What happens when the companies building the next generation of nuclear reactors suddenly realize the fuel they need is still largely theoretical? In the past few months that realization has turned into a quiet scramble. Several advanced reactor developers have lined up multi-year contracts with Centrus Energy for high-assay low-enriched uranium, better known as HALEU. The deals look solid on paper. The production schedule looks tighter once you start adding up the volumes.
Why Advanced Reactors Suddenly Need a Different Kind of Fuel
Most of the commercial reactors running today use low-enriched uranium at roughly five percent enrichment. The smaller, more efficient designs coming next—fast reactors, microreactors, some high-temperature designs—need fuel enriched between five and twenty percent. That range is HALEU. It packs more energy into less volume, which is exactly what these compact systems require.
Until recently the United States produced almost none of it at commercial scale. Centrus has been running a small demonstration cascade of sixteen centrifuges that turns out roughly nine hundred kilograms a year for the Department of Energy. That is useful for research and early testing. It is nowhere near enough for the fleet of reactors now moving toward construction.
I’ve been watching this space for a while, and the shift feels noticeable. Developers are no longer treating fuel supply as a problem they can solve later. They are writing checks now. Prepayments help Centrus fund the next set of centrifuges. In return the reactor companies lock in volume and, at least in theory, a place in line.
Oklo’s Multi-Year Agreement and Meta’s Indirect Role
In mid-November Oklo announced it had signed definitive multi-year contracts with Centrus. The fuel is intended for up to five Aurora powerhouses that form the first phase of a planned 1.2-gigawatt campus in Pike County, Ohio. Deliveries are scheduled to start in 2029 from Centrus’s Piketon facility. Oklo is prepaying, which helps underwrite the expansion of enrichment capacity.
There is an interesting financing loop here. Meta earlier committed to purchase power from Oklo. That prepayment for future electricity effectively flows downstream and helps pay for the fuel that will eventually generate the power. It is a tidy circle if everything stays on schedule. Whether schedules hold is another matter.
The same fuel agreement also covers future Aurora projects and, eventually, the refueling needs of Oklo’s first reactor planned for Idaho. That breadth makes the contract more significant than a simple first-core supply deal.
Other Developers Join the Queue
Oklo is not alone. Since August, Centrus has announced three additional definitive agreements, each involving prepayments.
- X-energy secured enrichment services for both LEU and HALEU to support its first Xe-100 reactors and the associated TRISO-X fuel fabrication.
- Radiant signed for HALEU destined for multiple Kaleidos microreactors, with deliveries expected before the end of the decade.
- Antares locked in “unobligated” HALEU that can also support certain national security missions, again with delivery targeted before 2030.
Centrus’s own materials still list TerraPower’s Natrium reactor under a 2023 memorandum of understanding, with a planned deployment around 2030. That relationship has not yet been converted into a definitive contract of the same type, but it sits in the background of capacity planning.
Taken together, these names represent a meaningful share of the advanced reactor projects that have attracted private capital and government support in recent years. They are placing bets that Centrus will be able to deliver.
Current Production Versus Future Demand
Today’s output is still limited to that single demonstration cascade. Centrus expects its next full cascade of new centrifuges to come online in 2029. The longer-term base-case plan targets roughly twelve metric tons of HALEU per year. How quickly the company reaches that rate remains less clearly defined.
Simple arithmetic already raises questions. Public materials suggest an Aurora core may require around five tons, each Xe-100 roughly one and a half tons, Antares about one ton, and Natrium in the neighborhood of fifteen tons. Add those initial loads and you are looking at more than twenty-two tons—nearly two years of the eventual full capacity—before counting multiple units or later reloads.
Oklo’s contract covers up to five Auroras. Dow’s Long Mott project alone contemplates four Xe-100 units. Radiant’s volume has not been disclosed in detail. The Department of Energy is also owed material under an existing large contract. None of these volumes appear fully reconciled against the same production timeline.
In my view this is the part of the story that deserves the most attention. Signing contracts is necessary. Matching those contracts to physical machines spinning at the right enrichment levels is the harder task.
