World model companies are keeping a lot of secrets
TechCrunch Russell Brandom ● Covered by 2 sources
World model labs are getting lots of money, but they won’t say what they’re building. That secrecy may be smart — until rivals and customers need a real product.
Based on reporting by TechCrunch, Russell Brandom — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
At the All In conference this week, a panel on world models turned into something rarer: a look at how little the industry is willing to say about itself. The big names are Yann LeCun’s AMI Labs and Fei-Fei Li’s World Labs. Both have buzz. Both have funding. And both still sound a lot more like research shops than businesses with a clear product to sell.
That was especially obvious when Michael Rabbatt, a co-founder of AMI Labs and its VP of World Models, was pushed on what the company is actually making. He wasn’t interested in giving much away. Over email, he said AMI is still in a research and building phase and isn’t talking publicly about product plans or a timeline. That is easier to accept when a company is less than a year old. But the same guarded tone shows up across the field.
World Labs’ Marble is probably the most developed product in the bunch, and even that feels more like a showcase than a finished business. Its demos cover media creation, explorable game-like environments, CGI effects, and some robotics uses. Useful, sure. Specific, not so much.
The secrecy goes beyond the labs themselves. Alex de Vigan, CEO of Physicl, a data supplier for the world model business, said his company knows its data has been useful, but not what it has been used for. He wanted more information so Physicl could build better data. Instead, he’s left guessing like everyone else.
That fuzziness is partly because world models can point in several directions at once. They might help self-driving systems, humanoid robots, interactive video, or software for doctors. AMI has already touched manufacturing, biomedicine, robotics, and a Nabia partnership in AI software for doctors. But there’s a reason these companies keep their cards close: if one lab shows its hand, the rest of the field — including the bigger AI labs — can start racing to the same target.
My take — AI-written commentary, not fact-checked reporting
This is what easy fundraising buys: silence. If nobody has to explain the product, nobody has to admit how far away it is. The whole world-model crowd is acting like secrecy is a strategy, when it’s often just a way to postpone the awkward part where the demo has to become a business.
Read more about this at: TechCrunch