Pim De Witte, founder of General Intuition and Medal.tv, turned down a deal reportedly worth $500 million from OpenAl, and instead raised $134 million to build an Al lab rooted in games, not words.
Context; EU AI Startup " While most 2025 AI and robotics rounds reported by EU-Startups have ranged between €3 million and €12 million, General Intuition’s raise underscores a sharp contrast in both scale and ambition – particularly notable given its early stage and cross-continental structure spanning New York and Geneva.
“When you play video games, you essentially transfer your perception, usually through a first-person view of the camera, to different environments,” added de Witte. “You get this selection bias towards precisely the kind of data you actually want to use for training work.”
Note: YT vid is 55 mins instead of 5,5 mins. My bad.
You must log in or # to comment.
Not my best choice for OP link, so here’s the original news NOS NL
It doesn’t happen every day that a Dutch entrepreneur says ‘no’ to $500 million. Pim de Witte did when OpenAI, the American company behind ChatGPT, reportedly offered him this astronomical amount for his company. But De Witte now wants to do what the Americans planned with his data himself.
" When it happened, I immediately thought: my God, did I make the right decision?" Pim de Witte
The fact that this game data proves so important for the development of robots is a “surprising development,” says Deborah Nas, professor of innovation at TU Delft. She explains why games are suitable as training material: “The footage has always been shot from the same perspective, so it’s very stable material. You also play such a game with a console. In the real world, robots are controlled with the same gaming console.”
OpenAI also saw this and reported to De Witte. At least, that’s how several media reports. De Witte cannot confirm that that company offered him a sum of $500 million, “but I can tell you what it’s like to turn down 500 million,” he says with a smile. “You look yourself in the mirror just as closely, and you slap your face a few times with some water and then you pick up the phone.”
In that telephone conversation he said ‘no’ to an amount that many entrepreneurs can only dream of. “When it happened, I immediately thought, my God, did I make the right decision?”
Nice to do it yourself’ De Witte decided: what the Americans can do, I can do myself. According to De Witte, the data that Medal.tv has is so unique that he expects to have a significant lead in further developing this AI technology. “We got to work and found out we can catch up with them. And yes, then you can become worth more than what they pay, right? And it’s just more fun to do it yourself.”
Six months after the million-dollar offer, De Witte still supports his decision. Investors are queuing up for his new company General Intuition. With that company, De Witte is now working on developing AI systems that mimic spatial understanding, the so-called ‘world models’. In a short time, De Witte has already raised more than 130 million dollars, and investors would still be willing hundreds of millions of dollars to invest.
“A lot is expected from these models, but it’s still a huge promise,” says Professor Nas. “And whether that promise is delivered in AI land is always something you need to see.”
Competition China and USA Developments in the field of AI have so far been mainly an American and Chinese matter. De Witte hopes to change that with his new company. “I think we have a really good chance from Europe, and we have one of the best opportunities from the Netherlands.”
Professor Nas is also hopeful. “I think this technology certainly offers an opportunity for Europe because we have a lot of talent. You need a lot of mathematics and physics to develop these types of models and we have very good university courses for that in Europe.”