OpenAI and the U.S. Department of Energy (DOE) have signed a memorandum of understanding (MOU) aimed at expanding collaboration on artificial intelligence and advanced computing in support of U.S. scientific research initiatives, including the Genesis Mission, a U.S. government–led effort that brings together federal agencies, national laboratories, and industry partners to apply advanced AI and high-performance computing to accelerate scientific discovery.

The agreement reflects a shared effort to explore how frontier AI models can help researchers generate ideas, test hypotheses more efficiently, and move from insight to validated results more quickly—capabilities that have implications across energy, health, national security, and fundamental science.

Under the MOU, OpenAI and DOE will establish a structured framework for information sharing, technical coordination, and exploratory collaboration, with the potential for follow-on agreements as specific projects are defined. The collaboration is part of OpenAI for Science, an initiative focused on pairing advanced AI models with real research environments, scientific tools, and domain expertise.

The Genesis Mission brings together government agencies, national laboratories, and industry partners to apply AI and advanced computing toward accelerating scientific discovery. The new agreement is designed to support that effort by creating pathways for deeper collaboration, with future project work defined through follow-on agreements.

OpenAI also submitted recommendations to the White House Office of Science and Technology Policy outlining how access to frontier AI models, compute resources, and real research environments could strengthen U.S. science and technology leadership. The company characterized 2026 as a potential “Year of Science,” positioning the DOE agreement as a practical step toward that goal.

OpenAI and the U.S. Department of Energy share a focus on advancing both basic and applied research while strengthening U.S. leadership in AI and advanced computing. The memorandum of understanding provides a formal mechanism for exchanging technical expertise, coordinating activities, and exploring collaboration in areas where DOE’s national laboratories offer unique capabilities, including world-class scientific facilities, modeling tools, and research data.

For OpenAI, the agreement reflects its approach to building OpenAI for Science: working directly with researchers to understand where AI meaningfully supports discovery, where its limitations remain, and how it can be safely integrated into real scientific workflows. That work depends on close collaboration with domain experts, access to advanced scientific infrastructure, and rigorous evaluation in operational research settings. The MOU creates space for that collaboration while maintaining the governance and accountability required for responsible scientific use.

In collaboration with DOE national labs, OpenAI convened the 1,000 Scientist AI Jam Session, an event spanning nine laboratories that engaged more than 1,000 scientists. Participants used frontier AI models to test domain-specific research challenges, evaluate model behavior, and provide structured feedback.

According to OpenAI, the session was designed to allow researchers to stress-test AI tools on real scientific problems and inform how future systems are developed for scientific use.

OpenAI has partnered with laboratories under the National Nuclear Security Administration (NNSA), including Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and Sandia National Laboratories, to support scientific and technical research.

As part of this work, OpenAI deployed advanced reasoning models on the Venado supercomputer at Los Alamos. The system serves as a shared resource for researchers across NNSA labs and is being used to explore how frontier AI models perform in high-performance computing environments tied to national research priorities.

OpenAI has also worked with Los Alamos National Laboratory to develop evaluation methods that examine how multimodal AI systems may be used safely in laboratory environments. The research aims to move beyond text-only benchmarks toward more realistic assessments of how AI tools influence outcomes in high-consequence scientific domains.

The evaluations emphasize expert oversight, careful study design, and risk-reduction practices, reflecting an effort to understand both the capabilities and limitations of AI in sensitive research settings, with a clear commitment to risk reduction.

OpenAI’s approach to scientific collaboration rests on two core principles:

DOE’s national laboratories sit at the intersection of these capabilities, operating some of the world’s most advanced scientific infrastructure while convening experts working on problems where improved reasoning and computation can translate directly into societal benefit.

The agreement was announced alongside a White House meeting related to the Genesis Mission, where government officials, national laboratory leaders, and industry partners discussed how AI and advanced computing could accelerate scientific discovery. Kevin Weil, Vice President of OpenAI for Science, joined Department of Energy officials and other collaborators at the meeting.

“When frontier AI meets the expertise of the national labs, it opens up new ways to explore ideas, test them faster, and accelerate scientific progress.”
— Kevin Weil, Vice President of OpenAI for Science

OpenAI and the Department of Energy say the MOU creates space for deeper collaboration while preserving the rigor, accountability, and governance required for responsible scientific use of AI. As specific projects emerge, the partners plan to define future agreements that translate the framework into concrete research outcomes.

Q: What is this agreement, exactly?
A: It is a memorandum of understanding that establishes a framework for OpenAI and the Department of Energy to share expertise, coordinate activities, and explore AI-driven research collaborations. It does not itself authorize specific projects, but enables future agreements as initiatives are defined.

Q: What kinds of research could this affect?
A: The collaboration may span areas such as energy systems, fusion research, bioscience, national security–related science, and other domains where DOE national labs operate advanced facilities, datasets, and computing infrastructure.

Q: How is this different from typical AI research partnerships?
A: Rather than focusing on isolated experiments or academic benchmarks, the work places frontier AI models directly into real scientific workflows, including national lab supercomputers and laboratory environments, with active involvement from domain experts.

Q: Are these AI models already being used inside national labs?
A: Yes. OpenAI has already deployed advanced reasoning models on systems like the Venado supercomputer at Los Alamos National Laboratory and collaborated with scientists across multiple DOE labs over the past year.

Q: What safeguards are in place for sensitive research areas?
A: According to OpenAI, evaluations are conducted with expert oversight, careful study design, and a focus on risk reduction, particularly in high-consequence domains such as bioscience.

This partnership matters because it places frontier AI inside the day-to-day work of U.S. national laboratories, rather than treating AI as an external tool or short-term experiment. These laboratories were created after World War II to provide permanent scientific capability, employing career scientists whose work continues across decades and presidential administrations. Decisions made inside these institutions influence energy systems, climate modeling, and national research priorities—so how AI is introduced there carries long-term consequences.

That context helps explain why consumer AI tools or document-based approaches like retrieval-augmented generation (RAG) are not sufficient on their own for national laboratory research. In sectors like finance or healthcare, AI is typically used to retrieve information or support decisions without directly interacting with underlying systems. In scientific research environments, AI may influence simulations, experimental design, or analysis pipelines that shape real-world outcomes, making governance, evaluation, and placement inside secure research infrastructure essential.

For policymakers and research leaders, the agreement reflects a practical constraint: advanced science increasingly depends on both high-performance computing and advanced reasoning systems, but neither is useful without clear rules for testing, oversight, and accountability. Embedding AI within national labs allows these tools to be evaluated under real scientific conditions, rather than relying on assumptions drawn from consumer or enterprise use cases.

For scientists, this collaboration affects how research work evolves. AI systems shaped through hands-on use inside lab environments are more likely to support meaningful tasks—such as exploring complex ideas, reviewing large bodies of research, or testing hypotheses—while also revealing limitations early. Over time, this approach influences which tools are trusted, how workflows change, and where human judgment remains essential.

At a broader level, the agreement reflects how governments and AI developers are beginning to treat frontier AI as shared scientific infrastructure, not just a commercial capability. The long-term impact will depend less on individual models and more on whether these collaborations produce systems that are reliable, transparent, and aligned with the public interest.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.