Three AI Labs Went Down Thursday. Nobody's Explaining Why — SkimNews
.gif)
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI attributed its outage to a "routing error" starting at 7:43 am PT on Thursday, September 3, which made ChatGPT and Codex unavailable across platforms; a fix was implemented by 8:17 am PT, per spokesperson Kathleen Chaykowski.
- Anthropic began alerting about a "partial outage" at 6:23 am PT that hit Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5, marking the issue resolved by 9:16 am PT, with Claude Sonnet 5 briefly experiencing similar errors shortly after 9 am PT.
- SpaceX, xAI's parent company, pinned Grok's outage on "an outage at our Memphis compute center this morning" and also publicly apologized to "impacted compute partners" — a phrase that gains weight given that Anthropic and xAI announced a compute partnership with SpaceX in May.
- xAI reported Grok outages across all platforms and services beginning at 6:30 am PT, marking the episode resolved at 10:05 am PT.
- Neither OpenAI nor Anthropic cited a shared external cause for the coinciding timing, and major internet infrastructure providers — Cloudflare, Amazon Web Services, and Microsoft Azure — did not report any outages on Thursday.
- There were scattered reports of a possible Google Gemini outage Thursday morning, but Google did not confirm any incident on its status dashboard or respond to WIRED's request for comment.
Why it matters: Three of the top U.S. frontier AI providers went dark within roughly a four-hour window without identifying any shared cause, and SpaceX's apology to "compute partners" intersects directly with Anthropic and xAI's public compute partnership announced in May. For enterprise customers running production workloads on these APIs, the absence of a unified explanation makes it impossible to tell whether the failures were coincidental or evidence of shared infrastructure concentration risk.
Ask SkimNews
