AMI Labs raises $1.03 billion to build world models

AMI Labs, a new venture co-founded by Yann LeCun, has raised $1.03 billion in funding at a $3.5 billion pre-money valuation. The company is focused on developing “world models”—a category of artificial intelligence designed to learn from reality rather than relying solely on language. This approach aims to move beyond the constraints of traditional large language models (LLMs).

The startup’s CEO, Alexandre LeBrun, argues that the transition to world models is critical, particularly in sectors like healthcare where LLM hallucinations could have life-threatening repercussions. Nabla, a digital health startup where LeBrun serves as chairman, is slated to be AMI Labs’ first disclosed partner. However, LeBrun noted that developing a viable alternative based on Joint Embedding Predictive Architecture (JEPA)—an AI architecture proposed by LeCun in 2022—will take time. “AMI Labs is a very ambitious project, because it starts with fundamental research,” LeBrun said, adding that it could take years for world models to transition from theoretical research to commercial applications. Unlike typical applied AI startups, AMI Labs is not expected to generate $10 million in annual recurring revenue within its first year.

Despite the long commercial timeline, investors are committing massive capital to the sector. European peer startup SpAItial recently raised a $13 million seed round, while World Labs secured $1 billion last month. AMI Labs’ own funding round significantly exceeded initial expectations. Last December, the company was reportedly seeking €500 million, but it ended up raising what is likely €890 million. The investment round was co-led by:

• Cathay Innovation
• Greycroft
• Hiro Capital
• HV Capital
• Bezos Expeditions

The capital will fund compute resources and talent acquisition. While AMI Labs has no immediate plans to generate revenue, it intends to deploy its models early with partners like Nabla to test them in real-world situations. True to LeCun’s research philosophy, the company plans to publish papers and open-source its code. It aims to build a collaborative research ecosystem across its hubs in Paris, New York, Montreal, and Singapore, allowing it to recruit talent and establish a presence near future clients in Asia.

Why it matters

AMI Labs is betting that “world models”—AI that learns from reality—will overcome the limitations of current LLMs, attracting massive capital despite the long timeline for commercialization.