Summary
Biohub, US federal agencies and industry partners announced a combined $1.8 billion commitment to develop open, AI-ready biological data and measurement technologies. The initiative aims to support predictive models of cell behaviour; the announcement describes planned work, not a completed virtual cell.
Biohub, the US Department of Energy (DOE), the National Institutes of Health (NIH) and industry partners announced an expanded Virtual Biology Initiative on October 7, 2026. Biohub says the partners are committing a combined $1.8 billion in funding, data, computing and measurement technology to develop an open resource for training predictive AI models of biology.
The initiative’s goal is to help researchers model how cells behave and respond to interventions. The announcement describes a coordinated data and technology effort intended to enable those models, rather than a completed “virtual cell” or a demonstrated medical treatment.
What the commitment will support
The contributions include different kinds of support, not just new cash. DOE plans to invest more than $500 million over five years in fundamental cell research, measurement, modelling and computation. NIH will coordinate relevant datasets, repositories and knowledge bases developed through more than $500 million in prior federal investment aligned with the initiative. Biohub will work with NIH to standardise those resources for AI training.
Biohub’s founding $500 million commitment anchors the work: $400 million is for technologies that expand biological measurement, and $100 million funds research outside Biohub. The technologies include cryo-electron tomography, which can resolve near-atomic detail inside cells; microscopy designed to image millions to billions of cells in living tissue; and tools for building and perturbing biological systems at molecular, cellular, tissue and whole-organism levels.
Google DeepMind, Isomorphic Labs and Meta are collectively investing $300 million in the initiative. DOE also plans to draw on national-laboratory capabilities, including exascale supercomputing, X-ray and neutron scattering, cryo-electron microscopy and tomography, and autonomous laboratories.
Why the data matters for predictive biology
Biological datasets often capture different parts of a system and use different formats. For an AI model to learn how a cell responds to a change, researchers need measurements across cell types, conditions and interventions, as well as ways to relate those measurements to one another. The initiative plans to expand such cell-response data and develop shared standards, common identifiers and a single access point so that researchers can use datasets together.
The intended “virtual cell” is a computational model of biological behaviour: researchers could use it to ask questions digitally and generate predictions to investigate. If the models become sufficiently predictive, they could help scientists decide which questions or experiments to pursue. That is the initiative’s proposed scientific value; the announcement reports a plan to build the underlying data and capabilities, not results from a validated model or a disease-treatment outcome.
Biohub says the resource is intended to be open to the research community worldwide. Its effectiveness will depend on whether the measurements are sufficiently broad and consistent to support useful predictions across different cells and conditions.