The OpenAI Foundation has launched Public Data for Health, its second science program focused on Life Sciences and Curing Diseases. Announced on September 15, 2026, the initiative aims to accelerate scientific progress by supporting the creation, preservation and broad availability of high-quality scientific datasets for researchers.
The Foundation says it is beginning the program with more than $125 million in grants across an initial group of nonprofit organisations and universities. The projects span multiple layers of scientific and health data, ranging from molecular information and drug development to epidemiology and regulatory knowledge.
The initiative is particularly focused on datasets that can help researchers use increasingly capable artificial intelligence systems to analyse biological information and support discoveries related to preventing and treating disease.
What Is Public Data for Health?
Public Data for Health is an OpenAI Foundation science program designed to support the development of scientific datasets that can be made broadly available to researchers.
The Foundation argues that while AI systems are becoming increasingly capable of analysing biological information at scale, many future breakthroughs will depend on having more high-quality observations and data available.
Some datasets with significant public value may not be created because individual institutions lack sufficient incentives or resources to fund them. The Foundation says its funding can help address this gap by supporting datasets that are difficult, expensive or otherwise unlikely to be developed through conventional funding models.
The program is focused on life sciences and health research across multiple diseases rather than concentrating on a single disease.
More Than $125 Million in Initial Grants
The OpenAI Foundation says it is supporting more than $125 million in grants through an initial tranche of nonprofit and university projects.
The grants cover different layers of scientific information, including molecules, epidemiology and regulatory knowledge.
Among the projects receiving support are OpenADMET, CTD Commons, and the University of North Carolina’s Initiative for Generative Immunotherapy.
These projects demonstrate the types of data initiatives the Foundation intends to support through the Public Data for Health program.
OpenADMET: Drug Development Data
OpenADMET will create open datasets, benchmarks and blinded competitions focused on predicting how small-molecule drug candidates are absorbed and distributed within the body.
The project aims to provide high-quality data that can be used to develop and evaluate AI models for predicting ADMET properties—absorption, distribution, metabolism, excretion and toxicity.
OpenADMET will collect extensive molecular data and connect it with information about drug transport. The project will also examine selected compounds in models of the human blood-brain barrier.
The resulting datasets are intended to give researchers around the world resources for developing and testing predictive models that could contribute to more efficient drug development.
CTD Commons: Preserving Regulatory Knowledge
Another initial grant will support CTD Commons, an initiative focused on preserving regulatory knowledge from failed or shelved drug development programs.
A Common Technical Document, or CTD, contains extensive information associated with an investigational drug, potentially including animal toxicology, manufacturing information and correspondence with regulators.
Much of this information does not appear in published scientific literature.
CTD Commons will investigate whether these records can be acquired and made openly available for research and analysis. The project aims to preserve information that could otherwise disappear and make it available to researchers and future drug development teams.
UNC Initiative for Generative Immunotherapy
The University of North Carolina will establish the Initiative for Generative Immunotherapy, another project supported through Public Data for Health.
The initiative will create multimodal, de-identified public datasets to support research into personalised cancer vaccines.
The project will examine information from hundreds of tumors and multiple cancer types. It aims to address gaps between tumor sequencing data and direct measurements of tumor surface proteins and immune responses.
By making the resulting data available to researchers, the initiative intends to provide resources that can support the development and evaluation of future personalised cancer vaccine approaches.
Three Areas of Data the Foundation Wants to Support
The OpenAI Foundation outlines three initial categories that will guide its approach to future Public Data for Health funding.
Connected Data
Connected datasets link information collected at different biological scales or through different methods.
The Foundation is interested in projects that deliberately connect multiple layers of biological information so researchers can better understand relationships between them.
The OpenADMET project provides an example, combining molecular profiling with drug transport measurements and other biological models.
Scarce Data
Scarce data refers to information that may be impossible to recreate once a particular opportunity has passed.
Examples can include biological samples from specific points in time, historical health information and records from research or drug development programs that may otherwise be lost.
The Foundation says it is interested in projects that preserve valuable data before it disappears or make existing private datasets broadly available while respecting privacy and consent.
Direct Data
The third area is direct data, referring to measurements that are closer to the biological or clinical states researchers ultimately want to understand.
The Foundation notes that scientific research has frequently relied on proxies because they were easier or cheaper to measure. Public Data for Health may support efforts to develop datasets or tools capable of measuring biological states more directly.
The UNC immunotherapy initiative illustrates this approach by focusing on direct measurements relevant to personalised cancer vaccines.
Data Accessibility and Privacy
A central principle of the program is making resulting scientific datasets as accessible as possible to researchers around the world.
The Foundation says data supported through its grants should be broadly available while maintaining appropriate protections for individual privacy and consent when human data are involved.
The initiative also encourages grantees to share data throughout their projects rather than waiting until the end, publish analyses through preprints and work with data users to assess the usefulness of datasets for important biological research questions.
Future Grant Opportunities
The Foundation describes the strategy outlined in its September 2026 announcement as an initial orientation and says it expects to update its approach as science and AI develop.
Researchers interested in potential future Public Data for Health opportunities are encouraged to consider whether their proposed datasets involve connected, scarce or direct data and whether the resulting resources could be broadly useful to the research community.
The Foundation specifically states that this section of its announcement is relevant to researchers interested in future grant opportunities.
Researchers with ideas for the program can contact the Foundation’s science team. However, the announcement does not provide a general grant application deadline or state that an open funding application round is currently available.
Final Thoughts
The OpenAI Foundation Public Data for Health program represents a major new funding initiative for scientific data creation and preservation. With more than $125 million in initial grants, the program is supporting projects involving drug development, regulatory knowledge and personalised cancer vaccine research.
Future opportunities are expected to focus on connected, scarce and direct datasets, with an emphasis on broad research accessibility and protection of privacy and consent.
Researchers interested in future funding should monitor official OpenAI Foundation announcements for additional information as the program develops.
