Tearing apart the uranium utility contracting gap

Uranium.

The financial press is calling it a bull market. CNBC has a graphic with a glowing atom. Every second newsletter is explaining the "nuclear renaissance" to you in 800 words that could have been four. Here's what they're not explaining: the utilities that actually buy uranium to fuel the reactors that supposedly make this a renaissance have spent thirteen consecutive years buying less uranium than their reactors require. Not a typo. Thirteen years.

That is the mechanism nobody wants to sit with. Because sitting with it requires you to do math that produces an uncomfortable number.

A utility buying below its replacement rate doesn't run out of uranium the next morning. It runs out on a schedule — a very predictable schedule — and the schedule is determined by the length of its existing contracts plus however long it can stretch its inventory buffers. Most utilities are now at a point where the buffers are thin, the existing contracts are rolling off, and the forward contracting pipeline looks like a road with a bridge out ahead.

Into this queue — which was already full of utilities that had put off their homework — walked Meta and Microsoft, signing agreements for 7.8 gigawatts and 800-plus megawatts of nuclear capacity respectively, in the first quarter of 2026 alone. They are not price-sensitive buyers. They need baseload power that doesn't carry gas price exposure, and they will pay whatever it takes to get it. They have never, to my knowledge, called a uranium broker at 5 AM from a loading terminal in Saskatchewan.

The market is pricing a spot price of around $86 per pound. The term price — which is where real procurement happens — just reached $90 per pound, the highest since 2008. Those two numbers are the tell. The spot market is sideways. The term market is telling you where utilities think this ends up. Those are not the same bet.

Story off. Numbers on.

The supply side of this machine has three broken parts running simultaneously. First: Kazatomprom, which controls approximately 43% of global primary uranium production, was projecting 85 million pounds out of Kazakhstan in 2026. The number they gave guidance on in early 2026 is 62 million pounds. That is a 27% shortfall from their own five-year-old targets. Not from some analyst's model — from their own. The primary culprit is Budenovskoye, a joint venture expansion that has missed every construction milestone since 2022. No new completion timeline was offered in the last press release. Just: construction is underway.

Second: Cameco at McArthur River, the highest-grade uranium mine on earth, delivered 10% less in 2025 than 2024 due to development delays in new mining areas. McArthur is supposed to be the supply backstop. It is currently the supply backstop with a maintenance window that ran long.

Third: SOMAÏR in Niger — zero production in 2025, the full year, under military junta control. Gone. Not disrupted. Gone.

The secondary supply cushion — the post-Fukushima overhang that let utilities avoid signing long-term contracts for most of the 2010s — is effectively gone. The World Nuclear Association puts secondary supply's contribution to reactor needs declining from 11–14% today to 4–11% by 2050. The tide that covered the rocks has gone out. The utilities that spent thirteen years not contracting are now contracting for replacement fuel in a market that can't produce meaningfully more before the early 2030s. The pipeline doesn't lie: in-situ recovery projects in the US take 7–10 years from permitting to first production. Athabasca Basin underground mines take 10–15 years from discovery. The supply that will deliver on 2026 contracts either already exists or is late-stage today. There is no other supply.

Long-term contract prices at $90 per pound are not the market getting ahead of itself. They are utilities accepting, finally, that higher prices are the price of being able to keep the lights on. I traded the thesis as a "when, not if" since 2023. The "when" appears to be now.

Nobody is properly pricing the Kazakhstan exploration lock.

On December 29, 2025 — a date that fell between Christmas and New Year's when approximately nobody in Western financial markets was paying attention — Kazakhstan quietly amended its subsoil use legislation. The amendment gave Kazatomprom priority rights over all future uranium exploration licenses in prospective areas. If another company finds uranium on a block Kazatomprom has its eye on, that company does not get to develop it. The state takes it. Extensions and production increases at existing joint ventures may now require Kazatomprom to hold at least 90% ownership, potentially compressing foreign partners to a 10% position.

Think about what this means for the supply response everyone is counting on. The bull thesis for uranium assumes that high prices eventually incentivize new mine development. The standard playbook: price rises, explorers drill, mines get built, supply catches up in 10–15 years. Kazakhstan just removed roughly 43% of global primary supply from that playbook. You can't incentivize exploration there in any meaningful Western-investor sense. You can't build a mine there and own it. You are a 10% passenger on Kazatomprom's schedule.

The second weak link is the queue itself. Utilities contract 2–5 years ahead of delivery, with uncovered requirements extending 7–10 years forward. The 13-year streak of undercontracting has stacked the queue so deep that even an aggressive contracting cycle starting today produces deliveries mostly in the 2028–2032 window. Every year utilities delayed, they pushed the delivery date further right. The gap between what the market needs and what it can produce in time is not a gap that higher prices close quickly. The lead times are baked in. The geology doesn't care about the price of uranium futures.

There is also the Russia factor, which I will not spend three paragraphs on because you have already read three paragraphs about it in every other uranium letter. The short version: Russian enrichment and conversion capacity — which handles fuel processing downstream of the mine — is being displaced by sanctions and export restrictions. Western enrichment cannot replace it on the same timeline as reactor demand growth. The fuel cycle has a chokepoint at conversion and enrichment that even abundant primary mine supply would not immediately solve. The mine is not the whole machine. This is a machine with multiple sequential bottlenecks, and right now three of them are constrained at once.

The trigger is the contracting cycle normalizing. Not a single dramatic event — no mine explosion, no export ban, no sovereign shock. Just utilities running the spreadsheet for 2028 and realizing their uncovered requirements are larger than anything they can source at today's prices in six months' time. At that point the stalemate breaks. The question is what breaks with it.

The first-order consequence is term prices moving above $90 in the back half of 2026 as contracting volumes normalize. Spot has been anchored near $85–87 because the Sprott Physical Uranium Trust — the main institutional buyer that drives spot episodic surges — has a 9-million-pound annual purchase cap and has already deployed most of it year-to-date. When that buying pauses, spot drifts. It's not bearish. It's a scheduled pause. The underlying term market is the signal, and the underlying term market is not pausing.

The second-order consequence is jurisdiction repricing. The Kazakhstan exploration lock changes the long-run supply thesis in ways the market has not fully repriced. If the 43% of global primary supply that sits in Kazakhstan cannot meaningfully expand outside of Kazatomprom's own timeline and budget, then the marginal pound at the end of the decade comes from Canada's Athabasca Basin or Australia's ISR operations — both higher cost, both longer lead time. The incentive price for Athabasca conventional mining is estimated around $70–80 per pound all-in. We are already there. Development pipelines in those jurisdictions are the incremental supply response that actually exists.

The broad uranium ETFs capture this thesis imperfectly. They hold everything — diversified miners, companies with significant exposure to Kazakhstan JVs now at risk of being squeezed to 10% partners, producers whose forward books are already sold at prices below where the market is heading. The surgical trade is in unhedged producers with western-jurisdiction assets and development-stage projects that benefit directly from a $90+ incentive price regime.

One honest caveat: the spot market has had a habit of disappointing in this cycle. It peaked at $101 in January 2026 and gave back 15% in seven days when geopolitical sentiment shifted. The paper market moves fast and erratically; the physical market moves slowly and in one direction. Thirteen years of undercontracting does not unwind in a quarter. The physical layer will keep winning. It just takes longer than you want it to — and then, when it goes, it goes all at once and the headlines are always surprised.