OpenAI Advancing the Next Era of National Science

OpenAI has announced a strategic commitment to integrate frontier AI models into the American scientific infrastructure, partnering with the U.S. government, National Laboratories, and universities. This initiative aims to compress decades of scientific progress into years by pairing frontier models with supercomputers, simulations, and specialized research facilities.

Strategic Commitments to the Genesis Mission

OpenAI is supporting the U.S. Department of Energy’s Genesis Mission—a collaboration between the DOE, 17 National Laboratories, universities, and industry designed to double the productivity and impact of American research within a decade. To achieve this, OpenAI is providing the following resources:

• Financial and API Support: OpenAI will provide $4 million in Codex access to approximately 2,000 Genesis researchers and $3 million in API support for two large-scale scientific campaigns.
• Usage Credits: Participating researchers can receive up to $10 million in API usage for every $2.5 million spent.
• Specialized Model Access: Selected national-laboratory researchers on eligible biology projects will have access to GPT-Rosalind's specialized bioscience capabilities.
• Cybersecurity and Early Access: Trusted national-laboratory leaders will receive early access to selected models and features, and selected cybersecurity researchers will receive expanded access to advanced cyber capabilities for defensive research.

Targeted Scientific Campaigns

OpenAI is focusing concentrated resources on two initial large-scale campaigns to establish repeatable methods for applying AI to difficult national challenges:

Breakthroughs for High-Temperature Superconductors

This campaign integrates frontier AI, simulation, materials expertise, and experiments to pursue superconducting materials that operate at higher temperatures and practical pressures. Potential applications include improvements in energy, transportation, medicine, scientific instrumentation, and national security.

An Atlas of the Machine-Accessible Frontier

This campaign aims to identify where AI can support meaningful scientific advances using existing knowledge, data, and computation, helping researchers distinguish between problems that are now tractable via AI and those that still require physical-world evidence.

Existing Collaborations and Infrastructure

OpenAI's current efforts build on several existing initiatives within the national-laboratory system:

• High-Performance Computing: Advanced reasoning models have been deployed on Venado, the supercomputer at Los Alamos National Laboratory, as a shared resource for National Nuclear Security Administration laboratories.
• Bioscience Evaluations: In collaboration with Los Alamos, OpenAI developed evaluations to study the safe use of multimodal AI systems in realistic laboratory environments to understand both opportunities and risks.
• Researcher Feedback: An "AI Jam Session" involving nine National Laboratories and over 1,000 scientists was conducted to test frontier models on domain-specific problems.

The Collaborative Model for National Science

OpenAI posits that realizing the potential of AI in science requires a coordinated effort across multiple sectors. In this model, the government coordinates priorities and invests in shared infrastructure, while laboratories and universities define the problems and uphold scientific standards. Industry provides the frontier capabilities and engineering expertise, and researchers remain central to the process by defining questions, choosing methods, and validating results.