ChatGPT, Claude and Grok go down simultaneously; Microsoft Azure comes under scrutiny

Several of the world’s leading artificial intelligence services experienced disruptions almost simultaneously on Thursday, leaving millions of users around the world unable to access or properly use the platforms.

OpenAI’s ChatGPT, Anthropic’s Claude and Elon Musk’s xAI’s Grok were all affected at roughly the same time, triggering widespread complaints from users. According to outage-monitoring platforms, more than 14,000 reports were recorded throughout the day, with complaints related to OpenAI’s services exceeding 12,000 at their peak.

The problems began to become more noticeable after the evening in Bangladesh. Users in India, the United Arab Emirates, the United Kingdom, Belgium, Turkey, Jordan, Greece, Singapore, Vietnam and Switzerland also reported similar difficulties, including being unable to send requests or receive responses.

What happened

OpenAI said error rates for ChatGPT and Codex had risen abnormally, affecting conversations, logins, file uploads, voice mode, search, deep research and image generation.

Anthropic’s status page reported problems with the Claude Mythos 5.1, Fable 5.1 and Opus 5 models at around 7:26 p.m. Bangladesh time (13:26 UTC). Mythos 5, Fable 5, Opus 4.8 and Opus 4.6 were subsequently added to the list.

The disruption affected Claude AI, Claude API, Claude Code and Claude Cowork.

xAI confirmed that the web version of Grok, as well as its Android and iOS applications, had been affected. Users attempting to send prompts through X were reportedly receiving messages indicating that the model was under unusually heavy load.

The disruption also affected services that rely on these AI platforms. AI coding tool Cursor said its services had also been disrupted because of the outages affecting Claude and Grok, interrupting developers’ ongoing work.

How the services recovered

Services began returning to normal at around 9:45 p.m. Bangladesh time. OpenAI said it had implemented a fix and that its services were recovering.

Anthropic said the impact of the incident had ended at 10:16 p.m. Bangladesh time (16:16 UTC), with all of its models back online.

Grok was among the last of the three services to fully recover. Even after OpenAI and Claude had returned to normal, xAI’s status page continued to display an outage notice. The incident was later marked as resolved.

Questions over the cause

None of the three companies has officially disclosed the root cause of the simultaneous disruptions. A technical official at Anthropic referred to the incident on social media as an “infrastructure issue”.

Attention has increasingly turned to Microsoft Azure. Azure provides cloud infrastructure for services involving all three AI platforms, while reports of disruptions affecting Azure also increased during the same period.

An infrastructure-monitoring platform recorded reports of ingress failures in Azure’s East US region at around 11:26 p.m. Bangladesh time.

By contrast, Google’s Gemini appeared to remain largely unaffected, despite operating on Google Cloud. Although user reports concerning Gemini also increased, Google did not officially acknowledge a service disruption.

Technology researchers, however, have cautioned against drawing a direct conclusion from the simultaneous outages. The platforms rely on complex infrastructure involving multiple cloud providers and service components. Therefore, disruptions occurring at roughly the same time do not necessarily establish the existence of a single common point of failure.