Mark Zuckerberg and Priscilla Chan’s nonprofit Biohub is partnering with the US government, Google DeepMind, Meta and Isomorphic Labs in a $1.8 billion effort to develop open AI biology datasets. The initiative, described as the “Virtual Biology Initiative,” aims to create a “virtual cell” that can help predict how human cells respond to diseases and potential treatments.
Outlets report that the collaboration focuses on generating data to train AI models, with the goal of speeding research and supporting new approaches to disease prevention and therapy development. The Wall Street Journal and others describe investment talks involving the US Department of Energy and the National Institutes of Health. The Verge and other coverage also notes that Google DeepMind, Meta and Isomorphic Labs collectively commit $300 million.
Biohub says it expects to release its first open dataset in about a year, with predictive modeling capabilities arriving in roughly the next five years. Across sources, the shared emphasis is on building digital tools and datasets rather than replacing laboratory testing immediately, by enabling researchers to “ask” and “predict” biological outcomes digitally.