The promise of artificial intelligence has always carried a little electricity. A machine that can code, reason, write, research and work beside us feels less like software and more like an extra mind at the desk.
For users willing to pay a premium, the attraction is obvious: if more access means more productivity, then paying more should buy substantially more freedom. That expectation now sits at the center of legal claims facing Anthropic over its Claude Max subscription plans.
The dispute concerns the company’s $100 and $200 monthly Max tiers and whether their advertised usage multipliers accurately represent what subscribers can actually consume.
Anthropic currently describes Max as providing either five times or 20 times the usage of Pro per five-hour session, while also acknowledging that paid plans operate under additional weekly limits. That distinction is crucial.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
A customer reading 20x may naturally imagine twenty times the overall capacity. But the legal complaint reportedly argues that the practical experience can look dramatically different, with the $200 tier allegedly providing only around twice the weekly usage of the $100 tier in some circumstances.
The result is a collision between marketing language and the mathematics of AI consumption. The deeper issue is that AI usage is not measured like minutes on a telephone plan. A short question may consume very little computational capacity, while a long coding session.
Complex reasoning task or enormous context window can burn through an allowance rapidly. Anthropic itself says usage depends on conversation length, model choice, features and the complexity of work.
That makes the phrase “20x more usage” both powerful and potentially confusing. For developers and professional users, the difference is not academic.
Claude Code, for example, draws from the same usage pool as Claude conversations. A user engaged in sustained software development can therefore consume significant capacity without necessarily feeling that the advertised multiplier translates into twenty times more working days.
Reports from Max subscribers have illustrated that frustration. One account described a $200 subscriber consuming 15% of a weekly allowance during a single five-hour coding session. Another published account described purchasing the $200 plan after seeing the promise of 20x usage.
Only to encounter restrictions that made the value proposition feel considerably smaller. But the controversy also exposes an important ambiguity in the lawsuit itself. Anthropic’s wording is specifically tied to per-session usage, not necessarily a promise of twenty times the weekly capacity.
Its pricing documentation explicitly states that weekly limits exist alongside rolling five-hour limits. And so the argument becomes a question of interpretation: what would a reasonable customer believe 20x means when deciding whether to spend $200 every month?
The answer could have consequences far beyond Anthropic. As AI companies increasingly sell premium subscriptions to programmers, researchers, writers and businesses, usage limits are becoming part of the economics of intelligence itself.
Consumers are no longer simply buying software. They are buying access to scarce computational resources, packaged in language designed to make those resources understandable.
If customers believe the language promises abundance while the architecture delivers scarcity, trust becomes the casualty.
The Anthropic dispute therefore represents something larger than one subscription plan. It is a warning that in the age of AI, numbers can be technically accurate and commercially misunderstood at the same time. Twenty times can sound like infinity—until the weekly clock starts ticking.