The Genesis Mission sounds large. More than $5 billion across 15 federal agencies, 278 selected projects, contributions from Google, Microsoft, and OpenAI, and a stated ambition the White House compares to the Manhattan Project. The projects span drug discovery, nuclear energy, materials science, and more. All of them share one requirement: use AI to speed up scientific discovery.
The number is real. The contradiction behind it is also real.
Since January 2025, the same administration has terminated or frozen over 7,800 research grants. It proposed cutting the National Science Foundation, which funds basic scientific research across the country, by nearly 57%. It asked for a 40% cut to the National Institutes of Health, the main funder of medical research. Across federal science agencies, more than 25,000 people left in 2025, many of them early-career researchers.
Universities, which train the scientists who run these AI-driven projects, are already responding. PhD programmes at MIT and Duke cut biology admissions by 20% in 2025. The NIH issued roughly a thousand fewer grants for early-career scientists that year compared to recent averages. Two-thirds of research scientists surveyed by the Boston Globe said they were steering promising students away from academic careers.
You cannot build an AI-powered scientific future while cutting the people who understand the science the AI is supposed to assist.
The man steering this is Michael Kratsios, the White House science adviser. He holds a bachelor's degree in politics from Princeton and spent seven years working at investment firms connected to Silicon Valley financier Peter Thiel before entering government. He has no science or engineering degree. Every previous holder of this role did. That is not a trivial detail: his entire framework for science funding comes from venture capital, not from research.
The accompanying policy document, titled "Science: A New Golden Age," proposes shifting money away from universities and toward individual scientists and private companies. It wants faster results, measurable targets, and political appointees with more power to block grants they dislike. It reads, in the words of researchers who reviewed it, like it was written without consulting a single working scientist.
The deeper problem is a category error. Venture capital works by funding many bets, cutting the ones that stall, and harvesting the winners quickly. Science often works the opposite way. The most consequential discoveries, from antibiotics to the internet's underlying architecture, came from research that had no obvious commercial application for years or decades. A funding system that demands fast, measurable returns will systematically defund the work most likely to matter in twenty years.
Meanwhile, China is doing the opposite. Its 2026 budget includes a 16% increase in central government spending on basic research. From 2020 to 2024, China's count of leading researchers, measured by influential publications, rose from around 18,800 to 32,500. The US count fell from about 36,600 to 31,800. China has now surpassed the US in the number of top-tier scientists, by that measure. At least 85 US-based researchers joined Chinese institutions between 2024 and late 2025.
Europe is also moving. The EU launched a 500 million euro programme specifically to attract researchers leaving the US. Canada, Australia, and others have created similar schemes.
What does this mean for a business operator who has nothing to do with university labs? More than it might seem. The drugs your company's health insurance covers, the materials your products are made from, the logistics software running your supply chain: all of it traces back, eventually, to basic research done in universities and federal labs. That pipeline is now under serious strain. The effects will not show up this year. They will show up in a decade, when the products that should have existed do not.
The Genesis Mission grants are real money going to real projects, and some of them may produce useful results. But spending $5 billion on AI-driven science while cutting the institutions that produce scientists is like buying a fleet of trucks while closing the mechanic schools. The trucks run for a while. Then they don't.