more inference servers deployable under a fixed power budget via priority-based oversubscription
~30%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 | POLCA oversubscription gain |
| As of | 2024 |
| Source | POLCA: Power Oversubscription in LLM Cloud Providers (Microsoft Research) · Abstract and §6.2: POLCA deploys 30% more inference servers while staying within performance SLOs |
| Review | checking…review by 2026-08-19 · standard cadence |
| Recorded changes | last 2026-06-29 |
| Claim id | more-inference-servers-deployable-under-a-fixed |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.