Origin Lab has raised an $8 million seed funding round to build a marketplace that supplies video game data to AI companies developing world models for robotics and physical simulations. The funding round was led by Lightspeed Venture Partners, with participation from SV Angel, Eniac Ventures, Seven Stars, and FPV Ventures.
Angel investors in the round included Kevin Lin and Kyle Vogt.
The startup is focused on solving a growing problem for AI companies building systems designed to understand and interact with the physical world. Unlike large language models, which can be trained using large amounts of internet text, world models require data that captures physical movement, environments, and object interactions.
Using Video Games As AI Training Data
Origin Lab’s co-founder and co-CEO Anne-Margot Rodde said video games contain large amounts of structured physical simulation data useful for AI training.
“The AI systems that are being built now need to understand how the physical world works and how things move,” Rodde told TechCrunch. “That data essentially lives in video games.”
The company’s other co-founders are Antoine Gargot and Colin Carrier.
Origin Lab plans to operate as an intermediary platform connecting AI labs with video game companies that own digital assets and gameplay environments.
According to the company, AI research groups building world models — including efforts connected to Yann LeCun’s AMI Labs and Fei-Fei Li’s World Labs — could purchase licensed datasets through the marketplace.
Video game companies, meanwhile, would gain additional revenue streams by licensing digital assets and gameplay data that already exist.
Origin Lab said it converts video game materials into formats usable for AI training. That process can range from rendering game assets to generating walkthrough footage automatically.
“It became clear that the video game industry was sitting on some incredibly valuable data, but there was no real way or infrastructure to basically connect AI labs and the video game industry,” Rodde said. “So essentially, we built that bridge.”
Growing Demand For AI Training Data
AI companies have shown increasing interest in video game footage and virtual environments as training material for world models and video generation systems.
The article referenced controversy surrounding an earlier version of OpenAI’s Sora video-generation model after observers claimed it appeared to reproduce footage from video games and Twitch streamers, suggesting possible training exposure to livestream content.
Amazon has also publicly discussed interest in using footage from Twitch for AI model training.
Licensing issues and concerns around data quality have historically limited broader use of video game content for AI systems.
Lightspeed partner Faraz Fatemi said the market for AI training data providers has grown rapidly as major AI labs continue scaling model development.
“We’ve seen how sharp the revenue scaling can be for data vendors that are serving the major labs,” Fatemi told TechCrunch. “These are very well-capitalized businesses, and the bottleneck for all of them is data.”
Featured image credits: Magnific.com
For more stories like it, click the +Follow button at the top of this page to follow us.