November 12, 2025 Yann LeCun, the Turing Award-winning AI scientist at Meta Platforms who helped build its research arm, is reportedly planning to leave the company in the coming months to launch his own start-up focused on “world models”.

LeCun joined Meta in 2013 and founded its Fundamental AI Research (FAIR) unit, becoming a flagship figure in modern artificial intelligence. His work on convolutional neural networks and energy-based models helped shape how machines process visual and spatial data — a foundation for many current AI systems.

According to reports from the Financial Times and others, LeCun has already begun raising capital for his next venture, which will push beyond large language models (LLMs) towards AI systems that simulate physical and causal understanding of the world.

Meta’s decision to reorganize its AI efforts may have played a role. The company recently acquired a stub-for-investment in Scale AI for $14.3 billion and installed its founder, Alexandr Wang, to lead a new “Superintelligence Labs” division. LeCun’s research unit now reports into this structure — a shift that many see as signalling Meta’s move from long-term foundational work to product-driven AI.

For the wider industry, LeCun’s departure would mark a significant inflection point. As one of the “godfathers of AI”, his exit hints at growing tension between research-first and product-first strategies in giant tech firms. It also opens a path for a new venture with the credibility to attract top talent and capital in the race for next-gen AI.

If Meta loses LeCun, it may underscore how the AI arms race is forcing even the most research-oriented organisations to reassess their priorities. For professionals tracking AI strategy, the message is clear: talent and vision are shifting fast — and the future belongs to those willing to bet on longer-horizon breakthroughs as well as the product sprint.