In an expanded class action lawsuit, a group of Claude subscribers says Anthropic deceptively advertised the limits of its Max subscription tier
Anthropic says power users are key to its business — it's prioritized them even when it means cutting off other popular applications, like OpenClaw.
Context & Ripple Effects
Anthropic’s Max tier had already drawn complaints from Claude Code users after restrictive limits appeared on $200 Max plans in 2025. In March 2026, Anthropic said peak-hour demand would make users reach session limits faster, tying the subscription experience directly to capacity management.
The company also removed subscription coverage for third-party tools such as OpenClaw while prioritizing capacity for Claude users. The expanded suit turns that operational trade-off into a dispute over whether Max’s advertised usage allowances matched the limits subscribers encountered.
First-order effects
- Anthropic must defend its Max marketing and usage-limit disclosures in an expanded class action, while affected subscribers seek to challenge the value of plans priced around higher usage access.
- Claude Max users face a sharper distinction between advertised usage multipliers and the session or weekly ceilings alleged in the complaint.
Second-order effects
- Anthropic’s product and billing teams face pressure to make capacity-driven limits, peak-hour rules, and plan entitlements legible before purchase rather than only when users approach a cap.
- Competing AI subscription services gain a clearer incentive to differentiate on enforceable usage terms, not simply headline access multipliers.
Third-order effects
- If similar disputes persist, high-compute AI subscriptions may move toward capacity-aware contracts with explicit peak, weekly, and tool-specific allowances instead of broad “more usage” promises.
- Consumer claims over AI plan limits could make subscription marketing a governance issue for model providers whose marginal compute demand changes with user behavior.
The trend: AI providers are testing whether flat-rate premium subscriptions can coexist with volatile inference capacity without turning usage controls into a customer-accountability problem.