 2026-12-31 — the first 10-Q, not any science result, is the repricing catalyst.
Earliest coverage: Bloomberg, "Quantum Startup Pasqal to List Via SPAC at $2 Billion Valuation" (2026-03-04) — closest Tier-1 coverage found; it covers only the generic redemption contingency, NOT the engine's specific claim. No consensus-grade source of the specific claim found. (2026-03-04)
Grader confidence: 0.70 · engine conviction 0.50
Timeline check: the Bleichroeder Acquisition Corp. II (BBCQ) extraordinary general meeting on the Pasqal combination was scheduled for 2026-08-25 (Pasqal/GlobeNewswire 2026-08-20; Form F-4 declared effective 2026-08-05; DEFM14A on EDGAR), i.e. one day AFTER the freeze date. Therefore no post-close 8-K, redemption tally, PSQL trading history, or 10-Q existed at freeze; the engine's claim is a forwa
Between now and 2027-03-31 the CMVP validated-modules list gains fewer than three additional FIPS 140-3 Level 3 HSMs with ML-KEM+ML-DSA inside the boundary (moat holds through the CNSA 2.0 2027-01-01 default), AND at least one US federal or NATO-member code-signing/PKI award naming Luna T7 or Luna T-Series is published (SAM.gov/TED/agency release) by 2027-03-31, AND Thales' FY2026 results (~2027-03) explicitly cite PQC/HSM demand in the Cybersecurity & Digital Identity segment. No THLEF-versus-b
Earliest coverage: Intelligence Community News / ExecutiveBiz / Thales TCT press release (underlying event only — Luna T-Series first PQC-enabled US-made HSM to achieve FIPS 140-3 Level 3; NOT the second-order claim) (2026-08 (event coverage; no consensus-grade coverage of the edge's specific claim found at/before 2026-08-24))
Grader confidence: 0.60 · engine conviction 0.55
The edge is a compound, falsifiable second-order claim: (1) CMVP validated list adds <3 rival FIPS 140-3 L3 HSMs with ML-KEM+ML-DSA in-boundary by 2027-03-31 (scarcity moat through the CNSA 2.0 2027-01-01 default); (2) a US federal/NATO code-signing/PKI award naming Luna T7/T-Series is published; (3) Thales FY2026 results explicitly cite PQC/HSM demand in Cybersecurity & Digital Identity. Searches
Between now and 2027-03-31, FormFactor's Systems segment posts y/y revenue growth in at least two of the next three reported quarters AND management explicitly cites quantum/cryogenic orders on two consecutive earnings calls, with the pipeline traceable to the verified fab conversions (IBM Anderon, HRL, GF QTS, Rigetti Fab-1) rather than to pure-play system counts. No basket-relative price leg below a 15% threshold (noise).
Grader confidence: 0.60 · engine conviction 0.45
Searches surfaced abundant coverage of the headline events (FORM Q2 2026 record revenue, Systems revenue $48.5M +43.9% y/y; Flatiron dilution refrigerator launch; IBM's HRL acquisition announced 07-23, closed 08-26; CHIPS quantum foundry awards to IBM/GF/Rigetti). But every source that explains Systems growth attributes it to engineering-prober recovery and co-packaged optics/silicon photonics, no
OCI Helios managed service is GA (dated Oracle or Quantinuum release) by 2027-03-31, AND the first US federal or allied-government (Five Eyes/NATO/Japan) procurement notice or award that specifies error-corrected/logical-qubit access is awarded through OCI rather than AWS Braket, Azure Quantum or Google, by 2027-06-30. Impact entity is the procurement channel and Quantinuum's trapped-ion suppliers, not ORCL's P&L; no ORCL price leg.
Grader confidence: 0.70 · engine conviction 0.40
The edge is a compound second-order claim: (1) OCI Helios managed service reaches GA by 2027-03-31 and (2) the first US/Five Eyes/NATO/Japan government procurement specifying error-corrected/logical-qubit access is awarded via OCI rather than AWS Braket/Azure Quantum/Google by 2027-06-30, with the impact entity being the procurement channel and Quantinuum's trapped-ion supply chain. All coverage f
By 2027-12-31 PsiQuantum, the Queensland government or the Commonwealth publicly re-dates Brisbane utility-scale operations to 2028 or later in a primary source AND that source attributes the delay to chip/wafer supply or foundry cadence (GlobalFoundries photonics capacity, wafer allocation, Malta ramp) rather than to error-correction results or Brisbane construction. Attribution to physics or site works = MISS even if the date slips; a Chicago-first re-sequencing with Brisbane held at 2027 = MI
Grader confidence: 0.70 · engine conviction 0.55
The headline slip is well covered pre-freeze: Startup Daily (2025-12-09) said "The timelines for both computers is now closer to 2030" and Forbes Australia (2026-05-20) said the project "has run months behind schedule"; InnovationAus/Business News Australia/Capital Brief covered the Brisbane Airport-to-Moreton Bay relocation, and one summary says the end-2027 target "has now been pushed out to 202
The moat is the CMVP QUEUE, not the product: FIPS 140-3 validations take on the order of 18-24 months in the CMVP pipeline, so rivals (Entrust, Utimaco, cloud-HSM silicon) physically cannot field an equivalently validated CNSA 2.0 HSM before the 2027-01-01 acquisition default — making Thales TCT effectively single-source for compliant NSS HSM procurement for roughly the mandate's first year. Claim: US federal/NSS HSM awards in 2026H2-2027 concentrate visibly to Luna-family devices.
Earliest coverage: Closest non-qualifying antecedent: PostQuantum.com (Marin Ivezic blog), 'CNSA 2.0 2027 Procurement Gate' — mechanism only (CMVP queue arithmetic), no entity, single blog, not tier-listed (2026-05-14)
Grader confidence: 0.70 · engine conviction 0.65
The underlying event was consensus by freeze: Thales TCT's 2026-07-29 PR claiming the first FIPS 140-3 validation containing all CNSA 2.0 PQC algorithms was echoed by Intelligence Community News and ExecutiveBiz before 2026-08-05. But the engine's edge is the second-order claim — that CMVP queue length (18-24 months) physically bars Entrust/Utimaco/cloud-HSM rivals from an equivalently validated C
The specific detail: Pasqal co-signed the ~50-author neutral-atom coalition roadmap (arXiv:2607.21554, 2026-07-23) that commits it to the SAME shared milestones as QuEra/Infleqtion/planqc — publicly erasing the technical-differentiation story a $2.0B standalone valuation needs, while the capability ladder places it a full rung behind QuEra. Evidentiary tightening per skeptic: the frozen record must cite the actual author/affiliation list of arXiv:2607.21554 showing Pasqal signatories (the abstra
Earliest coverage: None found making the specific second-order claim; nearest event-only coverage: The Quantum Insider, 'Neutral-Atom Researchers Lay Out Industry-Wide Roadmap Toward Practical Quantum Computing' (2026-07-27)
Grader confidence: 0.70 · engine conviction 0.60
Skeptic gate PASSED first: the abstract page hides affiliations, but the arXiv HTML full text (arxiv.org/html/2607.21554v1) shows three authors affiliated 'PASQAL SAS, 24 Rue Emile Baudot, 91120 Palaiseau, France' (Alexandre Dauphin, Lucas Lassabliere, Pascal Scholl), alongside QuEra (Tout Wang, Alexander Schuckert), Infleqtion (Mark Saffman), and planqc (Johannes Zeiher) — 56 authors, submitted 2
Tightened per skeptic: IonQ closing SkyWater (2026-07-31) removes SkyWater as a plausible fab for every ion program IonQ doesn't own — no rival sanely fabs its core IP at a competitor-owned foundry. That makes Infineon the default MERCHANT trap-chip foundry for non-IonQ ion programs that lack captive fab capacity — explicitly acknowledging Quantinuum's Honeywell in-house microfab heritage as the live alternative path. Q4-leader Quantinuum's Apollo fault-tolerant promise (dated 2027-12-31 in the
Grader confidence: 0.60 · engine conviction 0.60
The specific second-order claim is: SkyWater's competitor-ownership (closed 2026-07-31) makes Infineon the default MERCHANT trap-chip foundry for non-IonQ ion programs, predicting >=2 additional ion-trap foundry partnerships/expansions at Infineon within 12 months. Adversarial search found the mechanism half fully covered at Tier-1 before freeze: the FTC Ferguson statement (2026-07-31) analyzes fo
Tightened per skeptic: HRL is the leading PUBLISHED cryo-CMOS QEC control team, and post-close that capability is off the market — superconducting rivals (Google, Rigetti) face a multi-year build-vs-buy decision with only a handful of independent cryo-CMOS teams (SemiQon, Equal1) left to buy. Claim: rival roadmaps are forced to publicly address cryo-control within 12 months — announcements or acquisitions of the remaining cryo-CMOS teams are the gradeable trace (countable trigger: ≥1 rival cryo-
Earliest coverage: postquantum.com (closest non-qualifying analysis; no qualifying consensus source found) (2026-08-02)
Grader confidence: 0.72 · engine conviction 0.60
The underlying event (IBM's 2026-07-23 definitive agreement to acquire HRL from Boeing/GM) was widely covered by Tier-1 (Reuters via Gambetta interview) and trade press by the freeze date, but every consensus-grade piece frames it as IBM diversifying into silicon-spin qubits / hedging its superconducting bet / gaining dual-modality manufacturing at the Anderon foundry. The engine's specific second
The timing detail: control-and-test spend per LOGICAL qubit rises with code distance, so 2026's Q4-rung transitions convert into control/test purchase orders 2-3 quarters later — before consensus models quantum test as a line item; and hybrid quantum-HPC testbeds (TangleLab-style, NSF construction from 2026-09-15) procure exactly the integration layer KEYS sells (Quantum Benchmark/Labber stack). Claim: px.quantum_supply_basket closes its price-vs-event gap, outperforming the pure-play basket ove
Earliest coverage: none found for the specific second-order claim
Grader confidence: 0.65 · engine conviction 0.55
The engine concedes the generic "buy the shovels / KEYS is a quantum play" frame is consensus; the edge under audit is the second-order timing/procurement mechanism. Targeted refutation searches found (1) abundant consensus coverage of the underlying TangleLab event (Yahoo Finance, Seeking Alpha, CMU release, SDxCentral, DatacenterDynamics, TribLive, 2026-07-26..28) — but all name Rigetti and HPE,
The granular claim: the classical-refutation frontier follows the ledger's universal pattern — PEPS/belief-propagation methods get extended to the hardest reported instances within ~2 more papers, flipping the claim to debunked under the engine's grading; when it flips, the delta is commercial, not academic: the claim-linked pipeline (materials-simulation deals sold on 'beyond classical') stalls, and QBTS underperforms px.quantum_pureplay_proxy by ≥15% within 120 days of a full-scope reproductio
Grader confidence: 0.72 · engine conviction 0.50
The engine's edge is a conjunctive second-order claim: (i) the classical-refutation frontier (PEPS/belief-propagation) gets extended to D-Wave's hardest reported instances within ~2 more papers, flipping the supremacy claim to debunked, and (ii) the delta is commercial — the claim-linked materials-simulation pipeline stalls and QBTS underperforms a quantum pure-play proxy by >=15% within 120 days
The specific chokepoint: qubit wafers must be screened cold BEFORE dicing and packaging, and merchant supply of cryogenic wafer probe systems is essentially FormFactor (plus its HPD cryostat line). Foundry conversions show up as probe/cryostat tooling orders 2-4 quarters after announcement — meaning the July-2026 foundry news becomes FORM bookings visible around FY2027 guidance. Claim: FORM's quantum/cryo systems bookings inflect upward by mid-2027, sourced to the two new foundries.
Earliest coverage: None found meeting the consensus bar. Closest non-qualifying precursor: The Motley Fool, "This Quantum Computing Stock Has a Secret Weapon Nobody on Wall Street Has Priced In" (first-order theme only; itself asserts the thesis is NOT priced in) (2026-05-29)
Grader confidence: 0.70 · engine conviction 0.50
The engine's edge is a second-order timing/mechanism claim: cold pre-dice qubit-wafer screening is the chokepoint, merchant supply is essentially FormFactor (incl. HPD cryostats), foundry conversions (IBM Anderon spin-off + GlobalFoundries QTS, the July-2026 foundry news) convert to probe/cryostat tooling orders 2-4 quarters later, producing a FORM quantum/cryo bookings inflection by mid-2027. Ten
The tradeable detail is NOT NTT (too big for quantum to move its P&L). It is that the FY2027 supply-chain build-out must place DATED COMPONENT ORDERS with Japanese photonics makers roughly a year before any prototype exists — the first checkable Japanese quantum supply contracts. Watch detector/waveguide suppliers (Hamamatsu-class) for quantum-attributed orders or segment language in FY2027 disclosures. Second-order: as Western cryo export controls harden into industrial strategy (IISS analysis,
Grader confidence: 0.72 · engine conviction 0.55
The engine's edge has two second-order components: (A) Japanese photonics component suppliers (Hamamatsu-class) receiving dated, quantum-attributed orders from the OptQC/NTT FY2027 build-out, checkable in FY2027 disclosures; (B) relative de-risking of Japan's room-temperature optical route as Western cryo export controls harden. Nine targeted searches across English and Japanese news indexes found
The specific transmission channel: PSQL is not just a stock, it is France's procurement-champion model listing its credibility. A broken de-SPAC (high BBCQ trust redemptions read as a sentiment print, then post-listing fade — the 90-session relative-performance leg carries the grade) reprices the entire French/EU quantum venture pipeline at exactly the moment neutral-atom differentiation is eroding — Pasqal co-signed the ~50-author industry-wide neutral-atom roadmap (arXiv:2607.21554) alongside
Earliest coverage: None found matching the specific claim. Closest adjacent: Computer Weekly, 'European Union deep tech plan too late for quantum champions IQM and Pasqal' (different mechanism: EU funding gap forcing US listings, not broken-listing repricing the pipeline); Bloomberg 2026-03-04 covered only the underlying deal announcement.
