NERSC Opens Perlmutter Supercomputer to AI-for-Science Proposals
March 27, 2026 · 2 min read
David Almeida
The National Energy Research Scientific Computing Center has issued its 2026 call for AI-for-science proposals, offering research teams free access to the Perlmutter supercomputer for projects that advance machine learning in scientific domains. Proposals submitted by April 30, 2026, receive full consideration.
What Successful Teams Receive
Selected projects get up to 10,000 Perlmutter GPU node hours — each node equipped with four NVIDIA A100 GPUs — plus up to 20,000 CPU node hours for generating AI-ready datasets and storage allocations on NERSC file systems. Teams demonstrating effective utilization can request additional resources. All awards cover the NERSC 2026 Allocation Year through January 19, 2027.
Unlike traditional grants, there is no cash award. The value lies entirely in access to one of the Department of Energy's flagship computing platforms, representing compute resources that most university labs and small research teams cannot replicate independently.
Who NERSC Wants to Hear From
NERSC is seeking teams that combine deep learning expertise with strong domain knowledge and demonstrated proof-of-concept results. Any research area aligned with the DOE Office of Science mission qualifies — materials science, climate modeling, high-energy physics, genomics, chemistry, and beyond.
Current NERSC users have no advantage; the call is open to any qualified team. Proposals are submitted through an online form and reviewed on a rolling basis until resources are exhausted.
A Growing Federal AI Compute Pipeline
This call is part of DOE's expanding investment in AI-powered science, which includes the $293 million Genesis Mission and dedicated AI research grants across the national laboratory system. For teams that cannot compete for large monetary awards, Perlmutter access offers a practical alternative — world-class compute with a straightforward application process.
Researchers can contact nersc-ai@lbl.gov with questions. For tracking of DOE AI funding opportunities and application strategies, visit grantedai.com.