Yann LeCun, Meta’s chief artificial intelligence scientist and a Turing Award-winning pioneer of modern AI, is reportedly leaving the company to launch his own start-up. Sources say the move follows growing tensions with Meta CEO Mark Zuckerberg, after LeCun was made to report to new AI chief Alexandr Wang.

CTO at Hyperbolic, Yuchen Jin hit out at Zuckerberg commenting on X: “Yann LeCun is finally leaving Meta. An outcome that is inevitable after Zuck spent $15B to acquire Alexandr Wang and made Yann report to him. OpenAI’s LLMs and ChatGPT have made Zuck panic. Llama 4 flopped, and Yann never believed in LLM-to-AGI. Zuck’s patience ran out.”

LeCun founded Meta’s Fundamental AI Research Lab (FAIR) in 2013, focusing on AI systems capable of learning from real-world data and mimicking human reasoning, rather than concentrating exclusively on text-based large language models (LLMs). His research philosophy has increasingly clashed with Meta’s pivot toward LLMs, a shift driven in part by the rapid market success of OpenAI’s ChatGPT.

The release of Meta’s Llama 4 model, which under-performed compared with rival offerings from Anthropic, Google, and OpenAI, reportedly intensified internal disagreements. LeCun has publicly criticised the reliance on LLMs, describing them as “useful but fundamentally limited” for reasoning, planning, and replicating human-like understanding.

Meanwhile, Zuckerberg has accelerated Meta’s AI ambitions by acquiring a 49 percent stake in Scale AI for $14.3 billion, appointing founder Alexandr Wang to lead the company’s Superintelligence division. This organizational change placed LeCun, who previously reported directly to Zuckerberg, under Wang’s leadership, a shift that sources say made it increasingly difficult for him to pursue FAIR’s long-term research agenda.

LeCun’s upcoming start-up is expected to extend his FAIR research into “world models”, a class of AI systems designed to learn from visual and spatial data rather than text alone. The goal is to create machines capable of understanding and reasoning about the physical world, enabling them to think in ways more akin to humans.

Early fundraising discussions for the venture are reportedly underway, with significant investor interest anticipated given LeCun’s stature in the AI community. Analysts suggest his work on world models could eventually produce systems with advanced reasoning, spatial awareness, and decision-making capabilities -areas where conventional LLMs remain limited.

LeCun’s departure adds to a wave of leadership changes at Meta. Recently, Joelle Pineau, vice-president of AI research, left for Cohere, while Meta also laid off 600 employees from its AI division. At the same time, Shengjia Zhao, co-creator of ChatGPT, joined Meta as chief scientist of the Superintelligence Lab, signalling the company’s renewed focus on high-profile, commercially oriented AI talent.

The leadership shake-up occurs amid growing investor concern. In late October, Meta shares dropped 12.6 percent, wiping out nearly $240 billion in market value, after Zuckerberg disclosed that the company could spend more than $100 billion on AI next year. Analysts attributed the market reaction to concerns that Meta’s AI roadmap lacks coherence, struggling to balance long-term research initiatives with commercially viable AI products.

Related stories

AI boom could ‘break’ global power grid, Buffet, Musk and Zuckerberg warn

Scale AI cuts 14% of workforce following Meta investment and CEO departure