Anthropic has been dragged into court after claims it flogged pricey Claude plans with more usage than customers received.
A federal lawsuit filed on 15 June 2026 claims Anthropic oversold the allowances for its Max 5x and Max 20x AI subscription plans. The complaint, brought on behalf of Washington, D.C.-based customer Karl Kahn, seeks customer reimbursements.
The lawsuit seen by the Wall Street Journal says Anthropic misled punters about the limits on its premium plans and wants class-action status. It covers people like Kahn who bought the plans since April 2025.
Anthropic’s Claude AI offers free and paid generative AI chatbot services, with more compute promised as the monthly bill climbs. Claude Pro costs $17 to $20 a month, while Max 5x costs $100 and Max 20x costs $200.
The case, filed in the Northern District of California, says Anthropic markets Max 5x and Max 20x as offering 5–20 times the usage caps of Pro. Some users claim the real caps are hard to work out and appear lower than advertised.
The lawsuit leans heavily on emails it says Anthropic sent to different subscriber tiers in July 2025. Those emails reportedly explained how much weekly usage each tier could expect from specific AI models then in use.
“The actual usage provided by the Max 5x and Max 20x plans is far below the advertised amount of usage,” said the lawsuit.
Kahn began using Claude for personal jobs before turning it into a coding workhorse. He upgraded to Max 20x in April 2026, then hit weekly limits within weeks.
In one five-hour coding sprint, Kahn allegedly burned through 15 per cent of his weekly allowance. That is not much fun when the monthly tab looks like a small utility bill.
Kahn “found himself needing either to halt his work, ration his usage, or purchase additional usage to ensure that he could complete his work,” alleged the lawsuit.
The complaint wants a court finding that Anthropic’s marketing of the plans was fraudulent. Anthropic did not immediately respond to Seeking Alpha’s request for comment.
The grumbling from Claude power users has become a useful warning light for the AI bubble. Companies are still trying to decide how much cash to throw at AI tools, how many staff they need and how many jobs can be rewired around chatbots.
That has put Anthropic, OpenAI and Google under the microscope. It has equally exposed the awkward truth that everyone wants clever AI, but nobody wants to admit how much compute it eats.
AI outfits are scrambling for the computing power needed to run their systems. Anthropic and OpenAI are both chasing possible IPOs, which makes customer anger about rationed access rather inconvenient.
