SpaceX and Nvidia are collaborating on a space-optimized AI computing system designed for launch into orbit.
Elon Musk confirmed the project on Monday. He said SpaceX, working with Nvidia, has designed a space-optimized version of Nvidia’s Vera Rubin NVL72 system. The first launch is planned for the fourth quarter of 2027, with significant scale expected in 2028.
Musk described the design as “significantly simpler, lower cost, denser and lighter than a traditional rack.” The system is intended for SpaceX’s planned Starmind AI satellites, which would act as orbital data centers.
Nvidia said SpaceXAI will use its Vera CPUs to speed up orchestration, code execution, and data processing for next-generation agentic AI systems. The first-generation Starmind AI satellite will be based on the optimized Vera Rubin NVL72 rack-scale system.
The partnership aims to take the same accelerated computing architecture used in large AI data centers on Earth and adapt it for the conditions of space. Challenges such as power, cooling, and reliability differ greatly from ground-based facilities.
SpaceX has previously said it plans to expand its AI computing capacity sharply in the coming years. This orbital effort is part of that broader push.
Reference links
- https://nvidianews.nvidia.com/news/spacexai-adopts-nvidia-vera-cpu-to-accelerate-agentic-ai-at-massive-scale
- https://m.economictimes.com/tech/artificial-intelligence/spacex-nvidia-develop-space-optimized-ai-system-for-orbital-launch-next-year/articleshow/133488899.cms
- https://www.benzinga.com/markets/tech/26/08/61400018/elon-musk-says-spacex-targeting-q4-2027-launch-for-nvidia-powered-ai-system-touts-lower-cost-denser-and
