A group of Claude subscribers is dragging Anthropic into court, and the claim cuts right at a sore spot for anyone paying for AI these days. According to The Verge AI, an expanded class action lawsuit filed today accuses the company of deceptively advertising the usage limits on its Max subscription tier. The case is led by attorneys Monica Vaca and Kati Daffan, both former Federal Trade Commission staffers under Lina Khan.
This is significant because it’s a rare attempt to put an AI company on the hook legally for something users have been grumbling about for months: the industry’s tightening money squeeze.
What the lawsuit claims
Anthropic’s Max plan sits above the $20/month Pro tier and comes in two flavors:
- $100/month for “5x” the usage of Pro
- $200/month for “20x” the usage of Pro
Those big multipliers are the problem, the complaint says. The Verge AI reports that the “20x” and “5x” figures are technically real, but only for five-hour chunks of time that are also capped on a weekly basis. Add it up, and the plaintiffs argue the actual boost in capacity is far smaller than the marketing suggests.
To even find the catch, Vaca said, “you’ve got to dive deep.” It takes clicking two separate hyperlinks to reach the word “session,” then jumping to the Pro plan page to learn what “session” actually means.
The weekly allowance is what the pricing page makes you think you’re buying. The rolling 5-hour window is what actually controls whether you can work… It’s like giving someone a bigger gas tank while keeping the fuel pump limited to one gallon every five hours.
Why the timing matters
Here’s the detail that stands out. Anthropic announced the Max plan in April 2025, but the disputed weekly limits didn’t land until a few months later in August, as the company pushed harder to compete with OpenAI.
That lines up with a broader pattern. Over the past eight months or so, as pressure to turn a profit builds, AI customers have complained that top labs are quietly passing steep operating costs down to subscribers. Anthropic itself nodded to the issue in its recent Fable 5.1 release, saying it was “addressing the feedback we’ve received from customers on price.”
Anthropic’s defense
The complaint was first filed in July, then withdrawn and refiled as an expanded class action. In an earlier motion to dismiss, Anthropic argued it didn’t hide the session limits at all.
Accessing that fine print required nothing more than clicking hyperlinks available in the purchase process, the digital equivalent of flipping a product over to read the back label.
Anthropic did not respond to The Verge AI’s request for comment.
Vaca pushes back hard on the “read the label” logic.
This is hard for consumers, they don’t know what’s in that black box. There’s no way for them to audit it… They’re basically taking a leap, and they’re believing in an honest marketplace.
Why practitioners should care
If you pay for AI tools, this case touches you directly. A few things worth watching:
- Usage caps are getting complicated. Advertised multipliers mean little without the session and weekly math behind them. Read the mechanics, not the headline number.
- The pricing squeeze is real. Labs are balancing compute costs against subscriber expectations, and mid-plan rule changes like Anthropic’s August shift can happen again.
- Legal precedent is forming. Vaca notes she and Daffan spent 38 combined years at the FTC. “There is a long line of precedent on false advertising,” she said. “You can’t lie when you’re marketing a product.”
There’s a human angle too. Vaca said some plaintiffs felt pressured to buy pricey AI subscriptions just to stay relevant in the job market, then felt shortchanged on a service running $100 to $200 a month.
Whether the court agrees that a buried hyperlink counts as fair disclosure could shape how every AI company advertises its limits going forward. Expect other labs to watch this one closely. Full details are available in the original report from The Verge AI.