Azure is currently experiencing Compute capacity constraints in the West Europe region. Starting as early as 05:30 UTC on 13 March 2026, users attempting to create, update, scale, start, and/or restart Virtual Machines in West-Europe region may experience allocation failures.
Microsoft is experiencing regional deployment failures for AMD GPU SKUs. Constrained Virtual Machine series include:

At this time, you are not able to create or use VMs in myDRE with VM type: NcasT4v3. Please select another VM type.
This issue originates from Microsoft Azure. We are monitoring the situation and will provide updates as more information becomes available. You can also follow Microsoft’s official updates on
their Azure status pages.
We apologize for the inconvenience.
Current status: Due to an increase in demand for compute resources in the region, there are capacity constraints on available compute infrastructure. This in turn resulted in service management operation and allocation failures for VM resources.
We are currently working on platform levers to restore healthy buffers in relation to capacity. We will continue to monitor the situation and supply updates accordingly.
2026-03-19 14:00 CET: No new updates from Microsoft regarding this issue.
2026-03-20 14:48 CET: New update date provided. Added that this might affect starting Databricks services.
2026-04-07 13:00 CET: No new updates from Microsoft regarding this issue.
2026-04-14 13:00 CET: No new updates from Microsoft regarding this issue.
2026-04-21 12:00 CET: Status update regarding VM resources
2026-05-21 15:20 CET: Status update
Past updates:
- Starting at 05:30 UTC on 13th March-2026, you have experienced allocation failures in relation to performing service management operations - such as create, update, scaling, start. This issue affected the following virtual machines (NcasT4v3 Series). Microsoft will be providing their next update by 18 March 2026, or as events warrant.
- Microsoft observed an unexpected spike in usage that resulted in backend Virtual Machine components reaching an operational threshold. This lead to failures for selected customers deploying virtual machines. Microsoft is actively working on mitigating this issue. (13 Mar 2026 13:16 CET)
- Added a note that running VMs may not be directly affected (13 Mar 2026 12:27 CET)
- Starting at 10:11 UTC on 13 Mar 2026
- Starting at 05:30 UTC on 13 March 2026, you have been identified as a customer using Virtual Machines in West Europe who may have experienced allocation failures in relation to performing service management operations - such as create, update, scaling, start. This issue affected the following virtual machines (NcasT4v3 Series) and Azure Databricks Service experiencing issues.
During the impact window, we observed an unexpected spike in usage that resulted in backend Virtual Machine components reaching an operational threshold. This led to failures for customers deploying virtual machines.