On Thursday morning, the odds were good asking an AI chatbot anything delivered a three-lemon jackpot of bupkis. ChatGPT, Claude, and Grok – three giants of frontier AI – each went down within the same window, and each confirmed the outage on its own status page.
For its part, OpenAI’s status page listed 15 affected ChatGPT components (login, search, image generation, voice mode, Deep Research, and more) plus four Codex components, with the outage starting around 10:58 AM ET.
Over on Claude’s status page, Anthropic reported elevated errors on a specific list of models including Mythos/Fable 5.1, Mythos/Fable 5, Opus 5, Opus 4.8, and Opus 4.6. That spanned Claude.ai, the API, Claude Code, and Claude Cowork.
At xAI’s own status page, it listed an active Grok outage, hitting the web app and API in two regions. Even Cursor, the AI coding tool, got caught in the crossfire, citing the Grok and Claude issues. Google’s Gemini, notably, never officially copped to anything despite a bump in user reports.
By late morning, SPB noted OpenAI had applied a fix and services were “recovering”. Claude bounced back on most models except Opus 4.8 and Opus 5; Grok was still down as of the last check.
No confirmed root causes. One working theory floated by MacRumors points to a Cloudflare disruption that morning, since a wide range of sites and services route traffic through Cloudflare’s infrastructure. None of the three AI companies has confirmed that.
The timing was too good for the internet to resist.
OpenAI’s own account tweeted “The stars are almost aligned” mid-outage, which read either as a coincidence or a tease. That’s because the outage landed right as OpenAI’s next model, code-named Astra, is rumored to be close to launch.
OpenAI hasn’t said the two are connected, and it’s probably not. ChatGPT goes down often enough on its own.
Journalist Paris Marx summed up the temporary return to pre-chatbot life on Bluesky: “and for a brief moment, millions of people had to use their brains again.”
AI Rival Z.ai was less philosophical. While ChatGPT, Claude and Grok stumbled, its official X account posted just four words: “We’re still up.”
On Reddit, one commenter wondered whether the machines had simply made a break for it: “Maybe AGI left the sandbox.” – a reference to the idea of human-level, general-purpose AI escaping its digital confines.
Or maybe, it was just a rough Thursday for the robots.
