Early-stage startups attracted by the prospect of up to $150,000 in free Azure credits have discovered a painful gap in what those credits actually cover. Multiple founders report unexpectedly finding their credit cards charged hundreds or thousands of dollars after experimenting with third-party AI models, with neither Microsoft nor the model vendors accepting responsibility for the bills.
Riyaj Shaikh was billed for several thousand dollars after he mistakenly assumed startup credits covered his foray into Anthropic's products via Azure. Bogdan Sevriukov, a CTO with decades of experience and no stranger to cloud computing, created an Azure account in September 2025, tempted by the prospect of startup credits worth up to $150,000. When Microsoft announced its Anthropic partnership in November and advertised Claude availability for Foundry customers, he assumed his credits applied. They didn't, and Microsoft billed him nearly a thousand euros.
The core issue is straightforward, even if the user experience is not. Accessing Anthropic's Claude via Azure AI Foundry incurs charges that startup credits don't cover, something a Microsoft support forum moderator initially insisted wasn't the case. Startup credits cannot be used for Microsoft Azure support plans, third-party branded products, products sold through Microsoft Azure Marketplace, or products otherwise sold separately from Microsoft Azure. That rule exists in Microsoft's documentation, buried in the fine print.
What founders found objectionable was the execution. Azure AI Foundry displays both Microsoft-native models (such as Azure OpenAI) and third-party Marketplace models (such as Anthropic Claude) in a completely unified interface with no visual distinction, no warning, and no confirmation step before charges are incurred. The system provides few to no notifications warning users when startup credits are converted into real credit card charges. The first many founders knew about the problem was when an invoice appeared on their credit card statement.
The attribution tangle makes matters worse. Shaikh found himself trapped in a loop when he sought a refund: Microsoft pointed him to Anthropic and then Anthropic pointed him back to Microsoft. A technical lead for Microsoft Azure Subscription and Billing Management acknowledged to one founder that "while deploying the services, the system did not notify you that these credits could not be consumed with your available benefits," yet he received no warning when charges began to rack up.
The situation has prompted organised action. At least 20 participants in the Microsoft for Startups program have signed a Change.org petition calling on Microsoft CEO Satya Nadella to address what they describe as a "billing trap" inside Azure AI Foundry. The petitioners are urging Microsoft to introduce clearer labeling, explicit billing warnings, and confirmation prompts before developers deploy third-party models.
The complaints raise a legitimate design question. Cloud platforms have a responsibility to make pricing transparent, especially when the audience is cash-constrained startups trying to evaluate new technology. The question is not whether third-party marketplace charges should exist; they have their place. The question is whether a unified interface with no visual distinction between covered and uncovered services meets a reasonable standard for transparency. When Microsoft's own support moderators provide contradictory billing guidance, and users face a bureaucratic maze to resolve charges, the burden of clarity falls back on the company that owns the platform. Microsoft told reporters that it listens to customer feedback and continuously works to provide clear guidance in product documentation and pricing details. Clearly, not all customers agree.