“Anthropic ends subscription-based OpenClaw access April 4, forcing users to pay-as-you-go API rates or buy extra usage bundles”
“Heavy OpenClaw users ran up $1,000 to $5,000 in daily API-equivalent costs on $200/month Claude Max plans”
“OpenClaw creator Peter Steinberger, now at OpenAI, accuses Anthropic of copying features before locking out open-source competitors”
“Subscribers receive a one-time credit equal to their monthly plan cost, redeemable by April 17”
“Anthropic announced Friday evening that Claude subscriptions will no longer cover usage through third-party tools including OpenClaw, effective April 4 at 12 pm PT. Users who want to keep running OpenClaw with Claude must now purchase separate "extra usage" bundles or switch to API-key billing, according to Boris Cherny, head of Claude Code. The change breaks apart one of the most popular combinations in autonomous AI tooling and exposes a widening gap between flat-rate subscription pricing and the compute demands of agent-driven workloads.”
Anthropic then went on to immediately offer its own ‘Claude Managed Agents’ offering with a la carte pricing. With more revenue earning potential vs its ‘all you can eat’ subscription offerings. From $20/month to over $200/month.
Again, the battle between the two companies harkens behind the ‘all you can eat’ business conflict in ‘variable cost’ usage and pricing driven businesses rom times immemorial. From Chinese lunch buffets to mainstream users using up huge amounts of internet dial-up capacity on AOL in the nineties on flat monthly pricing.
And the battles are getting a bit personal, for the principals here, and customers. As Venturebeat explains further:
“The response from the developer community has been a mixture of analytical acceptance and sharp frustration.”
“Growth marketer Aakash Gupta observed on X that the “all-you-can-eat buffet just closed,” noting that a single OpenClaw agent running for one day could burn $1,000 to $5,000 in API costs. “Anthropic was eating that difference on every user who routed through a third-party harness,” Gupta wrote. “That’s the pace of a company watching its margin evaporate in real time.”
“However, Peter Steinberger, the creator of OpenClaw who was recently hired by OpenAI, took a more skeptical view of the “capacity” argument.“Funny how timings match up,” Steinberger posted on X. “First they copy some popular features into their closed harness, then they lock out open source.”
“Steinberger claimed that he and fellow investor Dave Morin attempted to “talk sense” into Anthropic, but were only able to delay the enforcement by a single week.”
“User @ashen_one, founder of Telaga Charity, voiced a concern likely shared by other small-scale builders: “If I switch both [OpenClaw instances] to an API key or the extra usage you’re recommending here, it’s going to be far too expensive to make it worth using. I’ll probably have to switch over to a different model at this point.”
.“I know it sucks,” Cherny replied. “Fundamentally engineering is about tradeoffs, and one of the things we do to serve a lot of customers is optimize the way subscriptions work to serve as many people as possible with the best mode.”
The IPO timing angle on the OpenClaw ban hadn't occurred to me before reading this. Heavy users generating thousands in API-equivalent costs per day on $200/month plans - at scale that's genuinely material to unit economics before a public offering.
The Steinberger-OpenAI connection adds a competitive layer that makes it harder to read as purely principled. Wrote about the subscription math alongside Managed Agents (which is metered from day one): https://thoughts.jock.pl/p/ai-opinions-april-2026-claude-mythos-meta-spark Anthropic building the official orchestration layer while banning the unofficial one is consistent strategy, at least.
The IPO timing angle on the OpenClaw ban hadn't occurred to me before reading this. Heavy users generating thousands in API-equivalent costs per day on $200/month plans - at scale that's genuinely material to unit economics before a public offering.
The Steinberger-OpenAI connection adds a competitive layer that makes it harder to read as purely principled. Wrote about the subscription math alongside Managed Agents (which is metered from day one): https://thoughts.jock.pl/p/ai-opinions-april-2026-claude-mythos-meta-spark Anthropic building the official orchestration layer while banning the unofficial one is consistent strategy, at least.