Three major AI chatbots suffer simultaneous outage ahead of OpenAI's GPT-6 Astra launch

The outage began around 10:58 AM ET on Thursday, September 3, with users reporting errors across nearly every major ChatGPT feature, including conversations, login, search, file uploads, Voice mode, image generation, and the Agent and Deep Research tools. OpenAI's Codex services were also affected. The company acknowledged the problem on its status page, stating it was investigating elevated errors across 15 components.

OpenAI later told The Register that a routing error starting at 7:43 AM PT had caused the disruption, with a solution implemented by 8:17 AM PT.

Anthropic's Claude also went down, experiencing a three-hour, six-minute outage due to what Anthropic described as an infrastructure issue. A spokesperson confirmed service was fully restored by 16:16 UTC, affecting Claude.ai, Claude Code, Claude Cowork, and the Claude API.

SpaceX's Grok status page reported issues beginning at 6:30 AM PT, with the company stating it was working to restore service. As of late Thursday, Grok had not yet fully recovered.

The coincidence of three separate outages raised the possibility of a common provider being at fault. However, Cloudflare — used by all three companies — firmly denied any issues, stating its services were operating normally. AWS, Google Cloud, and Microsoft Azure also showed no relevant problems on their status pages.

The outages come just as OpenAI is reportedly preparing to launch GPT-6 Astra, a major upgrade for ChatGPT and Codex. The company has not linked Thursday's disruption to the launch preparations, and OpenAI has said such outages are not uncommon.

Why This Matters

The simultaneous failure of three leading AI platforms highlights the growing dependency on a small number of AI providers for both consumer and enterprise users.
With OpenAI's GPT-6 Astra launch imminent, any lingering reliability concerns could affect user confidence ahead of a major product release.
The incident underscores the fragility of AI infrastructure — even brief outages can disrupt workflows for millions who rely on these tools for research, coding, and daily tasks.

Background

Major AI outages are relatively rare but not unprecedented. OpenAI's ChatGPT has experienced several notable disruptions in the past, often attributed to infrastructure scaling issues. The simultaneous failure of multiple providers is unusual, however, and typically points either to a shared external dependency (such as a cloud provider or CDN) or a coincidental clustering of separate incidents. Cloudflare, used by all three companies for security and content delivery, was quickly ruled out as a common cause. The outage also comes at a pivotal moment for OpenAI, which is preparing to release GPT-6 Astra — the successor to its GPT-5 family launched over a year ago.

Key Perspectives

[OpenAI]: The company has acknowledged the routing error and resolved it quickly, emphasising that service is recovering. They have not linked the outage to Astra preparations. [Anthropic]: Claude experienced its own infrastructure issue but has fully restored service. The company has not commented on broader implications. [SpaceX/Grok]: Still investigating the outage, with no estimated time for full restoration as of late Thursday. [Critics/Skeptics]: The coincidence of three simultaneous failures remains suspicious, even without a confirmed shared cause. Some users may question the resilience of AI infrastructure and the readiness of new models.

What to Watch

OpenAI's official announcement of GPT-6 Astra — expected soon, possibly within days.
Any follow-up postmortems from OpenAI, Anthropic, or SpaceX explaining root causes.
User sentiment and adoption rates for competing or self-hosted AI alternatives following the outage.
Cloudflare's ongoing status — if it turns out to be implicated, it would represent a systemic risk for the AI ecosystem.

ZOTPAPER