Yann LeCun's AMI Labs Raises $1.03B, Europe's Biggest AI Seed

Yann LeCun's Advanced Machine Intelligence (AMI) Labs closed a $1.03 billion seed round at a $3.5 billion pre-money valuation in March 2026, marking Europe's largest-ever seed funding for an AI startup. The Paris-headquartered company, founded just four months after LeCun departed Meta in November 2025 where he was chief AI scientist, is building world models based on his Joint Embedding Predictive Architecture (JEPA) framework—an alternative to large language models that learns abstract representations of the physical world rather than predicting tokens.

The round was co-led by five institutional investors: Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, with participation from NVIDIA, Samsung, and angels including Jeff Bezos, Tim Berners-Lee, Eric Schmidt, and Jim Breyer. LeCun had initially sought €500 million but demand exceeded that significantly, allowing the company to be selective. AMI named Alexandre LeBrun, former CEO of French health-tech startup Nabla, as CEO, with LeCun serving as executive chair. The company operates from Paris with offices planned in New York, Montreal, and Singapore.

For architects, AMI Labs represents the first major industrial bet that LLM-scale architectures have fundamental limits and that world models—systems that understand causality and physical constraints—are the next frontier. LeCun's credibility (Turing Award 2018, FAIR directorship) and the capital backing signal serious conviction. The JEPA approach has shown early results: V-JEPA 2 achieved 77.3% accuracy on physical motion understanding benchmarks and 80% success on robotic manipulation tasks. This is directionally important for AI infrastructure planning—track whether world models shift capital allocation in frontier AI away from LLM optimization.