Uranium at $90: Record Institutional Buying Meets a Tightening Market

The Sprott Physical Uranium Trust has just closed five consecutive quarters at its fastest capital-raising pace ever, buying roughly 7 million pounds of uranium in 2026 while the spot price sits at US$90 per pound.

That is the tension worth sitting with. Institutional money is accelerating into a market most retail observers still file mentally under the post-Fukushima collapse, buying physical material at prices management openly calls attractive rather than expensive.

The accumulation sits against a backdrop that sharpens the signal. Primary mine supply now covers only 74-90% of global reactor requirements, 70% of post-2027 demand is uncontracted (the highest in three decades), and a new set of demand actors, from AI hyperscalers to Asian state construction programmes, is entering at scale.

This piece lays out what the institutional accumulation data, the demand catalyst timeline, and the small modular reactor horizon actually mean for how seriously to take the uranium investment outlook right now, and where the genuine uncertainties still sit. The aim is a more precise view, not a cheerleading one.

What Sprott’s acquisition pace reveals about institutional conviction

Start with the raw scale of what one vehicle has absorbed. When the Sprott Physical Uranium Trust launched on 19 July 2021, it held 18.2 million pounds of U₃O₈. As of 10 September 2026, that figure stands at approximately 81.7 million pounds.

The trajectory of that growth matters more than the endpoint.

• At inception in July 2021: 18.2 million pounds held
• During 2025: just under 9 million pounds acquired
• During 2026 to date: approximately 7 million pounds acquired (6 million in Q1, a further 1 million subsequently)

The 2026 cadence is the tell. The trust operates under an annual acquisition capacity limit of 9 million pounds, with mechanisms to raise equity accretively beyond that ceiling if needed. Concentrating 6 million of this year’s buying into Q1 alone is not the behaviour of a fund drip-feeding capital as it arrives.

It is front-loading conviction.

That conviction shows up most clearly in the fundraising itself.

Over the most recent five quarters, SPUT raised more capital than in any equivalent period since its 2021 inception, and it did so through documented sentiment softness following the 2024 US election and the tariff turbulence that came with it.

Consider what that combination represents. Energy policy uncertainty and market volatility temporarily thinned retail participation, yet the capital kept arriving at a record clip. Sophisticated allocators were writing cheques into a period when the mood had cooled, not warmed.

Management frames the current US$90 spot price, up roughly 10% year-to-date in 2026, as an attractive entry level rather than a stretched one. The purchasing behaviour backs that framing rather than contradicting it.

Here is the read for you. When physical buyers accelerate into softness and treat a ten-percent-higher price as cheap, they are not chasing momentum. They are expressing a medium-term thesis that current prices undervalue where the market is heading. The pace and timing of SPUT’s 2026 buying give you the clearest available proxy for where patient, long-term capital has already positioned itself.

The structural supply gap that makes $90 uranium look cheap to long-term buyers

The conviction rests on arithmetic, so it is worth building the arithmetic up one figure at a time.

Global primary production runs at roughly 145 million pounds annually. Reactor requirements exceed 180 million pounds. That single comparison already leaves a shortfall of more than 35 million pounds before any second-order factor enters.

For 2025 specifically, the picture tightens once secondary sources are included.

The secondary supply line deserves a moment. Underfeeding at enrichment facilities and stockpiled inventories once masked the structural gap by topping up the market when mines fell short. Those buffers are finite, and they have largely been drawn down. The cushion that hid the deficit for years is gone.

The uranium supply deficit is rooted in a structural mismatch that accumulated over more than a decade of underinvestment following the Fukushima shutdown, leaving the global mine base far too thin to respond quickly when reactor demand began recovering at scale.

Then there is the forward view, which is where the picture turns from tight to pressured. The World Nuclear Association projects demand growth of 28% by 2030 and a doubling by 2040, while 70% of post-2027 demand sits uncontracted, the highest proportion in 30 years.

The World Nuclear Fuel Report 2025 projects demand growth of 28% by 2030 and a doubling by 2040 across multiple scenarios, providing the baseline forecasting framework that institutional buyers use when sizing long-duration uranium positions.

