Yann LeCun has officially launched Advanced Machine Intelligence (AMI Labs), a new AI venture that secured a massive $1.03 billion seed funding round at a $3.5 billion valuation — one of the largest seed rounds ever for a European company.
The Vision: Building AI That Understands Physical Reality
The Paris-based company aims to develop AI models capable of understanding the physical world, directly aligning with LeCun's long-standing theory of Joint Embedding Predictive Architecture (JEPA). Unlike traditional large language models that rely on pattern recognition in text, JEPA focuses on building "world models" that allow AI to understand cause and effect in physical environments — a fundamental shift from LLMs that LeCun has championed for years.
"The next frontier in AI is building systems that can learn, reason, and plan like humans do," LeCun stated at the launch. "We need AI that understands how the world works, not just how to predict the next word."
The Team: Recruiting Top Talent
LeCun personally recruited Xie Saining, a prominent AI researcher, after three persistent attempts. Xie, who joined the venture just two and a half months ago, serves as Co-founder and Chief Scientific Officer. The core leadership team also includes CEO Alex Lebrun and Chief Research Innovation Officer Pascale Fung, a renowned AI and natural language processing expert.
The company has already attracted several other top researchers, signaling strong confidence in its technical direction from the academic community.
The Funding: Europe's Largest Seed Round
The $1.03 billion seed funding represents a significant milestone for the European AI ecosystem. According to industry analysts, this is among the largest seed rounds ever recorded globally, let alone in Europe. The $3.5 billion valuation reflects investor enthusiasm for LeCun's unconventional approach to AI development.
Why JEPA Matters
LeCun has been a vocal critic of current LLM limitations, arguing that next-token prediction alone cannot lead to human-level intelligence. JEPA proposes that AI systems should build internal representations of how the world works, enabling them to reason about physical causality and plan actions accordingly.
This approach could potentially address key challenges in AI safety, reasoning, and generalization that have plagued the industry.
What's Next
AMI Labs plans to hire additional researchers and begin publishing technical papers on its JEPA-based approach in the coming months. The company will operate primarily from Paris, with plans to expand to other research hubs.
The launch marks a pivotal moment in the AI race, as LeCun — one of the field's most influential figures — puts his theoretical vision into practice with unprecedented resources.