US, Meta and Alphabet pour $1.8 billion into open AI biology data initiative
The US government, Meta Platforms and Alphabet are joining the nonprofit Biohub in a $1.8 billion effort to create open datasets for training AI models in biological research.
A partnership involving the US government, Meta Platforms, Alphabet and the nonprofit Biohub aims to build open biological datasets for AI training, bringing total investment to $1.8 billion. The Department of Energy will allocate more than $500 million over five years for laboratory measurement, modeling and computation, while the National Institutes of Health will manage datasets and repositories funded by over $500 million in prior federal grants.
Meta, Google DeepMind and drug-discovery startup Isomorphic Labs are contributing $300 million, and Biohub has already committed $500 million since April. The Virtual Biology Initiative will use techniques such as spatial transcriptomics to capture how cells react to a wide range of conditions, seeking to compress drug-development timelines. Data will be released publicly after an embargo period for commercial funders, and Biohub plans to approach additional pharma companies and philanthropies. The goal is to generate billions to trillions of cell measurements and develop accurate predictive models within five years.
Why it matters
Open AI-driven biology data could dramatically speed drug discovery and lower development costs.
In this story
