OpenAI, Anthropic, and xAI Hit Rare Simultaneous AI Outages, With Causes Only Partly Explained

ChatGPT/Codex, three Claude models, and Grok suffered overlapping outages on Thursday, Sept 3, 2026. OpenAI blamed a routing error and xAI blamed a SpaceX Memphis compute center outage; Anthropic resolved its Claude outage but declined to explain the cause,…

On Thursday, Sept 3, 2026, four major AI providers experienced rare overlapping downtime. Anthropic reported a partial outage with elevated errors on Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 starting at 9:23 am ET, identified the cause internally within about 15 minutes, and resolved it by 12:16 pm ET (9:16 am PT), with a brief Claude Sonnet 5 error spike after noon; per WIRED, Anthropic declined to publicly explain the cause. OpenAI reported elevated errors across ChatGPT and Codex from 10:43 am ET (7:43 am PT) and attributed the problem to a routing error; the two reports disagree on when service was fully restored, with WIRED describing a roughly 34-minute downtime window (7:43-8:17 am PT) and Ars Technica reporting resolution at 12:55 pm ET after a mitigation. xAI's Grok showed a user-facing error message starting at 6:30 am PT (9:30 am ET), which xAI attributed to an outage at SpaceX's Memphis compute center; DownDetector reports for Grok surged from fewer than 10 to 1,365 by 9:45 am ET before declining. Ars Technica also reported that Google experienced interruptions during the same window, with no cause given. Despite the overlap, no shared third-party infrastructure cause was confirmed: Cloudflare, AWS, and Azure all reported no simultaneous issues.

• Anthropic: partial outage with elevated errors on Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 from 9:23 am ET; cause identified internally within about 15 minutes; resolved by 12:16 pm ET (9:16 am PT); brief Claude Sonnet 5…
• OpenAI: ChatGPT and Codex showed elevated errors from 10:43 am ET (7:43 am PT); OpenAI attributed the downtime to a routing error. Sources disagree on restoration: WIRED reports a ~34-minute unavailability window (7:43-8:17 am PT), while…
• xAI: Grok outage began at 6:30 am PT (9:30 am ET) with a user-facing error message; xAI/SpaceX attributed it to an outage at SpaceX's Memphis compute center; DownDetector reports surged from fewer than 10 to 1,365 by 9:45 am ET before…
• Google: AI services experienced interruptions during the same window, per Ars Technica; no cause was given, and WIRED's report did not mention Google.
• No shared third-party infrastructure cause was confirmed; Cloudflare, AWS, and Azure reported no simultaneous outages.