Anthropic dropped a bombshell on Friday evening. Starting April 4th at 3PM ET, your Claude subscription no longer covers usage through third-party tools like OpenClaw. If you want to keep using Claude through OpenClaw, you will need to pay extra through a new pay-as-you-go system billed separately from your subscription.

The timing is brutal. OpenClaw has exploded in popularity this year, becoming the go-to way for power users to manage everything from inboxes to calendars to flight check-ins through Claude. That popularity, it turns out, is exactly the problem. Anthropic says its subscriptions "weren't built for the usage patterns of these third-party tools" and that it is "prioritizing customers using our products and API."

Translation: OpenClaw users were burning through too much compute, and Anthropic would rather those cycles go to Claude.ai and Claude Code users who stay inside the walled garden.

Here is where it gets interesting. OpenClaw's founder, Peter Steinberger, now works at OpenAI. He says he and OpenClaw board member Dave Morin "tried to talk sense into Anthropic" and the best they managed was delaying the change by a week. That is not a negotiation. That is a courtesy call before the execution.

Anthropic is offering subscribers a one-time credit equal to their monthly plan cost as a consolation. Discounted usage bundles are now available for those who want to keep using third-party tools. Full refunds are available for anyone who wants to walk.

The real story is not about capacity management. It is about platform control. Anthropic watched its competitor's former employee build the most popular way to use Claude, and now it is pulling the rug. The company is nudging users toward its own tools, including Claude Cowork, its answer to the agent-powered workflow that OpenClaw pioneered.

This is the classic platform play: let third-party developers prove the market, then reclaim the territory. Apple did it with apps. Google did it with search features. Now Anthropic is doing it with AI agents.

The developer community on Hacker News and Reddit is furious. Many feel they were sold a subscription with certain capabilities, and those capabilities are being removed mid-cycle. Anthropic's framing of this as a sustainability move will be tested by how many users actually leave versus how many grudgingly open their wallets for pay-as-you-go pricing.

Our take: Anthropic is making a rational business decision and an emotionally idiotic one simultaneously. The developers being cut off today are exactly the power users who evangelize Claude to everyone they know. Losing their goodwill to save on compute costs is the kind of short-term thinking that creates long-term regret. OpenAI is watching this unfold with a very large smile.