According to Sifted, Meta’s outgoing chief AI scientist Yann LeCun is in talks to raise a staggering €500 million at a valuation of about €3 billion for a new startup called Advanced Machine Intelligence (AMI) Labs, which he’s launching early next year. The project will be led by CEO Alexandre Lebrun, the founder of Paris-based AI startup Nabla and a former Meta research engineer who worked closely with LeCun. Lebrun is transitioning from his role at Nabla, which has now inked a strategic research partnership with AMI Labs. LeCun announced his departure from Meta last month to focus on this venture, which aims to build “world models” that understand the physical world, unlike current text-generating LLMs. The discussions with investors are still in early stages and could change.
The “World Model” Gambit
Here’s the thing: LeCun isn’t just chasing another generative AI chatbot. He’s making a foundational bet on what he sees as the next necessary step—systems that comprehend how the world works, not just how to string words together. This means AI that learns from video, spatial data, and actions, which is basically the holy grail for making robots that can actually function in unpredictable environments. It’s a wildly ambitious and expensive research direction that most VCs in Silicon Valley, “hypnotised by generative models” as LeCun put it, might not have the patience for. So he’s raising a war chest upfront and setting up shop in Paris. Smart move? It’s a huge risk, but if anyone has the credibility to pull in half a billion euros on a vision, it’s a Turing Award winner.
A Deliberate European Play
This isn’t just about the technology. LeCun is making a pointed statement about geography and talent. By planting his flag in Paris and talking about a global entity with research orgs particularly in Europe, he’s betting that the next big AI breakthrough won’t come from the well-trodden path in California. He’s been a long-time advocate for European AI, convincing Meta to open its FAIR lab there nearly a decade ago. Now, he’s using his clout to potentially create a new epicenter. The question is, can €500 million and LeCun’s reputation create a magnet strong enough to rival the concentration of capital and talent in the Bay Area? It’s one of the most interesting experiments in the industry right now.
What This Actually Changes
Look, a €3 billion valuation on a company that doesn’t exist yet, focused on tech that’s largely research, tells you everything about the frothiness of the AI market. But it also signals a potential pivot. The industry has been obsessed with language and images. LeCun is betting the real value—and the harder problem—is embodiment and reasoning. If AMI Labs makes tangible progress, it could shift investor and researcher attention back toward robotics and simulation. For businesses, the eventual applications in complex logistics, manufacturing, and physical automation could be profound. Think about the need for reliable, intelligent systems in industrial settings—the kind of environments where understanding context and physics is non-negotiable. It’s a different path from the current AI narrative, and that’s probably why it’s so compelling.
