OpenAI has formalized an expanded collaboration with the U.S. Department of Energy through a new memorandum of understanding to broaden federally funded AI and advanced computing research aligned with national scientific priorities.

Genesis, the DoE’s initiative, aims to apply AI to major scientific priorities; the MOU commits to identifying new projects, providing frontier models, and expanding compute access to researchers to advance discoveries and energy-related research, including fusion.

OpenAI emphasizes responsible governance and transparency by tailoring work to government data rules, sharing results through papers and blog posts, and using cycles of feedback from agencies, scientists, and the broader community to guide development.

Milestones for success include public updates on efficiency gains, faster experimentation, and accelerated discovery across material science and public health, with a focus on reducing barriers, improving compute access, and delivering tangible breakthroughs in the year ahead.

A notable initiative is the 1,000 Scientists AI Jam, which gathered a thousand scientists across nine national labs to apply OpenAI models to chemistry, physics, biology, and other disciplines, gathering feedback to improve tools.

OpenAI has deployed models on the Venado supercomputer at Los Alamos and conducted bioscience evaluations with national labs, underlining a long-standing partnership focused on safe, aligned AI deployment for scientific discovery.

Prism, a new product released by OpenAI, is designed to help scientific researchers collaborate on research, reflecting the ongoing integration of government partnerships into OpenAI’s broader toolset for science.

The collaboration builds on several years of joint work with the DoE and its national laboratories, including Los Alamos, Lawrence Livermore, and Sandia, and centers on applying OpenAI tools to accelerate government science across domains.