☢️ uranium-edge

Where US utilities actually source uranium — by country, by delivery type (spot vs contract), by year. ~98% import dependency in 2024 (677K domestic / ~40M demand) with 28% concentration in a single country whose production is structurally impaired. Source: EIA-858 annual survey.

The macro thesis in one report. Primary mine supply by region vs three demand scenarios. Demand grows 30% by 2035 while supply grows only 15% — the gap widens every year. Carry inventory exhausted as early as 2028 in the bull case. Source: IAEA PRIS + WNA + EIA.

Ten-year history of domestic production (EIA-851 quarterly) vs fleet consumption (EIA-923). US production collapsed 98% from 1980 peak to near-zero in 2019–2021. The ISR revival is real but covers only 4 cents on the dollar of domestic need.

The US fleet still buys uranium at an average $68/lb because most contracts were signed 2012–2018 at $30–50/lb. As those contracts expire through 2029, the fleet average converges toward spot. EIA-923 Schedule 2 tracks actual delivered cost per plant each month.

Q4 2025 output of 1,043K lbs was 3.2× Q3, driven by Energy Fuels' White Mesa Mill (685K lbs alone). ISR-only — Lost Creek (ur-energy), Alta Mesa (enCore), Willow Creek (UEC) — was ~352K lbs. 2025 full year ~2.16M lbs (est.), highest since early 2000s but still just ~5% of the fleet's ~40M lbs annual consumption — ~95% import dependent.

2025 is the strongest year since 2018. White Mesa Mill (UUUU, 1.0M lbs) and Alta Mesa (EU, 0.6M lbs) account for 75% of output. Uranium One Willow Creek / Christensen Ranch restarted in 2025. Covers just ~5.4% of the ~40M lbs/yr US requirement — White Mesa's licensed 8M lb/yr capacity is the binding constraint on domestic conventional supply growth.

Bottom-up demand from actual NRC generation data. Every unit, every month. The highest capacity-factor fleet in the world — every 1% CF increase = ~541K lbs incremental demand.

Remaining lifetime U₃O₈ consumption per unit through license expiry. Every license renewal adds years of locked-in demand with zero new mine supply needed.

The forced-buying mechanism: as old contracts expire and Russian supply is banned, coverage drops below 85% — utilities must buy at whatever price clears. Waterfall model shows the exact year coverage breaks. Source: EIA-858 + Public Law 118-67.

Fleet-wide: EIA-858 shows committed supply drops below 85% by 2027–2028. Russian ban waiver removes 6M lbs/yr at expiry. Per-utility: 10 operators ranked by exposure from SEC 10-K disclosures. Constellation most exposed by volume; smaller operators most exposed by coverage ratio.

Per-utility profiles: fleet size, estimated annual uranium consumption, implied procurement volume, country-of-origin mix. Constellation alone consumes more uranium than all of France. Duke and Southern complete the top 3.

Tracks known utility contracting activity — term contracts signed, MOUs, LOIs, and active discussions — against each utility's uncovered requirement. 82% of tracked developer production capacity remains uncontracted. Source: EIA-858 + company disclosures.

EIA-858 UMAR 2024 (Sep 2025) Table 22. Utilities are actively rebuilding inventory — 55.9M lbs purchased in 2024 vs ~40M consumed. But 184.2M lbs of requirements remain unfilled through 2034, with 2028 the cliff year where existing long-term contracts expire en masse.

Monthly pulse from EIA Electric Power Monthly (6 weeks lag) combined with per-plant capacity factor rankings and historical trend 1990–2025. The Apr–May outage trough removes ~2.2M lbs vs the August peak. Includes operator averages, seasonal demand model, and implied U₃O₈ by month.

Per-plant uranium efficiency: EIA-923 Sch2 delivered cost × Sch3 generation = actual $/MWh of nuclear fuel by plant. Even if all contracts rolled to spot today, nuclear fuel ($4.24/MWh) stays 84% cheaper than gas ($26.90/MWh).

