General Intuition Targets $6B Valuation in Oversubscribed Round Led by Valor, Point72

General Intuition is in talks to raise funding at a $6 billion pre-money valuation from new investors including Valor Equity Partners, Point72 Ventures, and Seven Seven Six, according to sources familiar with the matter (TechCrunch). The round is oversubscribed and still being finalized. Existing investors Khosla Ventures and General Catalyst are also participating.

The New York-based startup, legally incorporated as General Intuition US Inc., is building a foundation model that trains generalized AI agents to move through space and time. The company describes itself as "the frontier lab for acting in space and time," aiming to build models that can perceive, predict, and act in both virtual and physical environments (General Intuition).

This latest round follows a rapid capital trajectory. In June 2026, General Intuition raised $320 million at a $2.3 billion valuation, bringing its total disclosed funding to $454 million at that time. Prior to the June round, the company had raised a $134 million seed round to teach agents spatial reasoning using video game clips. TechCrunch's reporting on the current deal confirms earlier coverage by The Wall Street Journal.

The technical approach is distinctive. General Intuition uses hundreds of millions of hours of gameplay and "action labels" — records of what buttons a player pressed and when — harvested from Medal, the video game clip-sharing platform that CEO Pim de Witte spun the company out of in October 2025. Vinod Khosla told TechCrunch he believes action labels will be a key part of the "emergence of intuition," the ability for a model to generalize across tasks it wasn't explicitly trained on.

General Intuition intends to use the new funds to improve its general model with a focus on robotic embodiments. The capital allocation plan includes spending more on compute infrastructure and hiring talent. The company already has a partnership with CoreWeave for compute infrastructure.

The jump from a $2.3 billion valuation in June to a reported $6 billion pre-money in August reflects investor appetite for foundation models that bridge the gap between digital environments and physical robotics. The thesis is that action-label data from gameplay — a rich, high-frequency record of decisions made in dynamic, spatial environments — can serve as a training substrate for agents that must act in the real world.

The participation of both early backers like Khosla Ventures and General Catalyst alongside new investors signals continuity in the strategic vision. Khosla's framing of "emergence of intuition" as a generalization property positions General Intuition's work not as a narrow robotics play but as a broader bet on action-grounded model scaling.

For practitioners, the relevant question is whether action-label transfer from gameplay to robotic embodiments holds up under distribution shift. The company's pipeline from Medal's gameplay data to spatial reasoning models and now to robotics applications suggests a belief that the action representations learned in virtual environments carry sufficient common structure to bootstrap physical agent performance. The use of CoreWeave infrastructure indicates serious compute commitment to that training pipeline.

The broader context worth noting is that embodied AI has attracted intense capital interest throughout 2026, with multiple labs pursuing variations on the theme of training agents in simulation and transferring them to physical platforms. General Intuition's differentiator is the scale and specificity of its gameplay-derived action dataset — a data source that is both voluminous and temporally dense in a way that passive video or curated robotics demonstrations are not.

The deal, if it closes at the reported terms, would put General Intuition among the more richly valued private AI labs focused on action and embodiment. The speed of the valuation increase — roughly 2.6x in two months — also reflects a competitive funding environment for companies with credible claims to proprietary training data at scale.