Mark Zuckerberg is backing a major effort to use artificial intelligence to understand how human cells work. His nonprofit research organisation, Biohub, is leading a $1.8 billion initiative to build the biological data needed to train AI models that can predict how cells behave.
Announced on October 7, 2026, the initiative brings together Biohub, the US government, Meta, Google DeepMind and other research partners. The goal is to create AI models that can simulate biological processes and help scientists understand diseases, test possible treatments and speed up medical research.
The project could change how researchers study human biology. Instead of relying entirely on laboratory experiments, scientists may eventually be able to use AI to predict how cells respond to drugs, genetic changes and disease. However, building reliable digital models of living cells remains a major scientific challenge.
Zuckerberg’s Biohub Leads the $1.8 Billion AI Biology Project
Biohub, co-founded by Zuckerberg and his wife, Priscilla Chan, focuses on using scientific research and technology to help prevent and treat diseases. The organisation is now expanding its work through a large international effort to create datasets for AI-based biological research.
The initiative combines funding, existing scientific data, computing resources and new laboratory measurement technologies. These resources will help researchers collect detailed information about cells and train AI models to recognise biological patterns.
Biohub had initially committed $500 million to the Virtual Biology Initiative in April 2026. The expanded programme now represents a total commitment of $1.8 billion, bringing together several organisations and public research resources.
The partners include the US Department of Energy, the National Institutes of Health, Meta, Google DeepMind and Isomorphic Labs, a company focused on AI-based drug discovery. Meta, Google DeepMind and Isomorphic Labs are jointly investing $300 million in the effort.
The US Department of Energy will contribute more than $500 million over five years. The National Institutes of Health will also help organise and standardise biological datasets and research resources built through more than $500 million in earlier federal funding.
ALSO READ: U.S. Government and Google Join $1.8 Billion AI Biology Initiative
How AI Could Create Virtual Cells to Predict Human Biology
A virtual cell is a digital model that aims to predict how a real cell behaves under different conditions. It would use large amounts of biological data to learn how genes, proteins and other parts of a cell interact.
Human cells perform different functions across the body. Some help fight infections, while others carry oxygen, transmit signals or repair damaged tissue. Their behaviour can also change because of disease, medication or genetic changes.
Scientists want AI models to learn these relationships. If the models become accurate enough, researchers could use them to predict the effects of certain changes before testing their ideas in a laboratory.
For example, a scientist could investigate how a particular cell might respond to a drug. An AI model could estimate the likely response and help identify which experiments deserve further testing.
This would not eliminate the need for laboratory research. Instead, it could help scientists choose promising experiments and avoid spending time on approaches that are less likely to work.
Biohub is working towards a model that can predict cellular behaviour across different biological conditions. The broader ambition is to create a system that helps researchers study living systems through digital experiments.
How Biological Data Could Shape the Future of AI Research
One of the biggest challenges in developing AI for biology is the lack of sufficiently detailed and standardised data. AI models used for language tasks can learn from enormous collections of text. Biological systems are more complicated. Cells contain many interacting components, and their behaviour can change depending on their environment and condition.
Researchers need detailed measurements to understand these interactions. They must collect information about different cell types, gene activity, molecular processes and responses to external changes.
The Virtual Biology Initiative aims to address this problem by generating more biological data and making it useful for AI training. It will also support new methods for measuring and imaging biological processes.
The quality of this information will be important. More data alone will not guarantee accurate predictions. The datasets must capture the complexity of biology and allow researchers to test whether an AI model’s predictions match real experimental results.
Biohub plans to make the resulting datasets available to the wider research community. However, commercial partners will receive an initial period of exclusive access before the data is released publicly.
This arrangement could help attract private investment while supporting the project’s longer-term goal of making biological research resources available to scientists worldwide.
How AI Cell Simulations Could Help Medical Research
The project could have several applications if researchers succeed in building reliable predictive models.
- Drug discovery: Scientists could use virtual cell models to estimate how cells might respond to potential medicines. This could help researchers prioritise promising drug candidates before conducting more expensive experiments.
- Disease research: AI models could help researchers study how changes in genes and cellular processes contribute to diseases. Better predictions could offer new clues about why certain conditions develop.
- Understanding treatment responses: Different cells can respond differently to the same treatment. More accurate models could help scientists investigate these differences and identify possible reasons for treatment failure.
- Faster scientific experiments: Researchers could use digital simulations to explore multiple hypotheses before choosing which ones to test in a laboratory. This could help them use research time and resources more efficiently.
These benefits remain potential outcomes, not guaranteed results. Scientists will need to demonstrate that the models can make reliable predictions across different cell types and biological conditions.
The technology also will not automatically produce new medicines. Drug development involves many stages, including laboratory experiments, safety testing and clinical trials involving people.
ALSO READ: Claude AI Helps Anthropic Identify a Previously Unknown Enzyme System
Zuckerberg Is Expanding His Focus on AI and Biology
The initiative reflects Zuckerberg and Chan’s growing focus on biological research. Their philanthropic work has increasingly centred on combining AI, computing and experimental science to improve the understanding of disease.
In November 2025, Biohub announced a broader effort to combine AI research with advances in biology. Its goals included developing virtual cell models, improving biological imaging and studying ways to use the immune system to detect and treat disease.
The expanded $1.8 billion initiative takes that work further by bringing government agencies and major technology companies into a coordinated effort to build the data infrastructure required for biological AI.
Meta’s involvement also connects the project to the wider technology industry’s interest in applying AI beyond chatbots and software development. Companies are exploring how advanced models could support scientific discovery, including research into proteins, medicines and human biology.
However, biological research presents different challenges from those found in many conventional AI applications. A model that produces convincing answers is not necessarily one that correctly predicts what will happen inside a living cell. Experimental evidence will be essential to establish whether these systems can deliver useful results.
Can Zuckerberg’s $1.8 Billion AI Bet Transform Biological Research?
The $1.8 billion initiative represents a significant commitment to building the foundations for AI-based biological research. Its success will depend on the quality of the data, the accuracy of the models and the ability of scientists to verify their predictions.
If the approach works, researchers could gain new ways to investigate disease and test scientific ideas. It could also help them decide which laboratory experiments are most likely to produce useful results.
Still, creating a reliable virtual cell is a difficult task. Human biology involves complex interactions that change across tissues, environments and stages of disease. Capturing these processes in a digital model will require extensive research.
For Zuckerberg and Biohub, the immediate goal is to build the scientific resources needed to make such predictions possible. Whether those resources eventually lead to major advances in treatment will depend on what researchers can demonstrate through further experiments.
The project marks a major bet on the future of AI in science. Rather than using AI only to process information, researchers want to build systems that can help predict how living cells behave. If those predictions prove accurate, the technology could give scientists a powerful new tool for understanding human disease.
