Learn About Discovery Loop's AI Products
Create a free account to access in-depth lessons on each tool and model.
Start Learning Free📋About Discovery Loop
Updated August 6, 2026Discovery Loop is an artificial intelligence research company founded in 2026 by four longtime Google researchers: Jeff Dean, Sanjay Ghemawat, Quoc Le and Oriol Vinyals, with Dean as chief executive. Their collective work spans Google's core infrastructure, including MapReduce and Bigtable, along with TensorFlow, AlphaFold and Gemini. Dean joined Google in 1999 as its thirtieth employee and left after twenty-seven years, most recently as the company's chief scientist.
The company is structured as a public benefit corporation, and its stated goal is to automate the experimental loop in science and engineering. Rather than a researcher designing one experiment at a time, Discovery Loop wants systems that can initiate, run and refine thousands of experiments in parallel, then feed the results back into the next round. Its first target is machine learning research itself, which is both the domain the founders know best and the one where a faster experimental loop compounds into every other application. Broader scientific problems are described as a later phase.
Alphabet is a founding investor and cloud partner, an unusual arrangement for a company built by departing employees. The initial round is co-led by Radical Ventures and Khosla Ventures, with Kleiner Perkins, Lightspeed and Doerr Capital participating; the company has not disclosed the amount raised or the valuation. As of its launch, Discovery Loop has no product, no published research and no announced timeline. What it has is a thesis about compounding research productivity and a founding team with an unusually strong claim to being able to test it.
