Worldmodeldata raises £7M seed to build AI's scarcest resource: gameplay data
What's the deal? WorldmodeldataDealroom has a profile for this one. Try Dealroom →, a Cambridge-based deeptech startup, has raised a £7 million seed round led by Iona Star CapitalDealroom has a profile for this one. Try Dealroom → to build the largest library of AI training data drawn from video games. It aims to collect 1 million hours of gameplay data by the end of 2026.
Why now? The round closed in December 2025, but the company waited until July 2026 to announce it. Founder and chief executive officer Rhea LoucasDealroom has a profile for this one. Try Dealroom → told Tech Funding News, "We decided to stay still for a little bit," using the time to work out which data AI training truly needs.
What's the endgame? World models are AI systems trained to predict how environments respond to actions, not just react to inputs. That requires recorded actions and their outcomes — data that is rare outside video games, where billions of players generate a natural record of moves, frames, and engine states. Worldmodeldata positions itself as a data provider, more like Scale AI, selling to the labs that need it.
The hire: Lord Richard Allan, Meta's former vice president of public policy for Europe, the Middle East and Africa, joined the board as non-executive chairman — notable given rising scrutiny over data provenance and AI regulation.
What could go wrong? The company has no finalised customer contracts, a team of about ten including advisors and contractors, and mounting competition. Since it raised, San Francisco's Origin LabDealroom has a profile for this one. Try Dealroom → secured $8 million and New York's General Intuition raised $454 million for similar gameplay-data projects.
The signal: At £7 million, Worldmodeldata's cheque is modest next to billion-dollar peers such as Fei-Fei LiDealroom has a profile for this one. Try Dealroom →'s World Labs and Yann LeCunDealroom has a profile for this one. Try Dealroom →'s AMI. Its bet is on a data-rich niche: as investor Gerry Buggy of Iona StarDealroom has a profile for this one. Try Dealroom → put it, the "coupling between action and consequence is the scarcest resource in AI today." Whether Worldmodeldata reaches 1 million hours before rivals will depend less on this round than on locking in unique data pipelines.
Read more: Tech Funding News
Image credit: Worldmodeldata
