General Intuition, a robotics AI company, has raised $320 million（約510億円） to build AI models trained on video game data rather than traditional simulated or real-world footage. The company uses billions of clips from Medal, a gaming platform, because they contain both visual information and action labels that show how players make decisions in complex environments—a combination the company believes is essential for training robots that can act intelligently in the real world.
Summaries like this, in your inbox every morning.
Sign up free →What happened
General Intuition raised $320 million（約510億円） in Series A funding, bringing its total funding to $454 million（約730億円） and valuation to $2.3 billion（約3700億円）. The New York-based company uses billions of gameplay clips from Medal, a gaming platform, to train AI models that can perceive, predict, and act in physical and virtual environments.
Why it matters
Most AI training relies on text or simulated data, but General Intuition argues that video game footage is more valuable because it captures humans perceiving an environment and deciding how to move through it—with embedded action labels showing exactly when players press buttons. This approach may help bridge the gap between language-based AI and the embodied intelligence needed for physical robotics.
What to watch
General Intuition plans to scale its compute capacity and pretraining for its next model version. The company also intends to make its API more broadly available this summer.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
5 minutes a day. The AI essentials.
200+ sources · Email / LINE / Slack
