Key points
- Discovery Loop seeks a valuation near $50 billion
- It discussed $10 billion in August
- Alphabet (GOOGL) is a founding investor and cloud partner
- The company is hiring for one job
Jeff Dean left Google five weeks ago to start Discovery Loop, and investors are already being asked to value it at around $50 billion. That's five times the $10 billion figure the company was discussing in August, according to Business Insider, which reported the new round on Friday citing people familiar with the matter. Alphabet (GOOGL) backed the company at launch and supplies its cloud computing. The startup has no announced product, no announced customers and one open job listing.
The reported valuation target
Business Insider's August 13 story said Discovery Loop was in talks for $1 billion at a valuation of around $10 billion. Friday's story gives the new valuation target but not the amount, the lead investor or a timeline. It also carries a warning that the terms could change and there is no guarantee the company raises at that number. A representative for Dean declined to comment in August.
For comparison, the same outlet points to two other Google alumni. David Silver, formerly of DeepMind, raised $1.1 billion this year at a $5.1 billion valuation for Ineffable Intelligence. Misha Laskin and Ioannis Antonoglou, also from DeepMind, reportedly raised for Reflection AI at a $25 billion pre-money valuation. Business Insider does not say whether the $50 billion figure is a pre-money or post-money valuation, so the comparison with Reflection AI is directional rather than exact.
Five times in a month, for a company with one job posting. I've watched a lot of AI rounds get repriced, but I still read that twice. That doesn't mean the investors are wrong. It means the price of a Jeff Dean is being set by who wants in.
What the company says it will do
Discovery Loop is a public benefit corporation, a for-profit company whose charter lets it weigh a public purpose against shareholder returns. At a talk in August, Dean said he and his co-founders "might make decisions that are not in the company's financial interest but are in the broader societal good," according to Business Insider.
The mission, as Dean wrote on X, is "to automate machine learning, science, and engineering to accelerate discoveries and progress." The method is scale. "Our general approach is to automate the experimental loop," he wrote in a note for Radical Ventures. "We think this approach is broadly applicable across many different fields of science and engineering." The company describes running thousands of experiments in parallel, scoring the results and going again with less human involvement each cycle.
Neither the August 5 announcement nor Dean's note names a customer, a product or a revenue figure.
Who the founders are
Dean is the name that sells the round. He joined Google in 1999 as its 30th employee, co-founded Google Brain, worked on the Gemini models and left as chief scientist after 27 years, TechCrunch reported on the day he quit. He's Discovery Loop's chief executive.
The other three founders spent their careers building the systems Google runs on. Sanjay Ghemawat also joined in 1999 and built much of Google's distributed computing infrastructure alongside Dean. Quoc Le was a founding member of Google Brain. Oriol Vinyals was a leading researcher at Google DeepMind in sequence modeling and reinforcement learning. Dean wrote that the four have worked together for between 14 and 30 years. A pitch deck he posted on X lists Google Search, Ads, Gemini and Gmail among the products they touched.
Who is already in
The first round was led by Radical Ventures and Khosla Ventures. Lightspeed, Kleiner Perkins and Doerr Capital also participated, according to the company's August 5 release, which did not give an amount. Alphabet's role as founding investor and cloud partner was confirmed by Google chief executive Sundar Pichai.
Vinod Khosla has already told people what he thinks this is. In a post on X on August 5 he wrote that "this team can redefine how research in engineering and science is done. If successful, the impact won't be measured in software features or benchmarks but in new science to improve billions of lives." Eight minutes later he posted again: "Could be larger than the impact that Google had when we backed them in 1998 at Kleiner."
Five weeks after launch, Discovery Loop has a valuation target, prestigious backers and an ambitious mission. It has not disclosed a product, a customer, a revenue figure or the size of the round.
Cover photo: Purdue Engineering / Wikimedia Commons, CC BY 3.0, cropped.
