Meta’s Ex AI Chief Yann LeCun Backs AMI’s $1.03B Bet on World-Model AI

Paris: Advanced Machine Intelligence Labs, or AMI, has closed a $1.03 billion seed round at a $3.5 billion pre-money valuation, believed to be the largest seed round ever raised by a European startup. The company was co-founded by Turing Award winner and former Meta chief AI scientist Yann LeCun, who serves as executive chairman, alongside CEO Alexandre LeBrun, former founder and CEO of clinical AI startup Nabla.

The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions — Jeff Bezos’ personal investment vehicle. Strategic backers include Nvidia, Samsung, Toyota Ventures, and Singapore’s Temasek, alongside prominent angels including Mark Cuban, Eric Schmidt, and Tim Berners-Lee.

AMI Labs is building what researchers call world models, AI systems that learn from physical reality rather than predicting patterns in language. Where large language models generate plausible text by learning statistical relationships between words, world models aim to build internal representations of how environments actually function, then use those representations to reason, plan, and make decisions.

LeCun has argued publicly for years that LLMs represent a fundamental dead end on the path to human-level AI, citing their inability to reason causally or handle situations outside their training distribution.

AMI’s technical approach centers on JEPA, Joint Embedding Predictive Architecture, a framework LeCun developed during his time at Meta that processes inputs at a higher level of abstraction than token-by-token transformers, allowing the system to ignore irrelevant detail while retaining semantic structure.

The company has no product and no near-term revenue plans. LeBrun has been explicit about that timeline: AMI is a fundamental research venture that could take years to move from theory to commercial deployment. Target industries include healthcare, robotics, aerospace, and hardware design, domains where AI must exhibit genuine causal reasoning rather than pattern matching, and where hallucinations carry real-world consequences.

Nabla, LeBrun’s former company and AMI’s first disclosed partner, is already embedded in clinical workflows across hundreds of health systems and expects priority access to AMI’s early models for FDA-certifiable agentic applications.

The raise lands as investor appetite for world model research accelerates globally. Fei-Fei Li’s World Labs raised $1 billion last month, while New York and Geneva-based General Intuition secured $134 million at seed stage in October. AMI Labs initially targeted approximately €500 million when LeCun announced the venture late last year after departing Meta.

Investor interest pushed commitments well beyond that figure. The company is now operating across four cities, Paris, New York, Montreal, and Singapore with a research team that includes former Meta and Google DeepMind scientists.