AMI Labs Raises $1.03B in Seed Funding to Build AI World Model Systems

Artificial intelligence has spent the past few years talking a big game. Generating text. Drafting emails. Painting pictures. Enter AMI Labs, walking into the arena with $1.03B in seed funding and a $3.5B valuation, aiming for something far more ambitious. When Yann LeCun, Alexandre LeBrun, Saining Xie, Pascale Fung, Michael Rabbat, and Laurent Solly decide to build together, it signals a shift from clever software tricks to systems that actually understand the world they operate in. That is not a small upgrade. That is a different category.

Credit where it is due. Congratulations to CEO Alexandre LeBrun and the entire founding crew for pulling together a coalition that reads like a who’s who of capital that knows how to smell gravity before everyone else notices the pull. Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions stepped up to lead the round, with strategic muscle from NVIDIA, Temasek, Toyota Ventures, Sea, SBVA, Association Familiale Mulliez, Groupe Industriel Marcel Dassault, and Alpha Intelligence Capital. Sprinkle in individual backers like Jeff Bezos, Eric Schmidt, and Mark Cuban and suddenly the table feels less like a pitch meeting and more like a council of people betting on the next computing era.

Now the real intrigue. AMI Labs is chasing something deeper than another chatbot that writes emails and poetry about your dog. The mission is world models. Systems that understand situations, remember context, anticipate consequences, and plan actions inside the messy physics of the real world. Healthcare. Robotics. Industrial automation. Manufacturing. Places where the cost of being wrong is not a typo but a broken process, a stalled machine, or a real human outcome. That kind of intelligence does not come from predicting the next word. It comes from understanding the world the words live in.

The early signal is already on the board. Nabla announced an exclusive partnership that gives them first access to AMI’s world model technology to push agentic healthcare AI toward something closer to regulatory grade reliability. If you want to understand why investors opened the vault for this round, start there. Real world environments demand AI that can think in sequences, simulate outcomes, and operate with guardrails strong enough for regulated industries.

This is what happens when frontier research meets serious capital and a team that has already spent years inside the machine learning trenches. Paris, New York, Montreal, and Singapore now sit on the same chessboard as AMI Labs builds infrastructure, hires world class researchers, and chases the kind of AI that does more than talk about reality. The kind that actually understands it, and the industries watching closely already know the stakes just changed.