A startup called General Intuition has raised $320 million in fresh funding, betting that the millions of hours humans spend playing video games could teach AI agents how to navigate the real world.
According to TechCrunch, the round values the company at $2.3 billion and was led by Khosla Ventures. Techmeme, citing TechCrunch reporter Rebecca Bellan, reports the new capital brings General Intuition's total funding to $454 million.
The company's core idea is that gameplay footage is a rich, untapped source of "action data." Rather than learning only from text or static images, General Intuition trains AI agents on spatial reasoning — the sense of how to move through and interact with an environment — by studying how players act inside games. The pitch, as TechCrunch frames it, is that this could help AI develop something closer to human intuition.
The ambition stretches well beyond gaming. TechCrunch's headline captures the leap the company is making: "From Fortnite to robots." The same instincts an AI learns by watching someone maneuver through a virtual world, the theory goes, could one day guide physical robots operating in real spaces.
Techmeme notes the company is based in New York and was co-founded by a 31-year-old, and Bellan describes visiting its R&D floor for the reporting.
Why it matters: Most of today's headline AI models learned from text scraped off the internet, but teaching machines to act in physical space requires a different kind of training data — and a $2.3 billion valuation signals that major investors believe video games may be one of the largest available sources of it.