The plan would give researchers shared access to government data, AI models, and computing resources while expanding the use of automated laboratories.
The White House called for overhauling the federal research system to accelerate scientific discovery, with AI playing a central role in the proposed changes. The plan would give scientists broader access to government data, specialized AI models, supercomputers and research equipment.
The recommendations appear in Science: A New Golden Age, a report released Monday by White House Office of Science and Technology Policy Director Michael Kratsios. The report calls for expanding the use of AI to accelerate scientific discovery as part of a broader overhaul of how the federal government funds, organizes and carries out research.
A separate federal research priorities memo puts the first deadline in place. Agencies with at least $3 billion in fiscal 2026 research and development funding must submit action plans within 90 days describing how they will begin carrying out the administration’s priorities.
Building Shared AI Research Resources
The plan calls for federal agencies to make more scientific data available for AI training and analysis. That could include experimental records, negative test results, and operational data from laboratory procedures that researchers often do not preserve or share.
Agencies would also invest in AI models built for specific scientific fields. These models could simulate natural processes and support research in areas including biotechnology, advanced manufacturing, semiconductors, critical materials, and nuclear energy.
The government would make the data and computing resources available across the research community because the required data and computing power are often beyond the capability of a single researcher or institution.
Much of the work would take place through the Genesis Mission, an AI for science initiative President Trump launched in November 2025. The program is building a shared platform that connects federal supercomputers, AI systems, scientific datasets, and advanced research equipment at national laboratories and other federal facilities.
The administration says the goal of the Genesis Mission is to double the productivity of American science and engineering within a decade. The Department of Energy is also directed to identify at least 20 national science and technology challenges for the program.
AI Systems Would Run More Experiments
The report also calls for greater use of robotics and automated laboratories.
Under that model, an AI system could generate a hypothesis, design an experiment, operate laboratory equipment, interpret the results, and adjust the next test. The report argues that running this process continuously could allow researchers to conduct more experiments and change course without waiting for each result to be reviewed manually.
Agencies are instructed to propose investments in automated laboratories, AI-connected research equipment, and systems that can control large experiments.
The plan also envisions limited partnerships with universities and private companies that can contribute computing power, AI models, data or scientific expertise. The Department of Energy announced agreements with 24 AI companies, semiconductor manufacturers and cloud providers for the Genesis Mission in December 2025.
Access would be controlled. The executive order establishing the Genesis Mission requires agencies to protect classified, private and proprietary data and apply federal cybersecurity standards. Outside researchers and companies would also be subject to access reviews and restrictions on how they use government data, models and computing systems.
The 90-day agency plans must identify specific actions, timelines and the offices responsible for carrying them out. Agencies must address longer-term funding through their fiscal 2028 budget requests.
