Upon entering General Intuition’s R&D floor at its New York office, the company’s 31-year-old co-founder and CEO Pim de Witte drew my attention to a monitor perched on a standing desk. Someone seemed to be playing something like Fortnite. He was not a person.
“Our agent plays for 100 hours straight,” said Kent Rollins, the company’s chief product officer, speaking.
Before I could get absorbed in the sight of an AI navigating the game’s virtual environment, I heard the electronic footsteps of a large four-legged robot approaching.
“The same brain that powers the agent playing the game also powers the robot,” de Witte told me. Josh Duplantis, a data analyst carrying a laptop with a live stream from the robot’s single camera, was quick to explain that the robot’s default mode was “explore.”
Leaning on this camera, its only eye, the giant bug-like bot approached me, circled around me, and continued into the office. He would occasionally snag chair legs or fall into an errant trash can, like a toddler who hasn’t yet learned how his body relates to the world around him. Duplantis said it took just eight minutes of real robotics data to perfect an AI model for the quadruped. Furthermore, this data was collected on the street, not inside the office where the robot was currently navigating.
An agency model that can generalize from gameplay to simulation to integration is General Intuition’s raison d’être. And this model’s ability to understand its place in the world has secured the support of some heavy hitters.
On Thursday, General Intuition said it raised $320 million at a $2.3 billion valuation, confirming an earlier TechCrunch report. The round brings General Intuition’s total funding to $454 million, following the $134 million round it raised at launch last October.
The startup grew out of de Witte’s other company, Medal, which allows gamers to upload and share video game clips. Hundreds of millions of hours of downloaded gameplay provided the initial dataset to train General Intuition’s model on spatiotemporal reasoning — or understanding how to move through space and time.
But the key ingredient wasn’t the game footage. It was the action tags embedded in these clips: records of exactly which buttons a player pressed and when. Most competitors, de Witte says, try to infer actions from video alone, which he argues is insufficient.
“We see this as just the next stage of future pre-training,” de Witte said. “We have a unified model that can respond to Fortnite on-screen information and take action, but also real-world dynamics in a way that an LLM never could.”
At one point, de Witte set me up with a laptop running General Intuition’s world model, a simulation environment that is generated frame by frame rather than rendered by a traditional game engine. As I often do when testing world models, I ran straight into a series of walls. In other demos I’ve tried, the agents you control sometimes go through instantly, but this one didn’t. From millions of hours of gaming, he somehow learned that walls are walls, stairs are for scaling, and shadows lengthen as the sun moves.
For General Intuition, this world model is not the product. is the training environment (internally referred to as “the gym”). The company ultimately wants to sell the agent model itself, and de Witte argues that action data built into the game helps the model distinguish “self” from “environment” in a way that gives it a richer understanding of causality.
Impressive as General Intuition’s technology is on display, the company isn’t alone in trying to solve this problem. Moreover, such a model has yet to be implemented in the physical world, at scale. Most approaches of this kind require massive amounts of real-world data that are slow and expensive to gather. General Intuition’s bet is that the game is an extensible shortcut.
His investors are also fine with that bet. General Intuition’s latest round was led by Khosla Ventures, with participation from General Catalyst, Jeff Bezos, Eric Schmidt, Nico Rosberg and researchers at Google DeepMind and MIT.
The vast majority of the round will go towards scaling computing capacity. General Intuition has entered into an agreement with CoreWeave and plans to focus on pre-training the next version of the model. A division is slated to make its API more widely available by the end of the summer.
Vinod Khosla, whose firm led the round, says he was drawn to de Witte’s vision and the company’s proprietary data position.
“If you look at LLMs, when logic came along, it was a quantum leap,” Khosla told me in a phone interview. “In world models, I think the quantum leap is the emergence of intuition in AI, an ability that resembles human intuition. The human action and reaction data you have in games is the key part of the emergence of intuition.”
Vision is a generational company
General Intuition isn’t the only company noticing that Medal’s human action data is a key piece of the puzzle of building dynamic world models and general agents. Brianna Martin, the startup’s chief of staff, said the company was born, in part, after Medal turned down a takeover offer from a major lab. Since then there have been other offers.
De Witte and his co-founders, Eloi Alonso, Adam Jelley and Vincent Micheli, are not interested in being acquired, nor are the startup’s investors looking for an exit yet. The amount and quality of proprietary data General Intuition has through Medal is one of the reasons Khosla is convinced the startup is a generational bet rather than an M&A target. that it could become the backbone for generalized agents and global models in simulation and the real world.
“At this point, it would be a data acquisition, which is not interesting at all,” Khosla said.
Part of that bet also involves trusting de Witte’s values.
The businessman spent three years working in the humanitarian field, including with Doctors Without Borders. As such, he has drawn a clear line on how General Intuition’s technology will be used: no agents will be used to harm people.
“We don’t want to be a tiered part of the system,” de Witte said. “Let’s say I had to come out and say, ‘We’re doing lethal autonomy.’ What do you think would happen in other countries?’
This limit on military use cases comes as Silicon Valley becomes increasingly bullish on war, though De Witte says he’s happy to see his models being used for search and rescue missions.
De Witte is Dutch and much of his team is European, which shapes the company’s identity. She says she brought Martin in part because of her decision to do so publicly quit Palantir regarding its cooperation with the United States Immigration and Customs Enforcement.
“I don’t know why Silicon Valley does what it does,” he said. “There’s a reason I’m not there.”
De Witte’s ethics does not simply limit what the models will not do. As a gamer who made $1.5 million building and hosting a private RuneScape server in his teens, de Witte also thinks about what happens to humans who lag behind what AI models can do.
General Intuition recently launched a platform called Nerve, a job market that allows players to earn money using their existing setups. Those who sign up start with data tagging and can eventually progress to teleoperating robots and other tasks. Medal’s user base, de Witte noted, is precisely the generation most exposed to AI-based displacement and wants to have a stake in what comes next.
A data shuttle
De Witte wants General Intuition to be an ecosystem enabler, like Anthropic or OpenAI — a model provider that allows others to leverage its technology. Today, the startup has a few customers in gaming, simulation and robotics.
“We’re not going to create a self-driving car company,” de Witte said. “We’re going to make it 10 times easier for the next person to start a self-driving car company.”
The company says that once it gets its API into the hands of more customers, it will be able to test its ability with a variety of use cases — such as testing a robot on a digital twin of a factory floor, powering a humanoid robot inside a game studio, or sending a quadruped to navigate dangerous environments.
While the quadruped is the first physical integration General Intuition has tested in the real world, it has also tested drones and other devices, including testing the model in driving games.
“It works on anything you can control using a gamepad or a keyboard mouse,” de Witte said.
Being able to build a data flywheel is one of the goals.
“We will select customers where we can differentiate the integrations for which this generalized foundation model serves as the backbone,” de Witte said. “So we’ll prioritize choosing customers on whether they can offer real-world data that will be interesting and useful for research. And whether they have a flexible internal team where we can be real embedded partners and learn from each other.”
Khosla said that General Intuition’s proprietary data is what got it this far, and that its ability to continue to collect data that no one else has will be essential. Especially since, despite the impressive demonstrations, whether bringing the simulation to the real world can hold up at scale is an open question that no one has fully answered yet.
Correction: An earlier version of this article misstated the number of years Pim de Witte spent in humanitarian work.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
