Anthropic Sued for Overselling Usage on Its 200-Dollar Claude Max Plans

Welcome back to SavvyMonk, your one-stop for AI and tech news that actually matters.

Anthropic spent the past week fighting a government order over its newest models, and now it has a second fight on its hands. A paying customer says the company sold usage it never delivered, and he wants every Max subscriber paid back.

Let's get into it.

Every excavator, forklift, and crane on the planet runs on hydraulic fluid. It leaks. It fails. It burns through 60% of the energy you put into it. That's been true for a hundred years.

RISE Robotics built Beltdraulics™ to fix all of that. Their patented actuator swaps out hydraulic cylinders for a fluid-free electric system that runs up to 3X faster and cuts operating costs by 50%. No oil. Full digital control. Built-in sensors that hydraulic systems can't touch.

The U.S. military is already a customer. MIT-founded. $9.3M in revenue. 20+ patents protecting the core technology. Dylan Jovine of ‘Behind the Markets’ said RISE “has all the little ingredients to be one of those really big winners.” His readers have been backing it ever since.

You can invest today through the community round on Wefunder.

The Ingredients of a Blockbuster Winner

A federal class-action lawsuit filed on Monday accuses Anthropic of overselling the usage allowances on its two most expensive Claude subscriptions.

The case, Kahn v. Anthropic PBC, was lodged in the US District Court for the Northern District of California and takes aim at the Max 5x plan at 100 dollars a month and the Max 20x plan at 200 dollars a month.

Both tiers are marketed as multiples of the base Claude Pro plan, which runs about 17 to 20 dollars a month, and the suit argues the real limits land far below those advertised multiples.

The plaintiff is Karl Kahn, a Washington customer represented by Kati Daffan, founding partner at the consumer protection firm Vaca Daffan LLP. He is seeking refunds and damages on behalf of everyone who bought a Max plan since the tiers launched in April 2025, and Anthropic declined to comment.

The Five-Hour Session That Started It

Kahn says he came to Claude for casual personal use, subscribed to Pro in June 2025, then moved up to Max 5x in January and to the top Max 20x tier in April as his work shifted toward heavy coding. The problem, according to the filing, showed up almost immediately.

A single five-hour coding session consumed roughly 15 percent of his weekly allowance, which left him choosing between stopping work, rationing his prompts, or paying for extra usage to finish what he started. The complaint frames that experience as the core of the case, arguing that the actual usage delivered by the Max tiers sits well below what the marketing promised and that a 20x plan that runs dry in a handful of sessions never came close to twenty times anything.

The July Email at the Centre of the Case

The most concrete piece of evidence is an email Anthropic sent subscribers in July 2025 that spelled out weekly usage expectations for individual models. Kahn argues those figures fell short of what the public 5x and 20x labels implied, and that the real ceilings were nearly impossible for a customer to track.

Part of the confusion comes from how the caps stack. Usage resets on a rolling five-hour window, and a separate weekly limit caps total activity on top of that, so a heavy user can hit a wall without any clear sense of which limit they tripped or when it lifts. The suit contends that this opacity is not a side effect but the mechanism that let the advertised multiples drift away from the service people actually received.

Why the Caps Exist in the First Place

Anthropic did not set out to throttle its best customers for sport. The weekly limits arrived in late August 2025 after a wave of power users ran Claude Code almost continuously, and the company said at the time that the change would touch fewer than 5 percent of subscribers.

The deeper tension is structural. Flat monthly pricing collides with the real per-prompt cost of serving someone who codes through Claude all day, and that gap is exactly where the dispute lives. This is also not the first complaint of its kind.

Back in July 2025, Claude Code users reported sudden restrictive limits with no announcement, which left many convinced their plans had been quietly downgraded. The lawsuit gathers that long-running frustration into a formal legal claim.

A Second Access Shock in the Same Week

The timing makes the suit sting more, because access to Anthropic's products has been unstable on a second front.

Days earlier, a US government order barred foreign governments, companies, and individuals from using the company's newest models, Fable 5 and Mythos 5, and Anthropic disabled both worldwide to comply rather than risk serving a restricted user.

The directive followed research from Amazon showing that some of Fable 5's cybersecurity safeguards could be bypassed to pull information about software vulnerabilities. Anthropic dispatched senior technical staff to Washington to argue for restoring access, all while Fable 5 was already scheduled to leave Claude subscriptions on 23 June and move to pricier usage-based billing.

For a paying customer, the throughline is uncomfortable. In one week the company faced a lawsuit over usage that allegedly vanished behind hidden caps and a government order that made a flagship model vanish outright.

The Backdrop Anthropic Did Not Want

All of this lands at a delicate moment. Anthropic recently filed a confidential S-1 and is preparing for a closely watched public listing after a funding round that valued it near 965 billion dollars, which makes it the most valuable AI startup in the world. A class action built around transparency and honest advertising is precisely the kind of story a company does not want trailing it into an IPO roadshow.

Daffan has argued that consumer protection law is settled on this point, that firms have to be honest about what they sell, and that companies which mislead customers can be forced to change their practices and return money. Whether a court agrees, the case puts a spotlight on a question the whole industry has been dodging, which is what a flat-rate AI subscription actually guarantees.

The Bottom Line

Anthropic sells the Max tiers as simple multipliers, and this lawsuit argues the math never held once real work began. The company has a defensible reason for its caps, but reasons are not the same as disclosure, and the case will turn on whether a buyer could understand what they were getting.

Pair that with a model pulled offline by government order in the same week, and the message to subscribers is hard to miss. Paying more does not always buy you certainty, and Anthropic now has to prove in court that it sold what it advertised.

Category: Consumer Research

"I pay for a subscription to [Service Name] at [Price] per month. Here is what the plan advertises for usage or features, copied below. [paste it]. Ask me five questions about how I actually use the service, then tell me in plain language whether I am getting my money's worth, what limits I am most likely hitting, and whether a cheaper or different plan would serve me better."

For most of the AI boom, the limits on these tools have been a footnote buried in a settings page, accepted as the cost of access to something genuinely useful. This case asks a sharper question. When the headline number on a plan and the real number you can use drift this far apart, is that a fine-print problem or a marketing one. A court is about to decide, and the answer will shape how every AI company describes a subscription from here.

Hit reply, I read every response.

See you in the next one.

— Vivek

P.S. Know a developer or founder paying for a premium AI plan and wondering where their usage goes? Forward them this issue. They can subscribe at https://savvymonk.beehiiv.com/