Yann LeCun Leaves Meta to Launch New AI Startup

After twelve years at Meta, Chief AI Scientist Yann LeCun will step down at year-end. The Turing Award winner announced plans to launch a new startup focused on advancing his research into “Advanced Machine Intelligence” (AMI)—with Meta remaining a key partner.

In a LinkedIn post, LeCun confirmed his departure following a dozen-year tenure—five as founding director of FAIR (Facebook AI Research) and seven as Chief AI Scientist. He described FAIR’s impact on Meta, the broader AI field, and the tech community as “spectacular,” calling its creation his proudest non-technical achievement.

LeCun stated that his new venture will collaborate with colleagues from FAIR, New York University, and other institutions to pursue a long-term AMI research agenda. The goal, he explained, is to drive “the next major revolution in AI” by building systems capable of understanding the physical world, leveraging persistent memory, reasoning, and planning complex actions—moving beyond language-centric models toward perception- and interaction-based architectures.

He emphasized that Meta will remain a partner in this initiative, underscoring the company’s ongoing commitment to FAIR and AMI research. LeCun noted that operating independently would allow the technology to have broader impact across domains beyond Meta’s core business.

Departing after 12 years amid growing tension with Meta’s AI strategy

LeCun’s decision follows months of speculation and mounting friction between his vision for open, foundational AI research and Meta’s increasingly product-driven, commercially oriented AI priorities. While he has championed open-source development, long-term world-model research, and his Joint Embedding Predictive Architecture (JEPA), Meta has shifted focus toward large language models and more closed, product-centric AGI initiatives—a direction LeCun has publicly criticized as a “dead end.”

According to Bloomberg, Meta redirected billions of dollars toward new initiatives and high-profile hires, making it harder for LeCun to secure funding for basic research. His influence further waned with the formation of the “Superintelligence Lab” and the appointment of a new Chief AI Officer, effectively marginalizing him from Meta’s strategic core.

At the same time, Meta imposed stricter controls on research publications, which FAIR researchers say curtailed academic freedom. Insiders revealed that LeCun began considering departure around mid-2025 after organizational changes diminished his team’s clout. Political differences reportedly deepened the rift: while LeCun has been critical of the Trump administration, Meta CEO Mark Zuckerberg has expressed support for certain aspects of its economic policies.

Meta remains involved—but won’t lead the new initiative

Despite internal tensions, LeCun stressed that his next move isn’t a temporary break from Meta. He expressed gratitude to Zuckerberg, Andrew Bosworth, Chris Cox, and Mike Schroepfer for their continued support, noting that their collaboration will help extend AMI’s reach beyond Meta’s proprietary products.

LeCun plans to remain at the company through the end of the year and said he would share more details about his startup’s structure and timeline “when the time is right.”