# IBM Cloud — `ibm-regions`

| | |
|---|---|
| Source | `https://cloud.ibm.com/docs/overview?topic=overview-locations` (server-rendered docs page; source markdown mirrored at github.com/ibm-cloud-docs/overview `locations.md`) |
| Kind / mode | `cloud_provider` / `html`, implementation `cloud_regions`, parser `cloud_ibm_regions_v1` |
| Fetch | L1 direct, 0 credits (robots.txt only disallows `/docs/search?q=`) |
| Schedule | weekly |
| License / attribution | Publicly published provider documentation; facts only — "IBM Cloud — cloud.ibm.com/docs" |
| Verified live | 2026-09-11: 12 regions, 12 valid, 0 rejected |

Only the two-column `Region | Zone` tables under "Multizone regions" and "Single-campus MZRs" are read:
"Dallas (us-south)" → name "Dallas (us-south)", code `us-south`, zones `us-south-1 us-south-2 us-south-3` → 3 AZs,
`externalIds.kind = mzr | single_campus_mzr`, `externalIds.ibm_zones`. "Chennai - Airtel" / "Mumbai - Airtel" →
city Chennai / Mumbai with `externalIds.partner = Airtel`. City coordinates from the curated table (Washington DC,
Sao Paulo, Montreal aliases). Spot checks: `us-south` → Dallas, Texas US, 3 AZs; `eu-de` → Frankfurt DE; `jp-tok` → Tokyo JP.

**Not extracted (candidates for the operators group):** the "Classic infrastructure data centers" tables (DAL10,
FRA02, …) and PoPs on the same page — those are IBM colocation footprints, not regions.