Skip to content

National Artificial Intelligence Research Resource (NAIRR)

What is NAIRR?

The National Artificial Intelligence Research Resource (NAIRR) provides researchers with hardware, software, models, and other resources to advance their scientific inquiry. These resources include access to public and private compute capacity, such as supercomputing clusters hosted at universities or cloud resources from Amazon or Google. This resource availability is critical if you have computational workflows that require a different type of resource or need to operate at a larger scale than is available on YCRC’s clusters.

NAIRR’s Research Resources page has a complete set of resources.

Accessing NAIRR resources

You need to submit a successful proposal to access NAIRR resources. A NAIRR proposal differs from a standard scientific proposal in that it places more emphasis on detailed specifications for specific computational requirements and less emphasis on scientific justification, hypotheses, method descriptions, or other typical components of a standard scientific proposal. The most common reasons for proposal rejection are a lack of a detailed description of resource needs and insufficient proof that your project requires the requested resources.

Scientists will likely be most unfamiliar with developing a detailed resource request and clearly communicating the need for the resources. For example, the review committee prefers to see granular breakdowns, by task, of compute needs, such as GPU hours or total number of tokens. In addition, successful proposals often include results from scaling tests or benchmarking that clearly demonstrate how your workflows will benefit from NAIRR resources.

You have several options for developing scaling tests and benchmarks. NAIRR offers start-up projects, which have less stringent proposal criteria and allow you access to a NAIRR testbed. You can use these resources to develop your workflows and produce scaling and benchmarking results. If your workflow already runs on YCRC clusters, we can help you develop scaling tests and benchmarks on our systems.

How to find help

The NAIRR website features informational documents and videos that provide a detailed description of the proposal development and submission process.

YCRC staff are also happy to help you with the proposal process, develop a resource request, or run benchmarks on our systems to provide quantitative evidence that you will benefit from NAIRR resources. Please reach out to us for a 1-on-1 consultation at research.computing@yale.edu.


Last update: November 5, 2025