GitHub Copilot is now available to NASA-funded projects as a paid service managed by the Science Cloud. The Science Cloud handles the operational and security infrastructure, including identity and access management, operational controls, and subscription management. 

Security & Compliance Rules 

GitHub Copilot operates strictly as a FISMA Low system. Because Copilot learns from user inputs, it is critical that users keep their prompts and code clean.

  • PROHIBITED DATA: You must not process or commit Personally Identifiable Information (PII), ITAR, EAR, or Controlled Unclassified Information (CUI) to this system.
  • PERMITTED DATA: Content is strictly limited to unclassified, publicly releasable information, source code/documentation not subject to export controls, and non-CUI research and development work.

If you are in doubt about whether your data is safe to commit, stop and contact Science Cloud ISSO Tara Leonard or the Science Cloud support team.

Options & Pricing

We offer several subscription tiers to meet the specific needs of your project.

  • Pricing: Subscription tiers change often. Please contact support@sciencecloud.nasa.gov to discuss tiering options and obtain updated pricing.
  • Fees: All subscriptions are subject to a ~5% overhead fee to cover CCS (Cloud and Computing Services) and Science Cloud operations.
  • Term Length: A minimum 1-year subscription is required.

Requesting Access & Onboarding

To streamline processing, onboarding must be PI-led. Project members must coordinate with their Principal Investigator (PI) to submit bulk onboarding requests as a group, ensuring funding is confirmed prior to provisioning.

How to Sign Up

Submit a Ticket: Log in to the Jira Service Portal at https://jira.sciencecloud.nasa.gov/servicedesk/customer/portal/34 or send an email to support@sciencecloud.nasa.gov.

What to Expect After Requesting

  • Provisioning Timeline: Once funding is officially confirmed, access is typically provisioned within 1 to 2 business days.
  • Invitations: You will receive an organization invitation directly from GitHub. Please accept promptly; invitations expire after 7 days.
  • Authentication: Upon your first login to access enterprise features, you will be prompted to securely authenticate.

Funding & Billing Operations

Because GitHub Copilot is a managed cloud service, all funding must be properly obligated before resources are used.

  • Fund Transfers: Funds must be transferred to Ames Research Center (ARC), where CCS manages the Science Cloud contracts. Moving funds between centers typically requires your Resource Analyst (RA); we will provide an ARC point of contact to facilitate this.
  • Account Balances: Subscriptions operate on a monthly billing cycle and cannot be paused for partial periods. Organizations are not permitted to go into arrears; if funds are exhausted, the subscription will be suspended.
  • Notifications: Automated alerts will be issued to your team when funds drop below three months of coverage.
  • Overages: Exceeding the monthly premium quota (e.g., 300 requests on lower tiers) does not halt your subscription but will incur additional costs that reduce your remaining funded period. (Tip: The break-even point for upgrading from Copilot Business to Copilot Enterprise is 800 premium requests per month).
  • Early Termination: If a project ends early, our team will work with your PI and RA to de-obligate unused funds where possible.

NASA Success Stories & Use Cases

During the March 2026 showcase event, several NASA projects shared how they are leveraging GitHub Copilot to accelerate their pipelines:

  • NASA POWER (Streamlining Security and APIs): Used SI agents to identify and remove dead code for their Earth observation data project.
    • Results: Slashed loading times, saved weeks of developer time, reduced Docker image size by 50%, and improved hourly data response times by 62%. Cloud operating costs were reduced by 33% while serving 5% more daily requests.
  • Model Training & Manuscript Preparation (Langley Research Center): James Warner, PhD, profiled and optimized a model training workflow.
    • Results: Reduced execution time on four GPUs from 33 hours to 10 hours (3.3x speed-up). Standardized complex publication-ready visualizations in Python, reducing a 2+ hour task to 30 minutes (4x speed-up).
    • Tip: Reset expectations often as models improve, always verify outputs, and learn with your colleagues.
  • Infrastructure Modernization (GMAO): Alexey Shiklomanov, PhD, utilized Copilot’s OpenCode agentic coding tool in the terminal.
    • Results: In under 10 minutes, he analyzed project structure, created a public S3 bucket, uploaded a test image, generated a public URL, and destroyed the infrastructure—a process that normally takes hours.
  • Accelerating the Science Cloud: The internal Science Cloud team used SI-assisted workflows for development.
    • Results: Delivered 11 new features and ~33,000 lines of code, tests, and documentation in 40–75 engineer-days (down from a traditional 165–245 days). Cost and calendar time were reduced by 65–75% while achieving higher-than-typical documentation and test coverage.



  • No labels