Detailed Analysis
A Reddit user's attempt to upgrade from Anthropic's Max 5x plan to the Max 20x tier surfaced a payment-processing bug that has left them locked out of both changing their billing method and continuing normal usage. The core issue stems from Stripe's Link payment system, which the user had previously used for convenience. When they tried to add a new card during checkout, the interface displayed the new card's details but silently continued attempting to charge the old Link-associated card in the background. After that card was declined and the user emptied the Link wallet entirely, the system entered a broken state, throwing a persistent error: "The connection to the user's Link account has been closed." Attempts to reconnect Link, use a different card, create a new Link account, or switch to Google Pay all failed at the same final checkout step.
This is a fairly narrow technical edge case, but it exposes a structural weakness in how Anthropic's billing infrastructure interacts with third-party payment processors. Link, a Stripe product designed to streamline checkout by remembering payment credentials across merchants, appears to retain a stale reference to a closed or emptied account that the front-end checkout flow doesn't properly reconcile against newly entered payment methods. The mismatch between what's displayed at checkout (the new card) and what's actually being charged (the old Link account) is a classic state-management bug, but its consequences are significant for the user: rather than simply failing to upgrade, they are functionally locked into their current plan and rate limits with no self-service path to remediation.
The more consequential part of the story is the support response, or lack thereof. The user reports waiting more than five days without a human reply, relying instead on Anthropic's automated support agent ("Fin," a common third-party AI customer-support tool used by many SaaS companies), which they describe as unable to resolve the issue. In the interim, they've hit their weekly usage limits on the lower-tier plan and are stuck waiting for the reset cycle while being unable to pay for the higher tier they want. This creates a bind that directly penalizes engaged, paying users who are trying to spend more money with the company but are prevented from doing so by a bug in the payment pipeline itself.
This complaint fits into a broader pattern of user frustration around Anthropic's customer support scalability as Claude's subscriber base has grown rapidly through 2025 and into 2026, particularly among developers and power users on Pro and Max tiers who depend on consistent access for work. As AI companies compete heavily on model capability, they are increasingly differentiated by operational reliability, billing transparency, and support responsiveness — areas where fast-scaling AI labs, often originally built around research rather than consumer SaaS operations, tend to lag behind more mature software companies. Incidents like this one, where a billing glitch effectively throttles a paying customer's productivity with no clear escalation path, underscore a recurring criticism in the Claude user community: that infrastructure and support systems haven't kept pace with the product's growth, even as usage-based and tiered subscription models make billing accuracy and account recovery increasingly mission-critical for professional users.
Read original article →