TL;DR · 30-second read
The Short Version
Centrus Energy, a small American nuclear fuel company, has won a $900 million government contract to produce a stronger, more concentrated kind of uranium fuel.
That matters because many of the new, smaller nuclear reactors that tech companies hope will power artificial intelligence computer centers cannot run on ordinary fuel. They need this stronger version, and very little of it is made in America today.
The lesson: powering artificial intelligence with nuclear energy depends as much on building fuel factories as on building reactors.
Centrus Energy (NYSE: LEU) has secured a $900 million contract from the US Department of Energy to supply high-assay low-enriched uranium, or HALEU, and has been added to the S&P SmallCap 600 index, Simply Wall St reported on October 19, 2026. The company’s sales have been rising, but its profit margin has narrowed to 10.2% from 24% a year earlier, and its backlog has been cited at about $3.6 billion.
Separately, Stocktwits reported that Centrus has tapped Fluor, the engineering and construction contractor, for a major expansion of its uranium operations in Ohio, home to the company’s Piketon enrichment site.
Executive Summary
The Department of Energy has committed $900 million to buy HALEU from Centrus, the US-based enricher that analysts describe as uniquely positioned in this fuel. HALEU is uranium enriched to between 5% and 20% uranium-235, the isotope that sustains a nuclear chain reaction. That is higher than the under-5% fuel used in today’s large reactors, and it is the fuel many advanced and small modular reactor designs are built around. Those are the same designs often proposed as dedicated, round-the-clock power for AI data centers.
The award matters because it sits upstream of every reactor timeline. A reactor designed for HALEU cannot run until that fuel exists in volume, and enrichment plants take years to license, build and commission. Pairing the contract with a major engineering firm for the Ohio expansion marks a step from demonstration-scale production toward industrial supply.
The financial picture is more mixed. Revenue visibility has improved, but margins have compressed. Consensus forecasts compiled before the award assumed roughly flat revenue through 2029. Whether the contract turns into profit depends on how efficiently Centrus builds out capacity at Piketon and Oak Ridge.
The Reactor Is Not the First Constraint. The Fuel Is.
Much of the discussion about powering AI data centers with nuclear energy has focused on reactors: which developer, which site, which utility, which technology company signs the power agreement. Those announcements skip a step. Many advanced and small modular reactor designs promoted for large computing campuses are engineered around HALEU rather than conventional low-enriched fuel. A reactor built for HALEU cannot simply switch to the fuel that today’s fleet burns, so its first day of operation is tied to a fuel supply chain that barely exists at commercial scale.
Enrichment is the gate in that chain. Natural uranium is less than 1% uranium-235. Raising it into the 5% to 20% range requires cascades of centrifuges, a licensed facility and years of construction. Someone also has to commit to buy the output before that capital is spent. The reactors that would eventually buy HALEU commercially are not yet running in volume, so private buyers cannot underwrite the plants. The $900 million Department of Energy contract fills that role: the government is buying supply ahead of demand. Centrus has already demonstrated HALEU production at Piketon under earlier Department of Energy work. This award, together with the Fluor-led Ohio expansion, is the step from proving the process to producing at scale.
This shifts the scheduling question for several groups. Reactor developers depend on HALEU deliveries for their first units. Technology companies and data center developers have signed, or are evaluating, advanced nuclear supply deals. Utilities are planning grids around those reactors. For all of them, the question is no longer only when a reactor is permitted, but when its enriched fuel arrives and from whom. Enrichment is not the only link: enriched uranium must still be converted and fabricated into fuel. It is, however, the most capital-intensive link, and the one this award directly addresses.
Fluor’s Selection Moves the Risk From Technology to Construction
Hiring one of the largest US engineering and construction contractors signals an expansion at industrial scale rather than another pilot cascade. It also changes which risks dominate. At demonstration scale, the open question is whether the process works. At industrial scale, the questions are construction cost, skilled labor, supply of centrifuge components and the schedule. These are the problems that routinely stretch large energy projects.
Analysts already treat these as the swing factors. The contract still requires efficient capacity build-out, careful cost control and timely funding flows to avoid periods of underused assets that weigh on returns. Bullish analysts single out the Piketon and Oak Ridge build-outs as the variable that separates their forecasts from consensus. Fluor’s involvement does not remove those risks. It concentrates them in a construction program that can be tracked against milestones.
Revenue Visibility Is Not Yet Margin
The contract extends revenue visibility and fits within a backlog cited at about $3.6 billion. Profitability is moving the other way. Net income has weakened and margins have slimmed to 10.2% from 24% a year earlier, even as sales grew. If the costs of ramping enrichment and meeting nuclear compliance requirements stay elevated, new awards could add volume without improving profit much.
The analyst spread shows how much hinges on execution. Consensus projects $487.7 million of revenue and $50.7 million of earnings by 2029, essentially flat revenue and a gain of about $2.2 million from today’s $48.5 million in earnings. The most optimistic analysts had penciled in $677.3 million of revenue and $107.7 million of earnings over the same period. Both sets of forecasts predate this award. The gap between them amounts to a price on whether the build-out arrives on time and on budget.
The S&P SmallCap 600 inclusion widens Centrus’s investor base and is likely to bring index-fund ownership and trading liquidity. It does not fund centrifuges. The capital for capacity still has to come from contracts, government funding flows and the company’s own balance sheet.
Background
Centrus Energy, which trades on the New York Stock Exchange as LEU, supplies nuclear fuel and related services to the nuclear power industry in the United States, Japan, the Netherlands and other markets. The company descends from the United States Enrichment Corporation, the federal uranium enrichment business privatized in the late 1990s. It took the Centrus name after a restructuring in 2014. Its work centers on uranium enrichment at Piketon, Ohio, with additional operations at Oak Ridge, Tennessee.
Interest in HALEU has grown alongside advanced and small modular reactor designs, many of which need fuel more concentrated than conventional reactors use. Large technology companies have increasingly looked to nuclear energy as a source of steady, carbon-free power for AI data centers. As a result, the domestic supply of enriched fuel has become a question for the computing industry as well as for utilities. Source: Centrus Energy Stock In Focus After $900 Million Nuclear Fuel Contract (Simply Wall St), an analysis of Centrus’s $900 million Department of Energy HALEU award, S&P SmallCap 600 inclusion and earnings outlook; and Centrus Taps Fluor For Major Ohio Uranium Expansion Amid Nuclear Fuel Push (Stocktwits), a report on Fluor’s selection for Centrus’s Ohio uranium expansion.Sources