Jan Lekun, Meta’s chief artificial intelligence (AI) specialist, Turing Award winner and one of the pioneers of the AI field, plans to leave the company in the coming months and create his own startup, the Financial Times reported.
According to sources, Lekun is in talks with investors for a new project. His startup will most likely develop so-called “world models” — AI systems aimed at understanding the dynamics and spatial properties of the real world.
Lekun has led Meta’s Fundamental Artificial Intelligence Research Lab (FAIR) since 2013. He declined to comment, and Meta did not respond to a request for comment from the FT.
The Financial Times writes that Mark Zuckerberg has recently shifted the company’s focus from long-term research to AI models and products that produce quick results. In the summer of 2025, he appointed Alexander Wang, co-founder of the Scale AI startup, as head of the “superintelligence” team — as a result, Lekun became subordinate to him.
These changes, as the FT notes, came after the unsuccessful launch of the Llama 4 model and the failure of the Meta AI chatbot to gain the expected popularity.
Lekun has repeatedly said that large language models are “useful,” but they do not have the ability to think and reason like humans. He intends to continue working in this direction in the future — on AI systems capable of deep modeling of the real world.
