Stop Celebrating the $5B AI Science Plan. The Government Can’t Spend It.

You’ve probably seen the headlines by now. The Trump administration just announced a $5 billion “Genesis Mission” — a sweeping initiative to deploy AI across federal science agencies. Five billion dollars. For AI. For science. Sounds like a dream, right?

Here’s the problem: the agencies that are supposed to receive this money have been gutted.

Not slightly understaffed. Not “going through a transition.” Gutted. The very apparatus tasked with implementing this vision has been hollowed out to the point where it can barely process expense reports, let alone coordinate a multi-agency AI transformation.

Throwing $5 billion at a broken system doesn’t fix the system — it just creates a more expensive broken system.

Let me walk you through what’s actually happening here. The announcement is political theater at its finest. A big number, a bold vision, a press release that makes everyone feel like America is “winning” the AI race. But if you’ve watched how federal science agencies actually function over the past few years, you already know the punchline.

The agencies that would receive these funds have seen mass departures of career staff. Institutional knowledge has walked out the door. The people who knew how to navigate procurement, how to evaluate AI proposals, how to separate real innovation from vendor snake oil — they’re gone. And the public trust that once made these institutions effective? Erooded by years of politicization, budget cuts, and leadership churn.

One commenter nailed it: “My big question is where is that money going to go? On the surface, I don’t mind the program. However, the implementation is likely to be terrible as we have seen so far. The various agencies tasked with implementing this have already been gutted and dysfunctional.”

That’s not cynicism. That’s pattern recognition.

Here’s what nobody in the press is talking about: the bottleneck isn’t money. It never was. The bottleneck is administrative capacity. You could pour $50 billion into these agencies tomorrow and the outcome would be the same — a trickle of actual deployment, a flood of consultant fees, and a press release in 2028 claiming success.

The real AI race isn’t won by whoever announces the biggest number — it’s won by whoever can actually execute.

Think about what effective deployment of AI for science actually requires. You need technical staff who can evaluate models. You need procurement officers who understand the difference between a real AI system and a PowerPoint deck. You need program managers who can coordinate across agencies without spending eighteen months in interagency meetings. You need trust — from researchers, from the public, from the private sector partners who’d actually build the tools.

Right now, you have none of those things in sufficient quantity.

What you have is a hollow shell. An org chart with names crossed out. A procurement system so broken that even well-intentioned career staff can’t move fast enough. And now, a $5 billion spotlight pointed directly at that dysfunction.

The taxpayer is footing the bill for this. Tech workers and researchers who might benefit from real investment will instead watch the money get absorbed by the usual suspects — Beltway contractors, middleware consultants, and the infinite layers of oversight that ensure nothing actually gets built.

Announcing money is easy. Spending it well is the hard part nobody wants to talk about.

So yes, be skeptical. Not of the idea — AI for science is genuinely worth investing in. Be skeptical of the delivery mechanism. Because right now, we’re watching someone promise to build a skyscraper on a foundation that can barely hold a garden shed.

The most dangerous thing in policy isn’t a bad idea. It’s a good idea attached to a broken machine. And that’s exactly what the Genesis Mission is.