The AI Thesis
AI & Robotics Intelligence Brief — For the Corporate World

Five stories that matter in AI and robotics — selected for their long-term strategic consequence, not their Twitter virality. The kind of news that rarely makes the front page of the FT, but probably should.

Yann LeCun — the scientist who won the Nobel Prize equivalent in computing and spent a decade running Meta's AI lab — has left to start a company explicitly built on the thesis that large language models like ChatGPT, Claude and Gemini are a wrong turn. His Paris-based startup, AMI Labs (Advanced Machine Intelligence), closed a $1.03 billion seed round in March 2026, the largest seed investment in European startup history, valuing the company at $3.5 billion before it has shipped a single product.

Investors including Bezos Expeditions, Nvidia, Samsung, Temasek, and Toyota Ventures have backed AMI's core bet: that true machine intelligence requires “world models” — AI that learns by understanding physical reality rather than predicting the next word in a sentence. The architecture, called JEPA (Joint Embedding Predictive Architecture), does not generate text. It learns by watching video and building abstract representations of how the world works, making it inherently better suited to robots, hospitals, and factories where hallucinations could be fatal.

LeCun has been blunt: “Real intelligence does not start in language. It starts in the world.” AMI will focus on industrial process control, wearables, robotics and healthcare, with its first partnership already secured with Nabla, a clinical AI company.

Sources: TechCrunch (10 Mar 2026) · MIT Technology Review (22 Jan 2026) · Crunchbase News (10 Mar 2026) · AMI Labs website