US government, Google and Meta back $1.8B push for largest open AI-ready biology dataset
anshulkundaje · x · 2026-10-08
- The U.S. government, Google, Meta and the Zuckerberg-backed Biohub are launching a $1.8B push to build the largest open, AI-ready biological dataset ever assembled, aimed at models that predict how biology works and help treat disease.
- Breakdown: Biohub $500M; DOE $500M+ across measurement, modeling and compute; NIH $500M+ in federally funded datasets and repositories; Meta, Google DeepMind and Isomorphic Labs jointly adding $300M.
- Scientist Sri Kosuri praises the talent and compute (5X ENCODE resources in 5X less time) but remains wary of the 'virtual cell' direction, questioning whether such models can truly encode our understanding of cell behavior.
More from Venture
- AEO vs GEO: being the instant answer is not the same as being the source AI cites — MicahBerkley · 2026-10-08
- Investor: AI Today Feels Like 2021 DeFi — Doubt Benchmarks Until It Ships — subinium · 2026-10-08
- Peter Yang: AI Can Decompile and Rebuild Any Software, SaaS's Moat Is Eroding — petergyang · 2026-10-08
- Samsung posts most profitable quarter ever yet trades at 4x forward PE; analyst blames lazy models — firstadopter · 2026-10-08
- Will AI apps start training models for their day jobs? Genspark's Gen-1 Slides on MiniMax M3 hints yes — IdekWatName · 2026-10-08
- "You could've just built a neo lab": the app layer stacks every startup risk at once — signulll · 2026-10-08