周四上午，顶尖人工智慧公司 Anthropic、OpenAI 与 xAI 旗下的大型前沿模型罕见地同时发生服务中断，导致其对应的 AI 聊天机器人出现停摆。xAI 的母公司 SpaceX 指出，Grok 的故障是由于当天上午其孟菲斯计算中心发生的停机事故所引起，并向受影响的合作伙伴致歉，而 Anthropic 与 xAI 此前曾在五月宣布与 SpaceX 建立计算合作关系。

针对此次服务中断，OpenAI 发言人表示问题起因于太平洋时间上午 7 点 43 分左右的路由错误，导致部分用户无法存取 ChatGPT 与 Codex，并在约 8 点 17 分修复完成；Anthropic 则自上午 6 点 23 分通报部分服务中断，影响了多个 Claude 模型，随后部署修复措施并于 9 点 16 分排除异常。与此同时，xAI 的 Grok 在经历数小时故障后，于上午 10 点 05 分宣告服务全面恢复，此外也有零星关于 Google Gemini 当机的传闻，但未获官方证实。

虽然同一行业内的多家顶尖公司同时发生停机通常暗示底层云端服务商或内容传递网路等第三方供应商出现问题，但 OpenAI 与 Anthropic 均未提及外部共同肇因，且主要网路基础设施业者如 AWS、Microsoft Azure 和 Cloudflare 当天皆未记录任何重大异常，使得这场近乎同步的连环停机事件显得格外不寻常。

On Thursday morning, frontier models from major AI companies Anthropic, OpenAI, and xAI experienced rare simultaneous outages that resulted in downtime for their respective chatbots. SpaceX, the parent company of xAI, attributed Grok's failure to an outage at its Memphis compute center that morning and issued an apology to affected compute partners, referencing the compute partnership Anthropic and xAI formed with SpaceX in May.

OpenAI explained that its downtime was caused by a routing error around 7:43 am PT affecting ChatGPT and Codex users, which was resolved by 8:17 am PT; Anthropic alerted users to partial outages affecting several Claude models starting at 6:23 am PT before fully deploying a fix by 9:16 am PT. Grok was also reported down across all platforms beginning at 6:30 am PT until service was restored at 10:05 am PT, alongside unconfirmed reports of disruptions to Google Gemini. (Key numbers: 7, 43, 8, 17)

Although multiple simultaneous outages across competing companies in the same sector typically point to an issue with a shared cloud provider, CDN, or third-party infrastructure vendor, neither OpenAI nor Anthropic cited a shared external source, and major infrastructure platforms including AWS, Microsoft Azure, and Cloudflare reported no outages on Thursday.