Hyperscalers Drive SMR Procurement as Utilities Stay Cautious
Corporate power buyers, not utilities, are fronting the capital risk on small modular reactors, reshaping who controls the next wave of nuclear development.
The URA uranium ETF gained 3.26% on Tuesday (2026-09-08), reaching $47.50, as market attention sharpened around small modular reactor procurement dynamics that utilities are largely leaving to their corporate customers to fund.8
Advanced nuclear developers had long expected regulated utilities to anchor their order books. Instead, the early commercial pipeline is being driven by hyperscalers and large industrial buyers willing to absorb early-mover risk. Kairos Power's 2024 agreement with Google targets up to 500 MW by 2035 — a utility-scale commitment placed by a technology company, not a grid operator.8
The logic is straightforward enough. Utilities operate under rate-of-return frameworks that penalise first-of-a-kind technology risk. Hyperscalers, facing projected U.S. data-center power demand that the IEA expects to more than triple from 34.7 gigawatts in 2024 to 106 GW by 2035, need firm, low-carbon power that cannot be delivered at scale by solar and batteries alone. They have both the balance sheets and the urgency to write checks utilities will not.4
That shift in who bears development risk has broader consequences for uranium markets. Goldman Sachs, in its Nuclear Nuggets report, incorporated SMRs into its supply and demand framework for the first time, projecting cumulative SMR deployments of nearly 46 gigawatts by 2045. The addition lifts its 2045 nuclear generation forecast by approximately 6% and adds an estimated 62 million pounds of incremental uranium demand, a 17% upside against prior long-term estimates.1,3
Goldman's analysts also flagged a cumulative uranium supply deficit they estimate at approximately 2.3 billion pounds between 2025 and 2045 — a figure that encompasses both conventional reactor expansion and the SMR buildout. Timing remains SMR development's most persistent problem, and that deficit figure is sensitive to how far schedules slip.2
The U.S. leads the world in SMR siting activity, with 28 announced projects as of mid-2026, more than its four closest rivals combined, according to data published by the National Public Utilities Council. The Corpus Christi port selection announced on September 1 (2026-09-01) illustrates how SMR deployment is moving into industrial settings beyond data centers, with port operators now evaluating the technology for decarbonising heavy energy loads. But siting announcements are not megawatts. The gap between announced projects and operating reactors remains wide, and no first-of-a-kind SMR has yet entered commercial operation in the United States.6,7
Policy is accelerating around the technology even as the hardware lags. Washington has set a target to quadruple U.S. nuclear capacity by 2050 and is pursuing faster permitting for advanced reactors. In March 2026, European Commission President Ursula von der Leyen described Europe's retreat from nuclear as a strategic mistake and pledged €200 million for a new generation of SMRs. Nuclear investment globally has grown more than 70% over the past five years, and the IEA expects annual nuclear spending to exceed $100 billion under stated policies.4
The IEA puts more than 70 GW of new nuclear capacity potentially online by the mid-2030s, one of the strongest development pipelines in three decades. Yet that pipeline's output lies a decade away. The hyperscalers writing procurement agreements now are betting that developers can close the gap between current engineering timelines and their own load growth curves, which are far steeper.4
Miners are watching but not yet buying. A June 2026 (2026-06-18) assessment found that SMRs remain largely absent from miners' actual procurement lists despite the commercial noise, with the sector described as moving from exploration-stage uranium concepts toward positioning as fuel suppliers rather than reactor customers. Eagle Energy Metals has been developing a micro modular reactor concept rated at up to 3.3 MW for remote deployments at mine sites and military outposts, but that is an exception rather than a trend.5
The near-term signal is whether the hyperscaler-led procurement model produces a bankable project that reaches a final investment decision — and under what contract structure. If corporate offtake agreements prove sufficient to finance construction without utility backing, it resets assumptions about who the primary customer for advanced nuclear actually is, and what credit quality underpins uranium demand growth. Goldman's 62 million pound demand addition rests on SMRs reaching the grid; every year of delay compresses that figure and pushes the supply deficit calculus in directions that uranium markets are still pricing only tentatively.1,2