AMI Labs secures $1.03B as Yann LeCun pushes forward with AI world model research

AMI Labs, the new company co-founded by Turing Award winner Yann LeCun after he departed from Meta, has raised $1.03 billion at a $3.5 billion pre-money valuation. The company is focused on world models, a form of AI designed to learn from reality rather than only from language.

For now, this area has fewer companies working in it than generative AI, though that may not remain the case for long. “My prediction is that ‘world models’ will be the next buzzword,” AMI Labs CEO Alexandre LeBrun said. “In six months, every company will call itself a world model to raise funding.”

LeBrun remarked with a smile, believing AMI Labs is fundamentally different. Its objective, he said, is to understand the real world. That could eventually lead to uses in areas such as healthcare, where the company’s first partner will be Nabla, the digital health startup where LeBrun is now chairman.

As chief executive of Nabla, LeBrun said he had reached the same conclusion as LeCun about the limitations of large language models in settings where hallucinations could have life-threatening consequences. At the same time, he knows it will take time before the startup can offer a practical alternative based on JEPA (Joint Embedding Predictive Architecture), the approach proposed by LeCun in 2022.

“AMI Labs is a very ambitious project, because it starts with fundamental research. It’s not your typical applied AI startup that can release a product in three months, have revenue in six months, and make $10 million in [annual recurring revenue] in 12 months,” LeBrun said. By contrast, he said, it may take years for world models to move from theory into real commercial applications.

Even with that longer time frame, startups building world models have been attracting major funding. SpAItial raised a $13 million seed round, unusually large for a European startup, while Fei-Fei Li’s World Labs secured $1 billion last month alone. Now AMI Labs joins that group with more money than was originally rumoured.

The French AI lab had reportedly aimed to raise only €500 million last December but ended up raising around €890 million, likely helped by the strength of its team. In addition to LeCun’s role as chairman and LeBrun’s record as an entrepreneur, the startup also includes Meta’s vice president for Europe, Laurent Soll,y as COO, along with prominent researchers Saining Xie as chief science officer, Pascale Fung as chief research and innovation officer, and Michael Rabbat as vice president of world models.

According to LeBrun, the strong investor interest allowed the company to be selective about backers, both in terms of shared expectations and background. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, with participation from several other funds, strategic investors, and individuals, including Tim and Rosemary Berners-Lee, Jim Breyer, Mark Cuban, Mark Leslie, Xavier Niel, and Eric Schmidt.

Beyond any strategic value, the new funding gives AMI Labs substantial runway to support its two main areas of spending: compute and talent. LeBrun said he plans to prioritise quality over quantity as the company builds out its team across four main locations: Paris, where the company is headquartered; New York, where LeCun teaches at NYU; Montreal, where Rabbat is based; and Singapore, both for access to AI talent and to stay close to future customers in Asia.

Although AMI Labs does not plan to generate revenue for the moment, it still intends to work with potential customers early in the process. “We are developing world models that seek to understand the world, and you can’t do that locked up in a lab. At some point, we need to put the model in a real-world situation with real data and real evaluations,” LeBrun said.

When that stage arrives, AMI Labs plans to work through partners to explore deployments. Nabla is the first publicly identified partner expected to gain access to the company’s early models, but LeBrun said it will not be the last. “This may explain the presence and strong interest of certain industrial players and potential partners in the investment round,” he said.

Along with its lead investors and angel backers, AMI Labs also counts Nvidia, Samsung, Sea, Temasek, and Toyota Ventures among its supporters, as well as the French groups Association Familiale Mulliez, Groupe Industriel Marcel Dassault, and Publicis Groupe. Aglaé Lab, Alpha Intelligence Capital, Artémis, Bpifrance Digital Venture, New Legacy Ventures, SBVA, and ZEBOX Ventures also participated.

Those investments may take time to translate into commercial uses. But in keeping with LeCun’s long-held views, AMI Labs said it plans to publish research papers as it progresses.

“We will also make a lot of code open source,” LeBrun said, noting that he had also previously worked at Meta’s AI research lab, FAIR. While open research has become “increasingly rare,” he said the founders of this startup still believe in it. “We think things move faster when they’re open, and it’s in our best interest to build a community and a research ecosystem around us.”