Turing Prize winner Yann LeCun, who previously led AI research at Meta, is writing a new chapter in the AI world through his startup AMI Labs. The company has just raised $1.03 billion in funding at a pre-money valuation of $3.5 billion. This amount far exceeds earlier reports initially, the company was said to be raising just €500 million, but overwhelming investor interest nearly doubled that figure to around €890 million.
World Models: The Next Frontier in AI
AMI Labs is building what are known as "world models." Unlike the large language models (LLMs) powering tools like ChatGPT, world models don't just process text they aim to understand reality itself, much like humans learn from the environment around them. The approach is rooted in JEPA, the Joint Embedding Predictive Architecture that LeCun proposed back in 2022. CEO Alexandre LeBrun believes "world models" will soon become the next big buzzword in AI, with every startup slapping the label on itself to attract funding. But he insists AMI Labs is fundamentally different because its actual goal is to comprehend the real world, not just play with language.
A Star-Studded Team
AMI Labs' greatest strength is arguably its team. Yann LeCun serves as chairman. Alexandre LeBrun, who previously built and sold startups like Wit.ai to Facebook, is the CEO. Meta's former VP for Europe, Laurent Solly, has come on board as COO. The research bench is equally impressive: Saining Xie is chief science officer, Pascale Fung is chief research and innovation officer, and Michael Rabbat serves as VP of world models. These are names that carry serious weight in AI research, and their presence played a major role in building investor confidence.
A Who's Who of Investors
The funding round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Jeff Bezos's Bezos Expeditions. Major industrial players like NVIDIA, Samsung, Toyota Ventures, Sea, and Temasek also invested. The list of notable individual backers reads like a tech hall of fame: Mark Cuban, Eric Schmidt, Xavier Niel, Jim Breyer, and even Tim and Rosemary Berners-Lee. French institutional investors including Association Familiale Mulliez, Groupe Industriel Marcel Dassault, Publicis Groupe, and Bpifrance Digital Venture also participated. LeBrun says the high level of interest gave the startup the luxury of choosing investors not just for their money, but for alignment in expectations and relevant backgrounds.
Healthcare: The First Real-World Application
AMI Labs' first disclosed partner is Nabla, a digital health startup where LeBrun himself serves as chairman. While leading Nabla, LeBrun experienced firsthand how dangerous LLM hallucinations can be in healthcare if an AI gives incorrect medical advice, the consequences could be life-threatening. World models, which aim for a deeper understanding of reality, could offer a safer and more reliable path for medical AI. Nabla will get early access to AMI Labs' models, and more partners are expected to follow as the technology matures.
A Long Road Ahead, But a Clear Vision
LeBrun has been refreshingly honest about timelines. AMI Labs is not a typical applied AI startup that can ship a product in three months and hit $10 million in annual recurring revenue within a year. Building world models that move from fundamental research to commercial applications could take years. The company has no plans to generate revenue for now, but it does intend to engage with prospective customers early so that its models can be tested with real-world data and real evaluations. This isn't about rushing to market it's about getting the science right first.
A Commitment to Open Research
In an era when major AI companies are increasingly keeping their research behind closed doors, AMI Labs has made a bold promise: it will publish its research papers and open-source much of its code. LeBrun, who also worked at Meta's AI research lab FAIR, firmly believes that openness accelerates progress. Building a community and research ecosystem around the company, he argues, is not just idealistic it's strategically smart. Things move faster when they're open, and AMI Labs wants to be at the center of that movement.
Four Cities, One Mission
The startup will operate across four cities: Paris, which serves as headquarters; New York, where LeCun teaches at NYU; Montreal, where Michael Rabbat is based; and Singapore, chosen both for its deep pool of AI talent and its proximity to future clients in Asia. LeBrun says he will prioritize quality over quantity when it comes to hiring, building a lean but world-class team across these locations.
AMI Labs' story is just beginning, but with over a billion dollars in funding, a team packed with AI luminaries, and a bold vision for understanding reality itself, this startup has firmly positioned itself as a serious contender in the emerging world models space. Whether it can deliver on that promise remains to be seen but few would bet against this team.
