The same government that can't keep a website running just promised to rebuild American science from scratch using AI money and a philosophy borrowed from tech Twitter.
The Summary
- Trump's "Genesis Mission" deployed $5 billion across hundreds of AI-driven science projects, pitched as the new Manhattan Project
- Michael Kratsios, Trump's science adviser with zero science credentials, told Congress this marks a "New Golden Age" prioritizing AI, robotics, and nuclear energy over life sciences
- The shift signals a hard pivot from biology and climate research toward tech-bro priorities: automation, energy, and anything that sounds like it came from a Balaji tweet
The Signal
The federal science apparatus just got a product manager. Michael Kratsios, who previously worked at Palantir and as Trump's tech adviser, is now steering $5 billion in research grants toward AI projects with the kind of conviction normally reserved for Series B pitch decks. No PhD required, apparently.
The Genesis Mission isn't subtle about what it's optimizing for. Hundreds of projects focused on artificial intelligence, robotics, and nuclear energy. Life sciences, the field that gave us vaccines and CRISPR, gets relegated to the back bench. The message is clear: if your research doesn't involve neural networks or uranium, you're playing yesterday's game.
"This is government funding reimagined as venture capital, complete with the same blind spots about what actually matters."
Here's what the tech-first approach gets right: American AI research needs coordinated investment, and the private sector won't fund the unsexy infrastructure work that makes breakthroughs possible. Nuclear energy does need a renaissance if we're serious about powering data centers without cooking the planet. These aren't wrong priorities.
But here's the tell. The people making these decisions aren't scientists debating trade-offs in peer review. They're political appointees who learned about research funding from watching Elon argue on X. Kratsios comes from the Thiel-sphere, where "science" means whatever makes good copy in a16z's newsletter.
The $5 billion figure sounds massive until you remember:
- NIH's annual budget is $47 billion, mostly life sciences
- The original Manhattan Project cost $28 billion in today's dollars
- China spent $115 billion on AI research in 2025 alone
This isn't actually Manhattan Project scale. It's a down payment on a rebrand. The administration wants the aesthetics of transformative science without the patience for how science actually works. Real breakthroughs come from biology labs studying fruit flies for decades, not hackathons with GPUs.
The Implication
Watch what happens when the first Genesis grants produce incremental papers instead of fusion reactors. Either the program quietly shifts back toward normal science, or it becomes a subsidy program for AI companies dressed up as research.
The real test: will these projects publish in Nature, or will they announce on Product Hunt? If it's the latter, this isn't science funding. It's industrial policy for the agent economy, channeled through people who think peer review is what happens in Y Combinator's Slack.