Why the contracting clock is ticking for utilities

The contract coverage numbers are where the urgency becomes visible. At the end of 2025, existing purchase contracts provided maximum deliveries of 174 million pounds for 2026-2035. One year earlier, the equivalent figure for 2025-2034 was 234 million pounds.

That is a 60 million pound drop in a single year, not a gradual drift.

A decline that steep tells you utilities are approaching a procurement cliff. New mine supply cannot respond quickly, because development timelines and permitting run for years, not months. Geopolitical friction compounds the problem: Russian supply sanctions and Kazakh production uncertainty are pushing utilities to diversify sources at a premium, layering structural cost pressure onto an already tight market.

US operators offer a concrete data point here. They purchased 46.9 million pounds U₃O₈ equivalent in 2025, down from 55.9 million in 2024, with new long-term contracts averaging US$70.46/lb and spot purchases averaging US$76.01/lb. Buying volume fell even as coverage thinned, which means the contracting decisions are being deferred, not resolved.

Uranium term prices have held stubbornly below spot through 2026, a divergence that reflects utilities deferring procurement decisions rather than abandoning them, and one that historically resolves through a rapid repricing event once contracting urgency becomes unavoidable.

For you, the takeaway is that the deficit is not a speculative narrative. It is the gap between what mines dig up and what reactors consume, compounded by a contracting lag utilities will have to close under time pressure. That arithmetic is the foundation the whole institutional case is built on.

How AI hyperscalers, government procurement, and Asian construction programmes are reshaping the demand picture

Three distinct forces are pulling on uranium demand at once, and each would merit attention on its own. Ranked by how quickly the pounds actually move, the order is worth keeping straight.

1. Government procurement: already funded and executing now
2. Asian construction: reactors already pouring concrete
3. Hyperscaler PPAs: signed, but ramping toward 2030-2035

The technology names generate the headlines, so start there while remembering they sit last on near-term impact.

The hyperscaler pivot, signed but not yet consuming

Technology giants have shifted from prioritising renewables to actively contracting nuclear baseload for AI data centres, citing reliability that intermittent sources cannot match. These are committed agreements, not press releases.

AI data centre power demand is growing faster than any grid expansion programme in history, and the reliability gap between intermittent renewables and always-on nuclear baseload has made reactor-sourced electricity the default answer for hyperscalers running inference workloads at scale.

Amazon signed a 1,920 MW power purchase agreement with Talen Energy, drawing from the Susquehanna plant and ramping to full volume by 2032, alongside a feasibility study with Energy Northwest for an SMR project of up to 960 MW. Google signed a 50 MW PPA with the Tennessee Valley Authority sourced from Kairos Power, the first tranche of a pledge to buy 500 MW from six to seven reactors by 2035, plus a separate deal with a Finnish utility. Microsoft locked in a 20-year PPA with Constellation Energy from the restarted 835 MW Three Mile Island Unit 1.

The scale of where this could go is what fixes attention.

Goldman Sachs estimates global data centre power demand will grow more than 160% by 2030, potentially requiring 85-90 GW of new nuclear capacity.

The critical qualifier: most of that consumption ramps toward the back half of the decade. The commitments are real, but the pounds arrive later.

Government programmes and Asian reactors: demand that is already funded

This is where the near-term consumption actually lives, and it is not conditional on future decisions.

The US Department of Energy is running a US$2.7 billion programme to build domestic low-enriched uranium (LEU) and high-assay low-enriched uranium (HALEU) capacity. On 5 January 2026, it finalised US$900 million task orders each with American Centrifuge Operating and General Matter Inc. to expand HALEU enrichment. On 7 May 2026, the US secured its largest-ever HALEU shipment in partnership with Japan. In August 2025, the DOE convened a Defence Production Act consortium of nuclear companies, a structural signal of how seriously Washington is treating domestic supply security.