LEU Obligations Still Matter
Centrus is not only in the HALEU business. It currently supplies low-enriched uranium to utilities under long-standing arrangements that rely heavily on Russian and French material. Those supply contracts wind down at the end of 2027 and 2030. The company has said it retains flexibility to use planned HALEU capacity for various assays of LEU or slightly higher enrichments if needed.
That flexibility is useful. It also means the same centrifuges may face competing claims—legacy utility customers, new advanced reactor developers, and government commitments. Sorting priority among those claims will shape who receives material in the early years of expanded production.
Backlog Growth and Management Commentary
Centrus’s contingent backlog—sales that depend on new capacity at Piketon—rose from about three billion dollars in early August to 5.4 billion by the end of September. That jump reflects the wave of definitive contracts signed in a short window.
When asked whether the company might accelerate capacity deployment, leadership indicated there is no material change to the planned cadence of rolling out new centrifuges. The message was essentially that the company is already moving as quickly as practical given manufacturing, installation, and regulatory realities.
A partnership with a major data-analytics firm has been mentioned as a potential source of efficiency gains in the future. Whether those gains can compress the timeline by months or years is still an open question.
On the specific question of which customer receives the limited HALEU produced today, the answer given was simply that it will be supplied to the commercial market. That leaves the allocation details opaque for now.
Timing Mismatches and Project Schedules
One practical buffer exists. Oklo recently widened the expected startup window for its first Aurora from a firm 2028 to a range of 2028–2029. When reactors slip, fuel demand slips with them. A fuel shortage is easier to manage if the machines that need the fuel are also running late.
That dynamic will not last forever. At some point the reactors will be ready and the fuel will need to be there. The industry is effectively conducting a large, multi-year coordination exercise between enrichment capacity, reactor construction, and fuel fabrication.
What the Contracts Actually Achieve
These agreements do several useful things at once. They provide Centrus with prepaid capital to support the build-out of centrifuges. They give reactor developers a contractual claim on future output. And they signal to investors and policymakers that commercial demand for HALEU is real and growing.
They do not, by themselves, create the physical capacity. Manufacturing new centrifuges, installing them, licensing the expanded operations, and ramping production all take time. The 2029 target for the next cascade is the nearest concrete milestone. Everything beyond that still contains uncertainty.
Perhaps the most interesting aspect is how concentrated the early demand has become. A relatively small number of developers account for a large share of the announced HALEU needs. If even a few of those projects move faster than expected, or if additional customers appear, the pressure on the early years of production will increase.
Broader Context for Domestic Enrichment
The United States has spent years talking about rebuilding domestic enrichment capability. Geopolitical events and supply-chain concerns have made that conversation more urgent. Centrus is currently the only company with a licensed HALEU production pathway in the country. That position carries both opportunity and responsibility.
Other potential producers exist on paper or in early development. None of them are yet in a position to deliver commercial volumes on the same timeline. For the next several years, therefore, the industry’s HALEU hopes rest heavily on the pace of work at Piketon.
I’ve found that energy transitions often look smoother in presentations than they feel in execution. Fuel supply is rarely the most glamorous part of the nuclear conversation, yet it remains one of the more binding constraints. The current wave of contracts makes that constraint visible in a new way.
Looking at the Numbers More Closely
Consider the base-case target of twelve metric tons of HALEU per year once full capacity is reached. That figure sounds substantial until it is stacked against the initial cores already under discussion. Multiple five-ton loads, several one-and-a-half-ton loads, and larger requirements for bigger reactors quickly consume years of output.
Refueling adds another layer. Once the first reactors are operating, they will need periodic replacements. Those later requirements are not yet the focus of public discussion, but they will arrive. Planning only for first cores leaves future demand unaddressed.
The Department of Energy’s existing contract also sits in the mix. Government material has priority characteristics that commercial contracts may not share. How those priorities interact with the new commercial backlog has not been fully spelled out in public statements.
Financial Implications for Centrus
From Centrus’s perspective the surge in contingent backlog is clearly positive. Prepayments improve near-term cash flow and reduce the amount of external capital required to expand. The contracts also demonstrate that customers are willing to commit capital years before delivery.
