I got Error 715-123420 when deploying Anthropic Claude Opus on Azure AI Foundry. My deployment status is "Provisioning state: Failed". Please help to unblock my subscription from the internal Fraud Protection system (RTFP).

Question

I got Error 715-123420 when deploying Anthropic Claude Opus on Azure AI Foundry. My deployment status is "Provisioning state: Failed". Please help to unblock my subscription from the internal Fraud Protection system (RTFP).

LAURI HOWARD 20

Hello Support,

LAURI HOWARD 20 Reputation points

2026-06-02T13:10:50.5533333+00:00

Hello,

Thank you for the detailed breakdown. I want to be completely transparent regarding the root cause of this deployment failure (Error 715-123420).

The automated marketplace block was likely triggered because my subscription was flagged for security-related traffic. I am a vetted security researcher (HackerOne handle: offroadergabut), and I was utilizing this environment to write, analyze, and debug Proof-of-Concept (PoC) scripts, JWT tokens, and OAuth configurations for authorized bug bounty programs (including M-Pesa Africa, Marriott, and Worldpay).

I used the AI environment strictly as a pair-programmer and documentation assistant to structure vulnerability reports, not to generate malicious attack tools. All testing was non-destructive, strictly in-scope, and fully authorized via HackerOne.

However, I now understand that running or analyzing these security testing scenarios triggers automated safety flags and violates the Acceptable Use Policy for this promotional Microsoft for Startups infrastructure. I sincerely apologize for this oversight.

Moving forward, I commit to completely isolating my security research and bug bounty activities from this Azure subscription. This environment will strictly be used for standard application development and core product building for our startup.

Since my goal now is solely to deploy the Anthropic Claude Opus model for legitimate product development, could you please escalate this to the Risk/Subscription Compliance team to review my account and manually lift the Marketplace deployment restriction? I am more than happy to provide proof of my HackerOne profile and authorized program scopes if required.

Thank you for your time and understanding.
Karnam Venkata Rajeswari 5,255 Reputation points Microsoft External Staff Moderator

2026-06-02T14:32:14.8866667+00:00

Hello @LAURI HOWARD ,

We sincerely appreciate your honesty and transparency regarding the root cause analysis.

Thank you for trusting us with the information.

We can help you escalate the case to the respective team. Please provide us the following details for better assistance as requested over the priavte message.

Respective teams will review and provide the needful resolution.

Thank you
Karnam Venkata Rajeswari 5,255 Reputation points Microsoft External Staff Moderator

2026-06-04T04:35:37.8666667+00:00

Hello @LAURI HOWARD ,

Hope the response was helpful and the issue is resolved.

Since I’ve converted my earlier comment into an answer, could you please take a moment to mark it as Accepted with an upvote? This helps others in the community with the same question find the solution more easily.

Thank you!

Answer accepted by question author

0 additional answers

Your answer

LAURI HOWARD 20 Reputation points

2026-06-02T13:10:50.5533333+00:00

Hello,

Thank you for the detailed breakdown. I want to be completely transparent regarding the root cause of this deployment failure (Error 715-123420).

The automated marketplace block was likely triggered because my subscription was flagged for security-related traffic. I am a vetted security researcher (HackerOne handle: offroadergabut), and I was utilizing this environment to write, analyze, and debug Proof-of-Concept (PoC) scripts, JWT tokens, and OAuth configurations for authorized bug bounty programs (including M-Pesa Africa, Marriott, and Worldpay).

I used the AI environment strictly as a pair-programmer and documentation assistant to structure vulnerability reports, not to generate malicious attack tools. All testing was non-destructive, strictly in-scope, and fully authorized via HackerOne.

However, I now understand that running or analyzing these security testing scenarios triggers automated safety flags and violates the Acceptable Use Policy for this promotional Microsoft for Startups infrastructure. I sincerely apologize for this oversight.

Moving forward, I commit to completely isolating my security research and bug bounty activities from this Azure subscription. This environment will strictly be used for standard application development and core product building for our startup.

Since my goal now is solely to deploy the Anthropic Claude Opus model for legitimate product development, could you please escalate this to the Risk/Subscription Compliance team to review my account and manually lift the Marketplace deployment restriction? I am more than happy to provide proof of my HackerOne profile and authorized program scopes if required.

