SpaceXAI issued an apology on Thursday (3) to Grok users and other computing partners who were impacted by a failure at its data center located in Memphis, United States.
This incident resulted in the unavailability of Grok for over three hours and apparently affected other artificial intelligence (AI) corporations that rely on the company's processing infrastructure.
In a post, SpaceXAI regretted the inconvenience faced by Grok users due to the shutdown at its morning computing center, extending the apology to its affected computing partners. The company confirmed that all systems have been restored and are operating normally.
The timing of this interruption coincided with reports of problems across several other AI platforms during Thursday morning. Claude, developed by Anthropic, also showed instability starting at the same time, presenting an increase in errors across multiple models, as recorded on the company's status page. Anthropic declared the incident resolved at 1:16 PM, specifically affecting the Claude Mythos 5.1, Claude Fable 5.1, and Claude Opus 5 models.
Although Anthropic did not specify whether there was a connection to the SpaceXAI data center, the two companies signed an agreement earlier this year for Anthropic to lease computing capacity from the Musk-led company.
Additionally, OpenAI also registered an outage on Thursday morning, causing an increase in errors across ChatGPT and Codex, according to the company itself. Users began reporting difficulties with the services around 11:30 AM, according to Downdetector records. OpenAI communicated that the issue was resolved at 1:55 PM but did not provide details about the cause.