Claude outage: paid users left without access while government continued working — Anthropic's double standards
On June 23, 2026, an 85-minute outage of the Claude AI paralyzed public access to the Anthropic platform, yet the specialized service for government agencies remained fully operational. The incident revealed a deep technical and ethical divide in the infrastructure: paid subscribers and developers incurred losses, while government users did not even notice the problem.
Timeline and Scale of the Incident
Anthropic specialists began their investigation 35 minutes after the first error signals. Although a technical solution was found relatively quickly, the total downtime lasted 85 minutes. The incident status changed to "resolved" only two hours after the initial problem was recorded. The outage affected key platforms: claude.ai, Claude API, Claude Code, Console, and Cowork. Notably, the government version of Claude was not included in the official list of affected segments.
Two Levels of Access: Architectural Inequality
The infrastructure separation was intentionally designed by developers. Claude for government agencies operates entirely independently, based on the specialized federal cloud service Palantir, which has strict FedRAMP High certification. Last year, the General Services Administration (GSA) granted a similar level of access to all three branches of the U.S. government for just $1. Fully isolated environments are considered the classic standard for strictly regulated cloud services. This is why government AI systems continued to operate without interruption while consumer products faced widespread outages.
Context: June Was a Tough Month for the Public Segment
The incident was just part of a very challenging period for the company's public segment. According to data from Anthropic's official status page, between June 9 and 23, technical specialists recorded more than 20 incidents. Users regularly encountered errors or complete system unavailability, which primarily affected the new flagship Opus 4.8. To cover the rapidly growing demand, the company quickly reserved up to 5 GW of new computing capacity from Amazon and signed new lease agreements for modern data centers. Against this backdrop, the isolated government segment initially remained safe, as it was not part of this overloaded resource pool.
Expert Opinion. This case is a clear marker of how prioritizing government contracts can create systemic risks for retail users. Paid subscribers essentially subsidize the development of infrastructure that becomes inaccessible to them at a critical moment. For long-term trust in the Anthropic platform, the model for distributing computing power must be reconsidered; otherwise, reputational losses may outweigh the benefits of government contracts.