Grader confidence: 0.70 · engine conviction 0.55
Timeline is decisive: at the 2026-08-05 freeze the de-SPAC had not closed — the F-4 was declared effective that very day, the redemption deadline was 2026-08-21 and the vote 2026-08-25 — so no redemption print or post-listing fade existed for consensus press to cover, and the engine's claim is a forward-looking transmission thesis, not commentary on a completed event. Tier-1 coverage found (Bloomb
The NSI undertakings and SkyWater vertical integration are in direct economic tension: the rational path for IonQ is to migrate next-generation trap-chip fabrication to Bloomington, Minnesota — the checkable migration is (a) whether SkyWater DISPLACES Infineon as the named fab for next-gen Oxford Ionics trap chips, and (b) the UK-vs-US split in design/test/packaging headcount and IP filings. The tell arrives well before any official review: IonQ's capex and hiring split UK-vs-Minnesota. Material
Earliest coverage: TechTimes (nearest non-qualifying coverage): "IonQ Clears Last Regulatory Hurdle: SkyWater Closes Friday, Ending Foreign Chip Dependency" (2026-07-29)
Grader confidence: 0.65 · engine conviction 0.55
The underlying events were fully consensus by freeze: the NSI final order with undertakings (gov.uk, 2025-09-11, uniformly framed by press as sovereignty PROTECTION — matching the engine's stated consensus baseline) and the SkyWater acquisition close (Bloomberg 2026-07-29/Reuters, whose Tier-1 angle was the FTC deadlock over fair access for rival quantum firms — a US antitrust mechanism, not the U
Tightened per skeptic to the precise conjunction: NO Chinese program posts a verified BELOW-THRESHOLD (Λ>1, distance-scaling) QEC result on DOMESTIC cryogenic hardware before end-2027. Note USTC has already posted surface-code logical operations (d-3 lattice surgery, logical CNOT) on a 107-qubit superconducting processor (arXiv:2607.01473, 2026-07-01) — so 'pinned at Q2 sampling' is stale; the load-bearing observable is whether the fridge fleet for such runs is domestic or legacy imported Bluefo
Grader confidence: 0.65 · engine conviction 0.50
The specific second-order claim is the conjunction: no verified below-threshold (Lambda>1, distance-scaling) Chinese QEC result on DOMESTIC cryogenic hardware before end-2027, with fridge-fleet provenance (domestic vs legacy imported Bluefors) as the load-bearing observable. No Tier-1 outlet was found covering even the headline USTC below-threshold event (PRL 2025-12-22, Lambda=1.40(6) at d=7), le
The listed-basket weakness is a FALSE NEGATIVE for Finland: the dilution-refrigerator layer is private (Bluefors), so the tightening shows up in TRADE FLOWS, not tickers — Finland→US/Japan exports graded on the finest available subheading (HS 8418.69-class, since HS 8418 is dominated by conventional refrigeration) should inflect ahead of any listed signal, and the engine grades Comtrade vintages. Named threat to the Japan leg: ULVAC's domestic ~10 mK DR (ledger 2025-03-22, explicitly built to di
Earliest coverage: None found matching entity+mechanism (closest non-qualifying: IndexBox, 'European Union Dilution Refrigerator Measurement Systems - Market Analysis') (2026-07-16 (non-qualifying closest match only))
Grader confidence: 0.55 · engine conviction 0.50
The edge's second-order claim is: listed-cryo-basket weakness is a false negative because the dilution-refrigerator layer (Bluefors, Finland) is private, so tightening surfaces in Finland→US/Japan trade flows at the HS 8418.69-class subheading ahead of any listed signal — plus named ULVAC displacement threat on the Japan leg and an order-book shift toward allied government buyers after EU rule har
Three specifics no model carries, with the load rebalanced onto the leg that cannot fail on mechanics. (1) The RECURRING quarterly drag: 48% of Quantinuum's ~$136.6M quarterly net loss is ~$66M, and basis-difference amortisation on $1.1B of intangibles runs on top of it forever — so the settled run-rate is roughly $70–120M per quarter against the ~$0 the market is modelling. The $265M stub covering only ~26 days almost certainly contains a one-time step-up charge, so a FULL quarter printing belo
Earliest coverage: none found (nearest miss: Quantum Zeitgeist Substack, single newsletter, excluded by protocol) (2026-05-23)
Grader confidence: 0.60 · engine conviction 0.72
ADVERSARIAL POSTURE — I tried to refute novelty on each leg and failed on all three, but the failure is partly a coverage failure, hence only moderate confidence. WHAT THE RECORD SHOWS AT/BEFORE 2026-07-26 (freeze is 3 days after Honeywell's 2026-07-23 Q2 print): every retrievable secondary source treats the Quantinuum event exactly as the engine characterises the consensus — as a one-time headli
THREE independent merchant-foundry silicon-spin datapoints landed in ten days, which is a trend rather than a news cycle: imec + Diraq published eight-qubit operation of a 300mm SiMOS FOUNDRY-fabricated device (Nature Communications, 2026-07-13); NEDO selected Hitachi to design 100+ qubit spin processors on Intel's 18A node with a quantum-specialised PDK through March 2029 (2026-07-22); IBM signed a definitive agreement for HRL Laboratories to pair CMOS-compatible spin-qubit fabrication with the
Grader confidence: 0.70 · engine conviction 0.68
I tried to refute the engine and could not, on the literal protocol. Three separate refutation angles were tested and all failed the entity-AND-mechanism test. (A) The compound trend claim — that imec/Diraq (2026-07-13), NEDO/Hitachi/Intel 18A (2026-07-22) and IBM/HRL/Anderon (2026-07-23) constitute a merchant-foundry silicon-spin trend that creates a Si-28 demand narrative — has NO coverage at al
Half-contrarian and more specific than either side. There is NO funding cliff — ~$569M of liquidity against $16M/quarter is roughly nine years, and shorts betting on a raise-driven collapse will keep being wrong. The unmodeled fact is the margin: at ~30% gross margin, every 'record revenue' quarter is a WORSE cash quarter, because Rigetti is effectively selling hardware near cost to place systems. The roadmap leg is now a MEASURED slip magnitude (~4 months on the first dependent milestone) rathe
Grader confidence: 0.60 · engine conviction 0.66
The edge is a compound second-order claim about one entity (Rigetti) with a specific mechanism: liquidity is ample so the short funding-cliff thesis fails, and the load-bearing unmodeled fact is gross margin — at ~30%, incremental hardware revenue is close to cash-neutral or cash-negative, so "record revenue" quarters worsen cash, and combined with a measured ~4-month first-milestone slip the corr
The capability claim is NOT what I contest — but the ladder position must be stated correctly: Quantinuum is ONE OF TWO programs at the engine's Q4 rung (QuEra has been level 4 since 2026-01-14 — Harvard/MIT/QuEra, Nature DOI 10.1038/s41586-025-09848-5, 96 error-corrected logical qubits from 448 atoms, 2.14x below threshold — i.e. EARLIER), and it is the only Q4 program that is publicly traded. The non-obvious detail is the collision of best-in-class physics with worst-in-class revenue quality,
Earliest coverage: Quantinuum / SoftBank Corp. joint white paper 'Quantum Computing Frontiers' (company release, covered by trade press 2026-07-21/22) — underlying event only, NOT the engine's second-order claim (2026-07-21)
Grader confidence: 0.60 · engine conviction 0.62
ADVERSARIAL ATTEMPT TO REFUTE, AND WHAT IT FOUND. The underlying event is confirmed real and public before freeze: DuckDuckGo results confirm a joint Quantinuum (NASDAQ: QNT) / SoftBank Corp. white paper titled 'Quantum Computing Frontiers' dated 2026-07-21/22, described in coverage as mapping 'two commercially-relevant quantum computing application areas against Quantinuum's hardware roadmap' (ch
Shipping PQC is worth approximately zero incremental revenue — Microsoft just demonstrated that by giving it away in a Patch Tuesday, and Cloudflare gives it away on free plans. The revenue is in crypto discovery, inventory and certificate lifecycle, which is subscription and recurring, and PANW bought that layer in February while the market was watching the algorithm announcements. The correct metric is machine-identity/Venafi ARR inside PANW's subscription line — NOT 'PQC-enabled' product pres
Earliest coverage: (none meeting the bar; nearest miss = Palo Alto Networks "Next-Generation Trust Security" launch coverage cluster, event-only, mechanism mismatch) (2026-03-23)
Grader confidence: 0.58 · engine conviction 0.58
The engine's edge decomposes into four linked assertions: (i) shipping PQC in firewalls/SASE carries ~zero incremental revenue because it is being given away (Microsoft Patch Tuesday, Cloudflare free plans); (ii) the monetizable layer is crypto discovery/inventory/certificate lifecycle, which is recurring; (iii) PANW acquired that layer in February (CyberArk close, which the Feb-2026 "final report
The DOC equity structure puts a financing floor under exactly NINE named firms, which means the funding cliff arrives first and hardest for the UNNAMED — US micro-caps and non-US programs — rather than uniformly across the sector. Expect the named-versus-unnamed financing spread to widen BEFORE any capability milestone lands, because it is a capital-structure event, not a physics event. On the evidence side, the honest statistic is the eight advantage-DEBUNK EVENTS on the engine's ledger — four
Earliest coverage: Semafor — "Big names missing from Trump's quantum investment" (NEAREST MISS, not consensus-grade: non-tier-1, single source, different mechanism) (2026-05-22)
Grader confidence: 0.60 · engine conviction 0.68
DISAGGREGATION. The frozen claim has three conjuncts that must all be matched: (i) entity = the UNNAMED cohort (US micro-caps + non-US programs), not the named nine; (ii) mechanism = a CAPITAL-STRUCTURE effect — the DOC equity structure puts a financing floor under the nine, so the named-vs-unnamed financing spread widens; (iii) timing = the spread widens BEFORE any capability milestone. The under
The cap-table-emigration lament ('Europe funds the factories, America owns the companies') is a well-worn generalist complaint and is deliberately STRIPPED from this entry — it restates consensus. What is not priced is the capital structure: Pasqal is listing at 95.6x EV/revenue with only $250.0M of hard committed proceeds against a $2.0B pre-money mark, i.e. 12.5% — thin capital behind a legacy valuation — INTO a tape the engine independently measures at pure-play froth 0.393 with a 90-day rela
Earliest coverage: Stock Titan (nearest miss, NOT consensus-grade) — "Pasqal–Bleichroeder (NASDAQ: BBCQ) deal targets $2B value and $645M cash" / BBCQ 8-K investor-presentation summary (2026-06-30)
Grader confidence: 0.60 · engine conviction 0.66
The engine's retained edge is a capital-structure claim, not a sovereignty or cap-table claim (it says so and strips the generalist lament). The claim to refute is: Pasqal is listing at ~95.6x EV/revenue with only ~$250.0M of hard-committed proceeds against a $2.0B pre-money mark (12.5%), into a pure-play quantum tape the engine measures at froth 0.393 / -15.71% 90-day relative, making this a rede
Read what the onshoring money actually bought. Bloch's named line items are a shared sensing testbed, automation software and integration; QMEC is manufacturing ENGINEERING; and ZPC's Chicago hub is service, installation and integration — the company's own release says manufacturing stays in Alberta. Nobody funded a US dilution-refrigerator production line, because nobody in the US builds them. Service and integration capacity is a demand MULTIPLIER on imported hardware: every new US installatio
Grader confidence: 0.66 · engine conviction 0.65
I tried to refute the engine and could not find the mechanism claim anywhere at or before 2026-07-26. What IS widely reported at/before freeze is the underlying event and the engine's supporting FACTS: BetaKit (2026-07-22) quotes ZPC CEO Chris Cassin — "Our headquarters, engineering expertise, and advanced manufacturing continue to anchor our global growth from Alberta" — and describes the Chicago
For Australia-the-sovereign the DARPA expansion is a NEGATIVE. It relocates the funded, independently-verified critical path — modulator materials manufacturing and the IV&V that determines whether the architecture is real — to California, converting Brisbane from a co-equal build into the deployment endpoint of a US-verified, US-owned architecture. The schedule detail sharpens it: the Linde cryoplant alone delivers H2 2027, with commissioning and cryo-cabinet acceptance still to follow, so the
Earliest coverage: Reuters — "Quantum computing startup PsiQuantum signs $125 million DARPA deal" (event coverage only; does NOT contain the edge's mechanism claim) (2026-07-22)
Grader confidence: 0.68 · engine conviction 0.62
The edge makes a compound second-order claim with a specific entity+mechanism pairing: that the DARPA expansion is NEGATIVE for Australia-the-sovereign because it relocates the funded, independently-verified critical path (modulator materials manufacturing plus the IV&V that adjudicates whether the architecture works) to California, demoting Brisbane to the deployment endpoint of a US-verified, US
That the program 'runs on Intel 18A' is in the press-release headline, so it is not the edge. The edge is the ARTEFACT: Japan's flagship qubit-fabrication program has ZERO named alternative fab path for 32 months, and the binding dependency is a quantum-specialised PDK — a design kit is exactly the artefact that gets reclassified when quantum ECCNs tighten, and unlike an installed tool it can be revoked without anything shipping. EDA/design-kit cut-off is the historically used control lever (SMI
Earliest coverage: The Quantum Insider — "Hitachi, Intel and AIST Launch Silicon Quantum Computing Project Backed By Japanese Government" (closest entity-matching coverage; mechanism does NOT match, so it does not establish consensus) (2026-07-23)
Grader confidence: 0.72 · engine conviction 0.58
Adversarial attempt to refute the edge failed on the mechanism leg. (1) Entity coverage is saturated and Tier-1-adjacent (Nikkei Asian Review 2026-07-22, plus TQI, Quantum Computing Report, EE Times Japan, Sankei, Zaikei, Tech Times) — but every located item restates the press release: NEDO funding, Intel 18A, PDK creation, cloud service FY2027, 100-qubit FY2028, 1,000-qubit FY2030, program throug
Quantum sovereignty is collapsing into semiconductor sovereignty. The defensible statement — the overstated 'three doors' count is dropped, since CEA-Leti's 300mm SOI line and SEALSQ/GlobalFoundries' cryo-CMOS MoU (2026-07-08, on GF's US footprint) are live alternatives — is this: imec is the only NON-US 300mm line with a DEMONSTRATED multi-qubit spin device, not merely a capable one, and demonstrated-versus-capable is the distinction that decides whom a national program can actually contract in
Grader confidence: 0.70 · engine conviction 0.55
ADVERSARIAL POSTURE: I tried to refute the engine's novelty and could not, but the refutation attempt did narrow what is actually differentiated. WHAT IS WELL-COVERED (and therefore NOT differentiating): the underlying event is saturated. imec/Diraq's eight-qubit 300mm SiMOS result published in Nature Communications 2026-07-09 and was carried 07-13 to 07-21 by at least seven trade outlets. Critic
The killer verified number: Q1'26 organic revenue was ~$24K (the rest acquired) against ~225M shares and an active $118.5M S-8, with no rung at all on the Q0-Q5 board. It is a dilution funnel, not a laggard with a roadmap — the single most mechanically exposed name when the pure-play proxy reprices. Trigger is S-8/S-3 filing cadence + shares crossing 240M as the mechanical early warning that front-runs any pure-play drawdown.
