After twelve years at Meta, Chief AI Scientist Yann LeCun is stepping down at the end of the year. The Turing Award winner says he plans to launch a new startup to continue his "Advanced Machine Intelligence" (AMI) research - and Meta will remain a partner.
In a LinkedIn post, LeCun confirmed his departure after twelve years at the company - five as founding director of FAIR and seven as Chief AI Scientist. He called the impact of FAIR on Meta, the AI field, and the broader tech community "spectacular," describing its creation as his proudest non-technical achievement.
LeCun said his new company will build on his long-term AMI research agenda with colleagues at FAIR, NYU, and beyond. The goal, he explained, is to drive "the next big revolution in AI" by developing systems that understand the physical world, use persistent memory, can reason, and plan complex actions. The project aims to move AI beyond purely language-based learning toward models grounded in perception and interaction.
He also emphasized that Meta will be a partner in the new venture, reflecting the company’s ongoing support for FAIR and AMI research. Pursuing the work independently, LeCun wrote, would allow the technology to have a broader impact across multiple sectors, including areas outside Meta’s core business.
A departure after twelve years and growing tension with Meta’s AI strategy
LeCun’s decision follows months of speculation and an increasingly visible divide between his open-research philosophy and Meta’s commercial AI priorities. While he has championed open source development and long-term research into world models and his "Joint Embedding Predictive Architecture" (JEPA), Meta has shifted its focus toward large-scale language models and more closed, product-driven AGI efforts - a direction LeCun has called a dead end.
Reports from Bloomberg suggest he struggled to secure funding for fundamental research as Meta redirected billions toward new projects and high-profile hires. His influence also waned after the creation of the "Superintelligence Lab" and the appointment of a new "Chief AI Officer", further moving him out of Meta’s strategic core.
At the same time, Meta imposed tighter controls on research publications, which FAIR researchers viewed as limiting academic freedom. According to insiders, LeCun had already considered leaving in mid-2025 after structural changes reduced his team’s influence. Political differences reportedly deepened the rift further - while LeCun criticized the Trump administration, Meta CEO Mark Zuckerberg supported parts of its economic policy.
Meta remains involved but without steering the project
Despite internal tensions, LeCun stressed that his next venture isn’t a break from Meta. He thanked Zuckerberg, Andrew Bosworth, Chris Cox, and Mike Schroepfer for their continued support, saying their partnership will help extend AMI’s reach beyond Meta’s own products.
LeCun plans to stay at the company through the end of the year. He said he will share more details about the startup’s structure and timeline "when the time comes."
