DeployGPU offers GPU instances in datacenters across multiple regions worldwide. The region determines the physical location of your instance, which affects latency and data residency.
Available regions
GPU instances are available across these regions:
- North America — United States and Canada
- Europe — EU datacenters
- United Kingdom
- Asia-Pacific — APAC datacenters
Specific cities and locations are shown on each offer in the marketplace. Availability and GPU selection vary by region.
Filtering by region
In the Compute Marketplace, use the Region filter to narrow offers by continent:
- North America
- Europe
- Asia-Pacific
- South America
The filter shows how many offers are available in each continent.
Region and volumes
Volumes are region-locked — a volume created in one region can only be attached to deployments in the same region. Plan your region choice before creating volumes if you need persistent storage.
Choosing the right region
Pick a region close to:
- Your location — For interactive SSH sessions and Jupyter access
- Your data sources — For faster data transfer during training
- Your users — If serving inference endpoints
If latency isn’t critical and you want the best price, sort the marketplace by price without a region filter. The cheapest GPU is often in whatever region has the most supply.
Data residency
Your instance and any attached volumes exist physically in the region you selected. When a deployment is stopped, instance disk data is not retained. Volume data persists in its region until you delete the volume. 
