Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.coreweave.com/llms.txt

Use this file to discover all available pages before exploring further.

CoreWeave offers compute in multiple geographically diverse locations to ensure high availability, resilience, global coverage, and low-latency access for your applications. We organize these locations into Geos, Super Regions, Regions, and Availability Zones (AZs). Regions and Availability Zones (AZs) are the most important concepts to understand when deploying applications on CoreWeave. These are located within Super Regions and Geos, which are explained in more detail at the end of this page. Regions contain one or more AZs. Our AZs are categorized as General Access or Dedicated Access. For a list of every AZ and its access type, see All Availability Zones.

​
General access

General Access AZs are available to all CoreWeave customers, subject to capacity. Tenant workloads are isolated using VPCs with per-tenant network segmentation enforced at the hardware layer by DPUs.

​
Dedicated access

Dedicated Access AZs are managed by CoreWeave and reserved for select customers. Infrastructure in a Dedicated Access AZ is single-tenant: no other customer’s workloads run in the AZ, and no other tenant shares the underlying network fabric. This provides full network isolation in addition to the platform-wide VPC and DPU-based tenant segmentation.

​
Region features

All CoreWeave Regions offer these features:
  • Free data transfer: move data between Regions without incurring additional costs.
  • Fast and reliable Internet access: each Region is equipped with redundant, 200Gbps+ public Internet connectivity provided by Tier 1 global carriers.
  • High-speed, private dark fiber backbone: all Regions are interconnected with 400Gbps+ of dark fiber to facilitate rapid data movement and synchronization.
  • Distributed File Storage: each Region offers high-performance shared filesystems for your applications.
  • Virtual Private Clouds: all Regions are equipped with VPCs to provide network isolation and security.
  • CoreWeave AI Object Storage: available in select Availability Zones. See the availability by AZ table below.

​
Geographical hierarchy

To help you understand how our infrastructure is structured, here’s a breakdown of the terms we use to describe our global infrastructure:
TermDescriptionExample
GeoA large geographic area, such as a continent or a large country. CoreWeave uses Geos to manage global infrastructure, ensure compliance with regulatory requirements, and to optimize for large-scale redundancy and disaster recovery.All locations in the United States reside in the US Geo.
Super RegionA large section of a Geo that typically contains multiple Regions. Super Regions are used to group Regions together for administrative purposes and to provide a high-level view of the infrastructure in a geographic area.All locations in the eastern United States are part of the US-EAST Super Region.
RegionAn area within a Geo that contains multiple Availability Zones (AZs). Regions provide the infrastructure foundation for resiliency. By deploying across multiple AZs within a Region, you can design failover-ready workloads that minimize downtime. Regions are strategically placed to offer low latency, high-performance connectivity, and meet data residency requirements.Region US-EAST-05 resides in the US-EAST Super Region.
Availability Zone (AZ)A partition within a Region that hosts one or more data halls. AZs are physically and operationally independent from each other to prevent failures from propagating across them.Availability Zone A in the US-EAST-05 Region is named US-EAST-05A.
These are essential concepts in CoreWeave’s infrastructure. They provide the foundation for high availability, fault tolerance, and disaster recovery for your applications and workloads.

​
Region and AZ labels

CoreWeave exposes Region and AZ information through two different sets of labels depending on the context: Kubernetes Node labels for scheduling, and metrics labels for queries, APIs, and CRDs.

​
Kubernetes Node labels

Use these labels in nodeSelector, affinity rules, and topology spread constraints:
LabelValueExample
topology.kubernetes.io/regionRegion nameUS-EAST-04
The standard Kubernetes label topology.kubernetes.io/zone is present on CoreWeave Nodes but does not hold the AZ name. It holds a rack number for use by topology-aware schedulers. There is no AZ-level Kubernetes Node label. Since CKS clusters are single-AZ, AZ affinity within a cluster is not needed.

​
Metrics and API labels

Use these labels in metrics queries, API calls, and CRDs. They appear on kube_node_info and in CoreWeave CRDs, but are not Kubernetes Node labels.
LabelValueExample
regionRegion nameUS-EAST-04
zoneAZ nameUS-EAST-04A

​
CKS clusters are single-AZ

CoreWeave Kubernetes Service (CKS) clusters are zonal. Each cluster is provisioned within a single AZ. All Nodes in a cluster belong to the same AZ, so you do not need AZ affinity rules within a cluster. If you need rack-level placement within an AZ (for example, to co-locate Nodes on the same NVLink fabric), use the node.coreweave.cloud/rack label. See InfiniBand and RoCE node labels for details.

​
Speedtest

To test network performance, run speed tests with ping, curl, or another method of your choice. Each Availability Zone in the table above has speed and latency test endpoints listed in its Network information section. For example, to measure name lookup, connect, and total transfer times with curl, use the following command:
Example curl command
curl [SPEEDTEST-URL] \
       --write-out '\nDNS: %{time_namelookup}s\nConnect: %{time_connect}s\nTotal: %{time_total}s\n' \
       --output /dev/null
Result
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 1024M  100 1024M    0     0  37.2M      0  0:00:27  0:00:27 --:--:-- 38.9M

DNS: 0.001822s
Connect: 0.091322s
Total: 27.465658s
Last modified on May 4, 2026