Snow dusts a half-built turbine hall in Kemmerer. A federal stamp says: go. The problem is obvious enough to see from the road — the reactor can be built, but there’s nothing to put inside it.
I’ve followed power projects for years; you watch winners get permits and then stall on supply. You should care because this is no small hiccup: it’s a national puzzle where politics, industry, and Cold War legacies collide.
At Kemmerer’s construction site, cranes creak — The permit exists, but the fuel chain does not
By unanimous vote, TerraPower — the company founded and chaired by Bill Gates — just won the federal green light to build a commercial Natrium reactor in Kemmerer, Wyoming, slated to start producing power in 2031. That permit clears a huge regulatory hurdle: the reactor can be assembled and its systems installed. But the reactor is like a high-performance engine with no gas in the tank.
TerraPower has been building since 2024, but company spokespeople told me the work has focused on nonnuclear sections so far. Getting steel and concrete up is one kind of logistics; filling a fast reactor with the specific fuel it needs is another, and far trickier.
Who makes HALEU and why does it matter?
High-assay, low-enriched uranium (HALEU) is the fuel that Natrium and other advanced reactors require. In commercial quantities today, the primary manufacturer is Techsnabexport — Tenex — a subsidiary of Rosatom, the Russian state nuclear corporation. That single-source reality left TerraPower exposed after Russia’s 2022 invasion of Ukraine; officials and company spokespeople said using Russian fuel was not acceptable for moral and commercial reasons.
At Oak Ridge and in DOE briefings, technicians sketch conversion plans — The government’s stopgaps and their limits
The U.S. Department of Energy has built a plan called the HALEU Availability Program to bridge the gap. The short-term fix is “downblending”: taking highly enriched material, including surplus weapons uranium, and diluting it into the HALEU concentration range.
That work is real — it requires dismantling warheads, melting uranium, reworking isotope concentrations. It can supply early loads to get reactors online, but it is a finite source and not a long-term commercial solution. TerraPower’s initial need has been reported at roughly 150 metric tons to cover operations from roughly 2028 through 2037 — about 15 tonnes per year on average. By contrast, Centrus Energy in Ohio was projecting roughly 900 kilograms per year in 2024, or roughly 6% of what Kemmerer will need annually.
Can the U.S. produce HALEU without Russia?
There are firms racing to do just that. Centrus is the only U.S. company actively enriching uranium toward HALEU levels, but its 2024 output projections fall far short of demand. Abroad, companies such as South Africa’s ASP Isotopes have struck supply agreements with TerraPower and hope to scale production. The DOE plans to fund and coordinate domestic capability, but building commercial enrichment and processing at scale will take years and capital.
That gap matters because early reactor loading needs are immediate and impossible to ignore: you can build a plant and still be powerless without the fuel.
At news desks and congressional panels, spokespeople speak plainly — The politics of supply are now policy
TerraPower publicly said it would avoid Russian fuel after 2022; company representatives have argued alternative suppliers will emerge as demand grows. The government has signaled strong support for the project: DOE briefings and public statements make Kemmerer a priority alongside other advanced reactor efforts.
Still, policy support and a permit do not create tons of HALEU overnight. The market needs more centrifuges, more processing plants, and more secure supply lines — all built while projects expect to start operations within this decade. The supply picture is a chess match with missing pawns.
At the front lines of procurement, companies make deals — Who’s in the short list and what they promise
TerraPower has cut agreements and strategic pacts. In 2024 it signed terms with ASP Isotopes for HALEU supply and loan arrangements; ASP has said it hopes to build plants with partners such as Necsa in South Africa. Centrus is the U.S. hope for enrichment-based HALEU, while DOE programs will lean on downblending to seed reactors early.
Those are plausible paths. They are also fragile: timelines, capital, export rules, and geopolitics all shift and can blow windows of opportunity shut. If you’re betting on a 2031 start date, you’re implicitly betting the supply chain accelerates faster than history suggests it will.
At my desk, the numbers are small enough to matter — What TerraPower really needs and what stands between permit and power
TerraPower’s reactor can be assembled now; it cannot generate electricity at scale without HALEU. Short-term supplies from downblending can load an initial core, but long-term commercial operation requires steady enrichment-based production. The current single-vendor reality — reliance on Rosatom’s Tenex for commercial HALEU — is politically unacceptable to many and operationally risky.
So the company is hedging. It has public assurances, supply deals, and hopeful ramp plans. You can admire the engineering and still feel the unease: a reactor standing ready, waiting on fuel from systems that do not yet exist at the needed scale.
Gizmodo reached out to TerraPower for comment and for details about any unreported HALEU sources. I’ll update if they reply.
Will the U.S. and its partners scale a fuel supply chain in time to power Kemmerer, or will this permit become a cautionary tale about building reactors faster than fuels can be made?