Grader confidence: 0.70 · engine conviction 0.60
The engine's stated consensus (QUBT is a dangerous micro-cap quantum meme stock) and every underlying data point are consensus-grade and heavily covered. But the protocol requires matching the same second-order claim on entity AND mechanism, and coverage of underlying numbers does not count. The differentiated mechanism is threefold: (1) QUBT is a pure dilution funnel with no real quantum roadmap;
The migration is a tailwind for the STANDARD and a headwind for the proprietary alternative. ML-KEM is now free and built-in (opt-in in Windows/z17, default-on at Cloudflare), collapsing SKA's 'simpler-than-PKI' pitch regardless of default state, and it bites an already-negligible $623K base.
Grader confidence: 0.72 · engine conviction 0.50
The engine's edge is an INVERSION of Arqit's own narrative: instead of 'NIST migration = tailwind for the vendor,' it argues the migration is a tailwind for the STANDARD (ML-KEM, now free/built-in in Windows and z17, default-on at Cloudflare) and thereby a HEADWIND for Arqit's proprietary Symmetric Key Agreement, whose entire pitch is 'simpler than PKI' — a pitch that ML-KEM's zero-cost ubiquity c
Ignore the quantum-hardware timeline; the datable catalyst is the 2026-08-12 GA of PQC-standard mainframes forcing a compliance-driven zSystems refresh + high-margin migration services, sold into the exact regulated base facing the deadline — a bindable date, not a 2029 promise.
Earliest coverage: Bank of America sell-side note (relayed via Financial Modeling Prep, "IBM's z17 Mainframe: A Catalyst for Sustainable Growth") — closest match, but mechanism mismatch (AI/Spyre refresh cycle, not PQC-compliance-driven) (2026 (pre-freeze; z17 cycle commentary))
Grader confidence: 0.60 · engine conviction 0.50
The frozen protocol requires a match on BOTH entity and mechanism, with event coverage excluded. The engine's differentiated move is a reframe: ignore the 2029 quantum-hardware promise and treat the 2026-08-12 PQC-standard mainframe GA as the bindable, datable catalyst forcing a compliance-driven zSystems hardware refresh plus high-margin migration services sold into the exact regulated base facin
The non-obvious lever is the design-in qualification moat: a hardware secure element, once socketed into a BOM and certified, is sticky and slow for a software vendor to displace. Foreground the datable convert-or-fail triggers — MoU->binding definitive agreement and first named QS7001 design wins by 2026-Q4 as the leading tells, with FY26 revenue holding the $27M floor as the coincident check.
Earliest coverage: timothysykes.com (trading newsletter, EXCLUDED under protocol) — nearest miss using 'turn its post-quantum pipeline into a moat / own the hardware layer' language; no qualifying Tier-1 or 2x independent Tier-2 source found (2026-05-21)
Grader confidence: 0.68 · engine conviction 0.45
The engine's edge is a mechanism-level thesis: certified hardware secure elements are sticky design-ins with high switching cost versus software PQC vendors, plus datable convert-or-fail tells (MoU->binding definitive agreement; first NAMED QS7001 design wins by 2026-Q4; FY26 revenue holding the $27M floor as coincident check). Searches surfaced heavy coverage of the underlying events but not this
The gradeable edge is the He-3 SOVEREIGN-PRIORITY divergence: DOE domestic allocation lets US superconducting programs keep scaling while allied programs (Japan) ration, and only ~5% helium-from-China exposure insulates the US on He-4 too. GFS is NOT a material vehicle (cryo-CMOS immaterial to a ~$7B foundry, revenue only expected Q4 2026); LAES (SEALSQ) is the geared micro-cap and the DOE He-3 auction is the datable rail.
Earliest coverage: War on the Rocks, "The Supply Chain Chokepoints in Quantum" (closest on-point source; contradicts the divergence claim rather than establishing it) (2025-10-20)
Grader confidence: 0.70 · engine conviction 0.60
The edge is a compound claim; I tested each element. (1) He-4 insulation via ~5% China exposure: this specific number is consensus-grade — USGS (primary government) reports US got 5% of imported helium from China 2021-2024, and the July 10 2026 China helium export ban was covered at Reuters/Bloomberg grade with the 'US not heavily dependent' insulation framing. But this is a generic supply-share f
Grade MARGIN not volume: Finnish cryo-export UNITS rise (Comtrade HS8418.69/8419.89 YoY up) while per-unit margin compresses as the He-3 charge (100-300 L at $1,900-2,600/L) becomes the dominant BOM line. The scarce input has shifted from the fridge (Finland can build them) to the He-3 inside it (22-30k L/yr supply vs 40-60k L/yr demand). No clean listed instrument (Bluefors private; EFNL is Nokia-dominated).
Earliest coverage: Closest coverage of the specific margin claim (all NON-qualifying under protocol): gtaic.ai 'Helium-3 market in Finland' (2026) and postquantum.com 'Quantum Cryogenic Infrastructure and Helium-3 Guide'. Underlying-event consensus (does NOT match mechanism): Science/AAAS + Reuters-syndicated Bluefors-Interlune deal, Sept 2025. (2026 (gtaic.ai margin-framing report; non-qualifying))
Grader confidence: 0.60 · engine conviction 0.58
The engine explicitly concedes the consensus (He-3 is a quantum bottleneck; Bluefors is leader) and stakes its edge on a distinct second-order FINANCIAL claim: read the trade in MARGIN not VOLUME — Comtrade HS8418.69/8419.89 export units up YoY while per-unit margin compresses as the He-3 charge (100-300 L at $1,900-2,600/L) becomes the dominant BOM line, i.e. scarcity has relocated from the fridg
Pasqal's neutral-atom modality is STRUCTURALLY He-3-IMMUNE, making France the natural PAIR-TRADE against He-3-exposed Japan/superconducting. As He-3 tightens, laser-cooled platforms' relative cost/availability improves, and Q-PLANET localizes the laser/vapor-cell supply that is the neutral-atom bottleneck. EWQ is a null instrument (Pasqal private); grade on deployment/disbursement events.
Earliest coverage: none qualifying; nearest non-consensus coverage: postquantum.com 'The Tweezer Array's Hidden Supply Chain' (pre-2026-07-18)
Grader confidence: 0.70 · engine conviction 0.52
The claim decomposes into (i) a generic mechanism — neutral-atom is structurally He-3-immune while superconducting is He-3-exposed — and (ii) the engine's distinctive second-order synthesis: Pasqal/France as a PAIR-TRADE against He-3-exposed Japan/superconducting, with He-3 tightening improving laser-cooled relative cost/availability, and Q-PLANET localizing the laser/vapor-cell bottleneck. The g
The controls bite ASYMMETRICALLY BY MODALITY: they strand the millikelvin superconducting program (import-dependent on Western cryo) while China redirects capital to room-temperature photonic (QBoson CNY 1B / ~$145M Series B, April 2026, marketed as 'no cryogenic cooling'). China's own 2026-07-10 helium export ban is a tell of resource INSECURITY (>80% import-dependent), not leverage. The exposed sector is superconducting-QC hardware, not China's quantum program wholesale.
Earliest coverage: postquantum.com — "China's Quantum Supply Chain: How Export Controls Are Building What They Sought to Prevent" (single specialized blog; NON-qualifying under protocol) (2026-04-09)
“The dual-track strategy of simultaneously developing domestic cryogenics AND room-temperature photonics indicates deliberate industrial hedging. ... The core selling point: room-temperature operation with no cryogenic cooling required.”
Grader confidence: 0.74 · engine conviction 0.60
The engine's edge is a compound second-order claim: (i) controls bite ASYMMETRICALLY BY MODALITY — stranding cryo-dependent superconducting QC while China redirects capital to room-temperature photonic (QBoson CNY1B/$145M Series B, Apr 2026, 'no cryogenic cooling'); (ii) China's 2026-07-10 helium ban is a TELL OF INSECURITY (>80% import-dependent), not leverage; (iii) the exposed sector is superco
It is the He-3 (NOT the He-4/semiconductor) deficit that specifically throttles Japan's SUPERCONDUCTING roadmap scale-up: each added fridge is a fresh 100-300 L He-3 draw, and Japan lacks a domestic neutral-atom/photonic hedge to pivot to (unlike France) — the at-risk MIRROR of France. Grade on objective metrics, not self-reported delay.
Earliest coverage: No source establishes the SPECIFIC claim. Closest adjacent consensus (generic He-3 bottleneck for superconducting): War on the Rocks 'The Supply Chain Chokepoints in Quantum' and Science/AAAS 'As helium-3 runs scarce...'; closest country-specific analog: MP-IDSA 'A Helium Shock to India's Quantum Technology Ambitions' (India, not Japan). Consensus trade press on Japan actually frames it as a PHOTONIC/optical bet (TechWireAsia 'Japan bets on light instead of extreme cold,' Nov 2025), contradicting the 'Japan lacks a hedge' premise.
Grader confidence: 0.72 · engine conviction 0.45
The generic prior (quantum needs cryogenics) and the first-order fact (He-3 scarcity throttles superconducting qubits) are both fully consensus-grade before the 2026-07-18 freeze, backed by Tier-1 (Science/AAAS, NATO/EU flags) and multiple Tier-2 sources. However, the engine's differentiating second-order claim is a specific synthesis: Japan is the 'at-risk MIRROR of France' because He-3 (specific
The date-pairing is the claim: the December 2026 lockup (180-day, S-1-disclosed — the calendar claim hangs on that filed date, not an assumed standard) releases roughly 5-6x the current float into a tape where the pure-play proxy is already underperforming, and Honeywell is mid-breakup with a rational incentive to monetize. We claim QNT specifically underperforms its own pure-play cohort in the lockup window — quality name, worst near-term supply dynamics — the opposite of the 'flight to quality
Grader confidence: 0.78 · engine conviction 0.60
Adversarial attempt to refute novelty failed under the frozen protocol. What WAS consensus by 2026-07-18: the IPO/$1.68B raise (Bloomberg, Reuters-via-CNBC), the 180-day lockup's existence (S-1 — but that is disclosure of the underlying event, and the engine itself concedes 'everyone knows lockups exist'), Honeywell's general desire to monetize (CEO Kapur statements), and QNT-as-quality-name. What
Consensus conflates funded with on-time. Our claim: Lyra misses late-2026 GA, and — the finer detail — the binding metric will be median 2Q fidelity at GA, not qubit count; Cepheus already demonstrated the pattern (count delivered, fidelity 99.1% below the marketed 99.5%). Expect a count-first Lyra launch with fidelity below target, which the engine grades as a second slip. Two graded slips on one program historically precede the narrative break, independent of cash.
Earliest coverage: none found (closest partial-match, non-qualifying: techi.com 'Rigetti's Multi-Chip Bet and the 99.1% Fidelity Question')
Grader confidence: 0.70 · engine conviction 0.60
The underlying events (Cepheus Jan-2026 delay; April-2026 GA at 99.1% vs 99.5% target; Lyra 336Q late-2026 target at 99.7%) are broadly covered, but protocol excludes underlying-event coverage. The engine's specific compound second-order claim — Lyra misses late-2026 GA, median 2Q fidelity at GA (not qubit count) is the binding metric, expect a count-first launch below fidelity target graded as a
The specific claim: base-rate grading matters more than the live debate. Every decided advantage claim in the ledger has died on roughly a 12-24 month fuse, and the classical-methods community now has a working BP-TNS toolchain aimed at exactly the instance classes D-Wave's rebuttal named. We predict a follow-up classical paper extending coverage to those harder instances by end of 2026, converting 'contested' to 'substantially debunked' — and that annealing-specific procurement (national-lab Ad
Earliest coverage: None qualifying. Closest non-qualifying near-miss: arXiv:2607.07530 'The NISQ Trap: Eight Years of Demonstrations the Hardware Was Built to Lose' (Amit Hagar, single-author preprint consolidating a Substack series) — covers only the base-rate component (~18-month classical reproduction pattern), not the D-Wave-specific follow-up-paper prediction or the procurement/bookings mechanism. (2026-07 (near-miss preprint only; no qualifying consensus source found))
Grader confidence: 0.70 · engine conviction 0.55
The underlying dispute (Flatiron/BU Science paper, D-Wave rebuttal, QBTS story-stock framing) is thoroughly covered, exactly as the engine's stated consensus baseline says — that coverage does not count under the protocol because it matches the headline event, not the second-order claim. The engine's edge is a composite: (a) base-rate grading (decided advantage claims die on a ~12-24 month fuse),
The claim is the sequencing: the PIPE resale effectiveness date (watch EDGAR for the resale registration statement) starts a 90-day window in which IQMX underperforms the pure-play proxy, because unlike IPO lockups this supply arrives with holders at cost-basis ~$10 and no growth print in between — IQM's on-prem system sales (real, but lumpy) will not produce a positive surprise inside the window. The Helsinki dual listing adds a second exit venue the overhang can clear through.
Earliest coverage: None found matching the specific claim. Closest adjacent: issuer boilerplate in IQM Finland Oy Form 424B3 (SEC, merger prospectus, Reg. No. 333-295867) disclosing generic resale-price risk — excluded as underlying-event disclosure, not the second-order claim.
