GAMING STARTUP RAISES $300 MILLION TO TRAIN AI ON TWO BILLION VIDEO GAME CLIPS A YEAR The best training data for teaching machines to understand space, time, and motion might not come from robotics labs or self-driving car footage. It might come from ten million people playing video games. General Intuition, a New York startup that spun out of Medal, the viral gaming clip platform, is in talks to raise $300 million at a $2 billion valuation. The investors include Jeff Bezos and Eric Schmidt alongside existing backers Khosla Ventures and General Catalyst. Medal generates roughly two billion first-person video game clips per year from its ten million monthly active users. General Intuition’s pitch is that this dataset, interactive, immersive, and physically consistent, is the ideal foundation for training AI that can reason about the physical world in real time. This is called a world model, and it is one of the most contested areas in frontier AI research right now. The idea is that an AI trained to understand what happens next in a game environment can transfer those skills to robotics, navigation, simulation, and planning. OpenAI tried to buy a gaming dataset for this purpose but was outbid and outmaneuvered. General Intuition is sitting on the data OpenAI wanted. The company was founded just eight months ago with a $134 million seed round. This new raise would bring total funding to more than $400 million before shipping a public product. That product is expected by late summer. Keywords: General Intuition AI funding, world models AI training, gaming data AI, spatial reasoning AI startup