Claude outage: paid users left without access, while the government segment operated without interruptions
A 85-minute outage of the AI Claude on June 23, 2026, paralyzed public access to Anthropic's artificial intelligence, but the specialized service for government authorities continued to function without any disruptions. This incident clearly demonstrated how deeply the company separates its public and federal infrastructure.
Timeline and Scope of the Incident
Anthropic specialists promptly launched an investigation and reported finding a technical solution within 35 minutes. However, general disruptions persisted for 85 minutes, and the incident status changed to "resolved" only two hours after the first signal was recorded. Errors affected several key platforms simultaneously: claude.ai, Claude API, Claude Code, Console, and Cowork. Notably, the specialized Claude service for government agencies was not included in the official list of affected segments.
Two Levels of Access: Deliberate Separation
Such a deep separation of AI systems was intentionally implemented by the developers. The government version of Claude operates entirely independently from commercial users. The platform is deployed on a specialized federal cloud service, Palantir, which holds strict FedRAMP High certification. Last year, the General Services Administration (GSA) provided a similar level of access to all three branches of government for just $1. Fully isolated environments are considered the classic standard for strictly regulated cloud services. This is precisely why government AI systems continued to operate without interruption while consumer products faced widespread outages.
Consequences and Community Reaction
Users, especially paid subscribers of the service, reacted extremely sharply to such a technological divide. One popular social media post accurately captured the general pessimistic mood of the audience: "Claude is down for everyone except the government."
The incident turned out to be just part of a very challenging period for the company's public segment. According to Anthropic's official status page, technical specialists recorded more than 20 incidents between June 9 and 23. Users regularly encountered errors or complete system unavailability, which primarily affected the new flagship Opus 4.8.
To effectively cover the rapidly growing demand, the company promptly reserved up to 5 GW of new computing capacity from Amazon. Additionally, management 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.
My analysis: This case is a clear signal for retail users and businesses. While government contracts ensure priority access and stability, paid subscribers find themselves at risk. As Anthropic actively scales its infrastructure, the gap in service quality between the "elite" and "mass" segments will only widen. Investors and users should consider this imbalance when assessing the long-term reliability of public AI products.