They Trained AI on 2 Billion Video Game Clips. Now It's Worth $2 Billion.
Startup leveraging Medal’s 2B annual videos to teach AI spatial-temporal reasoning lands Bezos, Schmidt backing.
General Intuition, a New York-based startup building foundation models that train AI agents to navigate through space and time, is in talks to raise approximately $300 million at a valuation just over $2 billion, according to sources familiar with the matter. The raise comes just eight months after the company spun out of Medal—a platform for uploading and sharing video game clips—with a $134 million seed round. Backers for the new round reportedly include Jeff Bezos, Eric Schmidt, and existing investors Khosla Ventures and General Catalyst.
The startup’s edge lies in its unique training data: Medal’s dataset of 2 billion videos per year from 10 million monthly active users. This first-person gameplay footage is ideal for teaching AI deep spatial-temporal reasoning, enabling agents to perceive, anticipate, and interact in real time within simulations. General Intuition differs from peers like Runway, Decart, and World Labs by focusing on building world models to train agents—not to sell the models themselves. The agents become the product, with gaming and robotics as near-term commercial uses. The fresh funds will be used to scale compute capacity, with a new product planned for release by the end of summer or early fall.
- General Intuition is raising $300M at a ~$2B valuation, backed by Jeff Bezos, Eric Schmidt, Khosla Ventures, and General Catalyst.
- The startup uses Medal’s dataset of 2B annual gameplay videos from 10M users to train AI agents on spatial-temporal reasoning.
- Unlike rivals that sell world models, General Intuition builds the agents themselves; plans to launch a new product by early fall.
Why It Matters
A unique gameplay dataset could unlock real-world AI agents for robotics and simulation, challenging world model competitors.