Anthropic Temporarily Bans OpenClaw Creator Peter Steinberger from Claude

Anthropic has temporarily banned Peter Steinberger — the creator of OpenClaw and now an OpenAI employee — from accessing Claude, according to a TechCrunch report.

Steinberger posted on X that he’d received word from Anthropic his account had been suspended over “suspicious” activity. Within hours, after the post went viral, his access was reinstated. An Anthropic engineer publicly offered help, claiming the company has “never banned anyone for using OpenClaw.”

The ban comes on the heels of a pricing policy change announced last week, where Anthropic said third-party harness usage — including OpenClaw — would no longer be covered under standard Claude subscriptions and would require separate pay-as-you-go API billing. Anthropic called this necessary because AI agents generate different “usage patterns” than prompts or scripts.

Steinberger pushed back, noting he was complying with the new API requirement when the ban happened. He also accused Anthropic of poor timing: “Funny how timings match up, first they copy some popular features into their closed harness, then they lock out open source.”

One notable detail: Steinberger now works at OpenAI, Anthropic’s chief competitor. When commenters noted the irony of him testing on a rival’s model, he clarified that his work for OpenClaw Foundation is model-agnostic, while his OpenAI role involves future product strategy. He added he’s “working on” reducing Claude dependency — a hint his OpenAI job may involve exactly that.

The incident highlights the escalating friction between OpenClaw’s open-source ecosystem and the proprietary AI providers it relies on. OpenClaw now sits at 347,000 GitHub stars and is navigating increasingly complex relationships with model companies who both support and compete with it.