Several of the world’s most widely used artificial intelligence services experienced disruptions on Thursday, September 3, with users reporting problems accessing ChatGPT, Claude and Grok around the same time.
The simultaneous issues affected services operated by OpenAI, Anthropic and xAI, creating an unusual situation in which three major AI platforms were experiencing problems at once. Reports of failures also increased on outage-tracking service Downdetector.
Companies acknowledge service problems
OpenAI’s status page confirmed elevated errors affecting ChatGPT and Codex, with widespread problems becoming apparent at around 11 a.m. ET.
The disruption went beyond ordinary chatbot conversations. Users also encountered problems logging in and using features including file uploads, voice mode, search, deep research and image generation.
OpenAI later said it had implemented measures to address the problem and was monitoring the recovery. Some of its services continued to experience degraded performance as systems gradually returned to normal.
Anthropic experienced problems across Claude, Claude Code and its API at roughly the same time. A member of Anthropic’s technical staff attributed the partial outage to an infrastructure issue. According to the company’s status updates, the disruption was resolved at around 12:15 p.m. ET.
Grok’s problems began earlier, at approximately 9:30 a.m. ET, affecting xAI’s service on the web as well as its Android and iOS apps. Some users attempting to access Grok through X were told that the model was overloaded and to try again later or select another model.
xAI said it was working to restore the service as quickly as possible.
At the time of the initial reports, there was no confirmed explanation showing that all three outages had been caused by the same technical failure.
The problems appeared to extend beyond the three consumer-facing chatbots. Cursor, an AI-powered coding platform, also confirmed an outage that it linked to problems involving Claude and Grok.
oh my god.
> ChatGPT is down.
> Claude is down.
> Grok is down.what's going on? pic.twitter.com/5NKlmUe9Ra
— sui (@birdabo) September 3, 2026
Gemini appears to avoid major disruption
Google’s Gemini service appeared to remain operational during the broader disruption. Although Downdetector recorded an increase in reports involving Gemini, Google had not confirmed an outage at the time.
One possible factor raised in reporting was Microsoft Azure, which was also seeing increased outage reports. ChatGPT, Claude and Grok use Azure infrastructure alongside other cloud providers, although no definitive connection between Azure and the simultaneous AI service problems had been established.
The overlap between the outages raised questions about whether the disruptions might share an underlying cause, but there was no evidence at the time confirming that the incidents were connected.
The unusual episode also highlights the potential for infrastructure problems to affect multiple parts of the rapidly expanding AI ecosystem. OpenAI, Anthropic and xAI each acknowledged problems with their services, but the information released by the companies pointed to separate investigations rather than a confirmed single cause behind the disruptions.
