Yann LeCun says he is leaving Meta at the end of 2025 to build a new startup and continue his “Advanced Machine Intelligence research”, with Meta as a partner
The announcement follows signs that LeCun’s role inside Meta was changing, including coverage that he would remain FAIR’s chief scientist while reporting to Alexandr Wang. It turns a leadership transition into a distinct company-building effort around his research agenda.
The proposed Meta partnership was material to the initial framing, but later coverage said Meta would not back the startup, leaving the eventual degree of corporate involvement unresolved. Earlier reports had already described his plans to found a startup as taking shape.
First-order effects
LeCun’s departure removes Meta’s chief AI scientist from its internal research organization at the end of 2025 and shifts his Advanced Machine Intelligence work into a new venture.
Meta is initially positioned as a partner rather than the sole institutional home for that work; subsequent reporting makes the scope of that relationship uncertain.
Second-order effects
Meta must preserve continuity for FAIR and its research agenda without LeCun leading it, following the earlier change in his reporting structure.
An independent LeCun-led company gives investors, research hires and prospective partners a separate vehicle for pursuing his research direction, rather than engaging solely through Meta.
Third-order effects
If senior researchers increasingly move from platform-company labs into standalone ventures, frontier AI research may be organized through looser corporate partnerships rather than fully internal labs.
The later dispute over Meta’s backing underscores that affiliation, funding and research access can diverge; the exact structure of such spinouts will matter as much as the talent move itself.
The trend: This is one instance of frontier AI research leaders moving between corporate labs and independent ventures while preserving, or renegotiating, ties to their former employers.
Yann LeCun, a Pioneering A.I. Scientist, Leaves Meta Dr. LeCun's departure follows a shake-up in Meta's artificial intelligence efforts, as Mark Zuckerberg pushes his company to keep up in the tech race. https://www.nytimes.com/... @elitanjourno @nytimes
Yann LeCun confirms that he's leaving to build a new company. “The goal of the startup is to bring about the next big revolution in AI: systems that understand the physical world, have persistent memory, can reason, and can plan complex action sequences.” https://www.linkedin.com…
You find media coverage about Yann LeCun to be dishonest and unscholarly. And yeah, that's how media coverage works. Yann LeCun seems pretty easygoing and lacking in ego if you've ever actually spoken to him
> “have persistent memory” Finally! @ylecun i've been waiting for your lab at meta to release this forever, but as a separate independent startup, even better. i assume it will be open sourced? this is the one ai company i'm most excited about.
It's happening 🚀 @ylecun saw breakthroughs before anyone else: deep learning, self-supervised learning, open science, open review, etc Today, he is challenging the dominant LLMs, and rightly so. Can't wait to see his next chapter with World Models ! [image]
He put Paris on the AI map by creating Meta's FAIR Paris lab, Meta's largest AI research center outside the US. He helped build a vibrant AI ecosystem in France, attracting top talent and making Paris a key hub for AI innovation.
Earthquake in the AI world: Yann LeCun, French AI pioneer and one of the founders of modern AI, is leaving his role at Meta chief AI scientist to launch a new startup, Advanced Machine Intelligence. This is a chance for Europe, as his venture could have offices or operations here…
wow. Finally Yann Lecun just officially posted on his FB, an hour before, that he's leaving Meta and launching a startup. To continue the Advanced Machine Intelligence research program (AMI). And that he will be “sticking around Meta until the end of the year.” [image]
will they do research? will they build “products” and generate revenue? will there be a 2-tier class/caste-system between researchers vs engineers? paper-writers with intellectual output vs pull-request machines? true innovators vs incremental followers? who knows, but stay …
Can anyone please explain what it is anyone means by this? And no not the definition, I mean “what is the thing actually being created here” [embedded post]