
Discovery Loop reportedly targets a ~$50B valuation for new fundraising, five times the $10B figure floated weeks earlier.
The AMW Read
Pre-product Discovery Loop’s reported $50B ask (from $10B weeks earlier) plus named Google/DeepMind founder exits introduce a top-tier agentic-research entrant at frontier-lab capital scale.
Discovery Loop reportedly targets a ~$50B valuation for new fundraising, five times the $10B figure floated weeks earlier.
Business Insider, citing people familiar with the talks, reported that Discovery Loop is pursuing fresh capital at roughly $50 billion — about 67 trillion won. Terms can still change, and a close at that level is not assured. As recently as last month, the two-month-old, pre-product startup was said to be discussing a $1 billion raise at a $10 billion valuation. The company was founded by former Google chief scientist Jeff Dean with Sanjay Ghemawat, Quoc Le, and DeepMind’s Oriol Vinyals; it pitches AI that runs thousands of experiments in parallel to automate scientific and engineering discovery. An early round was led by Khosla Ventures and Radical Ventures, with Lightspeed, Kleiner Perkins, and Door Capital participating, and Google as a founding investor and cloud partner.
That jump extends August’s Google/DeepMind talent exit into a capital-market story: pedigree alone is being priced like a frontier lab before any product ships. Vinod Khosla has compared the firm to early OpenAI, arguing success would show up as scientific discovery rather than software features or benchmarks. Per the AI Market Watch index, Discovery Loop is already tracked under AI Agents with undisclosed funding among roughly 5,000 companies — coverage, not a census — which makes a $50B ask an extreme outlier for its stage.
For investors and builders, treat the headline as a soft ask until a round closes: if it sticks, peer discovery-automation and research-agent startups will face a new valuation floor set by researcher brand and hyperscaler partnership rather than revenue; if it slips, the episode still marks how fast Google-origin talent can reprice the agentic research layer.