Grader confidence: 0.70 · engine conviction 0.55
The engine's edge is a specific second-order sequencing claim: PIPE resale registration effectiveness (not yet filed as of freeze — EDGAR shows no F-1/F-3; filing due within 30 days of the 2026-07-01 closing) opens a ~90-day window of IQMX underperformance vs a pure-play quantum proxy, driven by 14,548,000 PIPE shares at $10.00 cost basis with no growth print inside the window, and Helsinki as a s
The specific claim: within 12-18 months Riverlane raises at flat-or-down valuation or is acquired below Series C marks, because hardware vendors will take NVIDIA's free decoders plus in-house FPGA glue rather than license Deltaflow — the QEC software decoder becomes a feature of the control stack, not a product. Nobody covering the NVIDIA release framed it as a private-market repricing event; that is the detail.
Earliest coverage: None found. Closest adjacent: Forbes contributor piece (Peter Cohan) on NVIDIA quantum strategy shifting market power — but it targets Infleqtion, never mentions Riverlane/Deltaflow, and contains no private-market repricing mechanism; also not Tier-1 under the protocol. (2026-04-20)
Grader confidence: 0.75 · engine conviction 0.50
The claim to check is second-order: Riverlane raises flat/down or is acquired below Series C marks within 12-18 months BECAUSE hardware vendors adopt NVIDIA's free Ising decoders plus in-house FPGA glue, turning the QEC decoder into a control-stack feature rather than a licensable product. Public coverage at/before the 2026-07-18 freeze covers only (i) the underlying event — the Ising launch, unif
Consensus trades the PQC theme; our claim is the queue position. SEALSQ is one of the few merchant vendors with a PUBLISHED, dated certification calendar landing inside the CNSA 2.0 window — the claim is queue position versus the incumbents (Infineon, Thales, ST, NXP all run PQC secure-element programs), graded by registry publication dates. Each certification completion in H2 2026 is a dated, independently checkable event (certificates appear in ANSSI/NIST CMVP/CC public registries) that conver
Earliest coverage: None found meeting consensus grade. Closest non-qualifying: SEALSQ's own GlobeNewswire press release publishing the certification roadmap (the underlying event, issuer self-attestation) (2026-03-06)
Grader confidence: 0.72 · engine conviction 0.60
The specific second-order claim (SEALSQ's certification queue position vs Infineon/Thales/ST/NXP graded by ANSSI/CMVP/CC registry publication dates, >=2 of 4 certifications by year-end, named US federal/defense-adjacent design win within a quarter) appears in NO consensus-grade source at/before the 2026-07-18 freeze. All discoverable coverage is (i) SEALSQ's own press releases and their mechanical
The claim is measurable order-book physics, not category membership: verified fridge-dependent capital raised in Q2-Q3 2026 plus the IBM fleet implies quoted dilution-fridge lead times extend from ~12 months toward 18+ by early 2027, with a capacity-expansion or price-increase announcement as confirmation. Lead-time quotes are obtainable by any operator requesting one — a real-world sensor most investors never poll; a dated baseline RFQ quote is archived at freeze (2026-07) so the comparison gra
Earliest coverage: War on the Rocks, 'The Supply Chain Chokepoints in Quantum' (nearest-miss, NON-qualifying: component-level 12-18mo lead times for Sumitomo pulse-tube cryocoolers, not the dilution-fridge lead-time-extension forecast; publisher not on frozen tier lists) (2025-10)
Grader confidence: 0.70 · engine conviction 0.60
The bottleneck CATEGORY is unambiguously consensus (NIST's $20M cryostat-bottleneck center launched 2026-06-29 is a primary-government acknowledgment; War on the Rocks Oct-2025 covers chokepoints; trade/blog coverage of He-3 scarcity, DOE allocation, and Interlune deals is abundant) — but the engine explicitly concedes category consensus and stakes its edge on a quantified forward order-book claim
Specific claim: within ~12 months PsiQuantum formally resets the Brisbane utility-scale date beyond 2029 or the AU deal's milestone-linked loan tranches get restructured, and the exposure surfaces in Queensland FY27-28 budget papers. Second-order: Australia's real quantum capability migrates from Brisbane photonics to Sydney silicon — Diraq's 8-qubit 300mm imec result (ledger 2026-07-13) and Silicon Quantum Computing's DARPA QBI Stage B seat (both hold Stage B seats) are the assets that survive;
Earliest coverage: none found (closest fragment: Quantum Zeitgeist 'Australia Quantum Computing Companies 2026' vendor guide noting Australia's national bet on silicon — single blog, not on the frozen Tier-2 list, and lacks the PsiQuantum-reset→rotation mechanism)
Grader confidence: 0.70 · engine conviction 0.60
Adversarial pass could not refute novelty under the frozen rule. The underlying delay narrative is abundantly covered (Startup Daily, Forbes.au, InnovationAus, DCD: build running late, 2027→2029 slip, Moreton Bay move, tranches undrawn per budget papers) — but the engine itself concedes that as the consensus baseline. The specific edge is the compound prediction: (a) a FORMAL reset beyond 2029 or
The differentiated claim is the ORDER, not the raise: IQM's first dated intermediate roadmap milestone slips BEFORE the dilutive raise (shelf/ATM/secondary) — the slip is the tradeable lead indicator the market will misread as merely technical. This is the wedge's slip-predicts-funding-cliff pattern applied prospectively, before the slip. A public, marked-to-market national champion transmits sector repricing into Finland's whole quantum ecosystem (Bluefors customers, VTT programs) in a way priv
Earliest coverage: None found for the specific claim. Nearest adjacent: 24/7 Wall St "Quantum Computing's Cash Crisis" (sector dilution, no IQM slip-ordering) (2026-07-01)
Grader confidence: 0.70 · engine conviction 0.50
The underlying headline events (IQM's Nasdaq listing 2026-07-02, its roadmap, and the sector truism that quantum SPACs raise dilutive equity) were all consensus by freeze, and the engine correctly conceded the raise itself is near-consensus. The differentiated element is the ORDER and the lead-indicator mechanism: IQM's first dated intermediate milestone slips before the dilutive raise, and the sl
Procurement-without-capability reverses: we predict a visible H2-2026/H1-2027 slowdown in new German public quantum-hardware awards versus the prior year's pace, plus consolidation among German pure-play startups whose revenue is predominantly German public procurement — Germany's spend shifts from buying whole machines toward software/HPC-integration and PQC compliance (where BSI is genuinely strong). The at-risk party is the German NISQ-hardware vendor ecosystem, not Germany's fisc. Counting u
Earliest coverage: None found matching entity+mechanism. Nearest miss: heise online / Tagesspiegel / Handelsblatt coverage of the 2023 DLR QCI €200M budget cut (underlying event, not the second-order claim, non-listed tiers); Sifted op-ed 2026-05-21 on European quantum consolidation (different mechanism: US acquisitions/VC pressure, not German procurement dependence)
Grader confidence: 0.70 · engine conviction 0.45
The engine's edge is a specific forward second-order claim: German public quantum-hardware award pace visibly slows H2-2026/H1-2027 vs prior year, procurement-dependent German pure-play hardware vendors consolidate, and German spend rotates to software/HPC-integration and PQC. Eight targeted searches (English + German) plus full-text inspection of the most plausible pre-freeze candidate documents
Canada is quietly becoming North America's quantum PACKAGING and bosonic-QEC layer — value that accrues regardless of which qubit modality wins. Specific claims: (1) the Bromont C2MI center discloses >= 2 non-Pasqal anchor customers or a phase-2 funding expansion within 12 months; (2) >= 1 Canadian QBI team advances beyond Stage B; (3) Nord Quantique's GKP result earns a capability-board rung move — hardware-efficient bosonic QEC needing ~10x fewer physical qubits is the cost-curve wildcard cons
Earliest coverage: Nearest miss (does NOT establish consensus): HPCwire, "Pasqal Expands Canadian Quantum Supply Chain with New PIC Packaging Center" — single non-listed trade outlet, event coverage only, no bosonic-QEC or modality-agnostic-value component (2026-07-02)
Grader confidence: 0.62 · engine conviction 0.65
Every COMPONENT event is publicly known in quantum trade press before the 2026-07-18 freeze (C2MI PIC packaging center 2026-07-02; Nord Quantique QBI Stage B Nov 2025; GKP hardware-efficiency results Feb 2024 and sub-0.1% SPAM 2026-07-13; $1.4B valuation May 2026), so the differentiation is thesis-level synthesis, not fact-level novelty — hence moderate confidence. But applying protocol v1 literal
Pilot lines create QUALIFICATION moats — via first-qualified incumbency, not captivity: Q-PLANET explicitly publishes open-standard PDKs and ADKs to lower entry barriers, but in qualification-driven markets the FIRST TRL6-qualified, standards-referenced supplier keeps the design-in even after standards open. Consortium suppliers' lasers, atom chips and vapor cells become those first qualified parts. Specific claims: (1) cycle-1 design-fabricate-test completes on schedule (~within 15 months); (2)
Grader confidence: 0.80 · engine conviction 0.55
The Q-PLANET launch (announced July 8-13, 2026, days before the 2026-07-18 freeze) received tier-1 (Reuters flash) and quantum trade-press coverage, but strictly as the underlying event: budget, consortium size, component roadmap, TRL4-to-TRL6 goal, open PDK/ADK plan. The engine's specific second-order claim — that pilot lines create qualification moats via first-qualified incumbency (first TRL6-q
The ISO rail is a real second market: conservative government/defense buyers and non-US-aligned states wanting non-NIST cryptography now have a standardized option with UK fingerprints. Specific claims: (1) first government procurement citing ISO/IEC 18033-2 Classic McEliece appears within 18 months (watch non-NATO and defense VPN/infrastructure tenders where large key sizes don't bind); (2) NCSC's 2028 inventory deadline converts into disclosed UK public-sector PQC-discovery contract awards sta
Earliest coverage: Businesswire press release by Post-Quantum (closest adjacent framing, underlying event only — NOT consensus-grade under protocol) (2026-07-15)
Grader confidence: 0.75 · engine conviction 0.50
The underlying event — Classic McEliece entering ISO/IEC 18033-2 (June 2026, PR wave 2026-07-15) — was covered before the 2026-07-18 freeze, but exclusively by outlets outside the frozen source lists (Quantum Insider, Quantum Computing Report, Quantum Zeitgeist, Industrial Cyber, ForkLog, Yahoo/Businesswire PR, OIST), and mostly as echoes of one non-independent company press release. Under the pro
The uncrowded leg: Indian IT majors (TCS, Infosys, Wipro, HCLTech) win the migration grunt work the vendors cannot staff — the Y2K playbook re-run with cryptography. Specific claims: (1) >= 2 of the four name PQC/quantum-safe migration in earnings commentary or announce dedicated practice wins by mid-2027; (2) cumulative publicly disclosed PQC-migration engagements by Indian majors reach >= 5 by Q3-2027, concentrated in EU-regulated banking/telecom clients forced by the end-2026 start deadline.
Grader confidence: 0.65 · engine conviction 0.50
Eight targeted searches found abundant consensus on the underlying event (PQC migration is a Y2K-scale, compliance-driven services problem with a consultant bottleneck — CNN May 2026, EFF Apr 2026, TechTarget, eMudhra Mar 2026) but zero consensus-grade coverage of the engine's specific second-order claim that TCS/Infosys/Wipro/HCLTech capture that services layer as the uncrowded leg of the PQC tra
Unconditional leg: by 2027-03-31 IonQ publishes no peer-reviewed or arXiv result showing logical error below physical error on its own hardware (Board stays at Q2), and its quarterly releases continue to lead with #AQ rather than a logical error rate. Conditional leg: if Quantinuum files an S-1/F-1 or is entered in the Advantage ledger before 2027-03-31, IONQ trails px.quantum_pureplay_proxy by >15% over the following 90 trading days.
Earliest coverage: arXiv 2606.06455 "Breakeven demonstration of quantum low-density parity-check codes" (all authors IonQ, Inc.); trade coverage Quantum Computing Report + PostQuantum 2026-06-05; Quantinuum S-1/IPO 2026-06-03/04 with CNBC coverage (2026-06-04)
“Our implementation exhibits breakeven performance, with some instances achieving qubit lifetimes comparable to or slightly exceeding that of our trapped-ion qubits.”
Grader confidence: 0.85 · engine conviction 0.45
The edge's unconditional leg asserts that by 2027-03-31 IonQ publishes no arXiv/peer-reviewed result showing logical error below physical error on its own hardware. That was already false at the 2026-08-24 freeze: IonQ-affiliated authors posted arXiv 2606.06455 on 2026-06-04 reporting a GB4 qLDPC logical lifetime of 3.95±0.68 s vs 3.3±0.9 s (paper) / 1.1 s T2* (trade coverage) physical, on IonQ's
Lead leg (tightened): CUSTOMER LEAKAGE — SkyWater's merchant defense-microelectronics and bio-MEMS customers now fab at a foundry owned by a competitor-adjacent quantum company; disclosed design-transfers/diversification away from the now-IonQ-owned fab, visible in customer qualification announcements and SkyWater merchant bookings before Q1-2027, are the genuinely differentiated early warning. Supporting leg: the first fully consolidated quarter (Q3-2026, reporting ~Nov) mechanically collapses
Earliest coverage: Statement of Chairman Andrew N. Ferguson, In the Matter of IonQ, Inc./SkyWater Technology, Inc. (FTC Matter No. 2610061); mechanism first reported publicly by Bloomberg on 2026-07-29 in FTC-deadlock coverage (2026-07-31)
“In the long run, companies concerned about IonQ's control of SkyWater can switch to other foundries to develop their projects.”
Grader confidence: 0.70 · engine conviction 0.60
The edge's lead leg is CUSTOMER LEAKAGE: SkyWater merchant customers diversifying/transferring designs away because a competitor-adjacent quantum company (IonQ) now owns their fab. That exact entity-and-mechanism pair (SkyWater's existing customers × departure-to-alternative-foundries driven by IonQ's competitive ownership) appears in a Tier-1 primary regulator document five days before the 2026-0
The non-obvious detail: the funding-cliff bear case is factually WRONG — the real signal is the slip pattern. Rigetti's flagship 2026 win is the NSF-funded TangleLab testbed ($5M, award #2537076, CMU/PSC with HPE and Rigetti as partners, start 2026-09-15); the NSF abstract does NOT name which Rigetti chip deploys, so the smallest-chip corroboration is held pending a PSC/Rigetti source. Core claim (kept per skeptic): the 100Q milestone gets re-dated again on or before the Q2 call and median 2Q fi
Earliest coverage: The Quantum Insider (delay + fidelity shortfall vs 99.5% target); fullest match incl. re-dated timeline: Quantum Computing Report 2026-04-07 / The Quantum Insider 2026-04-08; corroborated pre-freeze by Wedbush initiation note 2026-08-03 (non-bulge-bracket, not tier-1) (2026-01-13)
“The system is currently performing at a 99.1% median two-qubit gate fidelity ... Rigetti plans to continue to improve the fidelity of its individual chiplets and expects Cepheus-1-108Q to reach a median 99.5% two-qubit gate fidelity later this year.”
