The partnership aims to cut the time it takes to develop new reactors in half and reduce operating costs by more than 50 percent, according to the announcement.
The collaboration is part of the U.S. Department of Energy’s Genesis Mission, a national initiative designed to build a powerful scientific computing platform to accelerate discovery, strengthen national security, and drive energy innovation.
A key objective of the mission is to ensure the faster, safer, and more affordable deployment of nuclear energy. This effort is codenamed “Prometheus.”
The Prometheus program will apply AI tools to reactor design, licensing, manufacturing, construction, and operation.
The goal is to automate processes that traditionally require significant engineering resources and take years to complete, while maintaining human oversight in decision-making – a “human-in-the-loop” approach. The agencies state this approach could at least double the speed of new reactor deployment.
The collaboration also addresses the growing electricity demand driven by artificial intelligence infrastructure.
By combining AI-driven reactor design with data centers powered by nuclear energy, the initiative seeks to create a cycle where both technologies support and strengthen each other.
AI is Changing Reactor Development Timelines
At the heart of this initiative is the use of generative AI models, digital twins, and agent-based workflow systems to simulate and validate reactor systems before physical construction begins. This approach is gaining traction as companies seek to reduce costs and accelerate innovation in the energy sector.
Digital twins will be trained using decades of nuclear energy data, laboratory research results, and operational experience from experimental reactors at the Idaho National Laboratory.
“This partnership marks a transformational approach to one of our nation’s biggest challenges – how to ensure abundant and reliable nuclear energy at the speed and scale needed for our AI-driven future,” said Idaho National Laboratory Director John Wagner.
“By harnessing AI to design, license, and operate reactors, we can fundamentally change the timeline for deploying advanced nuclear energy.”
NVIDIA will provide its AI infrastructure and GPU-accelerated computing solutions to the initiative. The collaboration includes adapting nuclear simulation codes, such as MOOSE, BISON, Griffin, and Pronghorn, to NVIDIA GPU architectures to expand modeling capabilities and reduce computation times.
“NVIDIA is proud to partner with the U.S. Government, applying AI and accelerated computing to advance nuclear energy while lowering the cost of energy for Americans,” said John Josephakis, NVIDIA’s global vice president of sales and business development for high-performance computing, and supercomputing.
“By combining the Idaho National Laboratory’s decades of nuclear energy expertise with NVIDIA’s AI infrastructure, AI can be used to design, license, and operate reactors faster, safer, and at lower costs – ensuring abundant energy needed for scientific advancement.”
Supercomputers and Nuclear Codes
The initiative will leverage the Department of Energy’s leading supercomputers for training large-scale AI models and conducting reactor simulations.
The Idaho National Laboratory’s infrastructure, including the Neutron Radiography Reactor and the Microreactor Applications Research Validation and Evaluation (MARVEL) project, will provide real-world data for validating digital twin models.
The MARVEL project has not yet begun operations.
Officials state that the broader goal of the initiative is to support the modernization of the regulatory framework and the industry’s transition to AI-powered nuclear energy tools.
The program could be expanded to include reactor developers, utility companies, investors, and other national laboratories.
“Now is the moment to aggressively accelerate the deployment of AI-enabled nuclear energy, increasing energy affordability for Americans while simultaneously driving AI development in the United States,” said Ryan Bahran, Deputy Assistant Secretary for Nuclear Reactor Technologies.
“This public-private partnership is a purposeful approach to accelerating AI that goes beyond incremental, small improvements. It has the potential to fundamentally change the paradigm for deploying nuclear energy and how science and innovation are developed.”
