🌐 KB008: UCR Research Computing Guide to NAIRR Pilot Allocations & National AI Resources
Scope: Research Computing Services Audience: UCR Faculty, Postdocs, Researchers & Students Last Updated: August 19, 2026
Category: AI / Machine Learning / Federal Allocations
Service Tier: Tier 3 National Infrastructure (Federal Bridge)
Target Audience: UCR Faculty, Postdocs, and PhD Researchers seeking zero-cost H100 GPU clusters, Cerebras Wafer-Scale Accelerators, and Frontier/DeltaAI allocations.
1. Executive Summary: What is the NAIRR Pilot?
The National Artificial Intelligence Research Resource (NAIRR) Pilot is a $100M+ joint federal initiative led by the National Science Foundation (NSF) and the Department of Energy (DOE) that provides zero-cost access to high-performance AI computing, specialized hardware, and foundation model APIs for U.S. researchers.
Through UCR Research Computing, faculty can leverage the NAIRR Pilot to scale AI workloads beyond local campus resources (HPCC and GCP Ursa Major) into national exascale supercomputers and specialized AI accelerators at no charge to their grants.
2. NAIRR Pilot Resource Portfolio (What You Can Request)
| RESOURCE CATEGORY | AVAILABLE HARDWARE / PLATFORMS | IDEAL RESEARCH USE CASE |
|---|---|---|
| Exotic AI Accelerators | Cerebras CS-3, SambaNova SN40L, Groq LPU | Wafer-scale AI training, ultra-low latency inference, & non-GPU chip architectures. |
| Massive H100/A100 GPU Clusters | NCSA DeltaAI (1,200+ H100 GPUs), SDSC Expanse, TACC Vista | Multi-node distributed LLM pre-training, fine-tuning, and large vision models. |
| Exascale Supercomputing | ORNL Frontier (DOE Exascale), ALCF Aurora | Climate AI, molecular dynamics, & massive multi-physics simulation campaigns. |
| Foundation Model API Credits | OpenAI, Anthropic Claude, Vertex AI, Hugging Face | Prompt engineering, evaluation benchmarks, & fine-tuning proprietary models. |
3. Allocation Tracks & How to Apply
🔹 Track 1: NAIRR Startup Allocations (Fast-Track On-Ramp)
- Best For: Code testing, benchmarking, and gathering preliminary data for grant proposals.
- Review Time: Fast-tracked (1–2 weeks) w/ lightweight 2-page proposal.
- Allocation Size: Up to 5,000 GPU-hours or $5,000 in model API credits.
🔹 Track 2: NAIRR Research Allocations (Full Campaign)
- Best For: Multi-node research campaigns, large-scale LLM training, and paper publications.
- Review Time: Monthly peer-review cycle.
- Allocation Size: Up to 100,000+ GPU-hours or wafer-scale Cerebras cluster access.
4. UCR Research Computing Onboarding & Support
UCR Research Computing acts as your institutional bridge to NAIRR Pilot allocations:
- Proposal Review & Benchmarking: We assist UCR faculty in running scaling benchmarks on HPCC or GCP Ursa Major to include in your NAIRR application.
- Access & Credential Mapping: Once granted, we help configure Globus endpoint transfers, SSH key authentication, and Slurm job scripts for DeltaAI, Expanse, or Cerebras.
- Hybrid Workflow Integration: Run pre-processing on HPCC/Ursa Major, launch massive GPU training on NAIRR DeltaAI, and store results in CephRDS S3.
🚀 Get Started Today
- NAIRR Portal: https://nairrpilot.org
- Request UCR Proposal Assistance: Email Research Computing at
research-computing@ucr.eduor submit a request via UCR Support Portal.
Published by UCR Research Computing | UC Riverside Information Technology Solutions