The world's reactors need more uranium each year than its mines produce, and that gap is set to widen sharply into the 2030s. Reactors required about 68,900 tU (~179M lbs U3O8) in 2025. The World Nuclear Association expects that to more than double, to just over 150,000 tU (~390M lbs) by 2040 in its reference case.
- Persistent gap: mine output has run below reactor needs for most of the past decade. Inventories and other secondary sources that filled the gap are running down.
- Concentrated supply: Kazakhstan alone supplies about 39% of mined uranium, and it has repeatedly cut its output guidance.
- Slow response: new mines take 10–20 years from discovery to production, so higher prices cannot quickly bring on new supply.
- Utilities are under-contracted: requirements are largely covered to 2030 but only about a third covered by 2034. Uncovered needs total more than 3 billion lbs through 2045.
- Price signal: long-term contract prices are around $93–94/lb in 2026, near 18-year highs.
The demand side
Reactor demand for uranium is set to more than double by 2040 because the global fleet is expected to almost double. The World Nuclear Association's 2025 Nuclear Fuel Report puts capacity at 398 GWe in June 2025 (WNA).
- 2025: about 68,900 tU (~179M lbs U3O8) of reactor requirements.
- 2040, lower scenario: 552 GWe of capacity, needing over 107,000 tU (~280M lbs).
- 2040, reference scenario: 746 GWe, needing just over 150,000 tU (~390M lbs).
- 2040, upper scenario: 966 GWe, needing over 204,000 tU (~530M lbs).
All three scenarios were revised upward from the 2023 edition, by 35–66 GWe. Demand is very hard to cut once a reactor is running: utilities cannot switch fuel, and uranium is a small share of generating cost.
Demand growth comes from four sources:
- New builds: about 60+ reactors are under construction worldwide, led by China and India.
- Life extensions: many US and European reactors are being licensed to run 60–80 years instead of 40.
- Restarts and uprates: closed plants are being restarted (Palisades, Crane, Duane Arnold) and existing units are raising output.
- New sources of demand: AI data centres (Driver 1) and the 38-nation pledge to triple nuclear (Driver 3).
Conversion: 1 tU ≈ 2,600 lbs U3O8. Pound figures are our conversions of WNA tonnage.
The supply side
Mine production in 2024 covered about 90% of reactor requirements. That is a recovery from 74% in 2020, but still short (WNA). World output was 60,213 tU, up from 47,731 tU in 2020.
Supply is concentrated in a few countries:
- Kazakhstan: 39% (23,270 tU). The state-controlled Kazatomprom alone produces 21% of world output.
- Canada: 24% (14,309 tU), mostly from McArthur River/Key Lake (13% of world output) and Cigar Lake (11%) in the Athabasca Basin.
- Namibia: 12% (7,333 tU), with Husab alone producing 7%.
- Australia: 4,598 tU.
Three countries supply about three-quarters of the world's mined uranium, and a handful of mines carry the market. A single flood, acid shortage or policy change can take a meaningful share of supply offline.
Secondary supply is shrinking. The gap between mine output and reactor demand has historically been filled from several sources (WNA):
- Down-blended weapons uranium: the US–Russia "Megatons to Megawatts" programme ended in 2013.
- Re-enriched tails and underfeeding: this depends on spare enrichment capacity, which is tightening as the West moves away from Russian enrichment (Driver 4).
- Commercial inventories: utilities drew these down through the 2011–2020 price slump.
- Recycled fuel (MOX and reprocessed uranium): small and slow to scale.
Why supply can't respond quickly
Price alone can't close the gap quickly, because new uranium supply takes about a decade to reach the market. Counting exploration and permitting, the full cycle takes 15–20 years (IndexBox).
- A decade of underinvestment. After Fukushima (2011), prices stayed below what most new mines need to break even for roughly ten years. Exploration budgets collapsed, and major operations such as Cameco's McArthur River were idled.
- Kazakhstan keeps disappointing. Kazatomprom cut its nominal 2026 capacity from 85M to 77M lbs, removing about 8M lbs (~4% of global supply) (Discovery Alert). Its growth is limited by tight sulphuric acid supply, skilled labour shortages and a history of downward guidance revisions.
- Even expansions are slow. Fresh investment at existing Athabasca Basin operations takes 3–4 years to produce first output.
- Permitting is the bottleneck. Environmental review, Indigenous consultation and licensing add years before construction even starts, especially for conventional mines.
- Restarted mines are already counted. Most idled capacity that could restart cheaply came back during 2022–2025, so the easy supply response has largely happened.
The result is a deficit that IndexBox sees growing to about 18% of demand by 2027.
Utility contracting and price signals
Utilities have covered their needs to 2030 but face a cliff in the early 2030s, and that has to be contracted at today's prices or higher (Discovery Alert).
- Coverage cliff: utility requirements are roughly fully contracted through 2030 but only about 36% covered by 2034.
- Uncovered needs: cumulative uncovered requirements exceed 3 billion lbs through 2045, about 17 years of today's total reactor demand.
- Term prices: long-term contract prices sat at $93–94/lb from March to July 2026, with forward references near $96.50/lb. That is around an 18-year high.
- Spot prices: about $89.54/lb in late September 2026, with term trading roughly $6/lb above spot. A term price above spot is a sign that buyers are paying for security of supply.
- Cameco's view: the company has said a contracting cycle of this size has never been seen before, and that its peak "remains ahead".
Term prices matter more than spot. Most uranium sells under multi-year contracts, and a high term price is what lets developers finance new mines.
Risks and counterpoints
The deficit is structural, but these factors could soften or delay it:
- Kazakhstan surprises upward: if acid supply and labour improve, Kazatomprom could add pounds faster than expected. Mitigant: it has a long record of cutting guidance, and its output increasingly goes to China and Russia rather than Western utilities.
- Large inventories: WNA estimates civil stockpiles at about 132,000 tU in China, 40,000 tU in the US and 36,000 tU in Europe (end-2022). Mitigant: China's stockpile is strategic and unlikely to be sold to Western buyers, and utility inventories elsewhere have been drawn down.
- Enrichment underfeeding: spare enrichment capacity can stand in for some natural uranium. Mitigant: Western enrichment is tight as Russian supply is phased out, which pushes the opposite way (overfeeding).
- Reactor delays: if new builds slip, demand grows more slowly. Mitigant: the deficit exists today on the current fleet, before most new builds start up.
- Price volatility: spot prices can swing sharply on financial buying and selling. Mitigant: long-term contract prices, which drive mine financing, have been more stable.
Investment implications
A long-running deficit, slow-responding supply and an approaching contracting cliff reward the companies that can deliver new pounds in the 2030s. That favours advanced North American projects.
- The 2030s supply has to start now. Given 10–20 year development timelines, the mines that will fill the post-2030 gap must already be in exploration or permitting today.
- Jurisdiction matters. With about 39% of supply in Kazakhstan and Western buyers avoiding Russian-linked material, US and Canadian deposits carry a security premium.
- Term prices support the economics. Prices in the low-to-mid $90s/lb make projects viable that weren't during the 2011–2020 slump.
- Utilities need new suppliers. With roughly two-thirds of 2034 needs uncovered, utilities are motivated to sign offtake agreements with credible developers, which can help fund construction.
- Scarcity value for resources in the ground. As few new discoveries reach production, defined resources in safe jurisdictions become more valuable as acquisition targets.
This report is for information only and is not investment advice. Forward-looking figures are estimates drawn from third-party sources and our own conversions.
Sources