Yann LeCun's startup AMI Labs has raised $1.03 billion at a $3.5 billion pre-money valuation, marking one of the largest funding rounds for an AI venture. The move represents a high-stakes bet against the technology most of Silicon Valley is chasing.
LeCun, a Turing Award-winning scientist who spent years leading Meta's AI research, is pursuing what he calls the real path to human-level intelligence: machines that understand the physical world, not just language. The startup Advanced Machine Intelligence seeks to commercialise artificial intelligence systems built around reasoning, planning and "world models."
The funding validates a contrarian position in AI development. LeCun has argued that "we are not going to get to human-level AI just by scaling LLMs," because they simply predict text rather than truly understand the world. He has argued that true intelligence requires understanding physics, causality, and the three-dimensional world; capabilities that can't emerge from predicting the next word in a sentence. "Language is a shadow of reality," LeCun told audiences at a 2024 AI conference.
World models are built on artificial intelligence that learns how the physical world operates by processing continuous, high-dimensional sensor data like images, video, audio and LiDAR, rather than just predicting the next word in a sentence.
Why the Departure Matters
LeCun joined Meta in 2013 to found Facebook AI Research, later known as FAIR, and became one of the company's most prominent AI leaders before departing at the end of 2025. The departure comes at a time of disarray within Meta's AI unit, which was dramatically overhauled after the company's Llama 4 model disappointed developers.
In October, Meta laid off 600 employees from its Superintelligence Labs division, including some who were part of FAIR. Those layoffs and other cuts to FAIR over the years, coupled with a new AI leadership team, played a major role in LeCun's decision to leave.
Meta's departure coincides with Meta's strategic pivot toward more powerful LLM-based models under new chief AI officer Alexandr Wang, the twentysomething founder of Scale AI. The philosophical gap became too wide to bridge.
The Technical Argument
The debate hinges on a fundamental disagreement about how intelligence works. Physical world data is exponentially more complex than text: a single second of video from multiple camera angles contains more information than thousands of pages of text. Training systems to extract causal relationships from that noise has stumped researchers for decades.
AMI Labs CEO Alexandre LeBrun predicted "world models" will become the next buzzword, then joked that in six months every company will call itself a world model to raise funding. "But he thinks AMI Labs is fundamentally different: its goal is to understand the real world."
The startup's first partner will be Nabla, the digital health startup of which LeBrun is now chairman. As CEO of Nabla, LeBrun had reached the same conclusion as LeCun on the limitations of large language models where hallucinations could have life-threatening repercussions.
The broader question is whether genuine AI progress requires a wholesale rethinking of the architecture that has dominated recent breakthroughs. Sam Altman at OpenAI has argued that scaling language models will naturally lead to physical understanding. LeCun's counter-argument is that you can't learn to see by reading descriptions of vision.
The Reality Check
Success is far from guaranteed. AMI Labs starts with fundamental research and is not a typical applied AI startup that can release a product in three months and have revenue in six months. In contrast, it could take years for world models to go from theory to commercial applications.
The spectacular valuation for a pre-revenue startup has amplified concerns about an AI investment bubble. Industry leaders have warned that excitement around AI may be outpacing business fundamentals, and LeCun's fundraising could test whether even the most respected names in the field can command premium valuations without proven commercial traction.
The practical payoff could be substantial if the technical approach works. The company's near-term target customers are organisations operating complex systems, including manufacturers, automakers, aerospace companies, biomedical firms and pharmaceutical groups. LeCun said AMI wants to become the main provider of intelligent systems regardless of the application.
The company plans to establish its headquarters in Paris, with LeCun serving as executive chairman. LeCun noted that Europe has a very high concentration of talent and there is a huge demand from industry and governments for a credible frontier AI company that is neither Chinese nor American, which he thinks will be an advantage.
This funding round does not settle the underlying technical debate. It does, however, put real capital behind the proposition that Silicon Valley's rush to ever-larger language models may be a dead end. Whether that proves prescient or premature will take years to determine.