Origin Lab Secures $8M to Bridge Gaming Data and AI Models

Origin Lab has secured $8 million in seed funding to position itself as the primary marketplace for AI developers seeking to train world models using video game assets. The round was led by Lightspeed Ventures, with participation from SV Angel, Eniac, Seven Stars, and FPV, alongside angel investments from industry figures including Twitch co-founder Kevin Lin and Cruise founder Kyle Vogt.

As artificial intelligence shifts toward interacting with physical environments, labs are racing to develop world models capable of simulating physics and spatial dynamics. While large language models rely on text, developers building robotics and simulation tools face a scarcity of high-fidelity training data. Origin Lab, founded by Anne-Margot Rodde, Antoine Gargot, and Colin Carrier, aims to solve this by unlocking the vast, untapped data reservoirs within the gaming industry.

Connecting Gaming Assets to AI Labs

The company functions as a bridge between two distinct sectors:

• For AI Labs: Organizations like AMI Labs and World Labs gain access to licensed, high-quality data sets that simulate how objects move and interact in space.
• For Game Developers: Studios can monetize existing digital assets by licensing them for AI training, providing a new revenue stream for their work.

“The AI systems that are being built now need to understand how the physical world works and how things move,” explained co-CEO Anne-Margot Rodde. “That data essentially lives in video games. 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. So essentially, we built that bridge.”

Addressing the Data Licensing Gap

The demand for gaming footage as a training resource is not new, but previous attempts to utilize it have been fraught with licensing complications and quality concerns. In late 2024, Amazon’s interest in utilizing Twitch streams and the controversy surrounding OpenAI’s Sora model—which appeared to ingest gaming footage without clear licensing—highlighted the need for a formal, legal framework for data acquisition.

Origin Lab plans to address these friction points by converting raw game assets into structured training data. This process ranges from simple rendering outputs to the automation of complex, multi-hour walkthrough footage, ensuring that the data is optimized for model training.

The Rise of Data-as-a-Service

Faraz Fatemi, a partner at Lightspeed who spearheaded the investment, views the startup as a critical infrastructure provider in an AI-dominated economy. As major AI labs face a persistent bottleneck in acquiring high-quality training sets, companies that manage the supply chain of data are becoming increasingly valuable.

“We’ve seen how sharp the revenue scaling can be for data vendors that are serving the major labs,” said Fatemi. “These are very well-capitalized businesses, and the bottleneck for all of them is data.”

By formalizing the exchange of digital assets, Origin Lab aims to replicate the success of firms like Scale AI, establishing itself as an essential supplier for the next generation of physical-world artificial intelligence.