General Intuition Raises $220 Million at $6.2 Billion Valuation to Train World Models on Gameplay Data
General Intuition, the New York City artificial intelligence startup building "world models" trained on gameplay video, has raised $220 million at a $6.2 billion valuation, betting that billions of hours of video-game footage hold the key to smarter AI agents.
The funding round, announced on September 29, was led by Valor Equity Partners and Atreides, with participation from Seven Seven Six, Point72, Khosla Ventures and General Catalyst. The company is a sister firm of Medal, the popular gaming-clip sharing platform.
From games to general intelligence
General Intuition trains what it calls action foundation models — large AI systems that learn from billions of action-labelled gameplay videos. Chief executive Pim de Witte has said Medal is on track for 3 billion uploaded gameplay videos per year, giving the company a vast and continually growing training dataset.
The idea is that an AI which learns to navigate, plan and act inside game worlds could transfer those skills to the real world — from robotics to digital assistants. The company said the new capital will fund GPU clusters, model training and robotics pipelines, while it scales its teams in New York and Europe. Its models are now being made available to outside partners through a newly opened waitlist.
Funding follows a rapid climb
The raise comes just months after General Intuition closed a $320 million round in June 2026 at a $2.3 billion valuation. With the latest round, the company's total disclosed funding now exceeds $650 million — a remarkable trajectory for a firm that barely existed in the public conversation a year ago.
The investment also reflects a broader land grab in AI: with language-model valuations ballooning, investors are hunting for alternative data sources that could power the next generation of models. Gameplay footage — rich in physics, planning and goal-directed behaviour — is emerging as one of the most promising.
Sources: GamesBeat, Finsmes
General Intuition, the New York City artificial intelligence startup building "world models" trained on gameplay video, has raised $220 million at a $6.2 billion valuation, betting that billions of hours of video-game footage hold the key to smarter AI agents.
The funding round, announced on September 29, was led by Valor Equity Partners and Atreides, with participation from Seven Seven Six, Point72, Khosla Ventures and General Catalyst. The company is a sister firm of Medal, the popular gaming-clip sharing platform.
From games to general intelligence
General Intuition trains what it calls action foundation models — large AI systems that learn from billions of action-labelled gameplay videos. Chief executive Pim de Witte has said Medal is on track for 3 billion uploaded gameplay videos per year, giving the company a vast and continually growing training dataset.
The idea is that an AI which learns to navigate, plan and act inside game worlds could transfer those skills to the real world — from robotics to digital assistants. The company said the new capital will fund GPU clusters, model training and robotics pipelines, while it scales its teams in New York and Europe. Its models are now being made available to outside partners through a newly opened waitlist.
Funding follows a rapid climb
The raise comes just months after General Intuition closed a $320 million round in June 2026 at a $2.3 billion valuation. With the latest round, the company's total disclosed funding now exceeds $650 million — a remarkable trajectory for a firm that barely existed in the public conversation a year ago.
The investment also reflects a broader land grab in AI: with language-model valuations ballooning, investors are hunting for alternative data sources that could power the next generation of models. Gameplay footage — rich in physics, planning and goal-directed behaviour — is emerging as one of the most promising.
Sources: GamesBeat, Finsmes
