Biohub has expanded its Virtual Biology Initiative into a collaboration with the U.S. Department of Energy, the National Institutes of Health, Google DeepMind, Isomorphic Labs and Meta. The partners describe a combined $1.8 billion commitment to generate the measurements and standardized datasets needed to train AI models of cells.
That headline figure is not a single new grant. It combines Biohub’s $500 million commitment from April, $300 million from the three technology companies, more than $500 million that DOE plans to spend over five years, and biological datasets and repositories created through more than $500 million in prior federal funding that NIH will coordinate for the initiative.
Building data for a “virtual cell”
The project aims to measure how cells respond to interventions across more cell types and conditions than researchers have studied so far. Biohub says the resulting data could support predictive models that let scientists test biological questions digitally before choosing which experiments to run in a laboratory.
The planned work includes large-scale imaging, molecular and cellular engineering, and spatial transcriptomics, a method that maps molecular activity while preserving where it occurs inside tissue. DOE also plans to contribute national-laboratory computing and measurement systems, including exascale supercomputers and advanced microscopy.
The goal is an open, standardized data resource, but access will not be immediate in every case. Biohub science chief Alex Rives told Reuters that commercial funders will receive an embargoed period to work with relevant data before it becomes public. He said government-funded work running alongside it will not carry that restriction.
Rives expects the first dataset in about a year and accurate predictive models within five years. Those are targets for the collaboration, not results it has already achieved. For now, the initiative is a bet that coordinated experiments and much larger datasets can make the behavior of cells predictable enough for useful computer simulations.
