The US government, Google DeepMind and Meta put $1.8 billion behind Biohub's plan to build a "virtual cell"

Biohub, the nonprofit research institute backed by Mark Zuckerberg and Priscilla Chan, said on Wednesday, Oct. 7, 2026, that it, the U.S. Department of Energy, the National Institutes of Health and a group of tech companies are committing $1.8 billion in money, data, computing and lab technology to generate open biological data for training AI models. The goal is software that can predict how a cell reacts to a drug or a change before anyone runs the experiment.

AI got good at language because the internet handed it trillions of words for free. Biology has no internet like that, and this is the most serious attempt yet to build one: an open library of how real cells react when you poke them, so models can learn to predict it. The interesting part is who's paying. The federal government and rival AI labs (Google DeepMind and Meta, on the same check) are pooling money for data anyone can use, which is close to the opposite of how the AI race usually works. Read the math carefully, though: about $500 million of the headline is NIH data that already exists, so the truly new money is closer to $1.3 billion. Still a lot of money. And if a "virtual cell" ever works, drug testing could start in software instead of in a lab dish. That's a big if, and nobody here is promising a date.

On Wednesday, 7 October 2026, Biohub published “International, cross-sector collaboration commits nearly $2 billion to build foundational data for AI models to predict and treat disease.” The page is dated October 7, 2026, and datelined Redwood City, California. Biohub, the U.S. Department of Energy, the National Institutes of Health, and new funding partners announced a $1.8 billion expansion of the Virtual Biology Initiative. That $1.8 billion covers funding, data, computation, and new measurement technology. Biohub calls it the largest coordinated commitment to generating AI-ready biological data to date. The data is meant to be an open resource for researchers. Those lines are Biohub’s. The National Institutes of Health published a matching note the same day. It confirms the partnership. It does not reprint the $1.8 billion total.

What “virtual cell” means here. The aim is software that can predict how a human cell responds to a drug, a mutation, or another change, before anyone runs that experiment in a dish. Biohub says the datasets would let researchers ask, predict, and answer biological questions digitally. That is the plan. The announcement does not say a working model exists today.

The Department of Energy’s piece is new money. DOE will invest more than $500 million over five years in lab measurement, modeling, and computation. The work runs through Genesis Mission, a cross-agency effort DOE leads. Biohub says that money draws on national-lab supercomputers, X-ray and neutron facilities, cryo-electron microscopy, and autonomous labs. Cryo-electron microscopy is a way to freeze a sample and take extremely detailed pictures of the molecules inside it. An autonomous lab is a lab where machines run experiments with less hands-on work from people. Those lines are Biohub’s.

NIH’s piece is mostly data that already exists. NIH will coordinate datasets, repositories, and knowledge bases built with more than $500 million in prior federal investment, under a program Biohub calls the Bio Genesis Mission. Biohub says it will help standardize those datasets for AI training. That $500 million is earlier federal spending, already on the books. It is not $500 million of new NIH money announced on Wednesday. Those lines are Biohub’s.

Google DeepMind, Isomorphic Labs, and Meta are collectively investing $300 million. Isomorphic Labs is the drug-design company Google DeepMind launched. Those three names share one $300 million figure. The announcement does not split the check company by company. That $300 million is Biohub’s.

Biohub’s own founding commitment is $500 million, from when the Virtual Biology Initiative launched in April 2026. Of that, $400 million is for new measurement technology: cryo-electron tomography, which images the inside of a cell in fine detail; microscopy that can image millions to billions of cells in living tissue; and tools to build and perturb biology, meaning change a cell or an organism on purpose so researchers can see what happens. The other $100 million funds research outside Biohub. Those lines are Biohub’s. They describe the April pledge, restated on Wednesday. They are not a second $500 million on top of that pledge.

Who else is joining. The Allen Institute, Broad Institute, Gladstone Institutes, the Human Cell Atlas, the Human Protein Atlas, and the Wellcome Sanger Institute are coming into the effort. NVIDIA will provide computing infrastructure, software, and expertise. Renaissance Philanthropy is helping raise more funding for data generation. Those names are Biohub’s.

Alex Rives, Biohub’s head of science, said: “An accurate predictive model of biology could dramatically accelerate scientific discovery by enabling scientists to perform experiments digitally.” That sentence is his, on the announcement.

Pushmeet Kohli, vice president of AI for Science at Google DeepMind, said: “We will not solve this challenge without open, experimental biological data at an unprecedented scale, showing how living cells behave and respond to changes.” That sentence is his, on the announcement.

What the announcement does not say. It gives no timeline for when a working virtual-cell model would exist. It names no specific disease and no specific treatment. Those absences are the page’s.

The picture is Biohub’s official Virtual Biology Initiative graphic: the initiative’s name in green type on a dark background. It is the image Biohub released with the Wednesday announcement. It is not a photograph of a cell, a lab, or a person. The frame does not print a calendar date.

In plain terms, Biohub said on Wednesday that it, the U.S. Department of Energy, the National Institutes of Health, Google DeepMind, Meta, and other partners are putting $1.8 billion in money, data, computers, and lab tools behind an open library of how cells behave, so AI models can learn to predict a response before anyone runs the experiment. More than $500 million of that is new DOE spending over five years. Another $500 million is NIH data that already exists. Google DeepMind, Isomorphic Labs, and Meta are putting in $300 million together. Biohub’s $500 million pledge from April is still in the total. Nobody on the page promised a date for a working virtual cell, or named a disease it would treat.