DOE’s Prometheus Bets $60 Million That AI Can Help Nuclear Get Out Of Its Own Way

The Department of Energy is putting $60 million behind an effort to use AI to accelerate nearly every stage of nuclear deployment, from reactor design and licensing to manufacturing, construction, fuel fabrication, and operations.

The three-year initiative, named Project Prometheus, is the first major Phase II project selected under DOE’s Genesis Mission. Led by Idaho National Laboratory, it brings together four national laboratories, universities, major tech companies, and more than 20 nuclear industry partners.

Participants include NVIDIA, Microsoft, Amazon Web Services, Oklo, TerraPower, X-energy, GE Vernova, Westinghouse, Aalo Atomics, Antares, Valar Atomics, and Deployable Energy.

The goal is to create a secure digital thread connecting engineering models, regulatory requirements, manufacturing specifications, construction records, and operating data. DOE wants AI tools to help prepare traceable licensing documents, organize decades of legacy nuclear records, accelerate safety analyses, improve manufacturing workflows, and support reactor monitoring while keeping humans responsible for final decisions.

Prometheus is targeting a 50% reduction in reactor deployment timelines and long-term operating costs. The project has also attracted more than $200 million in industry cost sharing and another $30 million in private capital and technical contributions.

Most attention has focused on using nuclear reactors to power AI, such as NANO Nuclear’s agreement with Supermicro and Riot Platforms’ proposed collaboration with Terrestrial Energy.

Oklo, NVIDIA, and Los Alamos National Laboratory are also working together on AI-enabled nuclear fuel research, plutonium-bearing fuels, and grid studies for nuclear-powered AI facilities. That agreement directly supports the broader Genesis Mission.

We’ve covered the increasingly ridiculous AI financing “circle jerk”, where tech companies invest in AI developers that then turn around and commit to buying chips, cloud capacity, or data center services from those same backers. Amazon, for example, plans to invest billions in Anthropic while Anthropic commits to spending more than $100 billion on Amazon Web Services infrastructure.

It is an impressive growth engine, provided nobody asks where the money is actually coming from…

None of that kills the nuclear case, though. The resurgence of demand for nuclear energy was originally pushed by green energy goals and was further elevated by the increasing demands for national energy security. With neither of these demand signals waning as the AI bubble stands ready to pop at any moment, nuclear will likely continue to see support from governments across the globe in the years ahead.