Empire AI access (web)

Citation

Empire AI. “Who We Are.” Web page with mission, history and FAQ, date not stated. https://www.empireai.edu/about/.

Cornell University AI Initiative. https://ai.cornell.edu/empire-ai-cornell-call-for-compute-resource-proposals.

Both fetched 2026-10-15 by the agent. Read 2026-10-15.

Bottom line

Empire AI is a New York State consortium that shares an AI computer among member institutions.
Empire AI’s public site does not publish an allocation policy. Each member runs its own call for its share. Cornell’s call is the public example.
At Cornell, the unit is the Service Unit (SU). “One SU corresponds to one hour of H100 compute”. The Provost pays until November 2026. “After that, an SU is expected to cost $0.50”.

Background

Governor Kathy Hochul launched Empire AI in 2024. The first system, Alpha, is at the University at Buffalo. Beta launched in 2026.

Gamma, the permanent facility, is “expected to be completed by the end of 2027”. The site has a “Registered User Access” link to a help desk. That help desk is for account holders.

Key claims

Mission: “Empire AI’s mission is to give New York’s academic researchers the advanced computing resources they need to pursue ambitious AI research in the public interest.” (Who We Are, Our Mission)
Members: “The Empire AI consortium includes Columbia University, Cornell University, CUNY, Icahn School of Medicine at Mt. Sinai, New York University, Rochester Institute of Technology, RPI, University of Rochester, SUNY, and the Flatiron Institute.” (Who We Are, FAQ)
Funding: Empire AI brings together “New York’s leading public and private research institutions, state government and philanthropic partners to build shared AI computing infrastructure for academic research.” (Who We Are, Our History)
Use: Alpha supported “more than 130 projects and hundreds of researchers across Empire AI’s member institutions”. (Who We Are, Our History)
Demand: “With more than 300 research projects queued to use the system, Beta represents a significant expansion”. (Who We Are, Our History)
Status of the Cornell call: “The application submission forms referenced in this story have been removed because the application period has concluded. New application opportunities are expected to open in 2027.” (Cornell call, Editor’s Note)
Eligibility at Cornell: “Any faculty member with PI status from Cornell, Cornell Tech, and Weill Cornell Medicine can apply.” (Cornell call)
Application at Cornell: “To apply for an allocation, please submit the two short forms that you find linked below.” (Cornell call)
Unit and term: “An allocation is the maximum number of SU you can use between December 2025 and November 2026. One SU corresponds to one hour of H100 compute, and one hour of B200 compute corresponds to two SU”. (Cornell call)
Cost: “This cost is supported by the Provost until November 2026. After that, an SU is expected to cost $0.50, which will provide Cornell PIs with a rate that is well below a typical AWS rate”. (Cornell call)
Scarcity: “We expect that demand for Empire AI will exceed Cornell’s share of Empire AI, at least until the “Gamma” expansion of Empire AI in 2027.” (Cornell call)

Evidence and method

The first page is the operator’s own. It describes the mission and the systems, not the access rules. The second page is a member university’s call for its share.

Limits

Neither page states the review criteria or who reviews at Cornell. Cornell removed the forms.
Neither page states reporting, acknowledgement or acceptable-use rules.
Neither page states how Empire AI divides capacity among members.
The Cornell call is one member’s process. Other members run their own calls.
The $0.50 SU price is “expected”. The page does not say who sets it.