Request GPU Startup classification + GPU quota for Azure ML training

Onyango, David 31 Reputation points

Hello Azure Support,

I am part of the Microsoft for Startups program and I am training computer vision models in Azure Machine Learning. My subscription is a Microsoft for Startups Sponsorship subscription.

At the moment, my subscription does not show any GPU VM families (NC, ND, NV) in any region, including East US. This prevents me from creating GPU compute for model training.

Could you please:

Classify my subscription as a “GPU Startup” (This is required for GPU VM families to become visible in quota management.)

Enable access to GPU VM families in East US.

Grant quota for the following GPU SKUs:

NC A100 v4 (preferred)

 **NC T4 v3** (fallback option)

This is required for training a retail computer vision model on Azure ML.

Thank you for your assistance.

  1. Anshika Varshney 13,405 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    Thank you for clarifying this in the private chat. If the previous tickets indicated that the quota was successfully assigned but you still cannot see GPU VM families in Azure ML Compute, this usually means the subscription classification step might not have been completed. For Microsoft for Startups Sponsorship subscriptions, GPU access requires the subscription to be tagged as “GPU Startup” before GPU SKUs become visible in quota management.

    Next steps:

    1. Verify Subscription Classification
      • Check if your subscription is tagged as GPU Startup. If not, request this via the https://foundershub.startups.microsoft.com/ or contact your program representative.
      • Include your subscription ID and a short justification (e.g., AI/ML workload details).
    2. Confirm Quota Assignment
      • In the Azure Portal, navigate to Help + Support → Quotas and confirm if GPU SKUs (NC A100 v4 or NC T4 v3) are listed for your region.
      • If they are not visible, classification is likely pending.

    Note: If classification was not part of the previous requests, yes, you’ll need to log a new ticket specifically requesting GPU Startup tagging and quota assignment.

    • Use Issue type: Service and subscription limits (quotas), select Machine Learning, and specify the region and GPU SKUs.

    References for guidance:

    Once classification is complete, GPU VM families should appear in your quota management and Azure ML Compute selection.

    I hope this helps you get back on track! If you're still facing issues, could you share more details?

    Thankyou!

  2. Anshika Varshney 13,405 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    Did you get any chance to review the response.

    Thankyou!

  3. Anshika Varshney 13,405 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet. In case if you have any resolution please do share that same with the community as it can be helpful to others. Otherwise, will respond with more details and we will try to help.

    Thankyou!

  4. Anshika Varshney 13,405 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    Thank you for sharing the update in the private chat.
    We have escalated this issue to our Product Group (PG) team and are currently reviewing the details. The team is actively working on the investigation, and we will update you as soon as we have more information. Thank you for your patience.

  5. Manas Mohanty 17,185 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    I have synced with concerned support ticket owner to enable “GPU Startup” flag before raising GPU quota.

    Checking internally here on relevant channels on our side too.

    Thank you.

Sign in to comment

2 answers

  1. Manas Mohanty 17,185 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    Got update from respective support ticket owner. He mentioned Start up subscription are eligible GPU quota already. Simple creating quota raise request should help you.

    Have suggested support ticket owner on the same.

    Please create Subscription quota tickets mentioning your required GPU family quota.

    Thank you for using Azure Services.

    0 comments No comments
    Sign in to comment
  2. Anshika Varshney 13,405 Reputation points • Microsoft External Staff • Moderator

    Hi Onyango, David,

    Thank you for reaching out on the Microsoft Q&A.

    Azure subscriptions under programs like Microsoft for Startups often default to sponsorship tiers that don’t expose GPU VM families (NC, ND, NV). To unlock GPU SKUs, your subscription must be tagged as a GPU Startup. This classification ensures GPU VM families appear in quota management and can be requested.

    Steps to Enable GPU Access and Request Quota

    1. Confirm Subscription Classification
      • Check if your subscription is tagged as GPU Startup. If not, submit a request via the Microsoft for Startups Founders Hub portal or contact your program representative.
      • Include your subscription ID and a short justification (e.g., AI/ML workload details).
    2. Request GPU VM Family Access
      • Navigate to Azure Portal → Help + Support → New Support Request.
      • Issue type: Service and subscription limits (quotas).
      • Service: Machine Learning.
      • Specify region (e.g., East US) and GPU SKUs (NC A100 v4 preferred, NC T4 v3 fallback).
    3. Provide Business Justification
      • Mention your project type (e.g., computer vision training) and urgency.
      • Requests typically take 2–7 days for approval.

    Useful References

    Monitor your usage in Azure Portal → Metrics → OpenAI (Tokens or Requests) to plan ahead and avoid hitting limits.

    Please let me know if there are any remaining questions or additional details, I can help with, I’ll be glad to provide further clarification or guidance.

    Thankyou!

    0 comments No comments
    Sign in to comment
Sign in to answer

Your answer