Yann LeCun's AMI Labs and Fei-Fei Li's World Labs Won't Say What They're Actually Selling

Two of AI's most talked-about companies raised serious money and serious buzz on a technology called "world models." Neither one will say what they're going to sell.

TechCrunch's Russell Brandom laid out this picture after moderating a panel on world models this week at the All In conference. The two names dominating the space are AMI Labs, co-founded by Meta AI veteran Yann LeCun, and World Labs, founded by Stanford computer scientist Fei-Fei Li. Brandom writes that both companies have accumulated significant funding and buzz, but rank low on what he calls the "trying-to-make-money scale."

World models, at a basic level, are about giving AI systems a working map of physical space, according to Brandom's reporting. That's the same underlying idea behind self-driving systems like Waymo. Push it further and the same approach could theoretically train a humanoid robot to carry boxes or turn a few minutes of video into a fully explorable 3D environment. The technology has multiple possible applications, and that's exactly the problem for anyone trying to figure out what these companies actually intend to build.

"We'll Talk About It When We're Ready"

Brandom pressed AMI Labs co-founder and VP of World Models Michael Rabbat directly during the panel about what the company is working on. Rabbat's answer: "We'll talk about it when we're ready to talk about it."

Following up by email, Rabbat told Brandom the company is "still in a research and building phase, so we're not talking publicly about any product plans or timeline."

AMI Labs is less than a year old, which is a fair enough reason to stay quiet for now. But the secrecy isn't unique to AMI. World Labs' product, called Marble, is described by Brandom as the most fully developed offering in the entire space, with demos covering media creation, explorable video game environments, and CGI effects, plus some robotics use cases. Even so, Brandom notes the platform reads more like a capability showcase than a company that has picked a lane.

AMI, for its part, has already put a toe in manufacturing, biomedicine, robotics, and physician-facing AI software through a partnership called Nabia, according to Brandom's reporting. That's four different industries for a company that won't say which one, if any, is the actual business.

Even the Data Supplier Is Guessing

The opacity runs downstream too. Brandom spoke with Alex de Vigan, CEO of Physicl, a data supplier for world-model developers, on the sidelines of the same conference. De Vigan told Brandom he knows his company's data is useful to whatever these labs are building, but he doesn't know what that actually is. "I wish they would tell us more," de Vigan said. "We could build more useful data if we knew what they were working on."

A vendor is selling training data into these companies and can't get a straight answer on what it's for. That's not how supply chains typically function.

Why Stay Quiet

Brandom's reporting suggests a straightforward incentive: as long as fundraising stays easy, there's no pressure to commit to one commercialization path, and revealing a specific direction risks tipping off competitors. A summary of the story published by Gate.com went further, naming OpenAI and Anthropic specifically as potential rivals that a public roadmap might invite in. That specific framing doesn't appear in Brandom's original reporting and should be read as Gate.com's own inference, not something Rabbat or anyone at AMI or World Labs said on record.

A daily.dev summary of the same story compared the dynamic to the "dark forest" concept from Cixin Liu's science fiction, where staying silent is a survival strategy. That's a colorful gloss layered on top of Brandom's reporting, not a characterization AMI Labs, World Labs, or their executives offered themselves.

None of the seven accounts of this story, including the original TechCrunch piece, disclose specific dollar figures for how much AMI Labs or World Labs has raised, or from whom. That's a real gap. Investors and industry watchers are being asked to take "buzz and funding" on faith without a public business case behind either company.

A young company can stay in a research phase. AMI Labs is under a year old. But at some point, capital chasing a technology with no announced revenue plan runs into a simple test: either a product ships and generates money, or the hype outlasts the patience of the people writing checks. Rabbat's own words leave that timeline undefined. "We'll talk about it when we're ready" is not a business plan. It's a placeholder for one.