General Intuition in talks to raise $300M at $2B valuation to build AI agents that understand space and time

General Intuition, a New York-based startup developing a foundation model that trains AI agents to navigate and interact in three-dimensional space, is in advanced discussions to raise approximately $300 million, according to sources familiar with the matter. The funding round would value the company at just over $2 billion.

Fresh capital eight months after spinout

The raise comes just eight months after General Intuition spun out of Medal, a platform for uploading and sharing video game clips, with a $134 million seed round. Sources tell Bitcoin World that the new funds have attracted backing from high-profile investors including Jeff Bezos and Eric Schmidt, as well as existing backers Khosla Ventures and General Catalyst.

Pim de Witte, who co-founded Medal, founded and leads General Intuition alongside co-founders Eloi Alonso, Adam Jelley, and Vincent Micheli — researchers with expertise in world modeling and simulation. The startup trains embodied AI and world models using Medal’s dataset of 2 billion videos per year from 10 million monthly active users.

A unique dataset for spatial-temporal reasoning

The startup’s core pitch is that this dataset — drawn from interactive, first-person gameplay — provides an ideal foundation for teaching machines deep spatial-temporal reasoning. Unlike static image or text datasets, these videos allow AI to perceive, anticipate, and interact in real time within simulated environments.

That dataset has reportedly attracted attention from OpenAI, which previously attempted to acquire Medal. Sources say OpenAI is not the only major AI lab to have expressed interest in the data or the company.

The world model race heats up

General Intuition operates in the increasingly competitive world model space. Startups such as Runway, Decart, and World Labs have all released world models recently. Google’s Genie 3 has begun integrating Google Maps data for more realistic simulation capabilities. However, General Intuition takes a different approach: it builds world models to train agents, not to sell the models themselves. The agents are the product, and the company believes its unique dataset gives it a viable path to market.

According to a source familiar with the company’s plans, General Intuition will use the fresh funds to scale up its compute capacity and aims to release a new product by the end of summer or early fall.

Why this matters

The development of world models is widely seen as a critical step toward general-purpose AI systems that can operate in the physical world — from robotics and autonomous vehicles to advanced simulation and training. General Intuition’s approach, grounded in real-world gameplay data, could offer a distinct advantage in teaching AI agents how to move through space and time with human-like intuition.

FAQs

Q1: What is General Intuition building?
A: General Intuition is building a foundation model that trains AI agents to understand and navigate three-dimensional space and time, using a large dataset of first-person video game footage.

Q2: Who is investing in this round?
A: Sources say the round includes backing from Jeff Bezos, Eric Schmidt, and existing investors Khosla Ventures and General Catalyst.

Q3: How does General Intuition’s approach differ from other world model startups?
A: Unlike companies that sell world models as products, General Intuition builds world models specifically to train AI agents. The agents themselves are the product, and the company’s dataset from Medal gives it a unique training resource.