General Intuition announced that it has raised $320 million in Series A funding at a $2.3 billion valuation. The round was led by Khosla Ventures, with participation from General Catalyst, Hillspire, the family office of Eric Schmidt, and Jeff Bezos.
General Intuition describes itself as a frontier lab focused on building AI systems capable of acting in space and time. The company is developing large action foundation models designed to understand, predict, and operate in dynamic environments.
The company’s training approach is built around large-scale gameplay data. General Intuition said its models are trained on billions of ground-truth, action-labeled gameplay clips from Medal, a platform with 17 million monthly active users.
This data foundation gives General Intuition access to a large volume of real user actions across interactive environments, which the company is using to train models that can learn from sequences of decisions, movement, timing, and outcomes.
General Intuition is also pushing the frontier of world models. The company said it is working on systems that can generate infinite training environments, which could help AI agents improve by practicing across a broad range of simulated scenarios.
The company’s broader goal is to build AI models that move beyond text-based reasoning and into action-oriented intelligence. By combining action-labeled gameplay data with world-modeling technology, General Intuition is targeting AI systems that can learn how to act in complex environments rather than only respond to prompts.
The funding is expected to support continued research and development across large action foundation models, world models, data infrastructure, and training systems.
The financing also gives General Intuition significant backing as investor interest continues to build around frontier AI labs focused on agents, simulation, robotics-adjacent intelligence, and models that can operate across time-based environments.
The company’s use of gameplay data is central to its strategy. Games provide dense streams of action, feedback, and decision-making, making them a useful training ground for models that need to understand how actions affect outcomes over time.
General Intuition’s connection to Medal also gives the company a differentiated source of action-labeled data at scale. With 17 million monthly active users contributing gameplay clips, the platform provides a large base of real-world interactive behavior for model training.
The Series A round positions General Intuition to expand its team, increase compute capacity, and accelerate development of models designed for more capable AI agents.
The company said its mission is to build a frontier lab for acting in space and time, with the long-term aim of advancing AI systems that can operate more effectively in complex digital and physical environments.
