General Intuition
Series A
General Intuition builds embodied AI agents and world models that learn spatial-temporal reasoning from interactive gameplay data.
Founded
2025General Intuition builds embodied AI agents and world models that learn spatial-temporal reasoning from interactive gameplay data.
General Intuition is a frontier AI research lab founded by Pim de Witte in 2025, focused on embodied AI agents and world models that learn spatial-temporal reasoning from interactive gameplay. The company has raised more than USD 450 million, including a USD 133.7 million seed led by Khosla Ventures and General Catalyst and a USD 320 million Series A reported in mid-2026.
It trains models on Medal's dataset of billions of first-person gameplay clips from millions of monthly active users, aiming to build agents that perceive, predict, and act in virtual and physical environments. The company is preparing a commercial API and partner program for games, simulation, and robotics use cases.
General Intuition builds embodied AI agents and world models that learn spatial-temporal reasoning from interactive gameplay data. Its models perceive environments, predict the outcomes of actions, and select actions in real time, with intended transfer to physical systems such as robots and drones.
The offering is currently available through a commercial API and partner program for select games, simulation, and robotics companies.
General Intuition targets applications in gaming bots, simulation, and robotics, where agents must perceive environments, anticipate outcomes, and act in real time. The company views spatial-temporal reasoning as a crucial building block toward more general machine intelligence.
It is onboarding early partners across games, simulation, and robotics through a commercial API ahead of a broader model release, with a first public product expected in 2026.
General Intuition's primary advantage is access to Medal's dataset of billions of first-person gameplay clips from millions of monthly active users, creating a unique data moat for training spatial-temporal reasoning models.
The company differentiates by focusing on embodied agents and world models rather than competing with game developers, targeting applications in gaming, simulation, and robotics where interactive play provides a scalable training curriculum.
General Intuition is a newly established company with an unproven commercial track record. Its embodied agents and world models remain in development, and the company has not yet demonstrated production-scale deployment or a repeatable revenue model.
The company faces execution risk in translating research capabilities into market-ready products while competing against well-capitalized AI labs and foundation-model providers with established distribution, compute, and customer relationships.
General Intuition has not publicly disclosed pricing for its foundation models or AI agent services. As a frontier research lab focused on spatial-temporal reasoning, the company is likely in a pre-commercial phase where pricing and revenue models have not yet been established.
The company has raised more than USD 450 million, providing significant runway for research and development. Future monetization is likely to follow an enterprise licensing or API model for its agents targeting gaming, simulation, and robotics applications.