Ex-DeepMind Researcher Lands $300M Valuation for Visual AI Startup Before Shipping a Single Product

Andrew Dai left Google months ago. He has yet to ship a product. Yet his new startup, Elorian, just closed a $55 million seed round at a $300 million valuation, backed by strategic investors including Nvidia and Menlo Ventures.

The size of the round, and the valuation attached to it, stands out even in a funding environment where pre-launch AI startups routinely command outsized bets. Elorian has no live product on the market. What it has instead is Dai's pedigree and a technical thesis centered on visual AI, with the company describing its long-term ambition as building toward what it calls visual AGI, a system capable of general-purpose visual reasoning rather than narrow, task-specific image processing.

Dai's résumé is doing a lot of the persuading here. Before founding Elorian, he worked as a researcher at Google DeepMind, where his work contributed to research that later fed into the development of ChatGPT. That lineage, tying a founder directly to foundational work behind one of the most consequential AI products of the past decade, has become a powerful signal for investors trying to identify the next major platform shift before it becomes obvious.

Why Visual AI, and Why Now

Most of the largest language models have already absorbed enormous volumes of text, and returns on scaling text-only systems are showing signs of flattening. Visual understanding, by contrast, is viewed by many researchers as comparatively underdeveloped: current systems can caption images or answer narrow visual questions, but they struggle with the kind of flexible, general reasoning about the visual world that humans perform effortlessly. Elorian is betting that closing that gap represents the next major frontier in AI capability, and that whoever gets there first will control a foundational layer other applications build on top of.

What the Backing Signals

Nvidia's involvement as a strategic investor is notable beyond the capital itself. The chipmaker has increasingly used strategic stakes in AI labs to secure early visibility into compute-intensive research directions and to shape which architectures get optimized for its hardware. Menlo Ventures' participation adds traditional venture credibility to a round that otherwise leans heavily on a single founder's technical reputation rather than product traction, revenue, or even a public demo.

For now, Elorian remains a bet on potential rather than proof. The company has not detailed a product roadmap, a launch timeline, or how it intends to differentiate its approach to visual reasoning from well-funded efforts already underway at larger labs. What it has secured is enough capital and enough investor conviction to spend considerable time building before it has to answer those questions in public.