Detailed Analysis
Anthropic's Claude gift subscription system has developed a documented, recurring billing bug in which paid plans — primarily the Max 5x tier at $100 per month — silently revert to the Free plan after approximately one week of use, with no notification sent to the affected user. The pattern is substantiated across multiple platforms: at least five distinct GitHub issues filed against Anthropic's public repositories, several Reddit threads spanning multiple months, and a growing cluster of corroborating comments from users who independently encountered identical symptoms. In the most thoroughly documented case, a recipient activated a gifted Max 5x subscription on March 29, 2026, received an official confirmation email and a paid invoice, used the plan normally for seven days, and then found their account silently downgraded to Free around April 5–6 with no explanation. The invoice remained visible in account settings showing a "Paid" status, while the active plan displayed as Free — a direct contradiction between billing records and actual service access.
The structural problem compounds the technical one. Anthropic's primary support channel routes users through an AI bot called Fin, which — in a particularly damaging dynamic — confirms that the issue is a backend technical failure while simultaneously stating it has no capability to restore the subscription or escalate to the engineering team. Users who navigate past Fin to submit human support tickets report receiving no response for three or more days. This creates a closed loop in which Anthropic's own infrastructure acknowledges taking a user's money, acknowledges losing the associated service, and provides no actionable path to restoration. Gift subscriptions are particularly exposed to this failure mode because no recurring payment method is attached to them; when the system drops the plan, there is no re-authorization trigger and no automatic alert, leaving only a $0.00 invoice as evidence of a transaction that the system effectively treats as void.
The bug appears to have persisted for several months without a public acknowledgment from Anthropic, which is notable given the company's otherwise active developer communication posture. Anthropic's official documentation describes gift redemption as a straightforward process in which the plan activates immediately upon redemption at claude.ai/gift/redeem, granting full access to Pro, Max 5x, or Max 20x features for the purchased duration. The gap between that documented behavior and the reported reality represents a meaningful product reliability failure, particularly because the affected price points — $100 per month for Max 5x, with multi-month options available — place these purchases in a category where customers hold reasonable expectations of standard billing accountability.
The episode reflects a broader tension that emerges as AI companies scale their consumer subscription infrastructure rapidly alongside their model capabilities. Anthropic has invested heavily in safety research and frontier model development, and Claude's underlying capabilities are widely regarded as competitive at the frontier level. But billing reliability, support responsiveness, and transparent bug communication are foundational elements of a sustainable subscription business, and gaps in these areas carry disproportionate reputational risk precisely because they affect paying customers at the moment they are most engaged with the product. The gift subscription pathway, which functions as a word-of-mouth growth mechanism, is especially sensitive: a gifter who intends to introduce someone to the platform and instead delivers a broken experience creates negative signal in exactly the demographic most likely to become long-term subscribers.
The absence of any official acknowledgment from Anthropic, combined with the multi-month span of documented cases, suggests this is not a transient infrastructure anomaly but a systemic gap in how the gift redemption flow interacts with the subscription management backend — possibly related to the lack of an attached payment method creating edge cases in plan state management. Until Anthropic publicly acknowledges the bug, communicates a remediation path for affected users, and ensures that human support is accessible to customers whose accounts are incorrectly showing as Free-tier due to company-side failures, the issue is likely to continue generating new cases and accumulating additional public documentation across developer forums and community platforms.
Read original article →