New York, NY — General Intuition, a Bezos-backed AI startup, has announced it raised $320 million at a $2.3 billion valuation, bringing its total disclosed funding to $454 million. The round was led by Khosla Ventures, with participation from Jeff Bezos, Eric Schmidt, General Catalyst, Nico Rosberg, and researchers from Google DeepMind and MIT.
The Bet: Gameplay Data > Internet Text
Most AI models today — from ChatGPT to Claude — are trained on text scraped from the internet. They’re brilliant with words but struggle to understand how objects move through space and time. General Intuition CEO Pim de Witte believes this “spatial-temporal reasoning” gap is the biggest obstacle to achieving AGI, and his solution is unconventional: video game footage.
The company spun out of Medal, a gaming clip platform where players upload and share their best moments. Those hundreds of millions of hours of gameplay come with action labels — precise records of exactly what buttons a player pressed and when. This, de Witte argues, is the secret sauce that competitors trying to infer actions from raw video alone can’t replicate.
From Fortnite to the Real World
During a recent TechCrunch visit to General Intuition’s New York R&D lab, the company demonstrated an AI agent that had been playing a Fortnite-like game for 100 hours straight. More impressively, that same model was simultaneously powering a quadrupedal robot roaming the office — navigating furniture, bumping into obstacles, and learning its environment in real time.
“The same brain powering the agent playing the game is powering the robot,” de Witte said.
The team revealed it took just eight minutes of real-world robotics data to fine-tune the AI model for the quadruped. The training data was even collected on a public street, not inside the office — proving the model can generalize across vastly different environments.
What’s Next: The “Intuition” Layer
Vinod Khosla compared this moment to when reasoning emerged in large language models: “In world models, I think the quantum leap is the emergence of intuition — a human intuition-like capability. The human action data in games is the key to that emergence.”
General Intuition plans to use most of the funding to scale compute capacity (via a deal with CoreWeave) and pre-train the next version of its model. The company also launched Nerve, a jobs marketplace that lets gamers earn money starting with data labeling and eventually moving toward robot teleoperation — giving the gaming community a stake in the AI revolution.
By the end of summer, General Intuition’s API will be available more broadly, enabling developers to build smarter factory robots, more immersive NPCs, and AI agents that actually understand the physical world.
FAQ
How is General Intuition different from other AI companies?
While most AI companies train on text data, General Intuition uses video game telemetry — actual button presses and player actions — to teach models spatial-temporal reasoning. This allows their AI to understand cause and effect in the physical world, not just predict the next word.
Will this technology be used for military purposes?
CEO Pim de Witte has drawn a clear ethical line: no agents will be employed to harm humans. The company refuses lethal autonomy contracts, though it’s open to search-and-rescue missions. De Witte previously worked with Doctors Without Borders and keeps General Intuition based in New York rather than Silicon Valley to maintain this distance.