SpaceX Partners with Nvidia to Bring Rubin GPUs and Vera CPUs to Space
SpaceX is partnering with Nvidia to design the compute payload for its upcoming Starmind AI1 satellites. Each satellite in this constellation will be equipped with Nvidia's next-generation Rubin GPUs and Vera CPUs to enable datacenter-class computing in orbit. This partnership represents a major leap forward for space-based edge computing, moving beyond communication satellites to establish orbital AI data centers. By leveraging Nvidia's advanced hardware, SpaceX aims to run massive AI workloads directly in space, powered by solar energy and cooled by the vacuum of orbit. The Starmind constellation is envisioned to scale up to one million satellites acting as orbiting compute nodes rather than traditional communication relays. The hardware payload will utilize Nvidia's Vera CPUs, which feature Olympus cores designed for agentic AI workloads, alongside Rubin GPUs built for massive-context processing.
## BACKGROUND
SpaceX's Starmind project is an ambitious initiative to build an orbital AI data center network, distinct from its Starlink internet constellation. Nvidia's Vera Rubin platform is a highly integrated architecture where GPUs, CPUs, and cooling systems are co-designed to maximize efficiency. The Vera CPU is optimized for agentic AI workloads, while the Rubin GPU microarchitecture is built to handle massive-context processing like generative video and long-context reasoning.