US agencies and tech companies join Biohub’s $1.8 billion biology data effort

Biohub announced on 7 October 2026 that the US Department of Energy and National Institutes of Health had joined an expanded Virtual Biology Initiative alongside Google DeepMind, Isomorphic Labs and Meta. Biohub puts the combined contribution at $1.8 billion in funding, existing data, computation and measurement technology. The partners intend to develop datasets for AI models that predict how biological systems respond to experimental changes.

DOE puts laboratories and computation to work

Biohub says the Department of Energy’s Office of Science will put more than $500 million into laboratory measurement, modelling and computation over five years. The contribution sits within the department’s Genesis Mission, a cross-agency programme. Biohub describes the work as fundamental cell research involving data collection, AI analysis and imaging, supported by facilities across the National Laboratory system.

The department’s part of the initiative draws on exascale computing, X-ray and neutron scattering, cryo-electron microscopy and tomography, and autonomous laboratories, according to Biohub. Darío Gil, the department’s Under Secretary for Science, said in Biohub’s announcement that the collaboration combines federal computing and experimental facilities with Biohub’s models, tools and biological data. The department’s commitment covers work to be undertaken over five years, rather than data produced by an earlier investment.

Biohub says the intended resource will record how cells respond to interventions across more cell types and conditions than have been studied so far. The partners also plan to develop and validate ways of measuring cells and their interactions at greater scale. Those measurements are the proposed training material for models that researchers could use to investigate biological questions digitally.

NIH brings earlier datasets into the initiative

The National Institutes of Health has a different role. Biohub says NIH will coordinate datasets, repositories and information collections produced using more than $500 million in earlier federal funding. Biohub will work with the agency to standardise that material for AI training. That earlier investment forms part of Biohub’s account of the $1.8 billion effort, alongside the Department of Energy’s five-year commitment and the other partners’ contributions.

Biohub says NIH will organise its contribution through its Bio Genesis Mission, drawing together existing biomedical datasets, research programmes and national data infrastructure. The resources it identifies include repositories catalogued by the National Library of Medicine and the National Center for Biotechnology Information, as well as NIH Common Fund programmes developing biological atlases, shared standards and datasets prepared for AI use.

The two agencies are thus contributing different things to the same planned resource: new measurement and computation from the Department of Energy, and coordination of biomedical material developed with earlier federal funding. Biohub’s stated task with NIH is to make those existing datasets suitable for model training.

Commercial funders receive an embargo period

Google DeepMind, Isomorphic Labs and Meta are collectively investing $300 million in the Virtual Biology Initiative, Biohub says. Biohub announced the initiative in April 2026 with its own founding commitment of $500 million. Of that amount, it says $400 million supports measurement technologies, including imaging inside cells, microscopy in living tissue and tools for changing biological systems. Another $100 million funds research outside Biohub.

The planned public resource comes with an access distinction. Biohub head of science Alex Rives told Reuters that commercial funders receive an embargo period in which they can work with data before its public release. He said government-funded work proceeding in parallel carries no such restriction. Biohub also plans to approach pharmaceutical companies and philanthropies, Rives said.

For researchers outside the funding group, those terms make the route by which a dataset is produced consequential to when they can use it. The embargo helps Biohub bring private money into what it calls an open-science project, Reuters reported. Priscilla Chan said in an interview with Reuters that Biohub has treated the work as a resource for a broader community rather than for one group.

Rives puts the first dataset about a year away

Rives said existing cell datasets contain hundreds of millions of cells. In his assessment, accurate predictive models would require billions of cells and eventually trillions. He expects the first dataset from the partnership in about a year and accurate predictive models within five years. Those are Biohub’s targets, rather than measured results from the expanded initiative.

The intended use is to let scientists examine possible experiments on a computer before choosing which to run in a laboratory, Axios reported. That ambition concerns the behaviour of living cells, beyond models of individual proteins or other biological components. Biohub says its work will gather observations of cellular responses under a wider range of conditions to train such models.

The planned measurements include spatial transcriptomics, which maps molecular activity within intact tissue, and screens that record how cells respond to environmental changes. Biohub also says its $400 million technology allocation supports cryo-electron tomography, microscopy capable of imaging millions to billions of cells in living tissue, and engineering tools that alter biology at molecular, cellular, tissue and whole-organism levels.