Are NV12ads_A10_v5 hosts running out of capacity in North Central US?

This usually means Azure temporarily lacks physical capacity for that specific VM SKU in the region, rather than there being a problem with your configuration. Check the deployment logs for the exact error. Your main options are to retry later, allow the platform or management tooling to use a comparable SKU, or create a capacity reservation. Reservations provide more predictable availability, but you pay for the reserved VMs even when they are powered off, so they may not be economical.

West Europe has been especially difficult. At times even common A, B, and D-series machines have been unavailable, and we could not obtain A10 capacity for months. We ended up moving to NV v710 instances to keep the service operating. Rack space and the transition away from older NV v4 hardware may be contributing factors.

I have also seen provisioning failures in other regions, including problems creating AKS clusters with Av2-family CPUs. When capacity shortages affect multiple VM families, it is worth checking Azure Service Health and opening a support case to determine whether switching regions or SKUs is necessary.

The pattern can be pretty severe: a VDI pool may have plenty of hosts available one day, then after overnight shutdowns only a small fraction will start because capacity has been taken by other workloads. Retrying later or configuring fallback VM sizes can help avoid depending on a single constrained SKU.