Share via

Cannot deploy any Anthropic Claude models on production tenant zero quota despite working deployment on separate dev tenant

Tyler Straub 0 Reputation points
2026-02-10T22:23:24.6+00:00

We have Anthropic Claude models (Sonnet 4.5, Haiku 4.5, Opus) deployed and running on our developer tenant in East US 2 using Global Standard deployment at approximately 100K TPM. Everything works fine there.

On our production tenant, which is a completely separate Azure tenant with its own paid subscription and valid billing, every Claude model in the Foundry Model Catalog shows "no quota available." This includes Sonnet 4.5, Haiku 4.5, Opus 4.1, and the newly released Opus 4.6. We have attempted East US 2, the same region that works on the dev tenant. This has persisted for multiple weeks with no change.

The production subscription is not a restricted type. It is not CSP, not sponsored, and not credits-only. Marketplace purchases are permitted on the production tenant. We have accepted the Anthropic Marketplace terms on the production tenant. Checking Usage and quotas in the Azure Portal and filtering for Claude models shows 0/0 available.

We submitted a quota increase request through the standard form, but the documentation for Foundry Models quotas and limits states that priority goes to customers who actively consume their existing quota allocation. We cannot consume quota we have not been allocated. The form appears designed for increasing existing quota, not establishing initial allocation on a subscription that currently has zero.

We have seen several other Q&A threads describing the same issue, including "No quota for claude models" and "Unable to deploy Anthropic Claude Opus 4.5 in Microsoft Foundry due to insufficient quota," where the suggested resolution is to open an Azure Support ticket requesting Global Standard enablement rather than using the quota increase form alone.

We need clarification on the following:

First, is there a separate process for initial Anthropic Claude quota allocation versus increasing existing quota? The current process appears to be a catch-22 for subscriptions starting from zero.

Second, are there tenant-level enablement steps for Anthropic models that are independent from subscription-level quota? Since our dev and production environments are on separate tenants, we want to confirm there is no tenant-level registration or provider enablement we may have completed on dev but not on production.

Third, is this a temporary capacity constraint during Public Preview or a policy-level gate that requires explicit approval per subscription? We need to understand whether waiting will resolve this or whether specific action on our part is required.

Fourth, is an Azure Support ticket the correct path to resolve this for initial allocation? If so, it would be helpful if this were documented somewhere, as the current documentation does not address the zero-default-allocation scenario.

We are blocked from moving our AI workloads to production. Our application depends on Anthropic Claude models deployed via Azure AI Foundry, and we have validated the integration on our dev tenant. We cannot go live until the production tenant has quota allocated. A concrete resolution path would be appreciated.

Foundry Tools
Foundry Tools

Formerly known as Azure AI Services or Azure Cognitive Services is a unified collection of prebuilt AI capabilities within the Microsoft Foundry platform

{count} votes

1 answer

Sort by: Most helpful
  1. Anshika Varshney 8,200 Reputation points Microsoft External Staff Moderator
    2026-02-11T00:34:09.46+00:00

    Hi Tyler Straub,

    Thanks for sharing the details. This behavior is expected for some subscriptions and tenants when using Anthropic Claude models in Azure AI Foundry (public preview).

    Claude models are Marketplace-based offerings and quota is not auto‑assigned to every subscription. Even with valid billing and accepted Marketplace terms, a subscription can show 0/0 quota if it hasn’t been explicitly enabled on the backend. This is why you may see the models working in one tenant while showing “no quota available” in another, even in the same region.

    At this time, the quota increase request form alone is not sufficient when starting from zero quota. For initial allocation, the recommended path is to open an Azure Support ticket and request Global Standard enablement / initial quota allocation for Anthropic Claude models. Several similar Q&A threads have been resolved through support after confirmation that the subscription type and billing profile are eligible.

    Since Claude availability and capacity are still managed during preview, waiting typically does not resolve this without intervention. There are no additional tenant-side configuration steps beyond Marketplace acceptance, but support can validate subscription eligibility and apply the required enablement.

    I Hope this helps. Do let me know if you have any further queries.

    Thankyou!

    1 person found this answer helpful.

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.