Announcing Project ProtoStar: Making Publicly Funded Science AI-Ready

Hugging Face has announced Project ProtoStar, an initiative published October 29, 2026 that aims to make publicly funded datasets, models, and code easier to find and reuse, especially for AI-driven research. Founding partners are Carnegie Mellon, Purdue, Georgia Tech, Arizona State, and George Washington University, working through the Open Forum for AI, along with the Astera Institute, the Open Research Community Accelerator as community partner, and Hugging Face as infrastructure partner. Members sign a memorandum of understanding backing releases in AI-ready formats on open infrastructure.

The company notes that the U.S. government spends about $190 billion a year on research and development, yet shared results often sit in incompatible systems or lack documentation. It says its platform hosts more than 17,000 public science datasets, close to 29,000 counting medicine and mathematics, and that new science dataset uploads in the first nine months of 2026 more than tripled from a year earlier. NIH's NCBI repositories there have passed 11 million downloads.

ProtoStar will run a series of open challenges hosted by participating institutions. The first, led by Carnegie Mellon from November 2 to 13, asks teams to use weather records through 2018 from 210 U.S. cities to predict 2019 to 2026 conditions: average temperature, total precipitation, the hottest day, and cooling degree days. Entries are measured against simple statistical baselines, and teams also propose a short climate adaptation policy idea. Final results are due in early 2027.

The organizers say the challenges should show whether open scientific resources can be readily reused and reveal which barriers remain for future infrastructure, funding, and policy.