Yann LeCun, one of the founding figures of modern deep learning, has secured over $1 billion to prove his long-held thesis that language models alone will never achieve true intelligence.
AMI Labs, the Paris-based startup LeCun cofounded after departing Meta, announced Monday that it has closed a $1.03 billion funding round at a $3.5 billion pre-money valuation. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, with individual backing from Tim Berners-Lee, Jim Breyer, Mark Cuban, Xavier Niel, and Eric Schmidt.
A Different Bet on Intelligence
While the rest of the AI industry races to scale large language models, LeCun has spent years arguing that text prediction is a dead end for achieving general intelligence. AMI Labs is built around his alternative vision: world models that learn from visual, spatial, and physical data to develop an intuitive understanding of how reality works.
The company disclosed plans to develop systems that "understand the real world" — AI that can reason about objects, predict physical outcomes, and plan actions in three-dimensional space. It is a direct challenge to the approach championed by OpenAI, Google DeepMind, and Anthropic.
Leadership and Structure
LeCun serves as executive chairman rather than CEO. The chief executive role belongs to Alex LeBrun, previously cofounder and CEO of Nabla, a health AI startup with offices in Paris and New York. The team draws heavily from LeCun's research network, and the company had originally sought just €500 million before investor demand nearly doubled the round.
Why It Matters
The funding is remarkable for several reasons. At $1.03 billion, it ranks among the largest early-stage AI raises ever, rivaling the initial capitalization of companies like xAI and Mistral. More importantly, it represents a major institutional bet against the prevailing paradigm.
If LeCun is right that language models have fundamental limitations, AMI Labs could pioneer the next generation of AI architecture. If he is wrong, investors will have placed a billion-dollar wager on a contrarian theory championed by one of the field's most decorated researchers.
What Comes Next
AMI Labs has not disclosed specific product timelines, but the company is expected to publish research on its world model architectures in the coming months. The startup is hiring aggressively across its Paris and New York offices, with a focus on computer vision, robotics, and reinforcement learning researchers.
For an industry that has consolidated around transformer-based language models, AMI Labs represents the most well-funded challenge to the status quo in years.
