General Intuition raises $320 million for AI agent training

What happened

General Intuition, a spinout from Medal (a game-clipping platform), announced a $320 million（約510億円） funding round led by Khosla Ventures at a $2.3 billion（約3700億円） valuation. The company is using Medal's archive of billions of gameplay clips paired with player inputs to train AI agents that can perceive, predict, and act in virtual and physical environments.

Why it matters

Most AI agents learn from video alone, but General Intuition's edge is that Medal's clips come bundled with the human action that caused each result—the player's input and the consequence. This proprietary data structure, rather than text or images scraped from the web, could become a defensible moat if the company can convert customer deployments into embodiment data that competitors cannot easily replicate.

What to watch

General Intuition says a commercial API is already in use with first partners in games, simulation, and robotics, with broader availability planned after a selective rollout. The core unresolved question is whether action-labeled game clips transfer cleanly to robots, drones, simulations, and industrial systems at scale—or whether gameplay intuition remains a detour, not a bridge, to real-world AI.