Nobody Is Saying Why OpenAI and Anthropic Had Outages Today

image via WIRED

September 3, 2026, 9:56 PM

• •OpenAI said a routing error disrupted ChatGPT and Codex for some users across platforms.
• •Anthropic reported elevated errors affecting Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 before resolving the issue.
• •xAI said Grok's outage resulted from an outage at SpaceX's Memphis compute center.
• •The simultaneous failures initially suggested a shared third-party infrastructure problem.
• •Cloudflare, Amazon Web Services, and Microsoft Azure did not report outages during the incident.

OpenAI, Anthropic, and xAI all experienced service disruptions affecting ChatGPT, Claude, and Grok on Thursday morning. OpenAI said a routing error made ChatGPT and Codex unavailable for some users from about 7:43 am PT until a fix was implemented around 8:17 am PT. Anthropic described a partial outage affecting several Claude models and said it identified the cause and deployed a fix, while xAI blamed an outage at SpaceX’s Memphis compute center. WIRED reports that the overlapping timing suggested a possible shared infrastructure problem, but OpenAI and Anthropic did not identify a common external cause and major infrastructure providers reported no outages.

Read original article