Yann LeCun’s new enterprise, AMI Labs, has drawn intense consideration because the AI scientist left Meta to discovered it. This week, the startup lastly confirmed what it’s constructing — and several other key particulars have been hiding in plain sight.
On its newly launched web site, the startup disclosed its plans to develop “world models” with the intention to “build intelligent systems that understand the real world.” The deal with world fashions was already hinted at by AMI’s identify, which stands for Advanced Machine Intelligence, nevertheless it has now formally joined the ranks of the most popular AI analysis startups.
Building foundational fashions that bridge AI and the actual world has grow to be one of many area’s most enjoyable pursuits, attracting prime scientists and deep-pocketed traders alike — product or no product.
World Labs, a direct rival based by AI pioneer Fei-Fei Li, turned a unicorn shortly after popping out of stealth. After launching its first product, Marble, which generates bodily sound 3D worlds, World Labs is now reportedly in talks to lift contemporary funding at a valuation of $5 billion.
There’s little doubt that VCs can be equally desirous to spend money on LeCun, including credibility to rumors that AMI Labs is likely to be elevating funding at a $3.5 billion valuation. According to Bloomberg, VCs in talks with the startup embody Cathay Innovation, Greycroft, and Hiro Capital, to which LeCun is an advisor. Other potential traders reportedly embody 20VC, Bpifrance, Daphni, and HV Capital.
Regardless of who writes the checks, traders could wish to notice an necessary element: As LeCun has made clear, he’s AMI’s government chairman, not its CEO. Instead, that function belongs to Alex LeBrun, beforehand co-founder and CEO at Nabla, a well being AI startup with workplaces in Paris and New York.
LeBrun’s transition from Nabla to AMI is a part of a partnership introduced final December by Nabla, which develops AI assistants for medical care and to which LeCun has been an advisor. In change for “privileged access” to AMI’s world fashions, Nabla’s board supported LeBrun’s shift from CEO to chief AI scientist and chairman, clearing the way in which for his new function.
Techcrunch occasion
San Francisco
|
October 13-15, 2026
As AMI Labs’ CEO, LeBrun shall be surrounded by acquainted faces. After Facebook acquired his earlier startup, Wit.ai, the serial entrepreneur and AI engineer labored beneath LeCun’s management at Meta’s AI analysis laboratory, FAIR. According to studies, the duo may also be joined by Laurent Solly, who stepped down as Meta’s vp for Europe final December.
The expertise overlap between AMI and Meta possible received’t cease there. LeCun instructed the MIT Technology Review that his former employer may nicely be AMI’s first shopper. But he has additionally been publicly important of a few of Meta’s strategic selections made beneath Mark Zuckerberg’s path. More broadly, the Review interprets AMI Labs as a contrarian wager towards massive language fashions (LLMs).
The limitations of LLMs that LeCun has identified embody hallucinations, that are a severe concern in contexts like medication, as LeBrun additionally is aware of firsthand. AMI Labs’ CEO instructed Forbes {that a} huge motive he took the function was the prospect of making use of its world fashions to healthcare. But the startup may also goal different high-stakes utilized fields.
“AMI Labs will advance AI research and develop applications where reliability, controllability, and safety really matter, especially for industrial process control, automation, wearable devices, robotics, healthcare, and beyond,” it wrote in its mission assertion. “We share one belief: real intelligence does not start in language. It starts in the world.”
Unlike generative approaches, which LeCun and his crew see as poorly fitted to unpredictable knowledge akin to sensor enter, the startup guarantees that its AI methods won’t solely perceive the actual world, but additionally have persistent reminiscence, the power to motive and plan, and be controllable and protected.
The startup plans to license its know-how to business companions for real-life functions, however says it additionally plans to contribute to constructing the way forward for AI “with the global academic research community via open publications and open source.” LeCun stated he plans to maintain his professor place at NYU, the place he teaches one class per 12 months and supervises PhD and postdoctoral college students.
This signifies that the French-born researcher will stay based mostly in New York, however he instructed the MIT Technology Review that AMI Labs “is going to be a global company [that’s] headquartered in Paris.” The information was welcomed by French President Emmanuel Macron, who expressed his satisfaction that LeCun, who can be a Turing Prize winner, selected Paris. “We will do everything we can to ensure his success from France,” he stated.
The startup may also have workplaces in Montreal, New York, and Singapore, however its determination to choose Paris for its headquarters will assist consolidate Paris’ fame as an AI hub, the place it would be a part of the ranks of H, Mistral AI and several other worldwide labs, together with FAIR. It’s becoming, maybe, that AMI is pronounced a-mee — like “ami” in French, which implies “friend,” LeCun has identified.
