The Nuclear Fuel Chain's Hidden Chokepoint

Uranium spot has done almost nothing for four months. Since April 2026 it has sat in an $84 to $87/lb range, a long way below the brief spike above $100/lb in January, which was itself more a function of capital-raising by the Sprott Physical Uranium Trust than of utility demand. If you only watched the spot tape, you would conclude the uranium story has stalled.

You would be looking at the wrong number.

The long-term contract price (the price utilities actually pay when they sign multi-year supply agreements, and the number that determines whether new mines, conversion plants, and enrichment cascades get built) has kept climbing through the same period. Kazatomprom’s own Q1 2026 disclosure put its long-term price indicator at $91.50/lb, up more than $11 year-on-year, with some offtake agreements carrying ceiling prices as high as $140 to $150/lb. That gap between a flat spot price and a rising term price is the entire story right now, and it is worth understanding why it exists before deciding what it implies for portfolios.

Spot and term prices measure different buyers. Spot reflects financial players, traders, and trusts moving pounds around in the short term. Term prices reflect utilities, the world’s most conservative buyers, locking in fuel for reactors they intend to run for decades. When utilities are willing to pay up for delivery years out while the spot market goes nowhere, it usually means the people closest to the physical supply chain see something the trading tape doesn’t yet.

Three things, roughly in order of how much attention each deserves.

First, reactor demand has a new, fast-moving component. Every major U.S. hyperscaler now has at least one nuclear power deal on the books for AI data-center load. Microsoft signed a 20-year, $16 billion power purchase agreement to take the full 835 MW output of the restarted Three Mile Island Unit 1 (rebranded Crane Clean Energy Center), targeting service in late 2027. Amazon has committed roughly $700 million to X-energy for small modular reactor development alongside a separate multi-billion-dollar buildout at its Susquehanna, Pennsylvania campus. Google has contracted for 500 MW from Kairos Power’s small modular reactor fleet. Meta’s combined commitments across TerraPower, Oklo, Vistra, and Constellation could add up to roughly 6.6 GW of capacity. None of this uranium demand shows up in a reactor tomorrow (most of these projects land between 2027 and the mid-2030s), but it has changed how utilities and financiers think about the multi-decade fuel-supply picture, which is precisely the horizon the term market prices.

Second, the supply side of primary mining is not expanding to meet it, and this is a policy choice as much as a geological one. Kazatomprom, which accounts for roughly a fifth of global mine supply, is deliberately cutting its 2026 nominal output by about 10% (from 32,777 tU to 29,697 tU), and it has been explicit that this is not forced by any operational constraint: the company confirmed sulfuric acid supply, its historical bottleneck, is stable for 2026. Management has called it an “operate-below-100%,” value-over-volume stance. With the long-term price still well under levels the company considers attractive relative to depleting ore grades at its legacy deposits, Kazatomprom would rather hold pounds back than sell into a price it views as too low. That is a rational monopolist-adjacent pricing decision, not a sign the company is running out of ore, though it does compound a structural gap that was already there. The World Nuclear Association’s 2025 World Nuclear Fuel Report puts identified supply sources at covering only about 46% of projected 2040 demand under its reference case, leaving a gap analysts have sized at somewhere between roughly 200 and 212 million pounds of U₃O₈ by 2040, depending on the demand scenario used.

Third, and this is the part that gets far less attention than mining, the bottleneck has moved downstream, into conversion and enrichment, the two processing steps between a mined pound of uranium and reactor-ready fuel.

There are more than forty active uranium mines globally. There are only three Western commercial conversion plants: Cameco’s facility in Ontario, Orano’s operations in France, and Solstice Advanced Materials’ Metropolis Works in Illinois (spun off from Honeywell in late 2025 and the only U.S. conversion facility). By share of Western capacity, Cameco holds roughly 19%, Solstice roughly 18%. Solstice is running its Metropolis plant toward a 20% capacity increase in 2026, backed by an order book exceeding $2 billion, largely from domestic utilities locking in years of forward supply, itself a decent proxy for how tight utilities expect this link in the chain to stay. Historically, more than 3,400 tonnes of conversion demand shifted away from Russian suppliers toward these three companies as Western utilities decoupled from Rosatom-linked material, adding further load onto capacity that was never sized for it.

Enrichment is tighter still, and structurally different in kind: it is dominated by four entities, two of which, Russia’s Rosatom and China’s CNNC, serve essentially no Western utilities and are not building capacity for a market they’ve been cut out of. On the Western side, Urenco (UK-headquartered) and Orano hold most of the remaining capacity, and neither is unconstrained: any enrichment work tied to U.S. national-security missions legally requires U.S.-origin technology, which rules out both regardless of price or available capacity.

This is where U.S. policy has been most deliberately interventionist, and it is worth being precise about the sequence and the actual dollar figures rather than the rounder numbers that circulate in sector commentary:

• May 2024: The Prohibiting Russian Uranium Imports Act banned imports of Russian low-enriched uranium effective August 11, 2024, while allowing the DOE to grant waivers under strict, declining annual caps. Critically, any waiver must terminate no later than January 1, 2028, a hard deadline that forces U.S. utilities off Russian enrichment services within roughly eighteen months from today, waiver or not.
• Early January 2026: The DOE awarded roughly $2.7 billion, appropriated under that same 2024 law, to domestic uranium enrichment companies, including Centrus, Urenco USA, and Orano, specifically to expand conversion and enrichment capacity.
• Mid-January 2026: A Section 232 proclamation on processed critical minerals formally classified import dependence on materials including uranium as a national-security risk and directed the Commerce Department to negotiate remedies with trading partners, potentially including price floors and trade restrictions, though the administration stopped short of imposing tariffs outright.
• June 2026: A separate $17.5 billion DOE conditional loan program was announced, this one aimed at the reactor-construction supply chain (long-lead components for new Westinghouse AP1000 units), not enrichment directly, though it reinforces the same broader industrial-policy thrust toward domestic nuclear capacity.
• May 2025, implemented through 2026: Executive Order 14300 directed a “wholesale revision” of NRC regulations, with proposed rules due within nine months and final rules within eighteen, a genuinely unusual compression of a licensing regime that has historically operated on decade-plus timelines.

