Origin Lab raises $8M to bridge gaming data and AI training

Origin Lab has raised $8 million in seed funding to build a marketplace that connects video game companies with artificial intelligence labs. The seed round was led by Lightspeed Ventures. Other participants in the funding round include:

• SV Angel
• Eniac
• Seven Stars
• FPV
• Kevin Lin (angel investor)
• Kyle Vogt (angel investor)

The startup, co-founded by Anne-Margot Rodde, Antoine Gargot, and Colin Carrier, aims to address the data scarcity issues facing developers who build world models—AI models designed to understand physical space and robotics. Unlike large language models, which train on vast repositories of text, world models require data that demonstrates how physical objects move and interact. “The AI systems that are being built now need to understand how the physical world works and how things move,” said co-CEO and co-founder Anne-Margot Rodde. She noted that this physical-world data essentially lives within video games.

Origin Lab will operate as a marketplace where world-model-focused developers, such as AMI Labs and World Labs, can purchase licensed data. The startup will convert video game assets into training data, which could be as simple as a rendering run or as complex as automating hours of walkthrough footage. This allows video game companies to generate revenue from existing digital assets, while providing AI labs with a structured pipeline. Rodde explained that because the video game industry lacked the infrastructure to connect with AI developers, the startup had to build that bridge. Faraz Fatemi, a partner at Lightspeed, noted that major AI labs are highly capitalized, but face a severe bottleneck when it comes to acquiring data. Fatemi compared the opportunity to the sharp revenue scaling seen by other data vendors serving major labs, such as Scale AI.

The launch of this marketplace comes as AI developers face heightened scrutiny over how they source training sets. In December 2024, OpenAI experienced a minor scandal regarding its Sora video-generation model, which appeared to use training data from video game footage and streamers, presumably because it had been trained on Twitch streams. While AI labs have historically sought out video game footage, licensing and data-quality issues have consistently blocked formal agreements. Origin Lab plans to formalize this pipeline, establishing a clear infrastructure to connect the two industries and resolve licensing friction.

Why it matters

Origin Lab is positioning itself as a critical infrastructure layer, solving the data scarcity problem for world-model AI by formalizing the licensing of video game assets.