General Intuition in talks to raise $300M at around $2B valuation

General Intuition, a startup building embodied AI and world models, is in advanced fundraising conversations targeting $300M at a $2B valuation. The funding round represents a significant bet on the company's approach to training AI systems that understand and interact with physical environments.

The company leverages Medal's dataset, which generates 2 billion videos annually from 10 million monthly active users. This video corpus gives General Intuition a rare advantage in training embodied AI models that learn how the real world works through diverse, continuous visual data. Rather than relying on synthetic simulations or limited public datasets, the startup taps into actual human interaction patterns captured at scale.

Embodied AI addresses a core limitation in current large language and vision models. These systems excel at pattern matching in text and static images but struggle with temporal reasoning and physical causality. General Intuition's approach of training world models on real-world video data directly tackles this gap. The company aims to build AI that understands not just what objects are, but how they move, interact, and respond to forces.

The $2B valuation reflects investor confidence in the embodied AI thesis as a path toward more capable, reasoning-focused AI systems. This sits well above many Series B or C rounds but reflects the strategic value of both the technology and the data moat. The $300M raise would position General Intuition as one of the better-funded AI infrastructure plays focused on robotics and physical world understanding.

Competitors in embodied AI and world models include established players like Nvidia with its AI foundation models for robotics, alongside newer startups building similar infrastructure. General Intuition's combination of proprietary video data and focus on world models differentiates its position.

The timing suggests confidence from investors betting on embodied AI as the next frontier in AI capability. Rather than pursuing scaling laws on text alone, this capital commitment backs a different