AMI Labs, the new venture co-founded by Turing Award winner Yann LeCun after he left Meta, has raised $1.03 billion at a $3.5 billion pre-money valuation. AMI is working on world models, which are AI systems that learn from reality, not just from language.
This category currently has fewer players than generative AI, but that may change soon. AMI Labs CEO Alexandre LeBrun predicts that “world models” will become the next buzzword. He remarked that in six months, every company might call itself a world model company to raise funding.
LeBrun made this comment with a smile because he believes AMI Labs is fundamentally different. Its core goal is to build AI that understands the real world. This technology could have significant applications in fields like healthcare. AMI Labs’ first partner will be Nabla, a digital health startup where LeBrun now serves as chairman.
During his time as CEO of Nabla, LeBrun reached the same conclusion as LeCun about the limitations of large language models, where hallucinations could have life-threatening consequences. However, he acknowledges it will take time for the startup to offer a viable alternative. This alternative is based on JEPA, the Joint Embedding Predictive Architecture proposed by LeCun in 2022.
LeBrun describes AMI Labs as a very ambitious project because it begins with fundamental research. It is not a typical applied AI startup that can release a product in three months. In contrast, it could take years for world models to progress from theory to commercial applications.
Despite this long time horizon, companies developing world models are attracting substantial investment. SpAItial raised a $13 million seed round, which is unusually large for a European startup. Meanwhile, Fei-Fei Li’s World Labs secured $1 billion last month alone. Now, AMI Labs joins this club with more funding than was initially rumored.
The French AI lab was reportedly seeking just €500 million last December but ended up raising approximately €890 million. This success is likely due to its strong team. In addition to LeCun’s involvement as chairman and LeBrun’s entrepreneurial track record, the company boasts Meta’s VP for Europe Laurent Solly as COO, along with high-profile researchers Saining Xie as chief science officer, Pascale Fung as chief research and innovation officer, and Michael Rabbat as VP of world models.
According to LeBrun, high investor interest allowed the startup to be selective, choosing investors aligned with its expectations and background. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions. Several other funds and industry-tied backers participated, as did individuals including Tim and Rosemary Berners-Lee, Jim Breyer, Mark Cuban, Mark Leslie, Xavier Niel, and Eric Schmidt.
Beyond the strategic value, this funding provides AMI Labs with meaningful runway to finance its two main cost centers: compute power and talent. LeBrun stated he will prioritize quality over quantity while building the team across four key locations: its headquarters in Paris, New York where LeCun teaches, Montreal where Rabbat is based, and Singapore to recruit AI talent and be close to future clients in Asia.
Although AMI Labs has no immediate plans to generate revenue, it still intends to engage with prospective customers early. LeBrun explained that developing world models requires real-world interaction. The models cannot be built locked in a lab; they need to be placed in real-world situations with real data for proper evaluation.
When the time is right, AMI Labs will turn to partners to explore deployments. Nabla is the first disclosed partner expecting to access these early models, but it will not be the last. LeBrun noted that this potential explains the strong interest from certain industrial players and potential partners in the investment round.
In addition to its lead investors and angels, AMI Labs is backed by NVIDIA, Samsung, Sea, Temasek, and Toyota Ventures, as well as French players 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.
These investments may take a while to materialize into commercial applications. Staying true to LeCun’s beliefs, AMI Labs will publish papers throughout its research process. The company also plans to release a lot of code as open source. LeBrun, who also worked at Meta’s FAIR research lab, noted that open research is increasingly rare. However, the startup’s founders still believe in it, arguing that progress happens faster when work is open and that it is in their best interest to build a community and research ecosystem around them.