Video Games Could Be the Secret Ingredient for Achieving AGI
A New York-based startup backed by Jeff Bezos, Eric Schmidt, and Google DeepMind believes that the path to artificial general intelligence (AGI) doesn’t run through more text data — it runs through video games.
General Intuition, a company valued at $2.3 billion, just closed a massive $320 million funding round led by Khosla Ventures, with participation from General Catalyst, Jeff Bezos, Eric Schmidt, former F1 champion Nico Rosberg, and researchers at MIT and Google DeepMind. The company’s total disclosed funding now stands at $454 million.
The Core Idea: Action Data > Text Data
While large language models like ChatGPT and Claude excel at understanding and generating text, they struggle with something fundamental: understanding how things move through space and time. This spatial-temporal reasoning gap is exactly what General Intuition aims to fill — using data from video games.
The startup spun out of Medal.TV, a platform where gamers upload and share gameplay clips. The key insight? Medal’s hundreds of millions of hours of gameplay aren’t just video files — they come with action labels: exact records of what buttons players pressed and when. This is the secret sauce that sets General Intuition apart from competitors trying to infer actions from video alone.
“We view this as the next stage of future pre-training,” says CEO and co-founder Pim de Witte, 31. “We have a single model that can respond to Fortnite information on the screen and take action, but also to real-world dynamics in a way that an LLM could never.”
From Fortnite to Robot Dogs
During a visit to General Intuition’s New York R&D floor, a quadrupedal robot guided by the same AI model that plays Fortnite navigated the office autonomously — bumping into furniture like a learning toddler. The company says it took just eight minutes of real-world robotics data to fine-tune the model for this embodied task.
The startup’s ultimate goal? Sell a generalized agentic model that can power robots in factories, navigate hazardous environments, or drive autonomous vehicles — effectively becoming the “brain” that connects simulation to the real world.
The Ethical Line
Unlike much of Silicon Valley’s growing “war tech” enthusiasm, de Witte has drawn a firm ethical line: no lethal autonomous systems. The Dutch-born CEO, who spent three years working with Doctors Without Borders, wants General Intuition’s technology used for search and rescue, not combat.
“We don’t want to be an escalatory part of the system,” de Witte said. “Let’s say I were to come out and say, ‘We’re doing lethal autonomy.’ What do you think would happen in other countries?”
What’s Next
General Intuition plans to use the majority of its new capital to scale compute capacity through a deal with CoreWeave, focusing on pre-training the next version of its model. The company also plans to make its API more broadly available by the end of summer — opening the door for developers to build on top of its “world model” technology.
Vinod Khosla, whose firm led the round, put it simply: “In world models, the quantum leap is the emergence of intuition in the AI — a human intuition-like capability. The human action data and reaction data you have in games is the key part to the emergence of intuition.”
Frequently Asked Questions
What exactly does General Intuition do?
General Intuition builds AI models trained on video game action data (button presses and controller inputs) to develop spatial-temporal reasoning — the ability to understand movement through space and time. The goal is to power physical robots, autonomous systems, and simulation environments.
How is this different from other AI companies?
Unlike LLMs that train on text data, General Intuition trains on hundreds of millions of hours of video gameplay with embedded action labels. This allows the model to understand causality, physics, and spatial relationships in a way text-only models cannot.
Who is backing General Intuition?
Investors include Jeff Bezos, Eric Schmidt, Khosla Ventures, General Catalyst, Google DeepMind researchers, MIT researchers, and former F1 champion Nico Rosberg. The company is valued at $2.3 billion.