ChatGPT, Claude and Other Major AI Services All Went Down on the Same Thursday Morning
Four of the biggest AI platforms suffered outages within hours of each other. Here is what happened, what caused it, and what users should know.
Key points
- Anthropic reported a partial outage affecting multiple Claude models starting at 9:23 a.m. Eastern on Thursday.
- OpenAI flagged degraded performance across ChatGPT and Codex, its coding assistant tool, from 10:43 a.m. Eastern.
- All reported incidents were marked resolved by early Thursday afternoon, the longest lasting roughly three hours.
- The outages at OpenAI, Anthropic, xAI (Elon Musk's AI company), and Google overlapped in time but appear to have been separate events.
Four of the biggest names in artificial intelligence went down on the same Thursday morning, leaving users staring at error messages or spinning wheels for hours. Ars Technica first reported the overlapping nature of the incidents.
Anthropics's problems started first. The company posted to its public status page at 9:23 a.m. Eastern, saying it had spotted "elevated errors" on requests to three of its Claude models: Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5. Those are cloud-based AI assistants, meaning the software runs on Anthropic's servers and users access it over the internet rather than installing anything locally.
Anthroptic said it had found the cause about fifteen minutes later. A fix followed, and the company declared the incident resolved at 12:16 p.m. A second, shorter disruption hit Claude Sonnet 5 just after noon.
What happened at OpenAI?
OpenAI reported its own separate problem at 10:43 a.m. Eastern, affecting both ChatGPT, its flagship AI chat service, and Codex, a tool that helps software developers write and check code. The company described "degraded performance" across both products. A mitigation measure (a temporary fix designed to reduce the impact while a full repair is prepared) went in place roughly thirty minutes later. OpenAI marked the issue resolved by 12:55 p.m.
Should users be worried?
No serious data loss or security risk has been reported. These were availability problems, meaning the services were slow or unreachable, not compromised.
For most people, that means a frustrating hour or two. If you use Claude or ChatGPT for work, the practical takeaway is straightforward: keep a backup plan for critical tasks. A second AI service, a saved draft, or simply a scheduled break can cover short gaps like these.
The fact that xAI and Google also experienced disruptions in the same window is unusual. Cloud outages happen regularly across the internet, but four competing AI platforms going down on the same morning is rare. Neither xAI nor Google provided detailed public timelines comparable to the Anthropic and OpenAI status updates.
All four companies run their AI services on large clusters of specialised computer hardware, and many share underlying infrastructure providers. Whether Thursday's timing was coincidence or shared cause is not yet clear.
Common questions
Will this happen again?
Cloud service outages are a normal, if unwelcome, part of using any internet-based software. Major providers typically publish "uptime" records, and short disruptions like Thursday's are not unusual in isolation, though four at once is uncommon.
How can I check if an AI service is down?
Most major AI companies publish live status pages. Anthropic's is at status.anthropic.com and OpenAI's is at status.openai.com. Bookmarking both takes under a minute and saves guesswork next time.
