Key Highlights:
- Anthropic has removed OpenClaw from its subscription plan and shifted its usage to a separate pay-as-you-go option.
- This move reflects the shift toward usage-based AI pricing and rising infrastructure force.
- Starting from 4 April, 2026 noon, the company has announced that the users will no longer be able to use Claude subscription limits for third-party-harnesses.
The removal of all third-party harnesses from Claude subscription plans by Anthropic reflects an important shift in how AI platforms are evolving. Developers will now have to pay separately to run agent workflows, essentially changing how these systems are used and scaled.
Anthropic’s Reason behind Doing this
Boris Cherny, head of Claude Code said in a statement: “We’ve been working hard to meet the increase in demand for Claude, and our subscriptions weren’t built for the usage patterns of these third-party tools. Capacity is a resource we manage thoughtfully and we are prioritizing our customers using our products and API.”
Anthropic has been direct with its reason behind this move. The company says that agent-style workflows, like those enabled by OpenClaw, consume a lot more energy than what a typical chat uses. According to Anthropic, that mismatch makes it difficult to sustain under one subscription model.
Anthropic described these tools as putting an outsized strain on its systems; this move reflects the need for better cost alignment and long term sustainability as the subscriptions were designed for interactive use not high-volume workloads.
But is it actually a cost issue or a controlled move?
Anthropic has always been seen as user-aligned and safety-focused prioritizing reliability and transparency over aggressive monetization.
At face value, this is a cost issue. Moving to usage-based pricing means correcting the imbalance created by agent workflows (heavy users being financially supported by lighter ones).
There’s another hard to ignore aspect in this situation! Third-party tools like OpenClaw don’t just consume energy, they define the workflows’ structure, and output generation, basically shaping how people interact with AI. The company can gain leverage and more control over how their models are used by limiting these third-party harnesses.
The strategy to retain developers within native ecosystems, stop additional layers from taking over the user experience, and more closely connect monetisation with actual usage is just as important as the infrastructure pressure.
Reports say: Peter Steinberg, creator of OpenClaw, expresses his disappointment towards Anthropic’s move saying “while I think what Anthropic does is sad for the ecosystem” .
He also suggested that the decision could have been avoided. He stated “tried to talk sense into Anthropic” . Basically, from his perspective, this was a decision that could affect developer trust.
Garry Tan wrote on X : “Anthropic shutting down OpenClaw may turn out to be a strategic blunder, or strategic genius. The OpenClaw community will be the determiner of whether it is A or B. It’s an interesting moment in history.”
This comment shows a wider uncertainty. The move could either protect the company from unsustainable costs or it risks frustrating and losing the same developers who helped build momentum around these tools.
It ultimately comes down to a lot of questions and curiosity.
What happens now? Will the developers still trust Anthropic? Are investors still going to invest? What is it going to look like for Claude in the future?
Is this Really the End of “Unlimited” AI?
The “unlimited” AI was never going to last with all these third-party integrations and powerful agent tools.
AI is becoming more and more powerful with each update, hence it progressively gets more expensive to run these models. This is what is pushing companies like Anthropic towards the “pay-as-you-go” models. This change is a disadvantage for developers as they will have to make cost decisions for everything and not just technical ones.
Also read: OpenClaw Founder, Peter Steinberger, Joins OpenAI Amid Surge in Autonomous AI Agents
Wrapping Up
This move reflects the future of the AI industry and where it is heading instead of just portraying a small pricing change for monetization of these services. Anthropic’s official statement and reasoning behind this move is fair but the debate and outrage shows that it’s not just about the cost but also about leverage and control over how much freedom developers have.
The next big question is: what comes next? Will developers accept usage-based pricing? Or move forward from Claude to other alternatives? Either way, the idea of unlimited AI is slowly starting to fade and paid models are replacing it.
