Quota increase for A100/A10 GPUs in West US 3/East US 2 for Finance AI research (Azure Sponsorship).

Question

We are unable to create A100/A10 compute clusters for our Finance AI research. We have $9,000 in Azure Sponsorship credits specifically for this project. The automated quota tool is rejecting our requests without manual review.

Project: Fine-tuning ModernBERT-Large. Need: 24-48 cores of Standard NCads_v4 (A100 80GB) or NVadsA10_v5 (A10 24GB). Location: West US 3 or East US 2. Please manually override this limit so we can utilize our research sponsorship.

Answer 1

Hey Vo Van Nam,

Sounds like you’re running into the standard vCPU/GPU quota limits for the NCads_v4 (A100 80 GB) and NVadsA10_v5 (A10 24 GB) series in West US 3/East US 2. Here’s how you can get those bumped up—ideally with a manual review since you’re on Azure Sponsorship credits:

Check SKU availability

• Sometimes new regions lag on certain SKUs.
  • Run this in Azure CLI (or Portal “Available VM sizes”): az vm list-skus --location westus3 --size NCads_v4 --query "[].{Name:name,Zone:locationInfo[0].zones}"
    • Repeat for NVadsA10_v5 in both regions.

1. Submit a quota increase request a. In the Azure portal, go to Subscriptions > your subscription > Usage + quotas. b. Filter by “Virtual Machines vCPUs – VM family” and locate: • Standard NCads_v4 series (A100 80 GB) • Standard NVads_A10_v5 series (A10 24 GB) c. Click Request increase next to each SKU. d. Select the region (West US 3 or East US 2) and set your new vCPU quota—e.g., 48 cores for 24 GPUs or 96 cores for 48 GPUs. e. In the Justification field, mention your Azure Sponsorship research project (Fine-tuning ModernBERT-Large, $9K sponsorship). f. Submit.

If the automated tool rejects your request

• In the quota form’s “Problems?” section, pick Other and explicitly ask for a manual review citing your sponsorship credit and research need.
  • Alternatively, open a full support ticket: Help + support > New support request > Service and subscription limits (quotas) > Machine Learning Service: Virtual Machine Quota. Provide the same SKU/region details and sponsorship context.

(Optional) Use Azure ML Studio

• Go to https://ml.azure.com/ > your workspace > Manage > Quotas.
• Request increases for the NCads_v4 and NVads_A10_v5 SKUs there—again noting your sponsorship.

Backup plan

  - If West US 3 capacity is still constrained, consider East US 2 where availability may be better.
  
     - Or spin up NC_A100_v4 (1–4 x A100 80 GB GPUs) in either region if you can parallelize your workload across multiple nodes.
     

Hope that gets you unblocked—once support approves, you’ll be able to spin up your 24–48 GPU cluster and move forward with your Finance AI research!

References • NC and NV series VMs (NCads_v4, NVadsA10_v5) – https://supportability.visualstudio.com/HighComplexityLowVolume/_wiki/.../NC

NV-series Virtual Machines • How to increase VM quota and limit – https://learn.microsoft.com/azure/machine-learning/how-to-manage-quotas?view=azureml-api-2

Issues in Quota Management (Instance) – https://supportabilityhub.microsoft.com/solutions/.../Issues in Quota Management (Instance)

Resolve Quota Allocation Issues for Azure Regions – https://supportabilityhub.microsoft.com/solutions/.../Resolve

Quota Allocation Issues for Azure Regions

Answer 2

Hi Vo Van Nam DaNang Speech,

Welcome to Microsoft Q&A,

GPU quota increases are limited for customers using free, benefit, or sponsorship subscriptions. Due to high demand for graphics-enabled VM types, availability is limited for these subscription types, regardless of the credit balance.

This is the most likely reason the automated quota tool is rejecting your requests. GPU VMs are not eligible for Sponsorship subscription IDs as a general policy, which means the rejection is not a capacity issue in the traditional sense. It is a subscription-type eligibility block. Having $9,000 in sponsorship credits does not override this restriction because the credits and the quota system are governed independently.

Capacity in the specific SKUs and regions you have requested (Standard_NCads_A100_v4 in West US 3, NVadsA10_v5 in East US 2) is among the most contended in Azure right now.

A Technical Support ticket (not the quota request form) routes your case to a capacity team engineer who can manually evaluate research use cases. When opening the ticket:

• Set Issue type: Service and subscription limits (quotas)
• Set Quota type: Compute-VM (cores-vCPUs) increase
• In the description, include: research institution name, project purpose (ModernBERT-Large fine-tuning for Finance AI), specific SKUs needed (Standard_NC24ads_A100_v4 or Standard_NV36ads_A10_v5), vCPU count (24-48), preferred regions, and the sponsorship subscription ID

Please Upvote and accept the answer if it helps!!