The ex-DeepMind researcher who raised $300 million before his startup had a product
Andrew Dai spent over a decade helping build some of the most influential AI systems in the world. Now he thinks visual AI is where the next big leap happens, and investors are betting heavily that he is right.
Key points
- Andrew Dai raised a pre-seed funding round valued at $300 million before his startup had launched any product.
- Dai spent more than ten years at DeepMind, the AI research lab owned by Google, working on foundational AI research.
- His earlier research contributed to work that later informed ChatGPT, the AI chatbot developed by OpenAI.
- Dai's new company is focused on visual AI, meaning software that understands and works with images and video rather than just text.
Most startups raise a few hundred thousand dollars to build a first version of their product. Andrew Dai raised $300 million before his company had one.
Dai is a former researcher at DeepMind, the artificial intelligence lab owned by Google and known for building some of the most advanced AI systems in the world. He spent more than a decade there working on foundational research, the kind of deep scientific work that does not always have an obvious product attached. Some of that research later fed into the development of ChatGPT, the AI chatbot built by OpenAI that brought conversational AI to mainstream audiences in late 2022.
Now he is betting on a different corner of the field entirely.
Dai believes visual AI, software that can see, interpret, and reason about images and video the way a person can, is one of the next major frontiers in artificial intelligence. Most of the public attention over the past two years has gone to large language models, the technology behind text-based chatbots like ChatGPT and Claude. Visual AI has moved faster quietly, and Dai thinks it is close to a turning point.
The $300 million pre-seed valuation, first reported by TechCrunch AI, is striking because pre-seed is normally the very earliest stage of funding, when a company is little more than a founding team and a pitch deck. A nine-figure valuation at that stage reflects how much weight investors are placing on Dai's personal track record rather than any product they can test.
What does this mean for ordinary people?
For most people, visual AI is already quietly present: it powers the face recognition that unlocks your phone, the systems that scan luggage at airports, and the tools that help doctors read medical scans. If Dai's thesis is right, the next wave will be far more capable. Think software that can watch a video and answer detailed questions about it, or tools that help a shop owner automatically track inventory from a security camera feed.
None of that is available from Dai's company yet. What exists right now is a well-funded team with a strong research pedigree and a clear direction. The product still has to be built.
For anyone watching the AI industry, the headline number matters less than the signal it sends. Investors are willing to place enormous bets on researchers with deep expertise, even with nothing to show yet. That says something about where the competition in AI is heading: toward people who have spent years understanding the science, not just shipping software quickly.
