Yann LeCun's AMI Labs Raises $1 Billion to Develop Post-LLM AI Architecture

Summary

Yann LeCun, who spent a decade as chief AI scientist at Meta, has founded Advanced Machine Intelligence Labs (AMI Labs) in Paris to pioneer a new approach to artificial intelligence that moves beyond current large language models. The startup has secured over $1 billion in seed funding from major investors including NVIDIA and a fund managing Amazon founder Jeff Bezos' private wealth, making it one of Europe's largest seed-stage fundraising rounds.

The company is developing Joint Embedding Predictive Architecture (JEPA), a fundamentally different approach to AI that creates abstractions of the real world to reason about cause-and-effect relationships. Unlike current LLMs such as ChatGPT, Claude, and Gemini—which excel at text generation and pattern matching but lack genuine physical reasoning—JEPA is designed to tackle complex, real-world problems requiring flexible understanding and foresight.

The technology is particularly targeted at the robotics industry, where current AI models have proven inadequate for training humanoid robots to perform complex household tasks safely and efficiently. LeCun contends that scaling up existing LLMs will never achieve superhuman intelligence or enable robotics breakthroughs, as these systems are fundamentally incapable of reasoning about physical reality and can only generate statistically plausible outputs rather than genuine understanding.

• LeCun argues that current LLMs lack genuine intelligence and cannot be scaled to human-level understanding or solve real-world robotics challenges