SpaceXAI data center failure knocks Grok offline for hours
A Memphis compute center operated by SpaceXAI caused a multi-hour outage that disrupted Grok and other AI services on Thursday. The company apologized to users and partners, noting all systems were restored. Anthropic and OpenAI also reported similar issues around the same time, though no direct link was confirmed.
The outage began at 6:30 AM PT and lasted roughly three and a half hours, affecting Grok across X and mobile apps. SpaceXAI’s status page labeled it a “models outage,” and the company confirmed all systems were restored by early afternoon. Elon Musk stated corrective measures were underway, though the root cause in Memphis was not disclosed.
Anthropic’s Claude and OpenAI’s ChatGPT also reported elevated errors starting around the same time, with both resolving by mid-morning. SpaceXAI had previously signed a compute-leasing deal with Anthropic, but no direct link between the incidents has been confirmed. Neither company has commented on a possible connection.
This incident highlights how concentrated AI infrastructure has become, with a single data center able to disrupt multiple major services simultaneously. Businesses and individuals relying on Grok, Claude, or ChatGPT for daily tasks could face productivity losses and workflow interruptions during such failures. It may also prompt regulators and customers to demand greater transparency about system dependencies and redundancy plans, potentially reshaping how AI companies disclose operational risks.