At the same time, the company must still execute on a complex industrial project. Building and qualifying new centrifuge cascades is not a software exercise. Delays in manufacturing, supply-chain issues for specialized components, or regulatory hurdles could push the 2029 date further out. Each slip would compress the window between capacity availability and reactor need.
Investors appear to be watching the backlog growth carefully. The more interesting test will come when the company begins reporting actual production volumes from the new cascades rather than contingent sales.
Implications for Reactor Developers
For the developers the contracts reduce one category of risk while introducing another. Fuel availability is now contractually addressed, at least on paper. Counterparty and execution risk remain. If Centrus capacity comes online later than expected, the developers may face schedule pressure of their own.
Some projects have built flexibility into their timelines. Others are still targeting aggressive first-operation dates. Those differences will matter when limited early production has to be allocated.
There is also the question of price. Prepayments and long-term contracts typically lock in certain commercial terms. How those terms compare with eventual spot or alternative supply remains to be seen once more capacity exists.
The Role of Government Support
Public funding has played a significant role in getting HALEU demonstration work this far. Continued policy support will likely influence how quickly additional cascades can be financed and built. Tax incentives, loan guarantees, or further offtake commitments could accelerate the build-out. Their absence would leave more of the burden on commercial prepayments.
The current set of private contracts suggests that commercial demand is beginning to stand on its own. That is a healthy development. It does not eliminate the value of coordinated public-private planning for a fuel type that remains scarce.
Potential Bottlenecks Beyond Enrichment
Enrichment is only one step. Once HALEU is produced it still needs to be converted into the specific fuel forms required by each reactor design—metal fuel, TRISO particles, or other specialized forms. Fabrication capacity for those forms is also limited. A smooth enrichment ramp will still face constraints further down the fuel cycle if fabrication does not keep pace.
Transportation and security requirements for higher-assay material add another layer of complexity. These are manageable issues, but they require planning and investment in parallel with the centrifuge expansion.
How This Fits the Larger Nuclear Revival Narrative
Advanced nuclear has attracted considerable attention and capital in recent years. Data-center demand, decarbonization goals, and energy-security concerns all point toward more nuclear capacity. Many of the designs that look most promising from a commercial or technical standpoint rely on HALEU.
If the fuel supply cannot scale, those designs face a practical ceiling. The contracts signed this year represent an attempt to remove that ceiling before it becomes binding. Whether the attempt succeeds will be measured in kilograms produced, not in press releases issued.
I keep coming back to the same observation: the industry is trying to rebuild a capability that was allowed to atrophy for decades, and it is trying to do so on a timeline measured in years rather than decades. That is ambitious. Ambition is necessary. Execution will decide the outcome.
What to Watch in the Coming Quarters
Several markers will indicate whether the current optimism is justified. Progress reports on centrifuge manufacturing and installation will matter. Any revision to the 2029 target date would be significant. Clarity on how early production volumes will be allocated among competing customers would reduce uncertainty. Additional definitive contracts or expansions of existing ones would show whether demand continues to outpace the announced capacity plan.
On the reactor side, updates to construction schedules and regulatory milestones will show whether fuel demand is likely to arrive earlier or later than currently projected. The interaction between those two timelines is the real story.
For now the contracts are in place, the backlog has grown, and the first new cascade is still targeted for 2029. The next few years will reveal how well the physical infrastructure can keep up with the commercial commitments already made.
The sudden concentration of HALEU offtake agreements marks a shift from theoretical interest to contractual reality. Reactor developers are no longer waiting to see whether enrichment capacity appears. They are helping to finance it. Centrus, for its part, holds a rare position as the near-term domestic source of a fuel type that many next-generation designs require.
That position brings both commercial opportunity and operational pressure. Meeting the volumes already under contract while also serving legacy LEU customers and government obligations will test the company’s ability to scale. The developers, meanwhile, have reduced one major uncertainty only to remain exposed to the execution risks that still sit between signed paper and delivered fuel.
In the end the market is doing what markets often do: revealing scarcity through rising commitments and prepaid demand. Whether that scarcity eases in time for the reactors now under development is the question that will shape the early commercial phase of advanced nuclear in the United States. The contracts signed this year have made the stakes clearer. The centrifuges still have to spin.