← News

ResearchSources: huggingface.co · 29 Oct 2026

Hugging Face launches Project ProtoStar to make publicly funded research AI-ready

A new initiative with universities and open science groups aims to release publicly funded datasets, models, and code in formats AI systems can use, starting with a climate forecasting challenge at Carnegie Mellon University.

In plain words

Hugging Face and several universities have started Project ProtoStar, an effort to publish government-funded research data in formats that AI tools can find and work with. The first test is a challenge where teams use past weather records from 210 U.S. cities to predict later conditions. Results will show whether AI actually beats simple statistical methods on this task.

What was announced

Announced by Hugging Face, the initiative brings together founding partners including Carnegie Mellon, Purdue, Georgia Tech, Arizona State, and George Washington universities through the Open Forum for AI, plus the Astera Institute and the Open Research Community Accelerator. Participants sign a memorandum of understanding to release suitable scientific artifacts in AI-ready formats on open infrastructure. The organizers argue that public availability alone does not guarantee usability: datasets and code are often scattered across incompatible systems or lack the documentation and metadata needed for reuse, including by AI systems.

The first challenge, hosted by Carnegie Mellon, runs November 2–13, 2026. Teams use climate records through 2018 to forecast four indicators for 2019–2026, including average temperature and cooling degree days, with predictions checked against observed weather and simple statistical baselines. Participants also propose a short climate adaptation policy idea and submit outputs as a Hugging Face dataset. Final results are due in early 2027, once complete 2026 weather records exist. The organizers state the challenge series is meant to identify technical barriers and documentation gaps, and they cite growth on their own platform—over three times as many new science datasets shared in the first nine months of 2026 as in the same period of 2025—as evidence of demand.

What it does not show

  • The announcement comes from Hugging Face, an initiative participant and infrastructure partner, so the framing and platform statistics are not independently confirmed.
  • No challenge results exist yet; final forecasting outcomes are planned for early 2027, and the claimed benefits of AI-ready release remain untested.
  • The source describes the initiative's goals and early commitments but does not document how many organizations have signed the memorandum of understanding or what resources have already been released under it.

For security and IT teams

Scope
Project ProtoStar covers release of publicly funded datasets, models, and code in AI-ready formats, with challenges hosted sequentially by participating institutions; future challenges will be announced on the Project ProtoStar page on Hugging Face.
Access
Organizations interested in joining can contact [email protected], per the announcement.

Sources

Source-support review completed 14 Nov 2026. This story was drafted with AI assistance and checked against retained source passages. Automated checking is not independent verification.

Updates and corrections

No corrections so far.

First published 14 Nov 2026

New to the subject? AI from the Ground Up starts from the beginning.