Musk: SpaceX and xAI to exclusively use Nvidia GPUs, plans to launch Vera Rubin NVL72 into orbit next year

Published Aug 5, 11:50 AM UTC

Elon Musk announced that SpaceX and xAI will exclusively use Nvidia GPUs, specifically praising the Vera Rubin NVL72 rack-scale system as the best AI compute architecture available. During SpaceX's earnings call, Musk stated the company has decided to build exclusively on Nvidia due to the superior design of the NVL72 VR200 computer, which eliminates cables for improved reliability. Historically, xAI has used Nvidia's Hopper and Blackwell hardware for training Grok models. Musk also revealed plans to deploy an optimized version of the Vera Rubin NVL72 computer in space as part of the Starmind AI satellite, with launches expected to begin next year. This marks a significant endorsement for Nvidia over competitors like AMD, and represents an ambitious project to deploy a full-scale AI rack system in orbit, raising questions about cooling and reliability in space.