Google Chief Scientist Jeff Dean and AI Veterans Depart to Launch Discovery Loop
In one of the most significant talent defections in the history of Silicon Valley, Google Chief Scientist Jeff Dean has departed the company after nearly 27 years to co-found Discovery Loop, a public benefit corporation aimed at automating scientific research and engineering.
Announced on August 5, 2026, the new startup has assembled what is widely considered the most pedigreed founding team in the history of artificial intelligence, drawing directly from the upper echelon of Google and Google DeepMind’s technical leadership.
The Founding Powerhouse
Joining Jeff Dean (who will serve as CEO) are three of Google's most influential and historically significant technical minds:
- Sanjay Ghemawat: A Google Senior Fellow and Dean’s primary collaborator of nearly three decades. Together, Dean and Ghemawat co-created the foundational distributed systems that powered Google’s early growth and laid the technical groundwork for big data, including the Google File System (GFS), MapReduce, and Bigtable.
- Quoc Le: A Google Brain co-founder and deep learning pioneer who led Google’s AutoML-Zero project, which focused on using machine learning to autonomously build and optimize AI systems.
- Oriol Vinyals: Previously the Vice President of Research at Google DeepMind and a technical co-lead for the flagship Gemini model family.
The departure of this core group represents a massive structural blow to Google’s AI leadership at a highly competitive moment in the global frontier AI race, compounding recent organizational restructurings under new DeepMind SVP Koray Kavukcuoglu (see Google Overhauls DeepMind Leadership: Kavukcuoglu Takes Day-to-Day Reins as Hassabis Steps Back).
Automating the Scientific Method
Discovery Loop’s core mission is to replace human-intensive experimental loops in science and engineering with autonomous, high-velocity AI systems. The company plans to deploy algorithms that can propose hypotheses, design and execute experiments, evaluate results, and iterate thousands of times simultaneously in parallel without human intervention.
The startup plans to begin by automating machine learning research itself (recursive self-improvement) before expanding into physical sciences, including semiconductor chip design, molecular biology, drug discovery, and materials science.
$1 Billion Fundraising Target and Strategic Backing
Discovery Loop has secured substantial early backing, structured in a way that keeps the company connected to its former home:
- Strategic Investors: Alphabet is participating as a founding investor, and Google Cloud is signed on as a primary infrastructure partner, meaning the startup remains within Google's computing orbit.
- Venture Capital: The initial funding round is co-led by Radical Ventures and Khosla Ventures, with participation from Lightspeed, Kleiner Perkins, and Doerr Capital.
- Valuation Target: According to mid-August reports from Business Insider, Discovery Loop is in active talks to raise a massive $1 billion round at a $10 billion valuation, establishing a long development runway before a commercial product is even released.