General Intuition raising $300M at $2B valuation to train AI agents on gaming footage

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

General Intuition, the New York-based startup building a foundation model that trains AI agents how to move through space and time, is in talks to raise around $300 million, sources familiar with the matter told TechCrunch. The raise comes eight months after General Intuition spun out of Medal, a platform for uploading and sharing video game clips, with a $134 million seed round. The fresh funds would bring the startup's valuation up to just over $2 billion, sources say. Sources tell TechCrunch General Intuition has secured funds from backers including Jeff Bezos and Eric Schmidt, as well as existing investors 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 who bring 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. The startup's pitch is that such a dataset -- unique because it allows AI to learn from interactive, first-person gameplay -- is the perfect base to teach machines deep spatial-temporal reasoning, allowing them to perceive, anticipate, and interact in real time in simulation. That dataset has reportedly attracted the attention of OpenAI, which previously attempted to acquire Medal. And sources say OpenAI hasn't been the only big AI lab to come knocking. The world model space that General Intuition is playing in is heating up. Startups like Runway, Decart, and World Labs have all released world models recently, and Google's Genie 3 recently began integrating Google Maps data for more real-world simulation capabilities. All of these companies see gaming and robotics training as near-term commercial use cases, but General Intuition takes a different approach: it builts world models to train agents, not to sell them. The agents are the product, and the startup's unique dataset gives it a path to viability. General Intuition will use the funds to scale up its compute capacity so it can release a new product by the end of summer or early fall, according to a source familiar with the matter.

General Intuition raising $300M for AI trained on gaming data

General Intuition, the AI lab that spun out of gaming clip platform Medal after rejecting a reported $500 million offer from OpenAI, is raising approximately $300 million at a valuation of just over $2 billion. Backers reportedly include Jeff Bezos and Eric Schmidt. General Intuition, a New York-based AI lab that trains agents to reason about space and time using billions of video game clips, is in talks to raise approximately $300 million at a valuation of just over $2 billion, according to TechCrunch. The round would come just eight months after the company launched with a $134 million seed, one of the largest on record. Backers reportedly include Jeff Bezos, whose own physical AI venture Prometheus recently raised $12 billion, and Eric Schmidt, alongside existing investors Khosla Ventures and General Catalyst. The Medal origin story General Intuition exists because of a deal that fell apart. Late in 2024, OpenAI reportedly offered $500 million to acquire Medal, a platform where gamers upload and share short video clips of gameplay, because it wanted the data to train its models, according to The Information. Medal's founder, Pim de Witte, turned it down. Instead he spun out General Intuition in October 2025 with co-founders Eloi Alonso, Adam Jelley, and Vincent Micheli, three researchers who had previously published influential work on world modelling and diffusion-based simulation. The company's core asset is Medal's dataset: roughly two billion video clips per year generated by more than 10 million monthly active users across thousands of games. Unlike YouTube or Twitch footage, which shows gameplay from a spectator's perspective, Medal's clips are first-person and interactive, capturing the spatial reasoning, timing, and decision-making that games demand. Agents, not models General Intuition's pitch differs from most of its competitors in one important respect. Companies like Decart, which raised $300 million earlier this month, and Google's Project Genie, which recently connected its world model to Street View's 280 billion images, are building world models as products in themselves, selling simulation environments to developers and enterprises. General Intuition builds world models to train agents, making the agents the product and the world model the training ground. The distinction matters commercially because it ties the company's revenue to what its AI can do, not to how well it can render a virtual scene. A crowded and well-funded field The world model race has attracted enormous capital in 2026. Runway, valued at $5.3 billion, launched its first world model in December and has been adding $40 million in annual recurring revenue per quarter. World Labs, founded by Stanford's Fei-Fei Li, raised $1 billion at a $5.4 billion valuation in February. Odyssey AI raised $310 million with backing from AWS and AMD, and Yann LeCun's AMI Labs raised over $1 billion at a $3.5 billion valuation. The common thesis is that AI needs to understand physics, not just language, and that video data is the bridge. General Intuition's founders argue that their data gives them an edge most competitors lack. Alonso and Micheli developed DIAMOND, a diffusion-based world model that predicts future frames directly rather than compressing them into tokens. The Medal dataset provides the kind of dense, interactive footage that static video cannot match. What comes next The company reportedly plans to use the new capital to scale compute and release a product by late summer or early autumn. The broader question is whether the agent market OpenAI and others are chasing will materialise quickly enough to justify the valuations being attached to companies building the underlying reasoning. OpenAI wanted Medal's data badly enough to offer half a billion dollars, and the fact that General Intuition said no and is now worth four times that amount says something about how the market prices control of training data in an era when every major AI lab is hunting for it.

Game-clip AI startup General Intuition in talks to raise $300M at $2B valuation

Game-clip AI startup General Intuition in talks to raise $300M at $2B valuation General Intuition PBC is in talks to raise about $300 million at a valuation of just over $2 billion, TechCrunch reported today, citing people familiar with the talks. The New York-based startup uses video game footage to train artificial intelligence agents to navigate physical space. The price tag is a steep markup. General Intuition launched eight months ago with a $134 million seed round. The new valuation is about four times that. General Intuition spun out of Medal B.V. in October 2025. Medal operates a platform where gamers upload and share short clips of gameplay and that dataset is the reason the startup exists. The company trains embodied AI and what are known as world models on roughly 2 billion clips a year generated by more than 10 million monthly active users across thousands of games. What makes the data valuable, the company says, is that it is first-person and interactive. Clips on YouTube or Twitch show gameplay from a spectator's seat. Medal's clips capture the player's own view, including the timing and split-second choices a game forces. General Intuition says that is the kind of footage that teaches a machine to read a space and act in it. The approach also sets the company apart from others chasing world models. Most build the models as products in their own right, selling simulation environments to developers and enterprises. General Intuition instead uses world models to train agents, making the agents the product and the simulation the training ground. The distinction ties its revenue to what its AI can do rather than how convincingly it can render a scene. The startup was founded by Pim de Witte, who co-founded Medal, alongside Eloi Alonso, Adam Jelley and Vincent Micheli, researchers known for their work on world modeling and diffusion-based simulation. Alonso and Micheli developed DIAMOND, a diffusion-based model that predicts future video frames directly instead of compressing them into tokens. The Medal dataset has already drawn interest from larger players. OpenAI Group PBC reportedly offered $500 million to acquire Medal late in 2024 to get at the data, an offer de Witte turned down. Sources told TechCrunch that OpenAI has not been the only major AI lab to approach the company. The world model field has pulled in heavy funding this year. Decart.ai Inc. raised $300 million last month, World Labs Technologies Inc. raised $1 billion in February and Odyssey ML Inc. raised $310 million with backing from Amazon.com Inc. and Advanced Micro Devices Inc. Alphabet Inc.'s Google has connected its Genie world model to Street View imagery for more realistic simulation. General Intuition plans to use the new capital to scale its compute capacity and ship a product by late summer or early fall, according to a source familiar with the matter. Backers in the new round reportedly include Jeff Bezos and Eric Schmidt, alongside existing investors Khosla Ventures and General Catalyst. Including the seed round, the financing would bring the startup's total funding to more than $400 million in less than a year.