Origin Lab raises $8M for a video game data marketplace
Startup Origin Lab has raised $8 million to launch a marketplace for high-quality data. On the platform, video game companies can license their datasets to AI labs. Researchers use this data to train world models: systems that learn to model the behavior of the physical world through prediction.
AI-processed from TechCrunch; edited by Hamidun News
Origin Lab, a young startup, has raised $8 million to launch a marketplace meant to bridge video game companies and AI labs. The platform will let game studios monetize simulation data, while researchers get high-quality datasets for training world models.
Why AI labs want game data
World models are systems that learn to predict and model the behavior of the physical world. They require huge volumes of realistic data on object interactions, physics, lighting and motion. Video games contain exactly that: complex physics simulations, mature graphics engines and millions of possible scenarios.
Large AI labs (OpenAI, Deepseek and others) are already working on world models, but for now they have to collect data by hand or rely on cheap open sources. Origin Lab offers a better way: direct access to professional simulations built by game studios.
How the marketplace works
On the Origin Lab platform, video game companies can publish datasets containing scenes, physical parameters, state transitions and other information. Each dataset gets a description, metadata and licensing terms. AI labs browse the catalog, pick the sets they need and license them, paying for access.
The startup acts as an intermediary:
- Game studios earn revenue from existing content
- AI labs save months on data preparation
- The platform provides legal transparency on licenses
- Quality is guaranteed by the developers themselves
Origin Lab takes a commission on every transaction, while ensuring that all rights are respected and creators receive fair compensation.
Why this matters now
World models are one of the top priorities for major AI companies. They see them as a way to build more general and reliable AI systems that can plan and predict based on an understanding of physics. But without access to quality data, that is not possible.
Video game companies hold a unique resource: they spent years developing realistic simulations and invested millions in engineering. Previously, this data was valuable only inside the studio. Now it can be monetized.
"Video games are the ideal source of training data for world models," said one of Origin Lab's co-founders.
What's next
Origin Lab plans to expand. Over time, the marketplace may include simulation data from other sources: architectural simulations, robotics models and engineering simulations. For now the focus is on video games, because they are the most developed and accessible source.
What this means
Origin Lab shows that data has become a commodity. Companies with high-quality simulations can now earn revenue simply by providing access. For AI researchers, this means more democratic access to training data: you no longer need to be a megacorporation to get professional datasets. The video game industry gains a new revenue stream at a time when traditional game monetization is getting harder.
- Nov 2, 2026Researchers release AVID dataset: 68.1% of targets have a bounding box side of just 8–16 pixels
- Nov 2, 2026Karpathy Built HackerTube, a Catalog of 6,454 Videos, in Two Hours with Claude Code
- Nov 2, 2026Cosmos generated training video for drones: detector mAP@50–95 up 10.3%
- Nov 2, 2026Karpathy builds a YouTube clone with 6454 videos from Hacker News recommendations in two hours
Key Questions Answered
Why AI labs want game data?
World models are systems that learn to predict and model the behavior of the physical world. They require huge volumes of realistic data on object interactions, physics, lighting and motion. Video games contain exactly that: complex physics simulations, mature graphics engines and millions of possible scenarios. Large AI labs (OpenAI, Deepseek and others) are already working on world models, but for now they have to collect data by hand or rely on cheap open sources.
How the marketplace works?
On the Origin Lab platform, video game companies can publish datasets containing scenes, physical parameters, state transitions and other information. Each dataset gets a description, metadata and licensing terms. AI labs browse the catalog, pick the sets they need and license them, paying for access.
Why this matters now?
World models are one of the top priorities for major AI companies. They see them as a way to build more general and reliable AI systems that can plan and predict based on an understanding of physics. But without access to quality data, that is not possible. Video game companies hold a unique resource: they spent years developing realistic simulations and invested millions in engineering. Previously, this data was valuable only inside the studio. Now it can be monetized.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.