Yann LeCun launches AMI Labs in Paris with a $1B bet against LLMs
Meta's former chief AI scientist has set up his new lab in Paris. AMI Labs raised $1.03 billion in March on a contrarian bet that "world models", not larger language models, are the path to general intelligence.
Yann LeCun, who served as Meta’s chief AI scientist for more than a decade, has launched a new Paris-headquartered AI lab called Advanced Machine Intelligence (AMI). The company is run by CEO Alex LeBrun, with LeCun as Executive Chair and Saining Xie (NYU, ex-Google DeepMind) as Chief Scientist.
In March 2026, AMI announced a $1.03 billion funding round at a $3.5 billion pre-money valuation, co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions.
The bet is technical and contrarian. AMI is building world models rather than scaling large language models — systems that learn an abstract representation of the physical world and make predictions in that representation, trained on video, audio, and sensor data rather than just text. The core architecture, JEPA (Joint Embedding Predictive Architecture), comes out of LeCun’s research at Meta. Initial application targets include industrial process monitoring, robotics, autonomous vehicles, and smart assistants.
LeCun departed Meta in late 2025 after the restructuring of the company’s AI unit around the poorly-received Llama 4 release. AMI plans satellite offices in North America and Asia, but the headquarters and the talent center are in Paris — placing the lab alongside Mistral AI as the second European frontier lab in the city.