NSF AI grant $10M: "Bridges-3," a national supercomputer integrating AI, data, and HPC (Carnegie Mellon)
The NSF awarded about $10M to "Bridges-3," a national computing platform integrating AI, data analytics, and high-performance computing (HPC). With GPU-accelerated computing and large memory, it serves the U.S. research community broadly.
Grant overview (primary data)
- Award amount$10,000,000
- RecipientCarnegie Mellon University（PA）
- ProgramInnovative HPC
- Period2026-06-01 〜 2027-05-31
- FunderU.S. National Science Foundation (NSF) / NSF
Key points
- A national computing platform integrating AI, data analytics, and HPC (successor to Bridges-2)
- GPU-accelerated computing + large memory + all-flash parallel FS + InfiniBand
- Continues national cyberinfrastructure under the NSF ACSS program
- Architecture informed by OSTP FY2027 national investment priorities (M-25-34)
- About $10M, Carnegie Mellon University (PSC)
The NSF awarded about $10,000,000 to the national computing platform "Bridges-3," operated by Carnegie Mellon University (NSF Award 2614022; program: Innovative HPC; June 2026 – May 2027).
Per the abstract, Bridges-3 continues PSC's long-standing role of providing production-quality national cyberinfrastructure through the NSF Advanced Computing Systems & Services (ACSS) program. As the successor to Bridges-2, it delivers systems and services that adapt to rapid change in computing and data technologies while remaining reliable, accessible, and broadly usable by the U.S. research community.
Technically, it offers a balanced, converged environment integrating GPU-accelerated computing, high-performance large-memory CPUs, a hardened all-flash parallel file system, and a high-bandwidth InfiniBand fabric — a cohesive, scalable platform for simulation, artificial intelligence, data analytics, and complex HPC workflows. The overall architecture is informed by the science priorities in the Office of Science and Technology Policy memorandum M-25-34/NSTM-2 for FY2027.
Why it matters
Shows how national investment in compute (HPC + GPUs) is becoming central as AI advances. A useful reference for how research computing infrastructure is built and refreshed.
FAQ
What is HPC?
Why does AI need computing infrastructure?
Sources (primary)
Source: NSF Award Search (U.S. National Science Foundation, public domain). Amounts are the obligated amount. For privacy, we do not handle principal investigator names.
- NSF Award (original, official)
- NSF Award ID: 2614022