Trend

Claude Outages Surge as Anthropic Chases 2026 Revenue Lead Over OpenAI

co-written by newsrooms.ai27. March 2026, 12:22

“Unexpected capacity limitations” are said to be repeatedly preventing Claude users from receiving responses to their requests. Anthropic’s AI service has been struggling with repeated outages for several days. The company’s status page shows that disruptions have occurred repeatedly, especially in March and particularly in recent days, affecting various services.

Anthropic is likely under considerable pressure right now. Following OpenAI’s setback from the Pentagon deal, there was a mass migration of users toward Anthropic (more on that here), which already caused outages due to overload at the beginning of March. In the US, Germany, Canada, and France, the Claude app continues to rank 1st or 2nd in the app charts, indicating sustained demand.

At the same time, analyses suggest that this strong demand for Claude is growing so steeply that Anthropic could overtake OpenAI in revenue as early as mid-2026. This is driven primarily by B2B use cases in the productivity space, where Anthropic scores points with “Cowork,” “Code,” and integrations into work software. Anthropic is currently growing its revenue tenfold:

Which services and models are affected?

The disruptions affect several areas of the Claude ecosystem simultaneously. The most powerful model Claude Opus 4.6 appears most frequently in the outage reports. The following services are also affected:

Claude.ai: The web-based chat interface as well as the associated desktop and mobile apps
Claude Code: Anthropic’s developer-focused coding tool
Claude Cowork: The collaboration feature, which is experiencing connection drops, with a restart of the desktop app recommended
Developer Console: The developer console at platform.claude.com
MCP calls: Errors in the Model Context Protocol, which enables third-party integrations

The difference between Claude.ai and the API

An important technical aspect of the current outages is that not all Claude services are equally affected. Fundamentally, two access routes can be distinguished:

Claude.ai is Anthropic’s direct user interface — the website, the desktop app, and the mobile apps. Anyone using Claude there is accessing Anthropic’s infrastructure directly. When disruptions occur at this level, the service is unavailable or only partially available to end users.

The API (Application Programming Interface) is a programming interface through which external developers and companies integrate the Claude models into their own products and services. Anyone using an app or web service that accesses Claude in the background does so via this interface. According to Anthropic, the API was not affected by the recent disruptions and continued to function normally. Third-party products with API integration therefore generally remained stable.

Background: Mass influx following the ChatGPT migration

The repeated outages coincide with an unusually strong surge in user growth. Since the beginning of March, before the disruptions, a clear movement from ChatGPT toward Claude has been observed. The trigger was a publicly disclosed deal between OpenAI and the US Department of Defense, which Anthropic stated it had declined.

In response, download numbers for the Claude apps rose sharply in key markets such as the US and Germany. The Claude app climbed to number 1 in the app store download charts, displacing ChatGPT from the top position. Many users actively sought out guides on how to delete their ChatGPT account and transfer their data to Claude.

Whether this sudden surge in users is directly responsible for the infrastructure overload has not yet been officially confirmed by Anthropic. The company merely acknowledged that the outages occurred and have repeated over the past few days. A detailed root cause analysis is still pending.

Conclusion

The repeated disruptions at Claude show that Anthropic’s infrastructure is reaching its limits under growing user pressure. The flagship model Opus 4.6 and the direct user interfaces are particularly affected, while the API interface remained stable. Users who depend on stability can temporarily switch to API-based third-party services or keep an eye on the status page at status.claude.com.