An incident involving a user of the AI platform Claude from Anthropic has revealed serious failures in the startup's billing system. On July 7, a developer from South Korea under the pseudonym remy_notes received an initial invoice for $1.6 million. At first, he thought it was phishing, but the next day a new demand arrived — for $16.6 million. The sender and payment link led to the official Anthropic domain and its Stripe payment gateway, while the user's own dashboard showed zero API usage.
Attempts to charge the developer's bank card were rejected twice due to exceeding the per-transaction limit. The transfers came from a U.S. merchant named ANTHROP. The user checked all running processes on his Mac mini, including eight AI agents, but found nothing suspicious. As a preventive measure, he canceled his Claude Max subscription and blocked his card.
It took four days and about 18 emails to customer support to clarify the situation. Some responses came automatically — from an Anthropic AI agent named Fin. Only on July 12 did the company confirm the error, stating that not a single dollar had been charged. The cause was attributed to a misconfigured auto-reload balance feature, which was disabled as a precaution. However, Anthropic could not explain why the system generated exactly $16.6 million.
Not an isolated case: systemic billing errors
This incident is not isolated. The auditing firm Vaudit, after reviewing invoices totaling $34 million across 60 organizations from March to June, identified overcharges of $1.7 million (about 5% of the audited volume). Most invoices were related to the use of Claude Code by Anthropic. Specialists discovered several systemic errors: customers were billed at rates for new models, even though an older, cheaper version was actually used; charges were recorded for requests that ended in errors or yielded no results. A separate category consisted of "retry storms" — when an AI agent repeatedly retries a failed task in the background, with each attempt reflected on the bill.
After Vaudit and client complaints, Amazon, Google, Microsoft, Anthropic, and OpenAI refunded about 80% of the overpayments.
This case highlights a fundamental problem: as AI infrastructure becomes more complex, billing systems turn into a "black box" where errors can cost millions. Users should carefully monitor their bills and set strict limits on auto-reload, while companies should invest in auditing their payment systems — otherwise, reputational risks will outweigh any savings.