Centrus Energy sits at the center of this because it is currently the only U.S.-owned enricher with deployment-ready, U.S.-origin technology in commercial operation, which makes it the default eligible counterparty for national-security-linked contracts almost by elimination rather than by competitive advantage in the conventional sense. In early July 2026, Centrus finalized a $900 million fixed-price contract with the DOE (total value including options around $1.07 billion) to deploy commercial-scale HALEU production at its Piketon, Ohio facility, with a firm delivery obligation of one metric ton of HALEU UF6 by March 2032. That is a meaningful contract, but a fixed-price structure also means Centrus, not the government, carries the execution risk on cost and schedule as it scales from a completed demonstration program to genuine commercial output. It is worth noting that of the enrichment capacity the DOE selected for expansion in December 2024, most of it, including Orano’s and Cameco’s planned U.S. projects and Urenco’s New Mexico expansion, is not expected to come fully online until the early-to-mid 2030s. The bottleneck does not clear quickly just because the money has been committed.

It’s tempting to treat uranium purely as a commodity story: tightening balances, rising marginal cost, eventual price discovery. But that framing understates what’s actually happening. Two of the four global enrichers are wholly state-owned instruments of national industrial policy (Rosatom, CNNC), operating on capital-allocation logic that has nothing to do with a market-clearing price. The remaining Western capacity is being built out not primarily by private capital chasing returns, but by government-directed loans, fixed-price procurement contracts, and an accelerated regulatory unwind: a state substituting for a market that private capital declined to fund adequately for a decade after Fukushima. Who is allocating capital into enrichment capacity, and on what terms, will determine how quickly the chokepoint clears, more than the uranium price alone. A commodity bottleneck being resolved by directed state and quasi-state capital behaves differently from one being resolved by a free market responding to price signals: the timeline is less elastic to price, and more dependent on policy continuity, budget cycles, and geopolitical alignment.

None of the above is a reason to treat the sector as a one-way trade, and it’s worth separating the fundamental picture from what’s already reflected in prices.

Valuation has already moved a long way ahead of current cash flows for much of the supply chain. Centrus, for instance, was trading around a P/E in the mid-50s even before its most recent contract announcement. The market is pricing a multi-year commercial ramp that has not yet happened, on a fixed-price contract whose execution risk sits with the company. Analysts covering the stock have also flagged that near-term revenue growth is forecast to be modest even as earnings are expected to decline over the next few years, which is a genuinely uncomfortable combination to hold alongside a rich multiple. A compelling structural story and a mispriced security are two different claims, and conflating them is one of the more common ways sound thematic analysis turns into poor portfolio construction. The uranium story being right does not automatically make every equity trading on that story cheap, or even fairly valued.

Market structure matters as much as fundamentals in this corner of the market. Uranium miners and fuel-cycle names remain a small-float, thinly-traded corner of the equity market, heavily represented in a handful of thematic ETFs and indices. That kind of structure means passive and momentum-driven flows can dominate price action over any given quarter, in either direction, independent of what’s happening in the physical market. A sharp equity drawdown in uranium miners this year, for instance, can reflect index rebalancing or risk-off flows in speculative small caps as easily as it reflects any change in the physical fundamentals described above. Treat equity price action in this space as a noisy signal about the underlying commodity thesis, not a clean one.

The regulatory and geopolitical timeline still has real uncertainty in it. The January 2028 hard deadline on Russian LEU waivers is a genuine forcing function, but Kazakhstan, the single largest source of primary supply, tightened state control over its own subsoil code in December 2025, a reminder that supply concentration risk cuts in more than one direction. And Section 232’s remedies (price floors, trade restrictions) remain under negotiation rather than implemented; policy intent is not the same as policy delivered.

A handful of near-term data points will do more to confirm or challenge this thesis than any single research note, including this one. Cameco’s Q2 2026 results, reported this morning (July 31), already give a partial answer. Revenue and adjusted EBITDA came in well below year-ago levels on weaker deliveries and higher costs following spring flooding near its Saskatchewan operations, and the print missed consensus EPS estimates. But 2026 production guidance of 19.5 to 21.5 million pounds (Cameco’s share) was left unchanged, the average realized uranium price for the quarter actually rose to $93.13/lb, and CEO Tim Gitzel described long-term uranium prices as having reached decade highs in the first half of the year, with the company keeping its contracting “strategically patient” to preserve exposure to further upside rather than chase near-term volume. That is a company treating a cost and logistics problem as separate from the pricing thesis, which lines up with the spot-term divergence argument above rather than undercutting it. Still to come: Kazatomprom’s operational update, expected in the coming weeks, for whether the company deepens, holds, or reverses its production discipline; and the trajectory of DOE waiver caps for Russian LEU as they ratchet down toward the 2028 cutoff. Each will say more about whether the spot-term price gap closes through stronger spot buying (bullish confirmation) or through another quiet quarter of term strength without spot follow-through, which is closer to what Cameco’s print just showed.

The uranium fuel chain is tightening. That much is well supported by the data. Whether that tightening is already fully reflected in the equities sitting on top of it is a separate question, and typically the more important one for anyone deciding what to actually own.