Biohub and federal agencies expand virtual biology initiative

Biohub, DOE, NIH and new funding partners announced a major expansion of the initiative, which was announced in April 2026.[1] The partners say it will create an open research resource to support understanding and treatment of human diseases.[1] The goal is to let researchers use AI models to ask and answer biological questions digitally and make predictions.[1] Biohub describes the commitment as the largest coordinated effort to generate AI-ready biological data to date.[1] Rives called the effort the beginning of a new scientific paradigm with AI.[2]

The $1.8 billion figure covers a combined commitment of funding, data, computation and measurement technology.[1] Google DeepMind, Isomorphic Labs and Meta are collectively investing $300 million in the Virtual Biology Initiative, which Biohub says will help develop technologies and multimodal datasets for predictive models of life.[1] DOE plans to contribute more than $500 million over five years through the Genesis Mission.[1] Biohub’s founding $500 million commitment anchors the work, with $400 million supporting new technologies that expand what biologists can measure.[1]

The initiative will expand measurements of cell responses to interventions across more cell types and conditions than have previously been studied.[1] Partners also plan to build and validate technologies to study cells and their interactions at greater scale, speed and accuracy.[1] Biohub will work with NIH to standardize datasets for AI model training.[1] The initiative also plans to use shared standards, common identifiers and a single access point to make partner datasets work together.[1] NIH will coordinate datasets, repositories and knowledge bases developed through more than $500 million in prior federal investment aligned with the initiative.[1]

Commercial partners will have exclusive access for one year to data they develop before it is shared publicly.[2] Rives said that this embargo means data becomes broadly available to scientific efforts very rapidly.[2] He also said researchers should be able to train models and assess their capabilities within a year of receiving the first large-scale dataset.[2] NIH Deputy Director Nicole Kleinstreuer said universal cell models could predict how any cell responds to an intervention.[3] Biohub says the models could eventually help scientists investigate aging, regeneration and the molecular mechanisms behind Alzheimer’s disease.[2]