The clock is ticking.
Germany just dropped €125 million onto the table. Not for incremental upgrades. This is for frontier AI. The kind of tech that changes everything. SPRIND, the federal innovation agency, wants to fund the companies that will become Europe’s answer to OpenAI. Or DeepSeek.
Why now?
Because waiting isn’t an option. Jano Costard from SPRIND put it bluntly: we have no time. The global race isn’t pausing for a committee meeting. It’s accelerating.
Most of the heavy hitters? US-based. Billions in private backing. China is moving fast too. DeepSeek released their V4 model last April. Europe is feeling the squeeze.
“A competition globally is not waiting.”
So here is the play. Three stages. Two years.
First off, up to ten teams get €3 million. They build something. Six of those survive the cut for the second round, grabbing up to €8 million. If they’re still standing, three finalists could walk away with €15.5 million apiece. Costard expects hundreds of applicants. Maybe thousands.
But let’s be real.
Is €125 million enough to punch above its weight against Washington and Beijing?
Probably not. Costard admits it. That money is just the seed. The real goal is unlocking billions in private follow-on funding. It’s a bridge. Build the tech until it looks inevitable. Until the potential is undeniable.
The strategy? Stop chasing yesterday’s benchmarks.
“We cannot try to compete with today’sAnthropic,” Costard says. You can’t win a horse race if you’re training donkeys.
The idea is to create new paradigms. Completely different approaches to how AI works. Current methods have limits. Europe needs to find what they haven’t.
This isn’t just about money though. It’s about where the startups end up.
There is a long, painful pattern here. Great ideas are born in Berlin, London, or Paris. Then they run into bureaucracy. Funding drags on for months. Founders get tired of the red tape. So they move. Usually to California.
The European Commission tried to fix the structural mess in March. They proposed the EU Inc. One company law. One borderless legal structure for the whole continent. Makes sense, right?
Costard agrees it helps. But structure alone won’t save talent. Speed matters. Public funding needs to move faster. Less paperwork. More velocity.
So where is Europe actually good at something?
Data. Manufacturing. Privacy.
We don’t lack the brains. The research pedigree is there. The technical ability is fine. The gap isn’t intelligence; it’s execution. Translating smart papers into killer products. Into the next big name.
Maybe we can build an AI that respects privacy in a way Silicon Valley won’t touch. Maybe industrial know-how is the edge.
Who knows if it’s enough?
The experiment starts now.
