Biohub, the US Department of Energy and the National Institutes of Health announced on 7 October a $1.8 billion expansion of the Virtual Biology Initiative, an effort to generate biological data that AI models can learn from. Biohub calls it the largest coordinated commitment to AI-ready biological data so far.
The total is built from several parts. DOE will invest more than $500 million over five years in lab measurement, modelling and computation. Google DeepMind, Isomorphic Labs and Meta are together adding $300 million. Biohub’s own founding commitment of $500 million, announced in April, anchors the effort.
The NIH share works differently. NIH will coordinate datasets and repositories built through more than $500 million in prior federal investment, and Biohub will standardise them for model training. That counts existing work, not new spending, which is worth knowing when reading the headline figure.
The aim is a so-called virtual cell: a model that predicts how cells respond to drugs and other interventions, so some experiments can run on a computer first. Biohub says the result will be an open resource for researchers. The announcement gives no release dates and no access terms for the datasets, so how open it proves to be is still to be seen.