NVLink-5 per-GPU scale-up bandwidth, bidirectional (NVL72 = ~130 TB/s rack aggregate); ~18x a 400 Gb/s scale-out NIC per direction (900 vs 50 GB/s each way)
1.8 TB/s bidirectional (~0.9 TB/s per direction)observed
| Value kind | observed — Reported measurements, counts, and specifications keep the precision and scope stated by their source; an exact specification is not treated as a range. |
|---|---|
| Scope | Named NVIDIA generation/platform, bidirectional aggregate convention; not delivered collective bandwidth, cache coherence or arbitrary mixed-platform performance. |
| As of | 2026-10-16 |
| Source | NVIDIA GB200 NVL72 product page (undated; checked 2026-10-16) · Fifth-generation NVLink: 1.8 TB/s of GPU-to-GPU interconnect; NVLink Switch System: 130 TB/s. The page is undated, so 2026-10-16 is the date it was checked. |
| Review | checking…review by 2027-04-16 · standard cadence |
| Recorded changes | last 2026-10-16 · 5 revisions tracked |
| Claim id | nvlink-5-per-gpu-scale-up-bandwidth |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.