How Yann LeCun’s new startup turned an AI rebellion into a $1 billion wager

Advanced Machine Intelligence is not just another AI startup with a famous founder. It is a bet that the current large-language-model boom, for all its dazzle, is still missing something fundamental: a real understanding of how the world works.

Advanced Machine Intelligence, or AMI, said on March 10 that it had raised $1.03 billion at a $3.5 billion pre-money valuation, an enormous opening salvo for a company built by former Meta AI chief Yann LeCun. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, giving the startup one of the bigger debut financings in the current AI cycle.

What makes the company more interesting than the size of the check is the argument behind it. LeCun told Reuters that AMI is trying to build systems centered on reasoning, planning, and “world models,” rather than relying mainly on the next-word prediction methods that power today’s dominant large language models. In plain English, the pitch is that today’s AI may be fluent without being especially grounded, and that a more capable generation of systems will need a deeper internal map of reality.

That makes AMI a startup founded not just on technology, but on dissent. LeCun has spent years as one of the most prominent skeptics of the idea that scaling up text-prediction systems alone will produce broadly capable intelligence. His departure from Meta at the end of 2025 turned that critique into a company. Reuters reported that he joined Meta in 2013 to found Facebook AI Research, later FAIR, before leaving after Meta reorganized its AI push under Meta Superintelligence Labs, a division led by former Scale AI chief Alexandr Wang.

AMI’s near-term target customers also hint at a different strategy from the chatbot arms race. LeCun said the startup is going after organizations that run complex systems, including manufacturers, automakers, aerospace companies, biomedical firms, and pharmaceutical businesses. Over time, he said, the same technology could support consumer products such as domestic robots and possibly even applications inside Ray-Ban Meta smart glasses. That is a wider and more industrial vision than the usual AI startup promise of better content generation or cheaper software coding.

The timing is not accidental. AI has become the gravitational center of venture capital itself. Reuters reported in January that AI startups captured a record 46.4% of the $209 billion raised in U.S. venture funding in 2024, up from less than 10% a decade earlier. A separate Reuters report in January 2026 said almost two-thirds of global venture capital funding in 2025 went to AI companies, with annual AI investment surging from just over $70 billion to more than $210 billion in recent years. In that climate, a founder with LeCun’s pedigree does not merely launch a company. He rings a very expensive dinner bell.

Still, AMI’s rise also exposes the strange mood of the current startup market. Investors are pouring ever larger sums into companies that promise to rethink AI from the ground up, even as many of those companies remain years away from proving durable commercial models. Reuters noted that analysts have warned the AI funding boom may not be endlessly sustainable, particularly for foundation-model firms that demand huge amounts of compute and elite talent before the revenue arrives. That does not mean AMI is smoke. It means the market is still willing to finance very ambitious ideas long before it can fully price their odds.

That is what makes AMI such a revealing startup story. It is not just a founder leaving a tech giant to cash in on his reputation. It is a prominent AI dissenter trying to prove that the next big breakthrough will not come from making today’s models slightly larger or slightly chattier. It will come from building systems that can actually model the world they are supposed to operate in. Investors just handed him more than a billion dollars to try. The rest of the industry now gets to find out whether this is a rebellion with teeth, or just a beautifully funded counterargument.