The Chan Zuckerberg Biohub has expanded its Virtual Biology Initiative to a total of $1,8 billion. The project aims to develop AI models capable of predicting how living cells react to drugs, mutations, or diseases, effectively creating a virtual cell simulation.
The initiative secured $500 million from the US Department of Energy for technology support over five years. The National Institutes of Health is also involved, leveraging over $500 million in prior funding to standardize datasets for AI training.
Private sector partners, including Meta, Google DeepMind, and Isomorphic Labs, contributed $300 million. The broader collaboration includes NVIDIA, Tahoe Therapeutics, and the Arc Institute, focusing on integrating multimodal biological data.
In January 2026, researchers released a dataset containing over 120 million single cells and 225 000 perturbation interactions. This data is four times larger than the previous Tahoe-100M benchmark and will be made openly accessible to the global scientific community.
The project targets a five-year timeline for developing accurate predictive models. The initiative aims to significantly accelerate drug development, potentially reducing the process from decades to just five years.
