The Public Data for Health program will fund universities and associations to build, preserve and open scientific corpus, from molecules to epidemiology.
On September 15, 2026, the OpenAI Foundation launched Public Data for Health, its second science program, with more than $125 million in initial grants. The money will go to nonprofits and universities responsible for building, preserving and making freely available quality scientific datasets.
The selected projects cover several levels of the chain: molecular information, drug development, epidemiology, regulatory knowledge. OpenADMET will build open datasets, benchmarks and blind competitions to test whether the models can predict the uptake and distribution of small molecules in the body — a necessary step to reduce the failure rate of drug candidates. At the University of North Carolina, the funds will finance direct measurements of tumor targets and immune responses on several hundred tumors and various types of cancers, the results of which will be published in depersonalized form.
The foundation had opened its life sciences activity with AI for Alzheimer's in April 2026. Where this first program targeted a single disease, the new one is intended to be transversal to all life sciences.
Read the full article on the original source
Go to the source →