Advanced Machine Intelligence (AMI) Labs, a new AI startup founded by Turing Award winner Yann LeCun after his departure from Meta, has raised €890 million ($1.03 billion) at a $3.5 billion pre-money valuation.
LeCun, who is French, said it was “one of the largest seeds ever, probably the largest for a European company,” on a LinkedIn post. He serves as executive chairman.
The news was lauded by French President Emmanuel Macron on LinkedIn: “Yann LeCun is opening a new chapter in artificial intelligence. He has completed one of the largest fundraising rounds – €1 billion with AMI Labs – to revolutionize AI. This is the France of researchers, builders and bold innovators. Congratulations!”
Leading the round are Cathay Innovation, Greycroft, Jeff Bezo’s personal fund Bezos Expeditions, Hiro Capital and HV Capital. Other investors include worldwide web creator Tim Berners-Lee and his wife, Rosemary, Jim Breyer, Mark Cuban and Eric Schmidt. Additional backers include Nvidia, Samsung, Toyota Ventures, Publicis Groupe and Temasek.
AMI Labs CEO Alexandre LeBrun told TechCrunch that they are different from other fast-moving applied AI startups that release a product in three months, book revenue in six months and make $10 million in annual recurring revenue in a year. Instead, AMI could take years to launch a product commercially.
The startup seeks to pursue a new class of AI known as “world models,” systems that understand the real world rather than relying primarily on text data, a limitation of many large language models today.
Other AI startups focusing on world models include World Labs founded by AI pioneer Fei-Fei Li, which raised $1 billion in February for a total of $1.23 billion since 2024.
AMI’s research focuses on AI that learns from real-world sensor data – such as cameras and other inputs – which is continuous, high-dimensional and often unpredictable. Instead of generating outputs directly from raw data, the company is developing models that learn abstract representations of the physical world and make predictions.
AMI said the technology could eventually support applications in robotics, automation, health care and industrial systems. The Paris-based research lab plans to publish research and open-source code as it builds teams in Paris, New York, Montreal and Singapore.