Summary
On October 7, 2026, Biohub, the nonprofit research institute founded by Mark Zuckerberg and Priscilla Chan, announced with the U.S. Department of Energy and the National Institutes of Health an expansion of its Virtual Biology Initiative, a program to generate biological data for training AI models that predict how cells behave. The partners put the combined commitment at $1.8 billion in funding, data, computation and measurement technology: more than $500 million over five years from the Energy Department, datasets built with more than $500 million in prior NIH funding, $300 million jointly from Google DeepMind, Isomorphic Labs and Meta, and the $500 million Biohub committed in April. Biohub called it "the largest coordinated commitment to generating AI-ready biological data to date." Commercial funders get a period of exclusive access to the data they pay for before it becomes public.
What Happened
Biohub published the announcement at 13:00 UTC on October 7, and Reuters, Axios and The Verge reported it the same day. Biohub launched the Virtual Biology Initiative in April 2026 with a $500 million commitment of its own: $400 million for measurement technology, including cryo-electron tomography, microscopy that can image millions to billions of cells in living tissue, and tools to engineer and perturb biology, and $100 million for research outside Biohub. The expansion adds three kinds of partner: two federal science agencies, three AI companies, and a group of scientific institutions.
The Energy Department's contribution runs through the Genesis Mission, the cross-agency AI-for-science initiative it leads. According to Biohub, the department will spend more than $500 million over five years on cell research, drawing on exascale supercomputing, X-ray and neutron scattering, cryo-electron microscopy and tomography, and autonomous laboratories across the national laboratory system. The NIH contribution is mostly existing data: through what the release calls its "Bio Genesis Mission," the agency will coordinate datasets, repositories and knowledge bases developed with more than $500 million in earlier federal funding, including repositories catalogued by the National Library of Medicine and the National Center for Biotechnology Information, and Biohub will standardize them for AI training. Darío Gil, the Energy Department's Under Secretary for Science, described the partnership as "a critical step forward in leveraging artificial intelligence for public benefit."
Google DeepMind, Isomorphic Labs and Meta are jointly investing $300 million; the release does not say how the amount is divided among them. Isomorphic Labs president Max Jaderberg said the company was joining "as a founding member." Pushmeet Kohli, Google DeepMind's vice president for AI for Science, said the effort would help create "an open, standardized data commons." The Allen Institute, Broad Institute, Gladstone Institutes, Human Cell Atlas, Human Protein Atlas and Wellcome Sanger Institute joined to help organize the scientific strategy; Nvidia will provide computing infrastructure, software and technical support, and Renaissance Philanthropy is helping raise more money for data generation.
Biohub's head of science, Alex Rives, told Reuters that commercial funders get "embargo periods where there's a period of time where the groups can work on the data, and then it becomes available as a public scientific resource," and that the government-funded work will carry no such restriction. Axios reported that the commercial partners will have one year of exclusive access to the data they develop. Rives told Reuters that current cell datasets contain hundreds of millions of cells while an accurate predictive model will need billions and eventually trillions; that the partners aim to have a first dataset in about a year; and that he expects accurate predictive models within five years. He also said Biohub plans to approach pharmaceutical companies and philanthropies next.
Why It Matters
The initiative funds data generation, on the premise stated by its partners that predictive models of cells need experimental data "at an unprecedented scale," in Kohli's words, beyond what any single organization can produce. It brings the federal Genesis Mission, two frontier AI labs and an AI drug-discovery company into one data effort led by a philanthropy. Reuters placed it alongside other AI-lab moves into biology, including Anthropic's wet lab and a grant program of more than $125 million from the OpenAI Foundation for biological and medical datasets. When Biohub launched the initiative in April, Rives told Axios that one of the main open questions was whether cell biology would show the same scaling behavior seen elsewhere in AI, with models improving predictably as training data grows; the project is in effect a test of that question at a larger scale.
The $1.8 billion figure combines new money with in-kind and earlier spending: Biohub's April commitment, computation and lab work at the national laboratories, and NIH datasets paid for by more than $500 million of previous federal funding. The announcement does not say how much of the total is new cash. The one-year exclusive window for commercial funders sits alongside the description of the data as an open resource, and the terms of that access have not been published. Biohub's timelines for a first dataset and for accurate predictive models are its own projections, and whether a predictive "virtual cell" can be built from this kind of data remains unproven.
§ How to read the metadata
- Landmark
- Fundamentally alters the trajectory; 2–5 per year.
- Major
- Meaningfully shifts the landscape; 2–4 per month.
- Notable
- Worth documenting; significance can be upgraded later.
- Confidence
- High = primary sources corroborate. Medium = credible secondary only. Low = provisional. Disputed = credible sources disagree.
- Contestation
- Uncontested = no formal challenge. Contested = at least one challenge open. Superseded = replaced by a later entry. Unresolved = dispute still open.