Grader confidence: 0.80 · engine conviction 0.55
The engine's edge decomposes into (1) 'funding-cliff bear case is wrong' — public 10-Q cash facts, widely discussed (e.g., $600M cash-runway commentary, Wedbush Aug 3 initiation); and (2) 'the slip pattern is the real signal: 100Q milestone re-dated again on/before the Q2 call, fidelity stays below 99.5%.' By freeze (2026-08-05) the milestone had ALREADY been publicly re-dated at least twice (end-
Tightened per skeptic: the non-obvious layer is MERCHANT quantum foundry capacity — but the design-win trigger counts only NEW third-party quantum process qualifications announced AFTER the 2026-07-31 close (SkyWater had pre-existing quantum foundry customers; a legacy customer re-announcing does not satisfy it). Fab qualification is a switching-cost moat that binds 12-18 months later as design-wins (ZuriQ just chose Infineon's European lines instead — the qualification race is live and countabl
Earliest coverage: IBM Newsroom / U.S. Department of Commerce joint announcement of the Anderon quantum foundry and proposed $1B CHIPS award (2026-05-21)
“IBM and U.S. Department of Commerce Announce America's First Purpose-Built Quantum Foundry, Supported by Proposed $1 Billion CHIPS Award”
Grader confidence: 0.60 · engine conviction 0.60
The edge's load-bearing identification — merchant quantum foundry capacity as the investable layer — was consensus-grade well before the 2026-08-05 freeze under branch (a): a primary government document (Commerce's proposed $1B CHIPS award, announced 2026-05-21) exists precisely to fund merchant quantum foundry capacity (Anderon, spun off to offer 300mm quantum wafer manufacturing services to the
The revenue inflection is a purchased, loss-making photonic-components business — lasers and photodetectors for lidar and aerospace — not quantum. Strip LSI and quantum revenue is ~$200K a quarter, BELOW the FY2025 run-rate the market thought it was leaving behind. And the specific asset the acquisition is meant to validate, the TFLN foundry, is named by the company itself as a subject of the pending securities litigation. Buying revenue to authenticate a claim that is under litigation for being
“"In other words, revenue increase largely reflects the inclusion of sales from recently acquired photonics and quantum communications businesses rather than organic growth from QCi's legacy operations." (The Quantum Insider, 2026-05-12). Independent second Tier-2 source, same day: "This growth is primarily attributed to the acquisition of Luminar Semiconductor, Inc. (LSI) in February 2026, which added established hardware sales and photonic components to QCi's portfolio." (Quantum Computing Report, 2026-05-12). Primary-document corroboration of the litigation sub-claim, QUBT Form 10-Q filed 2026-05-11: "the complaint alleges omissions and misrepresentations related to Quad M, QPhoton, NASA, millionways, and the TFLN foundry."”
Grader confidence: 0.87 · engine conviction 0.70
The edge decomposes into four assertions, and each was already in the public record at or before the 2026-07-26 freeze. (1) "The revenue inflection is a purchased, loss-making photonic-components business — lasers and photodetectors for lidar and aerospace — not quantum." This is the load-bearing claim, and it is a same-entity/same-mechanism match to two independent Tier-2 trade-press pieces publ
D-Wave's own capital allocation contradicts its narrative, and the proof is in the cash-flow statement, not the press release: the company that says annealing delivers value today spent 39% of its cash on a GATE-MODEL startup and then published a dated gate-model roadmap at its 2026-06-01 Investor Day — so the pivot rests on the company's own roadmap, not on inference. That lands $342.6M of goodwill on a business whose revenue just fell 81%. Meanwhile the advantage leg is no longer a forward hop
Earliest coverage: Quantum Computing Report (tier-2 trade press); consensus threshold actually crossed 2026-01-09 with The Quantum Insider (2026-01-07) + EE Times (2026-01-09) as two independent tier-2 publishers (2026-01-09)
“"D-Wave Quantum to Acquire Quantum Circuits Inc. for $550 Million to Accelerate an Error-Corrected Gate-Model Roadmap" (Quantum Computing Report, 2026-01-28); second independent tier-2 publisher: "D-Wave Buys Quantum Circuits in Shift to Higher Gear" (EE Times, 2026-01-09). For the advantage leg: "D-Wave Systematically Rebuts Flatiron Claims, Reaffirming Beyond-Classical Simulation Milestones" (Quantum Computing Report, 2026-05-26), published alongside "Flatiron Institute Tensor Network Algorithm Advances Classical Simulation, Challenging Boundary of Quantum Advantage" (same date).”
Grader confidence: 0.72 · engine conviction 0.65
The engine's edge is explicitly a COMPOSITE ("the stack nobody is adding up"), so I decomposed it into its three mechanisms and tested each against the frozen definition. LEG 1 — gate-model capital reallocation contradicting the annealing-delivers-today positioning. The acquisition target is Quantum Circuits Inc. ($550M, announced 2026-01-07, closed 2026-01-20; ~$275M cash component per a 2026-02
FormFactor is the only public company supplying BOTH ends of the cryogenic test stack — cryogenic wafer probe STATIONS and dilution-refrigerator CRYOSTATS — and the 2026-07-13/22/23 events convert cryogenic test from a lab instrument (one per research group, sold once) into a production metrology tool (one per test line, plus consumable cryogenic probe cards per wafer). Qualification into a leading-edge foundry test flow takes years, so this is a mechanical moat, not a preference: whoever is alr
Earliest coverage: Semiconductor Engineering — "Heating Up Cryogenic Wafer Testing" (Thomas Funke) (2020-05-12)
“"High volume testing demands a large chamber for full wafer samples, multi-channel cryogenic probe cards, a software-controlled wafer stage for step-and-repeat testing" — and, of FormFactor's system, "We are ahead of the pack for high-throughput cryogenic wafer test, with this unmatched combination of powerful features."”
Grader confidence: 0.55 · engine conviction 0.55
ADVERSARIAL FINDINGS. The edge is a conjunction and it splits cleanly at the joint, which is why this verdict is close. COMPONENT 1 — "FormFactor supplies both ends of the cryogenic test stack, and cryo test is scaling toward high-volume production with consumable cryogenic probe cards." This is NOT new at freeze and is refutable on the public record. Semiconductor Engineering published the produ
The primary, load-bearing thesis is OXIG.L NanoScience — the one listed name directly geared to the fridge duopoly, graded on NanoScience order intake + OXIG price-relative to the FTSE-250. The McEliece leg is demoted: Classic McEliece is international academic IP (Bernstein et al.), NOT UK-origin sovereign IP; the UK angle is only Post-Quantum Ltd's commercial position, graded solely on dated defence/gov tenders citing ISO 18033-2.
Earliest coverage: Oxford Instruments plc — "Sale of Oxford Instruments' quantum business, NanoScience" (LSE RNS primary regulatory announcement); corroborated by The Quantum Insider (11 Dec 2024) identifying NanoScience as OXIG's dilution-refrigerator maker (2025-06-10)
“"its quantum-focused business Oxford Instruments NanoScience ... up to £3m of deferred consideration linked to future revenues from quantum scaling systems."”
Grader confidence: 0.85 · engine conviction 0.50
The engine's primary claim identifies OXIG.L NanoScience as the single listed name geared to the dilution-refrigerator ("fridge") duopoly. This is not a differentiated second-order insight but the long-standing default consensus identification of Oxford Instruments' quantum exposure. The entity+mechanism (OXIG NanoScience = maker of Proteox dilution refrigerators; one half of the Bluefors/Oxford I
The non-obvious hit is the SECOND-ORDER channel. China exports only ~1-2% of global helium, so the headline volume looks trivial — but that flow was disproportionately RE-EXPORTED RUSSIAN helium routed into Europe. The US takes only ~5% of imports from China and is insulated; the differential loser is Germany/Europe. Grade the US-vs-Europe DIVERGENCE, not the aggregate ban. SHL.DE (Siemens Healthineers) is the geared listed name.
Earliest coverage: Reuters (via US News / TradingView newswire): "China Temporarily Bans Helium Exports as US-Iran Tensions Flare Again" (2026-07-10)
“Chinese companies have increasingly acted as intermediaries, importing Russian helium and re-exporting some volumes to overseas markets, including Europe. ... The United States is unlikely to be significantly affected because of its own supplies ... In contrast, the impact will be much greater for Europe and other countries that previously relied on Russian or Qatari helium but increasingly obtained those supplies through China.”
Grader confidence: 0.95 · engine conviction 0.55
The engine framed its edge as the non-obvious US-vs-Europe DIVERGENCE driven by China re-exporting Russian helium into Europe, with the US insulated. Every leg of that specific claim was already reported by Reuters on 2026-07-10, before freeze. The Reuters wire explicitly says Chinese companies imported Russian helium and re-exported volumes to Europe, that Europe's impact would be much greater, a
The graded detail: over the next 12 months Chinese flagship announcements (Zuchongzhi/Wukong successors) keep pace on QUBIT COUNT but produce no peer-reviewed below-threshold QEC (no Chinese program earns Q3 on the capability ladder) while Western programs add Q4s — the gap is at the error-correction rung, not the qubit-count rung, and the binding imports are now He-3 and cryo wiring/LNA payload capacity, not the fridge shell. Second claim: Chinese fridge makers start EXPORTING down-market units
Earliest coverage: Merics (Mercator Institute for China Studies), "China starts exporting quantum computers as its systems become more competitive" (2025-10-02)
“Origin Quantum is also preparing to export its Origin SL400 dilution refrigerator, which cools quantum systems, to Belt and Road countries such as South Africa, Russia and Iran.”
Grader confidence: 0.85 · engine conviction 0.55
The edge fails on both graded claims. (1) The core prediction — 'no Chinese program earns Q3 / no peer-reviewed below-threshold QEC over the next 12 months' — was contradicted ~7 months BEFORE the 2026-07-18 freeze: USTC's Zuchongzhi 3.2 demonstrated below-threshold surface-code QEC (Lambda=1.40, distance-7), published as a Physical Review Letters cover paper and Editors' Suggestion on 2025-12-22
The granular call: NY becomes the named cryo-CMOS and photonic-qubit foundry corridor. Specific claims: (1) the SEALSQ-GF MoU converts to a definitive agreement with a disclosed fab work-scope within ~9 months — highly material to SEALSQ (LAES, revenue base ~$10-20M); (2) a dollar-denominated, explicitly quantum-tagged NY CREATES / state incentive package or federal award lands on the Albany corridor within 12 months (the corridor's existing federal EUV/NSTC anchor shows the state plays at 9-fig
Earliest coverage: IBM and U.S. Department of Commerce joint announcement (IBM Newsroom / PRNewswire); corroborated same day by Bloomberg ("IBM Shares Soar on US Funding for $2 Billion Quantum Push") and WSJ (2026-05-21)
“Headquartered in Albany, New York as a standalone company, Anderon will operate as a state-of-the-art 300-millimeter quantum wafer foundry.”
Grader confidence: 0.92 · engine conviction 0.55
The engine's claimed edge rests on the premise that "nobody is naming New York State as the specific quantum-foundry geography." That premise was refuted 58 days before the freeze date: on 2026-05-21, IBM and the U.S. Department of Commerce jointly announced "America's First Purpose-Built Quantum Foundry" (Anderon), headquartered in Albany, NY at the NY CREATES/Albany NanoTech corridor, supported
The chokepoint claim: Belgium becomes the qualification gateway for external spin-qubit programs — the 'ASML pattern' where one small-country institution taxes a whole technology path. Specific claim (aligned with the falsifier): >= 1 additional quantum program beyond Diraq publicly signs an imec 300mm quantum-flow agreement within 12 months (candidate universe includes DARPA Stage B silicon players like Quantum Motion, plus hyperscaler silicon efforts), and/or imec announces a dedicated quantum
Earliest coverage: imec press release (EU Chips JU co-funded SPINS pilot line launch), corroborated same-week by Fraunhofer IAF/IPMS partner releases and Quantum Motion's own announcement (2026-04-07) (2026-04-03)
“SPINS aims to establish a lab-to-fab route for this technology via multi-project wafers (MPW) and standardized quantum process design kits (PDKs), which lower entry barriers for startups and SMEs.”
Grader confidence: 0.85 · engine conviction 0.50
The edge's falsifier-aligned claim ('>=1 additional quantum program beyond Diraq signs an imec 300mm quantum-flow agreement within 12 months, and/or imec announces a dedicated quantum PDK or foundry-access program') was not a differentiated forecast at the 2026-07-18 freeze — its second arm had ALREADY happened publicly on 2026-04-03, when the imec-coordinated, 50M EUR, EU Chips JU co-funded SPINS
Time-to-recognition — was the engine early?
world_events · protocol v1 · 2026-08-26Nothing here is a “lead” yet. This is the backward pass: for each frozen edge, did consensus-grade coverage of the specific second-order claim already exist at freeze? An edge that clears that bar is differentiated at freeze — a necessary condition, not proof of value. It becomes a proven-lead only when consensus actually arrives later (lead-days = recognition − freeze). Those accrue over 12–24 months.
Consensus-grade (frozen v1) = the same entity+mechanism claim in ≥1 tier-1 source OR ≥2 independent tier-2 sources. Conservative bias: already-consensus when uncertain — the grader’s job is to refute the engine’s novelty. All 19 already-consensus rows carry the exact quote that sank them. Rules were frozen (RECOGNITION_PROTOCOL.md v1) before this data.
Everyone knows ORCL CDS is at a record; the un-priced claim is its BETA to the long end. Because hung project-finance debt reverts to Oracle's own balance sheet, ORCL 5Y CDS widens >=80bp in the 30 sessions after a 30Y UST close above 5.40% while an equal-weight MSFT/GOOGL/AMZN/META CDS basket widens <25bp; ORCL equity underperforms that basket by >=10% over 60 sessions; Moody's or Fitch takes a negative action within 60 days.
