Share via

AKS isn't letting me configure new node pools with VM sizes for which I have available vCPUs in my quotas list

William Roton 20 Reputation points
2026-03-20T18:07:36.3933333+00:00

A couple of months ago I requested a quota increase for NV series vCPUs in the Central US region. This request was denied, but I was offered to have the quota increased in the East US region. I accepted and have now migrated my entire application to East US. Now, when I attempt to setup the new node pool, not only is the NV series unavailable, but D series machines are also unavailable.

I have 0 out of 24 CPUs used for the NV series in the East US region, and 0 out of 65 used on the D series in the East US region, yet these options are not available when select a node size for my node pool.

Why is this?

If these sizes are somehow not compatible with AKS, then can someone please recommend images that make good substitutes. I need the cheapest AKS compatible machine I can get with a GPU (N-series), and whatever D-series machine meets the minimum requirements for being a system node.

Azure Kubernetes Service
Azure Kubernetes Service

An Azure service that provides serverless Kubernetes, an integrated continuous integration and continuous delivery experience, and enterprise-grade security and governance.

0 comments

Answer accepted by question author

Jilakara Hemalatha 13,750 Reputation points Microsoft External Staff Moderator
2026-03-20T18:56:19.84+00:00

Hello William,

Thank you for reaching out and for sharing the detailed context.

Based on what you’ve described, the behavior you’re seeing is expected and is not related to your available vCPU quota. In Azure Kubernetes Service, VM size availability depends on multiple factors beyond quota, which can sometimes make this a bit confusing.

Even though your subscription shows available quota for both NV-series and D-series in East US, AKS only exposes a subset of VM sizes during node pool creation. This is due to the following reasons:

  • AKS‑specific VM‑size support: AKS exposes only a subset of Azure VM SKUs in the node‑pool creation UI, even if those SKUs are available in your subscription quota.
  • SKU / region availability: Some SKUs may be restricted in AKS, phased‑out, or temporarily unavailable due to capacity in a given region (especially GPU SKUs).

Reference: Virtual machine (VM) sizes, generations, and features for Azure Kubernetes Service (AKS)

Quotas, virtual machine size restrictions, and region availability in Azure Kubernetes Service (AKS)

Specifically for GPU workloads, NV-series VMs are primarily designed for visualization scenarios and are not typically supported for AKS GPU node pools. For AKS, NC-series VMs are the recommended and supported option for GPU workloads

Use GPUs for compute-intensive workloads on Azure Kubernetes Service (AKS)

For your scenario, you may consider using NC-series SKUs such as Standard_NC4as_T4_v3 or Standard_NC6s_v3 for GPU workloads, as these are commonly supported and cost-effective options. For system node pools, D-series SKUs like Standard_D2s_v3 or Standard_D4s_v3 are generally recommended as they meet AKS minimum requirements.

To validate from your end, you can also check which SKUs are available in your region using the below command:

az vm list-skus --location eastus --output table

If the NC-series SKUs still don't appear after running these commands, please open an Azure support ticket requesting SKU eligibility for the NC VM family in East US. This is a separate step from a quota increase and is sometimes required for GPU VM families to become selectable in AKS.

Reference: Quickstart: Request a quota increase in the Azure portal

Similar thread:https://learn.microsoft.com/en-us/answers/questions/5733254/azure-plan-subscription-without-access-to-vm-sizes

Hope this helps! Please let me know if you have any queries.

Was this answer helpful?

1 person found this answer helpful.

0 additional answers

Sort by: Most helpful
Most helpful Newest Oldest

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.