GPU-cloud SLA example: node uptime / rack uptime, with penalties (ClusterMAX 2.0 example figures; SemiAnalysis sets no baseline)
99% / 95%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 | Illustrative provider contract structure reported by ClusterMAX 2.0 (November 2025); not a market baseline or recommendation. ClusterMAX 3.0 (2026-09-23) publishes standardized Bronze, Silver and Gold SLA terms with thresholds and credits; the tier percentages are not in the public text. Its common NVL72 rule (rack down when 3 or more of 18 nodes fail) matches this example's 16-of-18 rack boundary. |
| Caveat | Uptime definition, exclusions, measurement window, spare entitlement, service credits, and penalties remain contract-specific. |
| As of | 2025-11 |
| Source | SemiAnalysis ClusterMAX · SLA examples: one provider approach commits 99% node uptime and 95% rack uptime (rack defined as 16 of 18 nodes, or 64 of 72 GPUs); another commits 99% node uptime only and allocates 16 of 18 nodes. |
| Review | checking…review by 2027-03-23 · standard cadence |
| Recorded changes | last 2026-10-16 · 3 revisions tracked |
| Claim id | gpu-cloud-sla-baseline-node-uptime-rack-uptime |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.