Earliest coverage: Bloomberg, "Oracle-Tied $38 Billion Debt Takes Months To Spread Across The Market" (closest adjacent coverage: hung/slow-syndicating Oracle project-finance debt; does NOT state the rates-beta mechanism) (2026-04-15)
Grader confidence: 0.60 · engine conviction 0.45
Consensus-grade coverage AT/BEFORE 2026-08-26 exists for every UNDERLYING element: (1) ORCL 5Y CDS at records (~203-215bp, July 2026; Bloomberg/Morgan Stanley Nov 2025 note; CNBC, Seeking Alpha, Yahoo/GuruFocus 07-29); (2) Oracle widening more than Alphabet/Amazon/Microsoft/Meta (GuruFocus 07-29, ABC 08-06); (3) project-finance loans hung/slow to syndicate and banks hedging via ORCL CDS (Bloomberg
Ownership nationality decides who gets paid: the Houthi ban makes Saudi-flag tonnage the one part of the fleet that loses in a record market. Claim: 4030.SR underperforms Frontline (FRO) by >=15% from 2026-08-26 to 2026-11-30, and Bahri's Q3 report shows Bahri Oil segment revenue below Q2's despite record spot VLCC indices.
Earliest coverage: Lloyd's List / Lloyd's List Intelligence — 'Blanket exclusion for Saudi-linked tonnage under new Red Sea charterers' buyback' (17 Aug 2026), restated in Red Sea Brief 20 Aug 2026 (closest adjacent coverage; does NOT meet the consensus bar for the specific claim) (2026-08-17)
“Saudi-owned, operated and flagged vessels appear subject to blanket exclusions under charterers' buyback arrangements. [...] cover remains available for non-Saudi-linked tonnage.”
Grader confidence: 0.55 · engine conviction 0.45
Underlying events are heavily covered (Houthi ban on Saudi shipping from 20 Jul 2026; Bahri VLCCs Layla/Amzan struck; Yanbu disruption; record Q2 Bahri profit of SAR 2.75bn with Bahri Oil revenue +266% y/y; Bloomberg 23-24 Aug on Saudi logistics roiled and the Amzan strike). The engine's edge is a different, second-order claim: that ownership nationality makes Bahri the one loser in a record marke
The specific un-priced leg is the Yanbu closure forcing the SUMED shuttle — a ton-mile shock that persists even if Hormuz stays chronic and that the Saudi fleet cannot compete for. Claim: Frontline's Q3 2026 VLCC spot TCE EXCEEDS the Q2 2026 record ($181,700/day) and FRO outperforms 4030.SR by >=15% from 2026-08-26 to 2026-11-30. This is the long side of the Bahri pair, not a standalone thesis.
Earliest coverage: Bloomberg, "Saudis Shuttle Oil North on Sinokor Tankers to Evade Houthis" (closest Tier-1 coverage of the mechanism's underlying event; names Sinokor/Dynacom/DHT as shuttle operators, no Frontline, no Bahri-vs-independent framing, no rate/TCE claim) (2026-08-21)
Grader confidence: 0.60 · engine conviction 0.45
Refutation attempt: searched for pre-2026-08-26 consensus-grade statements of the specific second-order claim. Findings: (1) Underlying event is fully consensus — Bloomberg 08-03 (Yanbu bustles, dark transits) and 08-21 (Sinokor/Dynacom/DHT shuttling Yanbu→Ain Sukhna, ~16.3m bbl), Lloyd's List Red Sea Brief 08-06 (short-loading/part-discharge at Ain Sukhna, SUMED top-up at Sidi Kerir), Argus (Yanb
Not 'Lynas wins from China curbs' but a specific repricing: with JARE already holding 75% of medium/heavy output and Japanese magnet makers unable to source Dy/Tb from China since Nov 2025, the next Lynas-JARE change is a PRICE-MECHANISM move (floor above $110/kg or spot-linked pricing) rather than more volume. Claim: by 2027-01-31 Lynas or Sojitz announces a revised Japanese heavy-RE pricing mechanism, and Lynas' quarterly report discloses a heavy-RE/Dy-Tb price reference above $600/kg (if Lyna
Earliest coverage: None matching the specific claim. Closest (event coverage, not the claim): Argus Media, "Jare, Australia's Lynas expand rare earth supply deal", 2026-03-11 — reports NdPr $110/kg floor and HRE volume terms; states no HRE floor and no future pricing revision. (2026-03-11)
Grader confidence: 0.60 · engine conviction 0.40
Searched for Tier-1/Tier-2 coverage asserting that the next Lynas-JARE/Sojitz change would be a heavy-RE (Dy/Tb) pricing mechanism (floor above $110/kg or spot-linked) and/or a Lynas-disclosed heavy-RE price reference >$600/kg. Findings: (1) March 10-11, 2026 deal coverage (Argus T2; Mining.com/Northern Miner citing Canaccord; SMM; Metal Tech News) reports a US$110/kg NdPr floor with 30% upside sh
OCP's share gain is a contract shift, not a price spike: with China legally absent and the US CVD suspended, OCP's Brazil+US MAP/DAP tonnage in Jul-Dec 2026 runs >= +30% y/y (US >= 250kt from ~zero), and the un-owned expression is Morocco sovereign USD paper tightening >=25bp vs EMBI GD — a beneficiary DM fertilizer positioning (MOS/NTR) does not capture.
Earliest coverage: None found for the specific claim. Nearest non-matching coverage: Federal Register 2026-07-24 (CVD sunset review, primary doc; event only); Argus "Commerce recommends keeping Moroccan phosphate duty" (event only); DTN 2026-07-17 / World Fertilizer 2026-08-24 (first OCP cargo, event only); Ecofin Agency 2026-03 "S&P affirms Morocco at BBB- but Gulf supply shock puts key export earner at risk" (sovereign-OCP link, but opposite direction and pre-suspension); discoveryalert.com.au Brazil phosphate outlook (notes OCP incentive to prioritise US flows; not a consensus-grade publisher, no sovereign-spread mechanism).
Grader confidence: 0.62 · engine conviction 0.55
The headline event (Trump 29-Jun-2026 proclamation suspending CVD on Moroccan phosphate for 8 months; OCP's first 54kt TSP cargo at New Orleans ~Aug 18-24; China's zero-export mandate on DAP/MAP through Aug 2026) is thoroughly covered by primary government documents, Argus, DTN, World Fertilizer and farm press. That does not satisfy the protocol. The engine's second-order claim has two distinct le
Directionally against the crowd: MP is more exposed to the risk NOT firing than to it firing. The specific detail is that the DoD floor is also an effective ceiling on operating leverage — a floor converts MP from a commodity call option into a regulated-margin processor, so convexity from a genuine cutoff accrues to unhedged non-floor producers and to magnet fabricators, not to MP. Second detail: the grading event is a calendar date, which is rare for rare-earth calls.
Earliest coverage: none found matching entity+mechanism (closest non-qualifying neighbor: Seeking Alpha, "MP Materials: A Miner Trying To Become A Contracted U.S. Magnet Platform," single blog, neutral/positive framing) (2026-02-13)
Grader confidence: 0.70 · engine conviction 0.33
No source of any tier, at any date up to the 2026-07-22 freeze, was found asserting the engine's specific second-order claim: that the DoD $110/kg floor is also an effective ceiling on MP's operating leverage, converting MP from a commodity call option into a regulated-margin processor, with cutoff convexity accruing instead to unhedged non-floor producers and magnet fabricators. Tier-1 coverage o
Headline crack spreads overstate complex-refiner realized margin during a SOUR-supply shock, because the differential leg inverts against configuration advantage. The falsifiable detail is segment-level and disclosed: realized refining margin per barrel can be tested directly against the change in the benchmark Gulf Coast 3-2-1 crack, so the mechanism can be confirmed EVEN IF the stock rises. A simple refiner (PBF) is the comparator that isolates configuration.
Grader confidence: 0.60 · engine conviction 0.30
The underlying event (sour-supply shock; Mars flipping to a premium over WTI) was heavily covered by Tier-1 from March 2026, and Reuters even warned generically that sour premiums 'will eat into refinery margins.' But the engine's specific second-order claim — that headline Gulf Coast 3-2-1 cracks OVERSTATE complex-refiner realized margin because the differential leg inverts against configuration
TIGHTENED: state it explicitly as a PAIR TRADE graded only as a spread — long STNG / short Frontline (FRO) — with the claim being the PERSISTENCE differential, not the direction of tanker rates. Crude ton-miles reverse the day the strait reopens; clean-product ton-miles persist because refinery restart schedules, cargo programming and term contracts reset on a multi-month lag. The entry is right only if persistence shows up in the index data (BCTI holding relative to BDTI) as well as in the equi
Earliest coverage: NEAREST-MISS (does NOT establish consensus): Vortexa, "Tanker freight behaviour diverges after the conflict shock", republished by Hellenic Shipping News (2026-04-28)
Grader confidence: 0.60 · engine conviction 0.36
The underlying event (2026 US-Iran war, Strait of Hormuz closure from ~Feb 27-28, 2026, still disrupted at freeze with the blockade resumed per Reuters Jul 16-21) is saturation-covered, but coverage of the event does not count. The edge's specific second-order claim is the PERSISTENCE DIFFERENTIAL expressed as a spread (long STNG / short FRO; BCTI holding vs BDTI post-de-escalation). Adversarial f
Suez Canal Authority monthly transit receipts fall FURTHER during a Hormuz shutdown rather than recovering, because Gulf-origin northbound cargo — not Asia-Europe container traffic — is the marginal barrel through the canal. That is the opposite of the rerouting reflex. Second-order: it forces a renewed EGP devaluation step or IMF review slippage at a moment when the Gulf sponsors are themselves absorbing an energy-export disruption.
Earliest coverage: none establishing consensus. Nearest miss for the record: "Al-Sisi Warns Hormuz Disruption Could Further Strain Egypt's Economy and Suez Canal Revenues", Egyptian Streets / Al Arabiya, 2026-03-02 — direction-only overlap, fails mechanism and second-order match under protocol v1 (2026-03-02)
Grader confidence: 0.72 · engine conviction 0.58
The engine's specific second-order claim — SCA monthly receipts fall FURTHER in a Hormuz shutdown because Gulf-origin northbound cargo (not Asia-Europe boxes) is the marginal barrel, forcing an EGP devaluation step or IMF review slippage while Gulf sponsors are themselves stressed — appears in NO tier-qualifying source at/before the 2026-07-22 freeze. Stronger still: the at-freeze Tier-1 record as
TIGHTENED: the first measurable Taiwan-risk event of this cycle is an energy-cost event with no military content — and the test requires MAGNITUDE and ATTRIBUTION, not the mere existence of a Taipower loss (a base-rate event). Specifically: the MOEA Electricity Tariff Review Committee approves an industrial tariff increase of at least 10% at a scheduled 2026 H2 meeting, with the announcement or accompanying Taipower disclosure explicitly citing LNG procurement cost. This lets an operator underwe
Earliest coverage: none found (closest adjacent coverage: Reuters/Straits Times/Taipei Times on the 2026-03-27 rate FREEZE noting a September review; BofA 2026-03-27 note saying LNG shortage fears fade and tariffs stay stable — both contradict rather than match the claim)
Grader confidence: 0.72 · engine conviction 0.45
The underlying event stream is well covered (Iran/Hormuz LNG cost spike hitting Taiwan's grid; the semi-annual MOEA tariff review; the March 2026 freeze), but per protocol that does not count — the test is the SAME second-order claim: the MOEA Electricity Tariff Review Committee approving a >=10% industrial tariff increase at the scheduled 2026 H2 (September) meeting with explicit LNG-procurement-
TIGHTENED: the binding constraint is physical cargo ALLOCATION, not solvency, and it fires at a TTF level far below crisis pricing — around EUR90-100/MWh, not the 2022 EUR300. The unfalsifiable Western fast-fashion gross-margin coda is CUT. The TTF gate becomes a required leg, and the disarm is explicit.
Earliest coverage: none establishes consensus; nearest-miss: Bloomberg, "Iran War Forces Cash-Strapped Asian Nations to Buy Expensive LNG" (fails mechanism match — solvency/price frame, buyers still securing cargoes, no TTF gate) (2026-07-21)
Grader confidence: 0.60 · engine conviction 0.45
The underlying event (Hormuz/Iran conflict, Qatar force majeure, Pakistan's record $21.88/MMBtu spot buys, QatarEnergy halving Bangladesh deliveries per Reuters Jul-6) is saturated Tier-1 consensus — but the protocol excludes underlying-event coverage. The engine's specific second-order claim has a distinctive signature: binding constraint = physical cargo ALLOCATION (not solvency), gated at TTF ~
TIGHTENED with a magnitude and an acknowledged ceiling: the claim is that WCS-Brent compresses by at least $4/bbl versus its H1-2026 average for six consecutive weeks, verified at the producer level in disclosed realised pricing, with Pacific-destination non-US export volumes as corroboration ONLY — Trans Mountain nameplate caps the volume leg, so the differential is the load-bearing channel. Because the routing premium is separate from flat price, Canada can outperform even if crude falls, whic
Grader confidence: 0.68 · engine conviction 0.50
The underlying machinery is consensus: TMX narrowing WCS differentials, Asian buyers paying a premium amid the 2026 Iran war, and Korea tripling Canadian crude imports are all Tier-1/wire covered before freeze. But the protocol requires the SAME second-order claim, and the engine's edge is the tightened country-level composite: WCS-Brent compression of >=$4/bbl vs the H1-2026 average sustained six
Not all subprime lenders are the same trade, and the catalyst is decelerating not accelerating — CACC's Q1'26 forecasted-net-cash-flow revision was only -$9.1M (0.1%), the smallest in three years. So the differentiated expression is the RELATIVE pair: short CACC, not-short ALLY, whose auto credit actually IMPROVED in Q1'26 (60+ delinquencies and NCOs down YoY). CACC is uniquely exposed because its forecast-collection accounting books forward markdowns while ALLY's incurred-loss book improves. Th
Grader confidence: 0.72 · engine conviction 0.58
The underlying headline (subprime-auto delinquencies at multi-decade highs; CACC exposed) is saturated consensus at freeze, but that is the first-order event the protocol explicitly excludes. The engine's edge is a three-part second-order claim: the short-CACC/not-short-ALLY RELATIVE pair, isolated by an ACCOUNTING-MODEL divergence (forecast-collection/CECL front-loading markdowns vs incurred-loss
The only differentiated claim, and what to grade, is the Tier-1-vs-OEM decoupling plus renewal-fizzle symmetry: recycling + stockpile + allied frameworks together cover <5% of German magnet demand THIS cycle, so traction-motor suppliers ration BEFORE any OEM discloses an output cut — and because the base case is a China renewal (RFF's temporary-restriction read), crowded rare-earth-panic hedges are a paid-to-lose position if the risk fizzles. The market belief that 'frameworks and recycling have
Earliest coverage: RFF issue brief ib-25-12 'The Strategic Game of Rare Earths: Why China May Only Be in Favor of Temporary Export Restrictions' — nearest-adjacent (supports only the renewal-base-case half; does NOT establish the graded compound claim, so not consensus-grade for this edge) (2025 (RFF IB-25-12))
Grader confidence: 0.60 · engine conviction 0.50
The engine concedes the base ('Germany dependent on Chinese REEs; auto/wind exposed') is already-consensus — correctly; that is Tier-1 covered (S&P Global, Reuters) and not graded. The graded edge is a compound positioning call. I ran 6 targeted searches decomposing it: (1) <5% framework-coverage-this-cycle, (2) Tier-1 rations before OEM discloses cut, (3) RFF temporary-restriction/renewal base ca
Turkey is a near-pure levered play on the product CRACK, not crude, and the unique tell is the GOLD-funded reserve drain: with near-total product-import dependence the crack is the direct input, and the drain is being covered by gold sales — so the gradeable dashboard is CBRT net reserves + gold, not CPI. The reserves+gold combination distinguishes this from every generic Turkey-CPI short; a crack normalization below $30 that refills the buffer is the specific way the trade burns.
