AMI Labs is a contrarian bet against the prevailing large-language-model paradigm. Co-founded by Yann LeCun after he left Meta, the lab pursues 'world models'—AI systems that learn abstract representations of real-world sensor data and make predictions in representation space, the way humans build mental models to reason and plan.
The company argues that generative, token-based models work best for discrete, low-dimensional tasks but do not genuinely understand the physical world. Action-conditioned world models, by contrast, let agentic systems predict the consequences of their actions and plan action sequences subject to safety guardrails, which AMI sees as essential for factories, hospitals, robots and wearables.
AMI is led by CEO Alexandre LeBrun, with LeCun as chair and a research bench including Chief Science Officer Saining Xie, Chief Research & Innovation Officer Pascale Fung, and VP of World Models Michael Rabbat. The company is global from inception, operating across Paris, New York, Montreal and Singapore, and names digital-health startup Nabla as its first partner.
In March 2026 AMI announced a roughly $1.03 billion seed round at a $3.5 billion pre-money valuation—the largest seed round ever for a European startup. Investors include Bezos Expeditions, Cathay Innovation, Greycroft, Hiro Capital and HV Capital. The funding is meant to bankroll long-horizon research into world models before commercial applications mature.