Ai Engineering 3 min read

Origin Lab Raises $8M for Game Engine Telemetry Marketplace

Origin Lab has secured $8 million in seed funding to launch a platform that converts raw video game engine data into licensed datasets for world model research.

Origin Lab has raised $8 million in a seed round led by Lightspeed Venture Partners. The startup provides a specialized marketplace connecting video game publishers with AI researchers building world models. The platform converts raw game environments into a trademarked data class called Artificial World Intelligence, providing the physical and spatial grounding necessary for autonomous systems.

The seed round includes participation from SV Angel, Eniac Ventures, Seven Stars Cloud Group, and FPV. Twitch co-founder Kevin Lin and Cruise founder Kyle Vogt also participated as angel investors.

Co-CEOs Anne-Margot Rodde and Colin Carrier lead the company alongside CTO Antoine Gargot. The leadership team draws heavily from game development and streaming infrastructure, leveraging Carrier’s prior experience as Chief Strategy Officer at Twitch to navigate the intersection of gaming content and data licensing.

The HUD-Free Capture Pipeline

Standard web-scraped video is insufficient for systems that must understand physics and spatial relationships. Origin Lab addresses this data bottleneck using proprietary multi-input capture software. The system records continuous two- to eight-hour sessions directly from game engines.

A critical technical differentiator is the platform’s ability to remove Head-Up Display elements at the engine level. Health bars, mini-maps, and ammo counters are stripped out before capture. This yields raw 3D world data that researchers cannot extract from standard screen recordings or streaming platforms.

Data ComponentCaptured Signal
VisualHigh-fidelity, HUD-free video
TelemetryScene composition and spatial coordinates
InteractionSynchronized player inputs
PhysicsEngine state and camera movement

The final datasets pair the visual output with this synchronized metadata, providing the precise cause-and-effect mapping required for spatial reasoning.

Inventory and Market Positioning

Origin Lab currently holds over 500,000 hours of content across more than 250 games. The company has secured exclusive licensing partnerships with over 20 game publishers representing 50 titles. Meta is already listed as a select client on the official platform, alongside an active contract with an undisclosed frontier AI laboratory.

The marketplace specifically targets research groups building spatial and physics-aware AI systems, such as Fei-Fei Li’s World Labs and Yann LeCun’s AMI Labs. As the industry shifts from understanding language to understanding physical interactions, these labs require vast amounts of interactive environment data.

Origin Lab positions its marketplace as a licensed alternative to unauthorized data scraping. The commercial model provides a legal framework for developers executing continued pretraining for spatial models, avoiding the intellectual property disputes that previously complicated AI video and simulation generation.

If you are building robotics pipelines or autonomous systems, structured game telemetry offers a viable proxy for real-world physics data. Review your current spatial datasets to determine if engine-level physics metadata could improve your model’s understanding of motion and environment states.

Get Insanely Good at AI

Get Insanely Good at AI

The book for developers who want to understand how AI actually works. LLMs, prompt engineering, RAG, AI agents, and production systems.

Keep Reading