Hugging Face, universities and Astera Institute launch Project ProtoStar to publish publicly funded science in AI-ready formats
A coalition (Carnegie Mellon, Purdue, Georgia Tech, Arizona State, George Washington University via Open Forum for AI; Astera Institute; Hugging Face as infrastructure partner; ORCA as community partner) launched Project ProtoStar. Participants sign a memorandum of understanding to release suitable scientific datasets, models and code in AI-ready, machine-actionable formats on open infrastructure. A first open challenge, Climate Risk Intelligence hosted by CMU, runs Nov 2-13, 2026, asking participants to predict urban climate indicators for 210 US cities for 2019-2026 from records through 2018.
Entities: Project ProtoStar, Hugging Face, Carnegie Mellon University, Purdue University, Georgia Institute of Technology, Arizona State University
1 primary
What happened
Hugging Face, Astera Institute and five US universities (Carnegie Mellon, Purdue, Georgia Tech, Arizona State and George Washington, via Open Forum for AI) launched Project ProtoStar, with ORCA as community partner. Participants sign a memorandum of understanding to release suitable scientific datasets, models and code in AI-ready, machine-actionable formats on open infrastructure. The first concrete step is a Carnegie Mellon-hosted challenge, Climate Risk Intelligence, running 2 to 13 November 2026. Entrants must predict urban climate indicators for 210 US cities for 2019 to 2026 using records up to 2018. No datasets, funding amounts or release deadlines beyond that challenge have been published.
Why it matters
If it works, researchers and developers get publicly funded data that is easier to find and reuse, and funders and regulators get evidence about what blocks reuse. The problem is real: legacy formats, scattered systems and poor documentation limit reuse of research data. For now, though, the commitment is to release what is "suitable", which leaves partners wide discretion. Practical impact depends on what is actually released, and nothing is yet named.
What is noise
The roughly $190B a year of US federal R&D is context for the problem, not something ProtoStar touches. The initiative has no stated funding, targets or enforcement, and the partnership framing is largely promotional for now. Hugging Face is both announcer and infrastructure beneficiary, and no federal agency is listed as a participant despite NIH and NSF being mentioned.
Watch next
- 01Whether the Climate Risk Intelligence challenge (2 to 13 November 2026) draws meaningful participation, and whether CMU publishes the number of entrants, submissions and results afterwards.
- 02Whether partners actually publish named datasets, models or code on Hugging Face under the memorandum of understanding within six to twelve months, and how many are new rather than already public.
- 03Whether federal agencies such as NIH or NSF, or other funders, formally endorse, fund or adopt ProtoStar's formats, and whether more universities beyond the initial five join.
Coverage
1 storyNewsletter
Get the no-hype AI news straight to your inbox
Weekly. No spam, unsubscribe any time. Privacy
More infrastructure signals
Full feed →- Deepseek releases V4.1-Flash, an open-source model that sharply cuts KV cache memory and input-processing compute for AI agents10 Sept 202682
- Mistral releases Mistral Large 4, a ~1T-parameter open-weight multimodal MoE model, with day-0 Hugging Face support29 Oct 202680
- Anthropic files S-1 with SEC disclosing enterprise customer counts, cross-product adoption and H1 2026 revenue of $16.3B26 Oct 202680
- Anthropic files public S-1 with SEC for Nasdaq IPO under ticker "ANT"26 Oct 202680