Anthropic confirmed on Monday, March 2, that its AI chatbot Claude suffered widespread outages affecting users across multiple platforms.
At 11:49 UTC, the company posted an “Investigating” notice on its status page. A follow-up update at 12:06 UTC said the investigation was ongoing.
Users attempting to access Claude.ai, the company’s mobile apps, developer console, and its coding assistant reported error messages, including: “That’s not working right now. You can try again later.”
Anthropic said in a statement that “Claude is currently unavailable on our consumer-facing surfaces such as claude.ai and our apps. The Claude API that powers businesses remains unaffected. Our team is working to restore full service.”
More than 2,000 users reported problems within hours, according to outage tracking services. The disruption follows a similar “elevated errors” incident on February 28, 2026, which affected Claude’s Opus 4.6 model.
The technical failure came just hours after Claude reached number one on Apple’s U.S. App Store free app rankings, overtaking OpenAI’s ChatGPT. Industry trackers show Claude’s free user base has increased by more than 60 percent since January.
The surge is linked to growing public reaction to OpenAI’s newly announced partnership with the U.S. Department of Defense.
OpenAI confirmed it would provide AI tools to the Pentagon for use in classified military environments. Its chief executive, Sam Altman, acknowledged the agreement publicly following criticism.
The development triggered an online campaign encouraging users to switch from ChatGPT to Claude in protest of AI being used for military operations.
However, fresh reports indicate that Claude itself was recently used in U.S. military operations.
According to accounts published by major U.S. outlets, American military command used Claude during the joint U.S.–Israel bombardment of Iran that began on Saturday, February 29, 2026.
The system was reportedly used for:
- Intelligence analysis
- Target selection support
- Battlefield simulations
The reports surfaced hours after U.S. President Donald Trump ordered federal agencies to cease using Anthropic’s technology.
In a post on Truth Social, Trump described Anthropic as a “Radical Left AI company” and directed agencies to sever ties immediately.
The dispute follows an earlier incident in January 2026, when Claude was reportedly used during a U.S. raid targeting Venezuelan leader Nicolás Maduro.
Anthropic objected at the time, stating that its terms prohibit the use of Claude “for violent ends, weapons development, or surveillance.
U.S. Defense Secretary Pete Hegseth criticized Anthropic in a post on X, accusing the company of “arrogance and betrayal.”
He demanded “full and unrestricted access” to AI systems for lawful military purposes but acknowledged that Claude would continue supporting certain systems for up to six months to allow a transition to alternative providers.
OpenAI has since stepped into the gap, confirming a formal agreement to supply AI tools to the Pentagon’s classified networks.
As of Monday evening, March 2, 2026, Anthropic said it was still investigating the cause. No estimated time for full restoration has been announced.
