Yann LeCun Raises $1B for World Models — The Full Timeline
— by Vivax
The world's most prominent critic of large language models just raised $1.03 billion to build world models instead.
The man who spent years warning that large language models are a dead end just raised one of the largest seed rounds in history to prove it. In March 2026, Yann LeCun — Turing Award winner and Meta's chief AI scientist for seven years — closed a $1.03 billion seed round for AMI Labs, a new venture built entirely around world models. Here is exactly how it unfolded, in order, and why it matters for grounded clinical AI.
November 19, 2025 — LeCun confirmed he was leaving Meta after 12 years, including five as founding director of its FAIR research lab and seven as Chief AI Scientist. He told WIRED that as Meta rushed to 'catch up with the industry on LLMs,' the work drifted away from his real interest: world models. On the company's new AI leadership he was blunt — 'You certainly don't tell a researcher like me what to do.'
December 2025 — Bloomberg first reported on December 5 that LeCun was leaving to build a 'world model' startup, and on December 19 he confirmed it. The new lab — later named AMI Labs, for Advanced Machine Intelligence (pronounced 'ami,' French for 'friend') — would be headquartered in Paris, with LeCun as Executive Chairman and Alex LeBrun (founder of Nabla) as CEO. Early reporting pegged a target valuation north of $3.5 billion before the company had even formally launched.
January 23, 2026 — AMI's team and thesis came into focus. The lab is built around JEPA, the Joint Embedding Predictive Architecture LeCun proposed in 2022: instead of predicting the next word or pixel, it learns abstract representations of how the world works — the way humans and animals reason about cause and effect. AMI says it will pursue reliability, controllability, and safety across robotics, automation, wearables, and — notably — healthcare.
March 9, 2026 — AMI Labs closed a $1.03 billion seed round at a $3.5 billion pre-money valuation, among the largest seed rounds ever raised by a European company. It was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, with strategic backing from NVIDIA, Temasek, and Toyota Ventures, plus individuals including Jeff Bezos, Eric Schmidt, Mark Cuban, and Tim Berners-Lee. The capital funds AI built around reasoning, planning, and world models — LeCun's explicit alternative to today's LLM-dominated approach.
The signal for healthcare is hard to miss. The field's most prominent skeptic of LLMs — backed by NVIDIA and some of the most sophisticated investors in technology — is putting a billion dollars behind the same thesis Vivax is building on: that safe, grounded, decision-grade AI needs a model of the world, not just a model of language. AMI itself names healthcare as a target domain. It is early — LeBrun cautions that world models will take years, not months, to reach production — but the direction of travel is now unmistakable.