General Intuition raises $320M to train agents from gameplay data

General Intuition has raised $320 million to scale its approach to training AI agents from video game data. TechCrunch reports the company is valued at $2.3 billion and is betting that millions of hours of gameplay can teach models richer action patterns.

The idea is that games provide dense feedback and varied decision-making environments that may transfer to real-world agent behavior. That puts the company in the growing category of labs trying to move beyond static text and image data.

The funding shows investor appetite for training methods that could make agents more capable in dynamic, interactive settings.