Anthropic faces lawsuit over Claude Max 5x and 20x usage oversold for $100 and $200

Anthropic has been hit with a California lawsuit accusing it of overselling usage on its most expensive Claude subscription tiers. Filed this week, the complaint comes from Washington, D.C. customer Karl Kahn, who alleges that Anthropic’s “Max 5x” plan, priced at $100 a month, and its “Max 20x” plan, priced at $200 a month, deliver far less usage than the company advertised.

The core of the allegation is straightforward but high-stakes for how these businesses monetize AI: Kahn is claiming the “x” branding implies a quantity or capacity promise that customers do not actually receive. The complaint asks a court to compel Anthropic to address the alleged gap between what was sold and what was delivered. In other words, this is not a vague complaint about customer support or billing confusion. It is a direct challenge to product promises tied to money.

To understand why this matters beyond one customer’s bill, you have to look at how AI subscription plans work in practice. Most AI platforms sell some combination of access, rate limits, and usage-based capacity, often with tiers meant to translate technical constraints into simple consumer language. When companies market plans with multipliers like “5x” and “20x,” they are effectively telling buyers that more dollars equals more usable throughput, even if the underlying system has variables like demand patterns, scheduling, or how “usage” is defined.

That is exactly where lawsuits like this become leverage points. Even if the technical implementation is complex, the customer-facing claim is not. If a plan is marketed in a way that implies a consistent usage outcome, buyers organize their workflows around it: developers plan compute needs, teams forecast costs, and executives set budgets assuming the tier will behave like the promise. When a complaint asserts the delivered usage is “far less” than advertised, the alleged harm is not theoretical. It affects operational planning and unit economics for anyone buying at scale.

There is also a real incentive problem that boards and CFOs should care about. The priciest plans are where margin and retention pressures concentrate. If enterprise or power users are deciding between tiers, the exact mapping from marketing claims to delivered capacity is part of the competitive pitch. A customer paying $200 a month does not expect to be treated as if they bought the low end. If the “Max 20x” tier is alleged to underperform versus what was advertised, that undermines trust, forces customers to rerate their risk, and can accelerate churn or downgrade decisions.

Kahn’s decision to sue in California matters too, because it signals the complaint is aimed at a formal legal remedy and not just a request for customer service. The source notes that the complaint asks a court to make Anthropic do something, which means the issue could quickly stop being only a product dispute and become a question of enforceability and remedies. Courts do not just care whether something went wrong. They care whether what was sold can be proven, whether the claims were represented clearly enough, and what it means to “oversell” usage when the service is partially limited by system constraints.

For executives at AI companies, this is also a governance moment. When product teams tighten pricing, or sales teams push higher tiers, the company needs tight alignment on how “usage” is defined, measured, and communicated. If marketing uses “x” language, internal accounting has to match it precisely, including edge cases. Boards should ask basic but uncomfortable questions: Are there clear disclosures about what “5x” and “20x” mean? Are there measurable differences tied to these tiers? Are there audit trails that show delivered usage versus advertised claims across time?

Second-order effects are likely. If the complaint gains traction, it can encourage other customers to bring similar actions, especially for the highest-priced tiers where the financial stakes are largest. It can also prompt broader regulatory and consumer protection scrutiny of how AI services market usage, since these subscription models sit at the intersection of digital advertising, billing, and performance expectations. Even for companies not named in this filing, the reputational risk is real because customers will ask the same question: does the tier actually match the promise?

In the AI subscription economy, the product is the promise. Anthropic now faces a legal challenge over its most premium Claude plans, specifically the “Max 5x” at $100 a month and “Max 20x” at $200 a month, with the allegation that delivered usage is “far less” than advertised. For peers, the strategic takeaway is simple: if you sell capacity in plain-English multipliers, you need operational evidence and clear definitions that can survive a courtroom timeline.