Yann LeCun's AMI Labs: $1B Bet on World Models vs. LLMs

AMI Labs, a startup led by Turing Award winner Yann LeCun, is building "world models" with over a billion dollars in funding. This company is developing AI that learns how physical reality behaves, rather than just predicting the next word in a sequence. AMI Labs raised $1.03 billion in a seed round, which is reportedly the largest in European startup history. Its goal is to create AI systems that understand the real world, hold persistent memory, and can reason and plan. This approach is a direct challenge to the large language models currently dominating the field. The technical foundation for this work is called Joint Embedding Predictive Architecture, or JEPA. It focuses on predicting in abstract space, rather than trying to recreate every detail. This matters because it represents a significant, well-funded effort to develop a new path for artificial intelligence beyond current large language models.