Earliest coverage: Kitco News — "Turkey taps its gold reserves, sells 58.4 tonnes of gold in two weeks" (tier-1, closest consensus, but underlying-event only) (2026-03-26)
Grader confidence: 0.66 · engine conviction 0.50
The engine's stated consensus ("Turkey chronically fragile, high inflation") is correct and its edge is a specific analytic construction. Consensus coverage at/before 2026-07-14 is heavy but frames Turkey as (a) levered to CRUDE oil price spikes, (b) selling gold to defend the lira/fund imports, and (c) an inflation/CPI story — the exact opposite emphasis to the engine's edge. The engine's specifi
The residual mispricing is relative-value: INDA still carries India as a net-oil-importer VICTIM, so the gradeable claim is Reliance/INDA OUTPERFORMING EEM THROUGH a rising crack, confirmed by rising product-export volumes — the very diesel-crack shock that damages Indonesia/Turkey is a windfall for Indian export-refiners arbitraging discounted Russian crude into premium European product. Grade the relative outperformance plus the export-volume tell, not the crack windfall in the abstract.
Earliest coverage: JP Morgan note via IndiasNews/Tribune: 'Diesel crack surge, INR depreciation and retail tailwinds support Reliance outlook: JP Morgan' — Tier-1 sell-side, but covers only the excluded crack-windfall/absolute-earnings angle, NOT the INDA-vs-EEM relative-value claim. (2026 (pre-freeze; covers excluded angle only))
Grader confidence: 0.60 · engine conviction 0.55
The claim decomposes into a consensus part the engine deliberately excluded and a residual the engine wants graded. CONSENSUS (excluded): (1) rising diesel crack lifts Reliance — Tier-1, e.g. JP Morgan 'Diesel crack surge... boosts Reliance outlook', Goldman 'upcycle refining margins'; (2) India as Russian-crude refining hub with rising product-export volumes — CREA 13-Jul-2026 report, which the e
The big four's ¥15tn paper loss is the headline; the mispriced name is Japan Post Insurance because its J-ICS ESR (197% new-standard, target floor 150%) is exposed through the mass-lapse module that rising rates inflate, not through the HTM mark (offset by liability revaluation). Claim: if 30Y JGB trades above 4.40% for >=5 sessions, 7181.T underperforms 8750.T by >=5% over the following 20 sessions AND JPI's next ESR disclosure prints below 180% (a >=17pt fall from 197%) — the first big-four-or
Earliest coverage: Japan Post Insurance, 2026年3月期 決算・経営方針説明会 (FY2026/3 results and management policy briefing), primary issuer disclosure (2026-05-29)
“26.3末のESRは大量解約リスクの影響により181%と、25.3末から15ポイント低下したが、適正水準の範囲内にある。大量解約リスクを除いた場合のESRは220%と適正水準の上限水準にあり、足許では解約率の増加は見られていない … 金利変動による大量解約リスクの変動が大きいことにより、円金利上昇および低下時の感応度が大きい状況にある (ESR変動要因: うち大量解約リスクの増加 △16pt)”
Grader confidence: 0.72 · engine conviction 0.40
The engine's 'edge' (JPI's J-ICS ESR is exposed through the mass-lapse module that rising rates inflate, not the HTM mark) is the issuer's own publicly stated ESR bridge, disclosed three months before freeze: March-2026 ESR 181% (not 197% — the 197% figure is the Sept-2025 new-basis number and is stale), a 15pt fall of which -16pt is explicitly attributed to mass-lapse risk, with JPI itself flaggi
Japan-specific: China's Dy/Tb/Y zero to Japan is a bilateral (Taiwan-remarks) sanction running on its own clock, so any Nov 10 extension of the US-China suspension does NOT restore Japanese flow. First observable damage is a yttria-coated chamber-part delivery slip at a Japanese tool maker or coater (TEL/Screen/AGC/Shin-Etsu) disclosed in Q3-Q4 FY26 results, BEFORE any Japanese automaker reports a magnet-driven line stop. Grade on the disclosure legs only; drop the EWJ/ACWI leg.
Earliest coverage: MINING.COM / The Northern Miner (reporting Bloomberg's question to China MOFA and spokesperson Lin Jian's reply; original trigger a Nikkei report of a US request) (2026-06 (exact day not retrieved; MOFA briefing "on Tuesday"; NAI500 repost dated June 2026))
“China has doubled down on its months-long restriction on exports of rare earth products to Japan despite reports of the US asking Beijing to lift the measures. ... Lin's statement came as he was asked by Bloomberg about an earlier report by Nikkei claiming that the US government had asked China to allow the resumption of rare earth metal sales to Japan due to concerns over the global supply chain related to Japan's technology sector.”
Grader confidence: 0.60 · engine conviction 0.40
The edge has two legs. Leg 1 (Japan's Dy/Tb/Y zero is a bilateral Taiwan-remarks sanction on its own clock that a US-China Nov 10 extension does not restore): by June 2026 a Tier-1 exchange existed — Bloomberg asked China's MOFA about a US request to resume rare-earth sales to Japan and the spokesperson reaffirmed the ban as a dual-use/remilitarization measure (primary government statement). Reute
Against a record FY26 remittance base that the IMF program arithmetic now extrapolates, GCC-origin remittances turn negative y/y in Q4-2026 (Sep-Dec prints) with a 2-3 quarter lag from Gulf construction/services labor demand — visible first in Saudi non-oil PMI < 50 and Pakistani worker-visa issuance, not in Brent. The trade is the 2031 Eurobond vs an oil-hedged peer, graded on SBP country-level data.
Earliest coverage: IMF Country Report No. 26/101, Pakistan: Third Review Under the Extended Arrangement Under the Extended Fund Facility (staff report), as reported by Dawn 'IMF highlights Gulf exposure as biggest external risk for Pakistan' (2026-05-15)
“A significant disruption to the GCC economies and/or return of migrant workers could weigh on these flows, a major source of financing for consumption and the balance of payments”
Grader confidence: 0.70 · engine conviction 0.40
Entity match: Pakistan's GCC-origin remittances (IMF: 55% of remittances, ~9% of GDP, from GCC). Mechanism match: the IMF staff report frames the risk channel as GCC economic disruption / return of migrant workers (i.e. Gulf labour demand) weighing on remittance flows and the balance of payments — the same second-order transmission the engine's edge names, and explicitly distinct from the first-or
Under-application is worse than the industry's own -8% delivery forecast: ANDA Jul-Oct 2026 deliveries print <= -10% y/y, with the shortfall concentrated in phosphate (MAP/DAP) after China's end-August renewal, and the loss shows up in credit — Banco do Brasil agri-NPL +>=50bp q/q in Q3/Q4 2026 and FIAGRO/CRA spreads widening — before it shows in CONAB's 2026/27 yield estimate (Feb-Apr 2027). The DM read (EWZ on Selic/election) never prices the September application decision.
“"We hadn't seen the Brazilian market decline since 2001 while maintaining steady year-over-year growth, and this year all indications are that we will experience an estimated 12% contraction," Altieri said. [S&P sub-head: Phosphate demand loss drives decline; Altieri called for better access to farm credit, which has become scarcer and more expensive.]”
Grader confidence: 0.80 · engine conviction 0.45
The edge's load-bearing components were all in consensus-grade print before the 2026-08-26 freeze. (1) Deliveries worse than -8%, phosphate-led: S&P Global (Tier-2, 2026-07-22) carries Yara Brazil's -12% (49 -> ~43 Mt) with phosphate demand loss as the driver; The AgriBiz (2026-07-13) has the same executive at -14%/42 Mt; Mosaic's Q2 call (2026-08-05, widely reported incl. Argus/S&P coverage of Br
India's export refiners (RIL Jamnagar, Nayara) are long the stalemate, not the spike: the trade pays only while Hormuz stays shut AND Brent stays below ~$100 — Urals discount >= $4/bbl on >= 2.0 mb/d plus NWE diesel crack >= $30. Hormuz reopening (Windward transits back above ~20/day for 4 weeks) is a FAIL branch, Brent > $105 for 10 sessions is the VOID branch; grade INDA/EEM and RIL Q2 FY27 GRM on that split. The consensus India hedge loses money if Hormuz merely stays shut.
Earliest coverage: Bloomberg, "India Refiners Reap Fuel Export Windfall as War Drives Shortages" (syndicated by Business Today citing Kpler/Bloomberg) (2026-07-14)
“Oil-product shipments from India are on track to climb to their highest level since September as refiners cash in on robust profit margins after a Russian export ban and hostilities in the Middle East tightened supplies. ... Russian crude has meanwhile helped Indian refiners maintain high operating rates despite the disruption. Data showed India imported 2.6 million bpd of Russian crude so far this month, accounting for more than half of its total imports.”
Grader confidence: 0.70 · engine conviction 0.45
Entity + mechanism match at Tier-1 six weeks before freeze: Bloomberg (07-14) reported Indian refiners reaping an export windfall from Hormuz/Russian-ban shortages while running discounted Russian crude (quote reproduced from the Business Today syndication at https://www.businesstoday.in/amp/latest/economy/story/india-fuel-exports-10-month-high-refiners-russian-oil-middle-east-squeeze-542917-2026-
TIGHTENED: not 'OZK has CRE risk' but that OZK is the highest-beta listed equity to the LONG END specifically, decoupled from the credit cycle — pain arrives via exit-cap-rate widening on completed projects, not borrower distress. The gradeable detail is the SHAPE: because average REGA loan size is a large multiple of peer average, deterioration should appear as step-function non-accrual additions concentrated in a handful of individually disclosed large credits, NOT as a rising provision rate,
“Bank OZK "moved a pair of CRE loans — including one office credit — to nonaccrual status" and "reported three commercial real estate charge-offs, including another office credit." (nonperforming loan ratio jumped to 0.90% from 0.20% while the bank "reported solid profits"); corroborated by CRE Daily 2026-04-24: "Nonperforming assets doubled to 1.08% of the portfolio, attributed mainly to five loans totaling $409.5M."”
Grader confidence: 0.72 · engine conviction 0.40
The engine's tightened claim has two parts. Part 1 (OZK = highest-beta listed equity to the long end; pain via exit-cap-rate widening on completed projects, not borrower distress): no coverage articulating this transmission channel was found — searches for OZK + exit cap rate returned nothing, and pre-freeze coverage attributes deterioration to project/sector distress (life-science leasing woes, o
For THIS shock India is net long, because the binding global constraint is Gulf REFINING and product-export capacity, not crude availability, and India holds the arbitrage. Gradeable detail: PPAC monthly petroleum product EXPORT volumes rise y/y while Reliance's O2C segment EBITDA for the Sep-2026 quarter exceeds its trailing four-quarter average — i.e. India's oil trade balance improves during an oil shock, the opposite of the textbook. Honest residual risk: tightened secondary sanctions on Rus
Earliest coverage: Bloomberg — "India Refiners Reap Fuel Export Windfall as War Drives Shortages" (also syndicated by The Economic Times 2026-07-14 and MSN 2026-07-19) (2026-07-14)
“India Refiners Reap Fuel Export Windfall as War Drives Shortages”
Grader confidence: 0.80 · engine conviction 0.58
The engine's second-order claim is: for this (Hormuz 2026) shock the binding constraint is refined-product/refining-export capacity rather than crude availability, and India — via its refiners' product exports — captures the arbitrage, inverting the textbook 'oil up = bad for India' read. By the 2026-07-22 freeze this exact entity+mechanism pairing was in Tier-1 coverage: Bloomberg on 2026-07-14 r
TIGHTENED: an INTRA-GULF divergence claim, with the UAE REMOVED from the short leg because ADNOC's Habshan-Fujairah pipeline (roughly 1.5-1.8 mb/d to a Gulf-of-Oman terminal) is its own Hormuz bypass — shorting it would short the second-best-positioned country in the region. Short leg is Qatar only (instrument-clean). Stated explicitly: both Saudi and UAE bypass terminals discharge into waters inside the risk zone, so the claim is relative resilience, not immunity. Export CAPABILITY, not reserve
Earliest coverage: Bloomberg, "Gulf Economies at Risk of Worst Slump Since '90s on Iran War" (Goldman Sachs note, economist Farouk Soussa), verified via Moneycontrol republication archived on Wayback 2026-03-16 (2026-03-15)
“Economists say Qatar, Kuwait and Bahrain remain the most vulnerable because their energy exports depend heavily on the strait. By contrast, Saudi Arabia and the UAE may partly cushion the shock thanks to higher oil prices and their ability to route some crude exports through alternative pipelines and shipping routes.”
