Origin Lab has secured $8 million in seed funding led by Lightspeed Ventures to establish a marketplace connecting video game companies with artificial intelligence firms building world models. Additional investors in the round include SV Angel, Eniac, Seven Stars, FPV, and angel investors Kevin Lin and Kyle Vogt. This capital injection aims to solve a critical bottleneck in the development of AI systems capable of interacting with the physical world. Unlike large language models that thrive on text, world models designed for robotics and physical simulation require vast amounts of visual data depicting object movement and spatial interactions. Currently, there is no easy source for such training sets, leaving many research laboratories scrambling to assemble high-quality datasets. Origin Lab proposes an innovative solution by leveraging the video game industry, a sector that has long possessed rich, complex data environments. Co-founders Anne-Margot Rodde, Antoine Gargot, and Colin Carrier explain that the AI systems being built today need to understand physics and motion, capabilities that are inherently present in video game engines. Rodde stated that while the industry sits on valuable data, there has been no infrastructure to bridge the gap between game developers and AI labs. Origin Lab serves as this intermediary, converting digital assets from games into training-ready formats. This process ranges from simple rendering runs to automating hours of walkthrough footage, creating a scalable pipeline for data distribution. The platform operates on a two-sided model. On one side, video game companies can monetize existing digital assets, extracting additional revenue from content they have already produced. On the other, world-model-focused organizations, such as Yann LeCun's AMI Labs or Fei-Fei Li's World Labs, can purchase licensed, high-quality data to train their systems. This arrangement addresses previous licensing hurdles and data quality issues that have historically prevented the broader adoption of game footage in AI training. The timing of this funding highlights a growing market for data vendors serving major AI labs. Faraz Fatemi, a partner at Lightspeed Ventures who led the investment, noted that the success of companies like Scale AI has demonstrated the revenue potential of data supply. He emphasized that while AI businesses are well-capitalized, data remains their primary bottleneck. Consequently, startups capable of securing and processing this data are becoming essential suppliers to the industry. Interest in video game footage has been evident for some time, though often fraught with controversy regarding copyright and source attribution. For instance, in December 2024, OpenAI faced scrutiny when its Sora model appeared to regurgitate footage from video games and streamers, likely due to training on platforms like Twitch. Amazon has also publicly expressed interest in utilizing similar data streams. Origin Lab aims to formalize this relationship through a structured licensing framework, ensuring that game assets are used legally and effectively to advance AI capabilities in the physical world.