Renumbered. This article was previously published as KB001 (AI/Cloud Access Request). Old links forward here.
Use this guide if you want AI model access for research, exotic hardware not available on campus, archive storage, or a Google Cloud project.
The basic rule: projects belong to a lab
Ursa Major does not create individual accounts. Every cloud project is anchored to a PI's lab, for billing and for security. Students and lab members are added to their PI's project with the PI's approval.
How a request is handled
| Your situation | What happens |
|---|---|
| Your PI already has an Ursa Major project | Research Computing asks your PI, by email, to approve adding you. Once the PI approves, you are added to the project. |
| You are a student and your PI has no project yet | The request goes to Research Computing, who contact your PI. |
| You are a PI requesting a new project | Research Computing meets with you to understand the work and set the project up in the right tier. |
What to include in your request
- Your name, NetID and department, and your PI's name.
- What you want to do (for example, use AI models in an analysis pipeline, run work that needs TPUs, or archive raw data).
- The data involved, and its protection level if you know it.
- If GPUs or other recharged resources are likely, a funding source (COA).
Send it to research-computing@ucr.edu or through the UCR Support Portal.
After approval
AI access and any API keys are tied to the PI's lab. Tier 1 AI use counts against the lab's allowance; use in a recharged project is billed to that project. Treat an API key like a password. See KB002: When your project is ready.
Owner: Research Computing
Reviewed: 4 Oct 2026