Thank you for your time and understanding.
Karnam Venkata Rajeswari 5,255 Reputation points Microsoft External Staff Moderator

2026-06-02T14:32:14.8866667+00:00

Hello @LAURI HOWARD ,

We sincerely appreciate your honesty and transparency regarding the root cause analysis.

Thank you for trusting us with the information.

We can help you escalate the case to the respective team. Please provide us the following details for better assistance as requested over the priavte message.

Respective teams will review and provide the needful resolution.

Thank you
Karnam Venkata Rajeswari 5,255 Reputation points Microsoft External Staff Moderator

2026-06-04T04:35:37.8666667+00:00

Hello @LAURI HOWARD ,

Hope the response was helpful and the issue is resolved.

Since I’ve converted my earlier comment into an answer, could you please take a moment to mark it as Accepted with an upvote? This helps others in the community with the same question find the solution more easily.

Thank you!

Answer 1

Hello @LAURI HOWARD ,

Welcome to Microsoft Q&A .Thank you for reaching out to us.

Error code 715-123420 is a generic deployment failure message and the exact cause is not documented in the provided references. However, based on the documented failure modes for Foundry model deployments and partner/Marketplace models, the most likely categories are:

Subscription eligibility / model availability
- Third‑party and some first‑party models can be blocked at the subscription or region level even when the portal UI allows starting the flow.
- For partner models (like Anthropic Claude), common blocking reasons include:
- Offer not available for the subscription’s billing country/region.
- Marketplace purchase eligibility checks failing.
- Model not available in the selected region.
- For Azure‑offered models (like GPT o4‑mini), deployment can fail if the model is not available to that subscription or region.
- The documented guidance for “Model not available” is to list available models and regions for the subscription and resource:
- Run az cognitiveservices account list-models against the Foundry resource to confirm which SKUs and regions are actually deployable for that subscription.
Marketplace access / billing account region
- For partner and community models, the Marketplace layer can reject the deployment even when the terms page opens. Documented Marketplace‑related errors include:
- Offer not available in your country/region.
- Marketplace purchase eligibility check failed.
- Unable to create a model deployment because Azure Marketplace rejected the subscription request.
- These are tied to:
- The billing country/region of the subscription.
- Provider’s allowed countries/regions for that specific SKU.
- Marketplace purchase eligibility for the subscription type.
- CSP subscriptions are explicitly unsupported for third‑party offerings, but an Azure Plan / Pay‑As‑You‑Go subscription can still be blocked if the billing region is not allowed for that SKU.
Quota or capacity
- For Foundry model deployments, quota issues are a documented cause of deployment failure:
- “Quota exceeded” when the subscription has reached deployment quota for the selected SKU or region.
- Recommended action is to check quotas in the Foundry portal or request an increase through Azure support.
- For Claude specifically, rate‑limit and quota issues are documented at the API level (HTTP 429), but deployment‑time quota exhaustion can also surface as a generic deployment error.
Permissions / RBAC
- For model deployments via CLI/Bicep, “Authorization failed” occurs when the identity does not have the required RBAC role; the fix is to assign the Cognitive Services Contributor role on the Foundry resource.
- In this case, the subscription role is Owner, which normally implies sufficient permissions, so RBAC is less likely to be the root cause.
Backend or transient platform issues
- The references include multiple examples where:
- A deployment error was ultimately resolved by changing region, indicating undocumented region‑specific capacity or configuration issues.
- A Foundry publishing error (“Value cannot be null. (TenantId)”) was fixed by a backend hotfix from the product group.
- A Claude deployment that was stuck with “API deployment for the resource is not ready” was resolved only by deleting and recreating the deployment later.
- These patterns indicate that some opaque error codes are caused by backend issues or transient platform faults that require either:
- Retrying the deployment (sometimes after deleting and recreating), or
- Direct intervention from Azure support / product group

These combination of checks will distinguish between:

Subscription/model availability or Marketplace eligibility issues (models not listed or region/country unsupported), and
A backend/platform issue (models listed and apparently eligible, but deployments still fail with 715‑123420), which then requires Azure support or a product hotfix.

The following references might be helpful , please check them out

Please let us know if the response was helpful

Thank you