Andrew Dai, a former Google DeepMind researcher, has secured $55 million in seed funding for his new venture, Elorian. The company aims to advance visual artificial intelligence and is valued at $300 million pre-product.

The substantial seed funding and high valuation for Elorian underscore investor confidence in the potential of visual AI and the expertise of its founder, signaling a significant new player in the race towards artificial general intelligence.
Andrew Dai, a former researcher at Google DeepMind, has launched Elorian, a startup focused on advancing visual artificial intelligence. The company successfully raised $55 million in seed funding at a $300 million valuation, achieved just months after Dai's departure from Google.
Dai, who has over a decade of experience in developing influential AI systems, including research that contributed to ChatGPT, believes visual understanding and reasoning represent a significant frontier in AI. He stated that current models excel in areas like math, physics, and coding, but progress in visual comprehension has been uneven. Elorian aims to develop models that push towards artificial general intelligence (AGI) with a focus on visual capabilities.
During the fundraising process, Dai emphasized refining a technical vision into an understandable narrative for investors. He opted for strategic partners like Nvidia and Menlo Ventures, valuing their understanding of frontier AI development over maximizing the valuation. The conversation highlighted practical advice for founders in the AI space, including communicating complex technical concepts, the importance of speed as a competitive advantage, and strategies for recruiting top AI talent from major tech companies.
The podcast episode also explored what top venture capital firms seek in frontier AI startups, the benefits of prioritizing strategic investors over the highest valuation, and how founders can build lasting advantages in the evolving AI landscape.
Pick the topics you care about. Get only what matters, on your cadence.