Dr. Yann LeCun, the Turing Award-winning computer scientist who has served as Meta’s Chief AI Scientist for over a decade, is planning to leave the company to launch his own artificial intelligence startup.
The departure marks a seismic shift in the AI landscape, signaling the exit of one of Meta’s most foundational and celebrated researchers. LeCun, widely regarded as one of the three “Godfathers of AI,” has reportedly informed associates of his plans to depart in the coming months. His new venture, for which he is said to be in early discussions to raise capital, will reportedly focus on “World Models,” a research path he has long championed as the true key to artificial general intelligence (AGI).
LeCun’s new venture aims to build AI systems that can learn and reason about the world in the same way humans and animals do. This “World Models” approach, which LeCun has publicly advocated for years, stands in stark contrast to the industry’s current focus on Large Language Models (LLMs) like OpenAI’s GPT series and Google’s Gemini.
In LeCun’s vision, true intelligence requires an AI to have an internal, predictive model of its environment. These models would allow an AI to understand cause and effect, grasp “common sense” physics, and plan complex actions by simulating potential outcomes abilities that current LLMs lack, as they are primarily trained on text. LeCun has often compared this approach to how an infant learns, by observing and interacting with the world rather than just processing static data.
The planned exit comes amid significant internal restructuring at Meta and reports of growing friction between LeCun’s research-first ethos and the company’s product-driven direction.
Sources familiar with the matter point to several factors contributing to the decision. Earlier this year, Meta CEO Mark Zuckerberg announced a major reorganization of the company’s AI efforts, consolidating initiatives under a new “Superintelligence Labs.” Crucially, this move included hiring Alexandr Wang, the former CEO of the data-labeling startup Scale AI, to lead the new division.
This restructuring reportedly placed LeCun, who previously reported to Chief Product Officer Chris Cox, under Wang’s leadership. This change, coupled with recent layoffs that have impacted LeCun’s foundational Facebook Artificial Intelligence Research (FAIR) unit, is said to have diminished the group’s influence and autonomy.
Reports suggest LeCun and other senior researchers had become frustrated with new internal publication rules requiring more extensive review before research could be published. This was seen by some as a threat to the academic freedom and open-source model that LeCun himself had established at FAIR, making it a globally respected research lab.
LeCun, 65, has never been shy about his skepticism of the LLM-centric path to AGI. While competitors and, increasingly, Meta itself poured billions into scaling text-based models, LeCun often publicly described them as a “dead end” for achieving true reasoning. His departure to found a startup dedicated to his alternative vision is a clear statement of his conviction.
The loss of LeCun is a significant blow to Meta. Recruited by Zuckerberg personally in 2013, LeCun founded FAIR and built it into an academic powerhouse, attracting top-tier talent and producing breakthrough research. He is the inventor of the Convolutional Neural Network (CNN), a technology that revolutionized computer vision and underpins countless modern AI applications, from facial recognition to medical imaging.
His work, alongside Geoffrey Hinton and Yoshua Bengio, earned him the 2018 ACM A.M. Turing Award, the highest honor in computer science.
With Hinton leaving Google in 2023 to warn about AI’s dangers and LeCun now leaving Meta to pursue a different path to AGI, two of the field’s three “Godfathers” have now exited the major corporate labs they helped build.
LeCun’s new company, though still in its infancy, will immediately become one of the most closely watched ventures in Silicon Valley. It represents a high-profile, deeply principled challenge to the prevailing AI orthodoxy, led by one of the few figures in the world with the credibility and expertise to pull it off. The move signals a potential splintering of the AI elite, away from monolithic corporate labs and toward founder-led ventures driven by singular, competing visions for the future of intelligence.