Asia adds physical reactors, not paperwork. China has 36-38 reactors under construction, roughly 39 GW of new capacity, sustaining an approval cadence near 10 units a year across 2022-2025, with 10 approved in April 2025 and 8 in July 2026. India recently completed two large uranium acquisition transactions to feed its expanding programme.

Here is what makes this moment different for you. These three forces are not arriving in sequence. They are accelerating together inside the same 12-to-24-month window, and that simultaneity compresses the timeline for when aggregate demand reaches utilities’ contracting desks. Convergence is what turns a demand thesis into an observable trend, and it is what gives an allocator like Sprott the confidence to keep buying at US$90.

Where the SMR horizon creates both upside and timeline risk for uranium investors

Small modular reactors carry genuine promise, and the honest position is to separate that promise from what the technology can contribute this decade.

Start with the useful part for the bull case: current institutional demand models factor in almost no SMR-driven uranium consumption. That means the SMR pipeline is upside optionality, not a load-bearing assumption. The near-term case does not need SMRs to deliver on schedule.

Now the pipeline reality. The World Nuclear Association lists over 70 SMR projects, and the OECD Nuclear Energy Agency tracks 51 designs in pre-licensing or licensing across 15 countries. Yet only about 2.5 GW is actually under construction or in development, roughly 5% of total announced capacity. Two units are already operational: Russia’s Akademik Lomonosov and China’s HTR-PM.

Developer target dates cluster around 2026-2030. Energy research consensus, drawn from groups including Deloitte and the Brookings Institution, tells a slower story. The first major wave of commercial SMR fleets is not expected until 2030-2035, and nuclear is projected to supply only around 10% of incremental data centre demand by 2030.

The reasons for the gap are worth naming plainly.

• Regulatory lead times of 7-10 years per design
• Transmission bottlenecks constraining where new capacity can connect
• Financing hurdles for first-of-a-kind builds

The market projection captures both sides of this tension: the behind-the-meter SMR data centre market is forecast to grow from US$500 million in 2025 to US$20.1 billion by 2035, a 44.7% compound annual growth rate. Enormous long-run scale, minimal near-term pounds.

For you, the cleanest framing is this. Treat the SMR pipeline as a call option on the post-2030 demand curve, not as a variable in today’s supply-demand equation. That framing protects against two errors: over-discounting uranium by dismissing nuclear as unproven, and over-pricing it by assuming SMR demand lands before the structural deficit has done its work.

Reading the uranium setup at US$90 with clear eyes

Pull the three layers together and the shape of the case becomes clear. Institutional accumulation is running at a record pace, the supply gap is arithmetic rather than speculative, and the demand catalysts are converging inside a single window rather than arriving one at a time.

That is the genuine strength of the setup. The honest counterweight matters just as much.

The uranium term price has held near US$80-81/lb since August 2024, even as physical buyers accumulate at US$90 spot and the January 2026 peak of US$101/lb has faded.

That divergence is the market’s split personality. Utilities are not yet panicking; institutional buyers are not waiting. Read together, they tell you the current dynamic is driven by forward-looking conviction rather than present procurement urgency, which means this thesis plays out over years, not quarters.

The uncertainties deserve equal billing. A faster-than-expected Kazatomprom ramp-up could reverse speculative flows. Macroeconomic conditions could thin the very institutional demand that anchors the case. And the SMR timeline gap remains real.

Three indicators are worth watching as leading signals:

• The pace of new utility contracting activity
• SPUT’s Q4 2026 capital-raising trajectory as seasonal slowdowns ease
• Whether Chinese reactor approvals sustain their 2025-2026 cadence

Management expects a renewed capital-raising phase as sector news flow accelerates, with anticipated IPOs from OpenAI, Anthropic and a potential Westinghouse listing broadening equity access to the theme. Watch those three data points, and you will know in real time whether the institutional thesis is being validated or challenged.

Investors exploring how to translate the physical accumulation thesis into equity exposure will find our full explainer on uranium stocks for 2026 useful, covering producer leverage ratios, developer project pipelines, and the royalty structures that offer different risk-return profiles across the sector.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors, and forward-looking statements are speculative and subject to change based on market developments.