SpaceX 將使用 Nvidia CPU 芯片進行人工智能開發
我是 LongbridgeAI，我可以總結文章信息。SpaceX 的人工智能部門將部署 Nvidia Vera CPU 芯片，以支持 xAI 的自主應用，包括 Grok 聊天機器人。此次合作推出了首款專為 AI 代理設計的 CPU，旨在加快反應時間並擴展基礎設施。SpaceX 計劃通過其 Starmind 衞星將這一架構擴展到太空，利用 Nvidia 的 Vera Rubin NVL72 系統，支持其更廣泛的擴展軌道計算能力的雄心
By Katherine Hamilton
SpaceX's artificial-intelligence division will deploy Nvidia Vera CPU chips to develop its agentic applications.
The partnership brings the first CPU built for AI agents to xAI's deployments, which includes the Grok chatbot among other projects, Nvidia said Monday.
SpaceXAI will use the chips to manage application workloads and speed up agents' reaction times. It plans to expand its AI infrastructure behind Grok on Nvidia Vera Rubin, while extending an optimized Vera Rubin NVL72 into space with its Starmind satellite.
Vera Rubin provides a common architecture for SpaceXAI to scale its AI factories, which SpaceXAI plans to take beyond data centers on Earth and into orbital computing.
The company's Starmind AI satellite will be based on Nvidia's Vera Rubin NVL72 rack-scale system, extending the same computing architecture that powers AI factories on Earth.
SpaceX has been ramping its computing capacity as its AI development ambitions increase. Earlier this month, it announced in partnership with Tesla that it planned to build a semiconductor fabrication plant in Texas, with the goal of producing 1 terawatt of compute per year.
Write to Katherine Hamilton at katherine.hamilton@wsj.com
(END) Dow Jones Newswires
August 24, 2026 11:39 ET (15:39 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.