White House Steers Science Funding to AI and Individual Scientists

The Trump administration wants to rebuild how the United States spends roughly $200 billion a year on research, steering federal money toward individual scientists and the use of AI while pulling it away from the universities and consensus peer-review panels that have run the system since World War II.

The plan runs through a report the White House Office of Science and Technology Policy released on July 21, 2026, titled Science: A New Golden Age, and a companion budget memo its director, Michael Kratsios, co-signed with Office of Management and Budget Director Russell Vought. The Wall Street Journal first reported the overhaul, which the administration frames as a race to outpace China across AI, quantum computing, and other frontier fields.

Much of the document is direction-setting rather than law: Congress, not the White House, writes the research budget it seeks to reshape. But it is the clearest statement yet of how this administration wants the federal science enterprise rebuilt, and it reaches every agency that funds research.

From institutions to individuals

The report’s premise is that America’s research system — largely unchanged since Vannevar Bush’s 1945 blueprint Science: The Endless Frontier, which built the National Science Foundation and the modern university-government partnership — has grown slow and top-heavy. Kratsios pitched it as the first comprehensive rethink of that enterprise in more than 80 years. Its fix is to send more support to individual researchers and to widen how the government awards money.

The memo tells agencies to supplement traditional consensus-driven peer review, which it says rewards incremental work, with faster, higher-risk mechanisms. Among them:

• “Golden tickets” that let a single reviewer fund an unconventional proposal a committee would pass over;
• “Fast grants” with applications a few pages long and review times under a month;
• Long-duration awards of five years or more, funded in full up front; and
• Prizes and challenges structured to draw at least three private dollars for every federal one.

It also pushes grants that follow the scientist rather than the school — portable awards modeled on the National Science Foundation’s graduate fellowships, which a researcher keeps when switching institutions. For universities, that is the sharp edge: money they have long treated as core support would instead chase individuals, agile new outfits such as the NSF’s X-Labs, and expanded ARPA-style teams built for high-risk projects. The report asks agencies to rebalance toward the physical sciences and engineering, and to build “metascience” units that study which funding methods actually produce results.

AI at the center

The overhaul puts AI at the core of how discovery happens. Agencies are told to treat the Genesis Mission — the AI-for-science program Trump launched in November 2025 with the stated goal of doubling research productivity within a decade — as a central priority, and to fund work that uses AI as an instrument of discovery rather than a convenience. The report calls for domain-specific scientific foundation models, curated federal datasets to train them, and autonomous laboratories where AI systems propose hypotheses, run experiments, and iterate with limited human input.

That ambition carries commercial weight. The memo directs agencies to expand access to advanced compute, buy commercial capacity through public-private partnerships where it is cost-effective, and build domestic supply chains for AI-ready scientific instruments with standardized, interoperable interfaces — a demand signal for chipmakers, cloud providers, and lab-hardware vendors. It sits alongside a wider Trump AI agenda that already spans compute buildout and electricity commitments for AI.

The same mission-driven logic runs through a slate of national goals the report ties federal research to: a quantum computer built to open an era of scientific discovery, a commercial fusion power demonstration by the mid-2030s, a return of Americans to the lunar surface by 2028, general-purpose robotics, and next-generation semiconductor manufacturing. The organizing frame is competition with China, which the administration now describes as a peer-level rival in research spending.

Guidance now, appropriations later

For all its reach, most of the report is advisory. Its budget priorities are recommendations agencies are asked to weigh in their fiscal 2028 requests to the OMB — and lawmakers control what actually gets funded.

One provision has teeth. Within 90 days, every agency with at least $3 billion in 2026 research budget authority must send Kratsios and Vought an action plan detailing how it will adopt the new practices. That is where intent turns concrete: whether the National Science Foundation, the National Institutes of Health, and the Department of Energy actually stand up fast grants, golden tickets, and portable awards, or file the memo as a statement of direction while the budget fight plays out.

For the AI industry, the direction is unambiguous. A federal research budget rebuilt around AI-for-science, commercial compute, and domestic instrument supply chains turns the government into a large new customer for the same companies racing to sell those tools — and puts universities on notice that the terms of federal support are being rewritten.