The US is phasing in a ban on Russian enriched uranium while replacement capacity takes years to build.
85%likely
Well sourcedNext check
ArchivedArchived · already in the price
Why it is different
Already played out in the price. Kept for scoring and for its checkpoints, not as a live idea.
Why it matters
Fuel costs feed into every reactor running and planned, though most of this is in the price already.
Next check · 15 Nov 2026
Does Centrus’s Q2-2026 report say the average price it achieved for enrichment rose year on year? (Q1-2026: up 52%.)
How we scored it ·
from how similar cases went
If right, how big3/5How overlooked2/5Upside vs downside2/5Source trust5/5Independent sources5/5
The exact wording we score
The published TradeTech or UxC Western SWU spot assessment is at least $100/SWU on 31 Dec 2031. Resolves YES at $100 or above; NO below $100. Russian waivers, enricher utilisation and reactor demand are explanatory mechanisms, not alternative resolution legs.
PROOF
Evidence and provenance
The links below are the evidence recorded when this forecast was scored. Each note states source quality or bias; a citation is not an endorsement.
Recorded sources · 5 citations / 5 independent origins
2026-07-28 | sell-side: enrichment scarcity is established; mainstream press: Russian-supply constraints are widely covered; priced market: SWU near $200 already reflects the bottleneck; specialist: nuclear-fuel commentary treats scarcity as consensus | E2
Revision history6 entries
Probabilities and interpretations are never silently overwritten. This is the append-only record of what changed and why.
Two checkpoints resolved YES; P held at 85. Centrus Q2-2026 (2026-08-05): volumes -23%, average SWU price +3% — YES (P 70, Brier 0.09), but +3% after Q1's +52% is a decelerating roll-through, not pricing power. Urenco H1-2026 (2026-08-13): end-June spot $200/SWU, unchanged from December — YES at the threshold (P 60, Brier 0.16); order book €27.3bn. Neither moves the $100 end-2031 floor or the edge. New checkpoint: Centrus Q3-2026 price direction, P=55 — Q2 is the first sign the repricing is exhausting itself.
The parent was not scoreable as written because “structurally elevated,” “pricing power” and “sold out” were three undefined conditions across a four-year interval. It is now one exact end-date price claim at the existing $100 kill threshold. Mechanism evidence remains in the file but cannot rescue or kill the outcome.
(2nd pass) — Harvest confirmed by review, not by automation. Under v2.1 the E≤2 rule only flags. Reviewed: edge here really is gone — SWU at ~$200 is ~4x pre-war, Centrus has multi-bagged, and this file itself says the thesis 'has largely already played out in price'. Unlike lpa-lowering (where the flag was wrong because new research had raised edge), nothing here argues the edge back up. Status set to harvest deliberately. Checkpoints stay tracked for calibration; the P=85 reflects a leg that is already true, not a live opportunity.
(2nd pass) — P 76→85, and a dead falsifier removed. The load-bearing leg ('SWU stays well above pre-war levels', operationalised at ~$100 against a $200 spot and a ~$55 pre-war level) requires spot not to halve — it was already true at logging time, which this file admits ('the thesis has largely already played out in price'). A claim whose main leg is already satisfied belongs at ~85, not at 76 alongside genuinely uncertain forecasts. Separately, the kill-criteria block still led with 'DOE grants a broad 2026-2027 waiver extending Russian LEU' and the body still listed it as a pending H2-2026 catalyst — that waiver was granted on/around 2025-08-19. The lead kill condition had already fired without killing the thesis, and the body was stale for a year. Removed; the Ukraine-settlement / TENEX-return path is now the sole live bear. This matters beyond one file: it is an observed counterexample to assumption A1 (US statutory deadlines bind as written), on which pqc and smr both depend — see assumptions/README.md.
v2 rescore. P held at 76. Auto-demoted to 🌾 Harvest by the E<=2 rule — edge fell 3→2 on 2026-07-23 when closer reading showed the move was already made, and v1 left it ranking in the active table anyway. It now sits in Harvest by rule, with its checkpoints still tracked. Stale catalyst corrected: this file lists a DOE Centrus waiver decision as pending in H2-2026; that waiver was in fact granted on/around 2025-08-19 covering all committed Centrus deliveries in 2026 and 2027. The bear catalyst has already fired and SWU still printed $200 at end-2025 — mildly supportive. No sanctions relief: a revised Sanctioning Russia Act of 2026 was introduced 2026-07-14 (escalation, not relief).
created, then deepened same day: mechanism verified against legislation/EIA/Urenco accounts; edge cut to 2 (now consensus, SWU already ~4×), probability→76, trust→5. Flagged harvest/archive candidate.
Method challenge
Adversarial review record
Archived for accountability
Structured internal role review tied to this frozen claim. It is not independent human peer review. Independent specialist review not yet performed
credit or macro specialist
market-expression and risk reviewer
filing and accounting reviewer
forecasting-method reviewer
Strongest specialist challenge
The thesis is already widely recognized and the spot price was roughly double the future threshold at call time. A high probability can be statistically coherent while offering no remaining information advantage or attractive entry.
Outside view
Commodity bottlenecks tend to mean-revert as capacity, substitution, inventory, and policy respond. The 85 percent estimate is supported mainly by the large cushion to 100 dollars, not by assuming scarcity persists unchanged for five years.
Causal rival
Russian supply normalization, Western enrichment expansion, Chinese exports, demand disappointment, or changed waiver policy can cut the assessment sharply. Long-term contracting can also reduce the relevance of spot without changing physical availability.
Measurement risk
TradeTech and UxC assessments are proprietary. Urenco's annual report has historically quoted the exact year-end level, but publication occurs later and is not guaranteed; if neither assessment becomes publicly reproducible the call must be disputed, not guessed.
Residual risk
The forecast is retained for accountability and calibration, but the active investment edge has been consumed. Any future public display must keep its harvested status and must not imply that an 85 percent probability is a current recommendation.
COLDTELL