The White House wants nuclear reactors in orbit by 2028. The fuel those reactors need doesn't exist domestically at scale yet.
That's the gap inside NSTM-3, the six-page space nuclear directive the Office of Science and Technology Policy released on April 14 at the Space Symposium in Colorado Springs. According to SpaceNews, the policy calls for parallel NASA and Defense Department competitions to develop space nuclear power systems, sets a 30-day startup clock for NASA to begin work on a 20-kilowatt reactor, and targets a reactor in orbit by 2028 with a lunar surface version by 2030. Jared Isaacman, the NASA administrator, noted at the same symposium that the agency has spent more than $20 billion on nuclear power projects over the last several decades, none of which flew.
The fuel problem is straightforward. SR-1 Freedom, the 2028 orbital demo NASA announced March 24, runs on high-assay low-enriched uranium, or HALEU. SpaceNews reported that the reactor will generate 20 kilowatts of electrical power using HALEU. HALEU is uranium enriched to between 5% and 20% uranium-235, a specification that allows next-generation reactors to achieve smaller cores and longer fuel cycles. Per SMR Intel, the US has one facility producing HALEU as of April 2026: Centrus Energy's American Centrifuge Plant in Piketon, Ohio, which outputs roughly 900 kilograms per year at demo scale. Expanding that to full production takes 42 months, according to a DOE task order issued in January 2026.
Russia was the only commercial HALEU supplier globally until the 2024 ban on Russian enriched uranium imports. The ban, which fully takes effect in 2028, cut off that source at the same moment the NSTM-3 directive demands fuel delivery. SMR Intel noted that the DOE has committed over $2.7 billion to building domestic enrichment capacity since the ban, but that capacity is currently allocated to commercial ground-based SMR projects: TerraPower's Natrium, X-energy's Xe-100, Kairos Power's Hermes. The space program is not in the front of that line.
New enrichment capacity does not arrive in time. Urenco USA will have 700,000 separative work units of capacity by 2027, but is currently licensed only up to 10% enrichment, below the HALEU threshold. Orano USA's planned US facility has a license application target of the first half of 2026 and a production target of 2031. For SR-1 Freedom to make 2028, DOE would need to divert limited domestic HALEU stock to a space mission ahead of commercial customers, rely on the existing 900 kilograms of demo-scale production, or find a path the directive does not name.
Dr. Bhavya Lal, a former NASA associate administrator for technology, policy, and strategy, has argued a lunar reactor is achievable by 2030 at roughly $3 billion. Joseph Cirincione, a nuclear policy expert, told The Independent the sub-10-year timeline was cock-eyed. Both may be right simultaneously: a reactor is physically possible; a reactor with committed fuel supply by 2028 is not.
The NSTM-3 directive is not naive about past failures. According to SpaceNews, it explicitly cites four recurring failure modes from previous NASA nuclear projects: insufficient mission pull, overly ambitious scope, unrealistic schedules, and fragmented agency leadership. Whether the 2028 orbital target reflects genuine learning or repackages the same optimism in a different font is the question the fuel supply answer will settle first.
Michael Kratsios, OSTP director, called the policy the roadmap for permanent lunar and Mars presence. The roadmap has a gap.