Grader confidence: 0.72 · engine conviction 0.50
The edge's core second-order content — intra-Gulf divergence with the UAE taken off the exposed side because its pipeline bypass preserves export capability, and Qatar on the most-exposed side because its exports (LNG) have no bypass — was tier-1 consensus by mid-March 2026, four months before the 2026-07-22 freeze. Bloomberg's Mar 15, 2026 report of the Goldman Sachs note (Farouk Soussa) ranked t
TIGHTENED: the Vaca Muerta energy leg is STRIPPED (Argentina's improving energy balance is now consensus and dilutes the claim). The surviving claim is that El Nino is a NET POSITIVE terms-of-trade event for Argentina and therefore a cheap hedge against the very food shock that hurts EM importers — capturing the VOLUME leg that a long-grains position structurally cannot. Self-disarm stated honestly: at Nino-3.4 above roughly +2.0C the Pampas benefit can invert into flood and wheat-disease losses
Earliest coverage: Reuters, "How El Nino threatens emerging market economies" (quoting Morgan Stanley's Fernando Sedano); earlier partial corroboration: Economist Intelligence Unit, "How will El Nino affect Latin America and the Caribbean?" (2023-06-21) (2023-08-04)
“There are, however, exceptions - Argentina had a record soy harvest in previous El Nino episodes, according to Morgan Stanley. "El Nino tends to be negative in EM, though Argentina is an exception," the bank's Fernando Sedano wrote in a note, adding "Argentina is likely the only net winner of El Nino."”
Grader confidence: 0.93 · engine conviction 0.50
The engine's surviving claim — El Nino is a NET POSITIVE for Argentina and therefore an inverse exposure to the EM food shock — was stated almost verbatim in Tier-1 press during the prior (2023-24) El Nino cycle, long before the 2026-07-22 freeze. Reuters (2023-08-04) framed El Nino as broadly negative for EM food importers (India rice, Philippines, Thailand, Egypt-style CPI exposure) while explic
The channel that hurts CRWV EVEN IF AI demand stays strong is collateral repricing, not demand. Bulls assume robust GPU demand rescues the equity, and 2026 already saw an $8.5B DDTL at ~5.9% (A3, 'first IG GPU-backed financing') plus a $3.1B facility — but those were fresh customer-contract-wrapped tranches. The genuine test is whether the ~$4.2B 2026 maturity wall can be rolled WITHOUT a new hyperscaler-contract wrapper, on pure H100 collateral value, at sub-10%. If CoreWeave can only refinance
Earliest coverage: Global Data Center Hub — "Is CoreWeave's $8.5B Deal the GPU Asset Class Moment?" (2026-04-07)
“Rating agencies are underwriting the hyperscaler customer, not the GPU hardware. Long-term take or pay agreements from Meta or Microsoft function as near sovereign credit support.”
Grader confidence: 0.79 · engine conviction 0.60
The engine's edge decomposes into (1) a collateral-repricing channel that bites even if AI demand stays strong, and (2) the diagnostic that CoreWeave's cheap/IG financing depends on fresh contract wrappers rather than bare GPU collateral value (the "uncontracted-tranche" test). Both were consensus-grade before freeze. Global Data Center Hub (Apr 7, 2026) states verbatim: 'Rating agencies are under
Two specifics past the consensus: (a) the pain is the distillate CRACK, not crude — carriers and analysts anchor to Brent/WTI, but jet fuel tracks a $110-140-equivalent crack while crude is ~$72, so fuel cost surprises up even in a flat-crude world; (b) AAL is the highest-beta expression because it is both fully unhedged AND the most leveraged major, so it is hit on the fuel channel AND the rates/refi channel at once — a convergence clean-balance-sheet peers (DAL/LUV) do not share.
Earliest coverage: IATA Economic Report — "A renewed rise in the jet crack spread adds to cost pressure" (2026)
“A renewed rise in the jet crack spread adds to cost pressure”
Grader confidence: 0.70 · engine conviction 0.52
The edge packages two specifics. Specific (a) — 'the pain is the distillate crack, not crude; jet fuel tracks an elevated crack while crude is ~flat' — is the explicit 2026 specialist consensus, not a differentiation. IATA published an economic report titled literally 'A renewed rise in the jet crack spread adds to cost pressure,' matching entity (airlines/jet fuel) and mechanism (crack/refining m
Everyone is hunting for the credit BLOWUP (defaults). The earlier-firing, differentiated channel is fee-engine reflexivity at the manager level: OWL's valuation embeds continued perpetual-vehicle inflows, and redemption gates choke inflows LONG BEFORE realized credit losses appear — so OWL de-rates on FLOW/redemption data, not default data. Best as short OWL vs diversified alts (BX/APO) whose fee base is less non-traded-BDC-dependent.
Earliest coverage: Sell-side downgrades/target cuts on OWL retail private credit flows (Deutsche Bank/Barclays to Hold, Jan 2026 onward); concretely captured in a dated April 7, 2026 report quoting Goldman Sachs on flow/redemption-driven fee pressure at OWL's non-traded BDCs. (2026-04-07)
“Goldman Sachs flags weaker flow trends and higher redemptions at a peer vehicle as a headwind for Blue Owl's non traded BDCs, citing potential pressure on management fees and fee related earnings if retail private credit sentiment stays soft.”
Grader confidence: 0.90 · engine conviction 0.50
The edge's specific second-order claim is: OWL (Blue Owl) de-rates on FLOW/redemption data rather than default data because its valuation embeds continued perpetual/non-traded-BDC inflows, and redemption gates choke inflows before realized credit losses appear. This exact entity+mechanism was consensus-grade before the 2026-07-14 freeze. (1) Tier-1 sell-side: an April 7, 2026 report states 'Goldma
The mispriced variable is DURATION, not level. Consensus treats the crack as a transient geopolitical spike that mean-reverts; the structural facts (Russia flipped to net diesel importer, permanent Gulf capacity offline, rationing spreading) make the distillate crack elevated-for-longer. VLO's distillate-heavy, heavy-sour-capable slate captures the diesel-SPECIFIC crack better than the gasoline-weighted 3-2-1 the market models, so forward EPS is under-modeled even after the run.
Earliest coverage: FinancialContent / MarketMinute — 'Valero Energy Surges 5.8% as Analysts Signal Structural Shift in Refining Profits' (citing Raymond James/Justin Jenkins, Bank of America, Goldman Sachs) (2026-03-27)
“the refining industry has entered a 'structural, not just cyclical' period of heightened profitability ... the current margin environment is likely to persist due to a lack of new global refining capacity and the high cost of entry for new players.”
Grader confidence: 0.72 · engine conviction 0.60
The edge's central wedge is that DURATION is the mispriced variable: consensus supposedly treats the distillate crack as a transient geopolitical spike that mean-reverts, while the engine argues structural facts make it elevated-for-longer, leaving VLO forward EPS under-modeled. Adversarial searching refutes the premise that this was non-consensus at the 2026-07-14 freeze. By March 27, 2026 — near
The market models GEV on backlog volume; the under-appreciated variable is PRICING/MARGIN on that backlog. H1-2026 new-order pricing is already running 10-20 percentage points higher per kW than Q4-2025. With slots sold out to 2030 and no competing capacity able to arrive inside the transformer bottleneck, GEV is booking scarcity-rent price escalation — the earnings surprise comes from equipment gross MARGIN expansion, not order count. The transformer shortage is itself the moat that lets GEV ra
Earliest coverage: GE Vernova Q1 2026 earnings call / SEC 8-K (CEO Scott Strazik), corroborated by Utility Dive "GE Vernova gas turbine backlog hits 100 GW as prices rise" (2026-04-22)
“"We continue to be in that 10% to 20% growth in price on new bidding and winning activity today relative to where we were in the backlog in the fourth quarter of last year," Strazik said during the earnings call. New pricing for 2026 Power equipment orders is 10%-20% above Q4 2025 levels, with immediate margin expansion.”
Grader confidence: 0.97 · engine conviction 0.58
The engine framed its edge as: the market models GEV on backlog VOLUME while the under-appreciated variable is PRICING/MARGIN, with H1-2026 new orders priced 10-20 points/kW above Q4-2025 and the earnings surprise coming from equipment gross-margin expansion. This is not a differentiated second-order read at freeze (2026-07-14) — it is management's own explicitly-disclosed guidance from three mont
The mispriced feature is the ASYMMETRY the $110 floor creates. The market still prices MP on spot rare-earth volatility, but the floor sits at roughly 2x prevailing spot NdPr, so the classic bear case — 'China floods the market and crushes MP's price' — is contractually neutralized: MP's realized price is downside-protected while upside stays open. MP is being valued as a beta commodity name when it now has bond-like downside plus magnet-margin optionality.
Earliest coverage: JPMorgan (analyst Bill Peterson) initiation/note, reported via Benzinga / SahmCapital: "MP Materials Backed By $110 DoD Price Floor As JPMorgan Sees 33% Upside" (2026-02-13)
“that creates asymmetric risk-reward — downside protection with leverage to higher pricing.”
Grader confidence: 0.96 · engine conviction 0.55
The engine's edge is the ASYMMETRY the $110 floor creates: bond-like/downside-protected realized price while upside stays open, neutralizing the 'China floods the market' bear case, versus a market still pricing MP as a beta commodity. The freeze date is 2026-07-14, but the underlying DoD Price Protection Agreement was announced in July 2025 (floor effective Oct 2025), giving the market ~12 months
The second-order, non-obvious link: the grid-EQUIPMENT bottleneck is BULLISH for incumbent generators, not just equipment makers. Because transformers/turbines are sold out for years, new dispatchable capacity cannot arrive to compete, so the scarcity rent accrues to EXISTING baseload (CEG's nuclear) as a margin uplift on assets already in the ground. Consensus frames the bottleneck as a cost/risk; for incumbent generators it is a supply moat.
Earliest coverage: Data Center Knowledge — "Why AI Data Centers Make Existing Power Plants More Valuable" (2026-06-30)
“Combined with permitting delays, interconnection backlogs, and supply-chain constraints, those higher costs have made existing assets more attractive.”
Grader confidence: 0.65 · engine conviction 0.52
The engine's edge combines two threads. The genuinely narrow differentiator — attributing the incumbent-generator moat SPECIFICALLY to the equipment (transformer/turbine) bottleneck rather than to permitting/interconnection, and reframing that bottleneck from cost/risk to moat — was not found cleanly asserted in that exact form (the primary.vc turbine piece, 2026-02-10, explicitly frames the bottl
Markets are anchored on a benign Brent tape (~$85) while the fiscal gap widens with the product CRACK the government won't pass through — Jakarta has politically frozen subsidized pump-diesel prices despite the rupiah slide, so every ~$10 on the crack widens the 2026 gap regardless of where Brent sits. The gradeable tell is the crack + BI reserves DIVERGING from a calm crude tape, not crude itself.
Earliest coverage: Bloomberg, "Indonesia Commits to Low Fuel Prices to Defend Growth, Social Stability" (entity+policy); BNP Paribas note "Indonesia Fuel Subsidies Risk 2026 Deficit Breach, Rupiah Pressure and Funding Strain" (second-order claim) (2026-04-02)
“Subsidised fuel prices — Pertalite and Bio Solar — will remain frozen through end-2026, with the government absorbing the widening gap between international prices and domestic pump prices ... the cost will ultimately hit the government purse.”
Grader confidence: 0.68 · engine conviction 0.55
The engine's substantive second-order claim — Jakarta has politically frozen subsidized pump-diesel prices despite the rupiah slide, so the government-absorbed gap widens the 2026 fiscal deficit and pressures the rupiah/reserves — is deeply consensus-grade well before the 2026-07-14 freeze. Tier-1 coverage matching the SAME entity (Indonesia subsidized diesel/Biosolar) AND the SAME mechanism (froz
The under-priced link is second-order: a Gulf refined-product/energy squeeze hits Pakistan twice — a higher diesel-import bill AND softer GCC remittances that fund ~62% of external inflows — a shared-shock correlation the market prices as two independent risks. Grade only the Pakistan USD sovereign curve / EMBI Pakistan sub-index, not a broad EM-bond ETF.
“After the Middle East war started, Pakistan faced some capital outflows and sovereign spreads temporarily widened to around 500bps ... A significant disruption to the GCC economies and/or return of migrant workers could weigh on these flows, a major source of financing for consumption and the BOP ... Access to short-term commercial financing, which is largely from GCC banks, could also be impacted if risk sentiment deteriorates significantly.”
Grader confidence: 0.82 · engine conviction 0.44
The engine's specific second-order claim is that a Gulf energy squeeze hits Pakistan via two correlated channels (diesel-import bill + GCC remittances funding external inflows) and that the Pakistan USD sovereign curve underprices this shared-shock correlation as two independent risks. Every component was consensus-grade before freeze: (1) The twin-channel mechanism (higher import bill AND weaker
This is a positively-convex bet whichever way Beijing decides on Nov 10: the A$ price-floor + strategic reserve structurally de-risk heavy-REE project financing precisely as Lynas becomes the sole ex-China Dy/Tb source, so Australian heavy-REE equities should decouple UPWARD from the China-suppressed REE price on tightening — and if China renews and keeps prices low, the price-floor STILL fires as a subsidy backstop. The price-floor-as-downside-hedge is the load-bearing, differentiated claim.
Earliest coverage: AMEC Design Paper for Australia's Critical Minerals Strategic Reserve (government-commissioned), reported by The West Australian / The Nightly (2025-12 (design paper) / 2026-01-02 (mainstream reporting))
“If prices fall, the government tops up. If prices spike, Canberra shares the upside.”
Grader confidence: 0.68 · engine conviction 0.60
The engine explicitly names 'price-floor-as-downside-hedge' as the load-bearing differentiated claim, conceding the upside/decoupling story is well-worn. But the downside-hedge/convex structure it treats as novel is precisely the publicly-stated DESIGN of Australia's price-floor scheme: the government-commissioned AMEC design paper (Dec 2025) frames the floor as de-risking rare-earth (incl. Dy/Tb)