Idaho National Laboratory leads Prometheus, a 32-partner project using artificial intelligence to speed nuclear reactor design, licensing, manufacturing and operations. (Photo: Idaho National Laboratory)
IDAHO FALLS — A federal research project led by Idaho National Laboratory will receive $60 million over three years to apply artificial intelligence to nearly every stage of building and running a nuclear reactor, a decision that places eastern Idaho at the center of Washington’s most ambitious attempt yet to shorten the timelines that have made new nuclear plants slow and costly to build.
The project is called Prometheus. According to Idaho National Laboratory, the Department of Energy selected it as the first Phase II award issued under the Genesis Mission, a broader federal initiative to use artificial intelligence to speed up scientific research. The selection was announced July 22 at a Genesis Mission summit in Washington, D.C., and the laboratory highlighted the award again this week on its official social media accounts.
The federal money is the smaller half of the arithmetic. The laboratory says industry partners have committed more than $200 million in cost share and more than $30 million in direct cash, capital and proprietary design data — roughly four dollars of private commitment for every federal dollar awarded.
The problem it is aimed at
American nuclear construction has rarely failed on physics. It has failed on calendars and budgets. Reactors take years to design, years more to license, and are built one at a time by a supply chain that has little opportunity to learn from repetition. Prometheus is aimed squarely at that pattern rather than at any new reactor technology.
The laboratory describes five lines of work: an AI platform and supporting infrastructure; design and licensing; manufacturing and construction; autonomous operations; and the unglamorous but foundational task of curating decades of Department of Energy nuclear data into a form AI systems can actually use. In each case the stated design keeps a person in the loop — the systems are meant to generate, analyze and verify, and to recommend actions during operations, with human oversight retained.
INL Director John Wagner framed the award in terms of speed and cost discipline, saying the nation’s nuclear future depends on moving with more urgency and driving costs down without loosening safety or technical standards. Prometheus brings together nuclear science and artificial intelligence
to do that, he said in the laboratory’s announcement.
Thirty-two partners, four national labs
Idaho National Laboratory leads a consortium that the laboratory counts at 32 partners. Oak Ridge, Argonne and Sandia national laboratories are the other federal participants, each bringing distinct capability — Oak Ridge in advanced manufacturing and its Frontier supercomputer, Argonne in digital twins and its Aurora supercomputer, Sandia in nuclear safety and security test beds.
Four universities are involved: North Carolina State, Penn State, the University of Tennessee at Knoxville and the University of Texas at Austin. The commercial side runs from reactor developers such as X-energy, TerraPower, Oklo, Westinghouse and GE Vernova to technology firms including Microsoft, Amazon Web Services and NVIDIA.
On the Idaho side, the laboratory lists the Advanced Test Reactor, the Materials and Fuels Complex, the Transient Reactor Test Facility and the Neutron Radiography Reactor among the facilities supporting the work.
What it means for Idaho
The award lands in the middle of a run of federal decisions pointing at Idaho Falls. Two days before the Prometheus announcement, the laboratory said it would play a leading role in a new Nuclear Energy Tech Hub. INL also received separate Phase I funding under the Genesis Mission for GeoFinder AI, a project on critical mineral discovery and recovery that includes Boise State University and Idaho State University alongside Mississippi State.
Taken together, the announcements suggest the federal government is consolidating a significant share of its nuclear-and-AI portfolio in eastern Idaho rather than distributing it evenly across the national laboratory system. For a region whose economy is already heavily tied to the site, that concentration carries both opportunity and exposure.
Timeline and open questions
The $60 million is subject to appropriations, a caveat the laboratory states plainly. A prototype platform has already been built and validated under earlier seed funding; the production version is targeted for initial deployment on the Genesis Mission Platform by March 2027, with specifications for broader commercial and departmental adoption expected later that year.
The laboratory says the platform will eventually be offered as a service through the Genesis Mission’s American Science Cloud and through commercial cloud marketplaces, and that it will be containerized so utilities and vendors can run it inside their own firewalls. It also commits to publishing benchmark suites and evaluation protocols, with every claimed AI advantage measured against a documented baseline established before AI was used.
What the announcement does not address is how regulators will treat the output. Prometheus intends to prepare traceable licensing documentation for the Department of Energy and the Nuclear Regulatory Commission, but the laboratory’s materials do not describe any commitment from the NRC about how AI-assisted submissions will be reviewed. That question — whether a regulator built around human-authored, human-defended engineering analysis will accept work products generated with machine assistance — is likely to determine whether the design and licensing gains materialize on the schedule the project envisions.
Source: Idaho National Laboratory news release, July 22, 2026, and the laboratory’s Prometheus program page. Photo: Idaho National Laboratory.
About the Author
