Origin Lab raises $8 million to help video game companies sell data to global model builders

As artificial intelligence begins to interact with the physical world, new types of laboratories are being built World models Which can be used to operate physical robots or object models in physical space. Unlike large language models, there is no easy source of data for these models, which has left many labs scrambling to assemble the necessary training sets.

Now, a startup has emerged with an unexpected data source: the video game industry.

This is a hypothesis Origin laboratorywhich just announced an $8 million seed funding round led by Lightspeed Ventures. SV Angel, Eniac, Seven Stars and FPV also participated, with angel funding from Twitch co-founder Kevin Lin and Cruise founder Kyle Vogt.

“The AI ​​systems being built now need to understand how the physical world works and how things move,” Anne-Margot Rudd, co-CEO and co-founder, told TechCrunch. “This data basically lives in video games.” The other founders of the company (pictured above) are Antoine Jargot and Colin Carriere.

In simple terms, Origin Lab will be a marketplace where labs focusing on global models e.g Yann LeCun’s AMI Laboratories or Fei-Fei Li Global Laboratories High quality licensed data can be purchased. On the other side of the trade, video game companies can generate additional revenue from the digital assets they have already created. In between, Origin Lab will convert video game assets into a model that serves as training data — something that could be as simple as running a demo or as complex as automating hours of walkthrough footage.

“It became clear that the video game industry was relying on some incredibly valuable data, but there was no real way or infrastructure to fundamentally connect AI labs and the video game industry,” Rudd says. “And basically, we built this bridge.”

Labs have long been interested in video game footage as a data source, but licensing and data quality issues have often gotten in the way. In December 2024OpenAI caused a minor scandal when the first version of its Sora video generation model appeared to replay footage from popular video games and streamers — perhaps because it was trained on Twitch streams. Amazon has been open Her interest in using Twitch streams To train models.

Origin’s success in raising money is a sign of a growing market — not just for training data, but also for startups that can serve as key suppliers to major AI labs. The success of companies like Scale AI makes this opportunity impossible to ignore, says Faraz Fatemi, the partner at Lightspeed who led the Origin investment.

“We’ve seen how severe the revenue volume is for data vendors serving large labs,” Fatemi told TechCrunch. “These are very well-capitalized companies, and the constraint for all of them is data.”