According to Financial Times, Yann LeCun, Meta’s Chief AI Scientist and one of the most influential figures in artificial intelligence, is preparing to leave the company to launch his own startup. The move marks a major shift in Meta’s AI strategy amid internal restructuring and growing competition from OpenAI, Google, and Anthropic.
Sources familiar with the matter told FT that LeCun is already in talks with investors and intends to unveil his new venture within the coming months. The startup will reportedly focus on developing “world models” — advanced AI systems designed to build internal representations of their environment and predict cause-and-effect relationships. Similar research is currently being conducted by Google DeepMind and World Labs.
LeCun, a professor at New York University and recipient of the prestigious Turing Award, has been a central figure in Meta’s AI division. As the founder and head of the Fundamental AI Research (FAIR) lab, he helped establish Meta’s long-term scientific foundation in artificial intelligence.
His planned departure comes during a period of strategic upheaval for Meta. In 2025, CEO Mark Zuckerberg reorganized the company’s AI initiatives, creating Meta Superintelligence Labs (MSL) and appointing Scale AI’s CEO, Alexandr Wang, to lead it following a $14.3 billion investment. Over 50 engineers and researchers were recruited from rival firms, but insiders say the shake-up has deepened internal tensions.
According to FT, new hires have expressed frustration with Meta’s corporate bureaucracy, while the existing generative AI teams have faced restricted access to resources. FAIR’s research, which traditionally focused on long-term AI development, has been sidelined in favor of more product-driven projects. The underwhelming performance of the Llama 4 model family underscored Meta’s lag behind industry leaders, analysts noted.
LeCun has been an outspoken critic of the hype surrounding large language models, arguing that the industry overstates their cognitive capabilities. “We should at least build a system smarter than a house cat before talking about superintelligence,” he once remarked.
Meta has yet to officially comment on reports of LeCun’s departure. If confirmed, his exit would mark a significant loss for the company’s research ambitions as it pivots toward applied AI development.
