SpaceX commits to Nvidia as its exclusive GPU supplier

SpaceX plans to rely exclusively on Nvidia GPUs, with Elon Musk saying the company has decided to build its compute capacity around Nvidia’s Vera Rubin architecture. Musk called it the best AI computer and said SpaceX values its close cooperation with Nvidia across multiple levels.

The move deepens Nvidia’s role as a core technology partner for Musk’s companies, including Tesla and SpaceX. Tesla has developed custom AI chips but still uses Nvidia GPUs for AI and data center workloads. Musk later reiterated the decision on X, writing that SpaceX has committed to using Nvidia GPUs exclusively because they are the best.

An exclusive Nvidia arrangement sidelines rival chipmakers including Intel, AMD, and Broadcom, and runs against the diversification strategy many large tech companies use to limit supplier risk. Musk also said SpaceX will receive a significant percentage of Nvidia’s GPUs next year, suggesting the company could become a notable contributor to Nvidia revenue.

SpaceX reported that second-quarter sales rose 92% year on year to $7.8 billion, driven by AI infrastructure and Starlink growth. Musk said the company aims to reach a $100 billion annual revenue run rate by year-end, while investors focused on AI spending and a $541 million loss, sending shares down more than 7% after hours.