Anthropic restricted access to its Claude AI models for users of the OpenClaw agent tool earlier this month, requiring them to pay separately for usage through the company’s API.
The change affects millions of users who previously accessed Claude AI via OpenClaw under Anthropic’s standard subscription plans, which did not cover third-party tool usage.
Anthropic cited unsustainable usage patterns from tools like OpenClaw as the reason for the shift, stating its subscriptions were not designed to handle the compute-intensive nature of such agents.
Boris Cherny, head of Claude Code at Anthropic, said the company aims to manage growth intentionally to serve customers sustainably long-term, framing the restriction as a necessary step.
The move reflects broader pressure on AI labs to monetize their services after years of offering low-cost or free access to advanced models, driven by investor expectations for returns on massive infrastructure investments.
Anthropic, like OpenAI and other leading AI firms, has introduced new pricing tiers and revenue strategies, including in-platform advertisements and enterprise-focused plans, to offset the costs of scaling AI systems.
Industry analysts note that the era of near-free AI access may be ending as companies face the financial reality of sustaining large-scale model development, and deployment.
Why did Anthropic restrict OpenClaw access to Claude AI?
Anthropic said its subscriptions were not built for the usage patterns of third-party tools like OpenClaw, which can be more compute-intensive due to continuous reasoning loops and integration with other tools.
What must OpenClaw users now do to continue using Claude AI?
OpenClaw users must pay for Claude AI usage separately based on consumption through Anthropic’s API, as subscriptions no longer cover third-party tool access.