Supercomputers/Systems/Nexus (Georgia Tech)
USA · operational · mixed
Nexus (Georgia Tech)
Also known as Georgia Tech Nexus; NSF Nexus
Operated by Georgia Institute of Technology at Coda Data Center (Georgia Tech, Atlanta) .
Researched with AI assistance, checked by scripts. These rows were researched by AI agents reading public operator, laboratory and vendor sources, appended by scripts and checked by automated validation; no subject-matter expert has reviewed them (how this site is made). A mechanical second-reader pass has since fetched every cited source: 1350 of 1540 systems have a readable citation that names them, and 64 are genuinely weakly sourced. Every claim carries its source and a confidence tier. Treat anything below verified as a lead, not a citation.
Measured performance
- Rmax
- -
- Rpeak
- -
- Rmax ÷ Rpeak
- -
- Cores
- -
- Power
- -
- Per watt
- - GF/W
Figures are as last publicly reported for the configuration described below, not a live measurement. Where a system was upgraded in place, the post-upgrade configuration is shown and the earlier one appears in the timeline.
What this machine is made of
One row per supplier relationship. “Supplier at build” is the company that shipped the part at the time; where that company has since been acquired, the parent it rolls up to today is shown beside it. That distinction is what makes ticker-level aggregation possible across a thirty-year dataset.
| Role | Part | Supplier at build | Quantity | Confidence | Source |
|---|---|---|---|---|---|
| Accelerator | NVIDIA B200 SXM Accelerators CUDA (Blackwell) · TSMC 4NP · 1,000 W 16 NVIDIA DGX B200 nodes, 8 B200 SXM GPUs each (128 GPUs) | NVIDIA | 128 reported | Reported | docs.nexus.gatech.edu |
| Accelerator | - NVIDIA RTX PRO 6000 Blackwell Server Edition: 304 GPUs in 38 scheduled nodes plus 48 GPUs in 6 persistent-service nodes | NVIDIA | 352 reported | Reported | docs.nexus.gatech.edu |
| CPU | Intel Xeon Platinum 8570 CPUs x86-64 · 56C · 2.1 GHz · Intel 7 Dual Intel Xeon 8570 (56 cores each) in each of the 16 DGX B200 nodes | Intel | 32 estimated | Reported | docs.nexus.gatech.edu |
| Interconnect | NVIDIA InfiniBand NDR (Quantum-2) Interconnect Fat tree or dragonfly+ · 400 Gb/s per port · Under 0.6 microseconds 8-rail NDR InfiniBand compute fabric on 12 NVIDIA QM9700 switches, 4 OSFP 800G ports per DGX node | NVIDIA | - | Reported | docs.nexus.gatech.edu |
| Storage | VAST Data Platform Storage Disaggregated shared-everything 8 VAST Ceres V2 appliances, about 10.8 PB raw NVMe flash | VAST Data | 8 reported | Reported | docs.nexus.gatech.edu |
AI datacenter metrics
Figures beyond an accelerator count, each with who stated it. What can and cannot be compared across AI datacenters.
| Metric | Value | As of | Basis | Note | Source |
|---|---|---|---|---|---|
| Accelerators | 480 accelerators | 2026-10 | Operator stated | Georgia Tech hardware page: 480 installed GPUs (128 B200, 352 RTX PRO 6000 including 48 in persistent-service nodes) | docs.nexus.gatech.eduatlantanewsfirst.com |
The read
NSF-funded AI and HPC platform built and run by Georgia Tech with the University of Illinois Urbana-Champaign (NCSA hosts two service nodes). A ribbon-cutting was held on 12 October 2026 at the Coda building in Midtown Atlanta; the NSF award is 20 million dollars, first reported in 2025 with a spring 2026 completion target. Georgia Tech's hardware page lists 72 nodes and 480 GPUs: 16 NVIDIA DGX B200 nodes (128 B200 GPUs), 38 scheduled RTX PRO 6000 Blackwell Server Edition nodes (304 GPUs) and 6 persistent-service nodes (48 GPUs, 4 at Georgia Tech and 2 at NCSA), on an 8-rail NDR InfiniBand compute fabric and a VAST Data store of 8 Ceres V2 appliances (about 10.8 PB raw NVMe). Press coverage in 2025 quoted 'over 400 quadrillion operations per second', an AI-precision figure with no stated basis that is not recorded as FLOPS. At least 90 percent of capacity goes to the national community through ACCESS and the system feeds NAIRR; the docs also describe federation with Delta and DeltaAI. No HPL run is published.
Site & facility
Change history
- 2026-10-18 Added Nexus (Georgia Tech) (Georgia Tech)
Further reading
Inside a node
Why accelerators exist, what the memory hierarchy costs, and how the CPU and accelerator merged onto one package.
The interconnect
Topologies, why latency and tail behaviour matter more than bandwidth, and how the fabric market consolidated.
Storage and I/O
Parallel file systems, why checkpointing dominates the write load, and the metadata failure mode nobody expects.