AI is being used to help design a new generation of nuclear reactors — systems that could one day help power energy-hungry data centers across the United States that support AI.
The Idaho National Laboratory, part of the U.S. Department of Energy and the nation’s leading nuclear energy research site for the past 75 years, is partnering with Oklo and NVIDIA to accelerate advanced reactor and fuel system design work.
The project, called Prometheus, is part of the U.S. government’s Genesis Mission, an initiative intended to “build the world’s most powerful scientific platform to accelerate discovery science, strengthen national security, and drive energy innovation.”
Jacob DeWitte stated in a May 12 press release, “This work brings together advanced reactor design, AI-enabled engineering tools, and INL’s deep technical expertise. Applying AI to reactor design workflows can accelerate development, improve engineering efficiency, and support progress on advanced systems, including on Oklo’s Pluto reactor.”
As part of a Strategic Partnership Project, Oklo will collaborate directly with Battelle Energy Alliance to integrate the Prometheus AI platform with Oklo’s “multiphysics design and analysis infrastructure to streamline engineering workflows and support development of Pluto, Oklo’s reactor system designed to use plutonium-bearing fuels.”
The Pluto reactor is part of the Department of Energy’s Reactor Pilot Program, an effort intended to speed development of advanced nuclear systems that supporters say could provide reliable, carbon-free electricity more quickly and efficiently than traditional reactors.
John Wagner said the partnership addresses one of the country’s biggest challenges: developing reliable energy sources to meet the growing demands of an AI-driven future. “By leveraging AI to design, license and operate reactors, we can fundamentally change the timeline for bringing advanced nuclear energy online.”
The announcement marks the latest step in a broader effort to combine AI with advanced nuclear technology, driven in part by the rapidly increasing energy demands of AI itself.
In December 2025, Oklo and INL conducted a series of plutonium-fueled fast reactor experiments at the Department of Energy’s National Criticality Experiments Research Center. The tests generated benchmark data for Pluto.
The effort expanded in February, when INL and NVIDIA announced a public-private partnership to use AI to accelerate reactor design, licensing, construction and operations. The collaboration became part of the Department of Energy’s Genesis Mission, a national AI initiative aimed at speeding scientific research and technological development.
The nuclear-focused component of that initiative, Prometheus, aims to cut reactor development timelines in half and reduce operating costs by more than 50% through AI systems, digital twins and automated engineering workflows with human oversight.
Under the project, AI systems will assist engineers by running simulations, monitoring design workflows, processing technical data and generating regulatory documentation, while human operators remain responsible for oversight and final decisions.
The United States is home to more than 4,000 data centers, according to industry estimates, with large concentrations in Virginia, Texas and California. Those facilities — especially AI-focused hyperscale centers — consume massive amounts of electricity to power advanced computer chips and the cooling systems needed to prevent overheating.
According to estimates from the International Energy Agency, U.S. data centers consumed about 183 terawatt-hours of electricity in 2024, more than 4% of the nation’s total electricity use. By 2030, that demand is projected to more than double. Some of the largest hyperscale data centers under development are expected to consume as much electricity as hundreds of thousands of homes.
That rising demand has renewed interest in nuclear energy as a steady, carbon-free source of electricity capable of operating around the clock. Several technology companies have already signed agreements with nuclear startups, while efforts are underway to restart dormant nuclear plants in states including Pennsylvania and Iowa to help meet future data center demand.
Supporters of advanced nuclear systems argue that smaller, faster-to-build reactors could eventually help stabilize electric grids strained by the rapid growth of AI infrastructure. Critics, however, continue to raise concerns about costs, nuclear waste and long-term safety. Once popular from the late 1960s through the mid-1970s, nuclear power growth slowed because of high construction costs, lower energy demand and safety fears following accidents, including the Three Mile Island accident and Chernobyl disaster.