EPA eGRID 2023 operational emissions + IPCC AR6 lifecycle comparison. Nuclear's carbon advantage vs gas is worth more per MWh than nuclear's entire fuel cost. Sensitivity table from $30 to $200/ton CO₂. Source: EPA eGRID 2023 · IPCC AR6 WG3.

Form 1 Account 518 captures the full nuclear fuel cycle (U₃O₈ + conversion + enrichment + fabrication) — 3× what EIA-923 shows. Uranium is only 34% of total fuel cost. Even if uranium doubled to $136/lb, nuclear stays 50% cheaper than gas. Source: FERC Form 1 + EIA-858 + 10-K.

Zero Emission Credit programs by state — rate formulas, expiry dates, plant-level ZEC revenue, and policy renewal risk. NY uses a dynamic formula (SCC minus LBMP); IL is market-sensitive with a $16.50 cap; NJ fixed $10/MWh. NY cliff expiry is the biggest fleet economics risk in the near term.

Nuclear generation as a percentage of total electricity by RTO/ISO region and state. Built from EIA Electric Power Monthly + EIA-923. Grid dependency map — where nuclear goes offline, gas fills the gap. Illinois runs on nuclear majority.

When the long-term indicator exceeds spot, utilities are paying a supply-security premium — the classic re-contracting signal. In the 2003–2007 bull cycle the inversion peaked at +$44/lb. Source: UxC Long-Term Price Indicator, cited via Cameco quarterly MD&A.

Policy, regulatory, and production milestones that move the uranium market. Section 232 tariff window, DOE Reserve rounds, CNSC licensing decisions, KAP quarterly production, utility contracting seasons, NRA milestones. No earnings noise.

The Dec 2027 TENEX waiver expiry forces US utilities to re-source 8.5M SWU/yr from a Western enrichment market already at 85–90% utilization. $7B in subsidies cannot build centrifuge capacity in time. Tracks all four federal programs and the UF₆ conversion gap.

CCJ, NXE, DNN, UEC, UUUU, EU, PALAF scored across 4 pillars: resource quality (grade, M&I confidence, jurisdiction), stage & timeline, cost structure (AISC vs spot margin), and valuation (EV/lb, P/NAV). Normalized ranking updated each quarter. Source: SEC filings · KAP IR · CNSC.

Margin and free cash flow sensitivity across four spot price levels for each producer. Shows who prints cash at $70/lb, who breaks even, and who needs a sustained bull market to survive. Operating leverage quantified per producer. Source: SEC 10-K MD&A · company cost guidance.

Sprott Physical Uranium Trust NAV premium/discount tracker. When SPUT trades at a premium, at-the-market issuance buys physical uranium and tightens spot supply. Historical premium/discount vs spot price correlation. A leading indicator of physical market sentiment. Source: Sprott daily NAV disclosures.

Sprott (SPUT), Yellow Cake (YCA.L), and Uranium Royalty Corp (URR) compared by lbs held, NAV premium/discount, fee structure, and liquidity. Physical trusts hold real uranium off-market — their combined inventory is a meaningful fraction of global above-ground stocks.

Cameco's legacy contract portfolio repricing model. As below-market contracts signed 2010–2018 roll off, CCJ's realized price converges toward spot/term. Each maturing contract adds $10–15/lb to realized price. The repricing story is the core CCJ earnings driver through 2029. Source: Cameco MD&A · annual reports.

Pre-earnings model vs actual results, line by line. EPS within 7.5% of actual. Realized price within $0.07. Westinghouse was the main miss. Full methodology and what we'd adjust for Q2. Free — open now.

Q2 2026 earnings teardown — updated when Cameco reports. Key metrics to watch: realized price progression vs legacy contract roll-off schedule, new contract volume signed in H1 2026, and any revision to full-year production guidance. Source: Cameco Q2 2026 earnings release + MD&A.