Yann LeCun is preparing to leave Meta after twelve years, but his next move is not a clean break from the company. The Turing Award winner plans to launch a startup built around his long-running "Advanced Machine Intelligence" (AMI) research agenda, while Meta remains involved as a partner.
The shift matters because it separates one of Meta’s most visible AI research figures from the company’s internal AI strategy at a moment when priorities around open research, language models, and product-driven AGI work have been under pressure.
What LeCun is building next
In a LinkedIn post, LeCun confirmed that he will step down at the end of the year. He said the new company will continue work connected to AMI, a research direction he has pursued with colleagues at FAIR, NYU, and beyond.
The aim is not simply to make larger language systems. According to LeCun’s description, the project is focused on developing AI that can understand the physical world, use persistent memory, reason, and plan complex actions.
That framing points to a broader ambition: moving AI beyond systems that learn mainly from language. The startup’s research agenda is instead centered on models grounded in perception and interaction, with capabilities that are meant to support more flexible behavior.
LeCun described the goal as driving "the next big revolution in AI." He also said more information about the startup’s structure and timeline will come "when the time comes."
A twelve-year chapter at Meta
LeCun’s departure closes a long period inside Meta’s AI organization. He spent five years as founding director of FAIR and seven years as Chief AI Scientist.
He characterized FAIR’s influence on Meta, the AI field, and the broader tech community as "spectacular." He also described creating FAIR as his proudest non-technical achievement.
That history is central to why the move is notable. FAIR has been associated with open research and long-term AI work, while Meta’s broader AI efforts have increasingly been shaped by commercial priorities and large-scale systems.
LeCun’s public research interests have included world models and his "Joint Embedding Predictive Architecture" (JEPA). The source article describes a visible divide between that open-research philosophy and Meta’s shift toward large-scale language models and more closed, product-driven AGI efforts.
Why the split reflects broader tension
The departure follows months of speculation about LeCun’s position inside Meta. 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 reportedly declined after the creation of the "Superintelligence Lab" and the appointment of a new "Chief AI Officer." Those changes moved him further away from Meta’s strategic core.
Another source of tension was Meta’s tighter control over research publications. FAIR researchers viewed those restrictions as limiting academic freedom, according to insiders cited in the source article.
The friction was not only technical or organizational. The source also says political differences reportedly deepened the divide: LeCun criticized the Trump administration, while Meta CEO Mark Zuckerberg supported parts of its economic policy.
None of that means the new company is positioned as an opponent to Meta. LeCun’s own framing is more complicated: the startup gives AMI a more independent path while preserving a relationship with Meta.
Meta stays close, but not in control
LeCun emphasized that Meta will be a partner in the venture. He thanked Zuckerberg, Andrew Bosworth, Chris Cox, and Mike Schroepfer for their continued support.
The logic is that independent development could let AMI reach areas beyond Meta’s own products. LeCun wrote that pursuing the work outside Meta would allow the technology to have broader impact across multiple sectors, including areas outside Meta’s core business.
That arrangement gives the move two meanings at once. On one hand, Meta loses its Chief AI Scientist at the end of the year. On the other, it maintains a connection to the AMI research program he wants to advance.
For AI research, the central question is whether an independent startup can give LeCun’s agenda more room than it had inside Meta’s changing structure. The answer will depend on details he has not yet shared, including how the company will be organized and when it will begin operating publicly.
What to watch next
The immediate timeline is clear: LeCun plans to stay at Meta through the end of the year. After that, the focus shifts to how his startup defines AMI in practice and how closely Meta’s partnership shapes the work.
The most important signals will be practical ones:
- How the startup describes its research program beyond language-based learning.
- Whether AMI work remains tied to FAIR, NYU, and other collaborators.
- How Meta participates without steering the project.
- When LeCun shares more details about the company’s structure and timeline.
For now, the move is best understood as a strategic relocation of a research agenda rather than a simple exit. LeCun is leaving Meta’s formal AI leadership, but the AMI project he wants to build will still have Meta beside it.