NERSC Opens AI for Science Call With Free Supercomputer Access
March 28, 2026 · 2 min read
Claire Cummings
The Department of Energy's National Energy Research Scientific Computing Center has opened its 2026 AI for Science call for proposals, offering researchers free access to one of the world's most powerful supercomputers to push the boundaries of AI-driven scientific discovery.
Selected projects will receive up to 10,000 GPU node hours on the Perlmutter supercomputer — each node packing four NVIDIA A100 GPUs — plus 20,000 CPU node hours for generating AI-ready datasets and dedicated storage quotas. For research teams that lack institutional computing infrastructure, this represents hundreds of thousands of dollars in equivalent cloud compute costs at zero charge.
What NERSC Is Looking For
Proposals must demonstrate how they will "leverage the Perlmutter supercomputer to push the state of the art in AI for science and produce novel science outcomes." Research areas aligned with the DOE Office of Science mission — energy, materials science, climate modeling, high-energy physics, genomics — will score highest.
NERSC evaluates projects on scientific significance, technical feasibility, readiness for scaling, and team track record. Crucially, applicants do not need to be current NERSC users. Teams with deep learning expertise, strong scientific domain knowledge, and proof-of-concept results from smaller-scale experiments are encouraged to apply.
Deadlines and How to Apply
Submissions are reviewed on a rolling basis, but the April 30, 2026 deadline applies for full consideration in the first review cycle. Proposals must include a defined scope with scientific background, computational resource requirements and usage strategy, team qualifications, and a timeline with deliverables.
Awardees receive consulting support from NERSC staff on system optimization — a significant advantage for teams new to large-scale GPU computing.
A Rare Free Resource for AI Researchers
Federal computing allocations like this are among the most cost-effective resources available to academic and nonprofit research teams. Researchers tracking funding opportunities on grantedai.com should note that NERSC proposals can complement traditional grant applications by providing the compute backbone that reviewers often ask about.
For more on federal research computing opportunities, visit the Granted blog.