🖥️ Building, not selling

Yann LeCun hasn’t moved. He still teaches at NYU, still lives in New York. But when he co-founded AMI Labs with Alexandre Lebrun in March 2026, Paris became the headquarters. €890m seed at €3.1bn pre-money, the largest seed round ever closed on the continent. A Turing Award winner who could have planted his flag anywhere chose a European capital for the articles of incorporation, not for the tax breaks.

The detail matters more than it appears. For fifteen years, the continent exported brains and imported finished products. The talent that stayed published papers while ARM got sold to SoftBank, DeepMind to Google, Skype to Microsoft. AMI Labs reverses the sequence. A frontier lab on world models, a ten-year horizon, no product before 2027, a fundamental research thesis at the core. You don’t do that from a place you treat as an antechamber.

Lebrun, the CEO, put it plainly: AMI will be a long project, scientific first, applied second. Healthcare first, then robotics and industrial applications. Some of the code will be open source, the papers published. It’s a research-lab positioning financed like a startup, not the reverse. At this capital level, in this field, on this continent, it has no precedent.

What makes a European hub credible in 2026 didn’t exist in 2016. Strong AI research clusters in Paris, Zürich, Amsterdam and Munich. ETH and EPFL pumping out PhDs who no longer drain automatically to DeepMind or Meta. BioNTech and Helsing showing that frontier science can scale from Mainz or Munich without relocating. Fifteen years of quiet cycles, public money, private money, offices closing, founders trying again. That’s the mesh LeCun looked at before signing.

One objection: the capital remains largely non-European. True. European funds co-lead, but Greycroft, Bezos Expeditions, Nvidia and Temasek weigh more heavily on the cheque side. The project is European, the money funding it is global. Uncomfortable to admit, but it doesn’t change the thesis: what came back isn’t the cheques, it’s the researchers. And that’s harder to replicate.

The question for a European founder in 2017 was: where do I need to be for my ambition to be taken seriously? The answer was mechanically San Francisco or Palo Alto. In 2026, AMI Labs has just said you can answer Paris, London, Amsterdam or Berlin without it sounding like a concession. That’s the signal.