AMI Labs Secures $1.03 Billion to Pioneer AI That Learns from Reality

In a major bet on the next frontier of artificial intelligence, AMI Labs has raised $1.03 billion in its first funding round. The company, co-founded by Turing Award winner Yann LeCun after his departure from Meta, is pursuing the development of 'world models' - AI systems designed to learn from the physical world, not just text.

CEO Alexandre LeBrun acknowledges the term may soon become fashionable. "In six months, every company will call itself a world model to raise funding," he told TechCrunch. Yet he insists AMI's approach is distinct, aiming to build a foundational understanding of reality. This long-term research effort, based on LeCun's Joint Embedding Predictive Architecture, contrasts with quick-turnaround AI startups. "It could take years to go from theory to commercial applications," LeBrun said.

The substantial funding, which exceeded initial targets, underscores investor confidence. The round was co-led by Cathay Innovation, Greycroft, and Bezos Expeditions, among others, with participation from NVIDIA, Toyota Ventures, and individual backers like Tim Berners-Lee and Eric Schmidt.

AMI's first known partner is digital health company Nabla, where LeBrun serves as chairman, highlighting potential in healthcare. The capital will fuel compute costs and talent acquisition across hubs in Paris, New York, Montreal, and Singapore.

While revenue is not an immediate goal, the company plans early engagement with potential clients to test models with real-world data. True to its academic roots, AMI Labs commits to publishing research and releasing open-source code, betting that openness will accelerate progress in this nascent field.

Source: TechCrunch