NVIDIA and SpaceX Collaborate to Develop Satellite Computing Payload; Starmind AI1 Project Announced
SpaceX announced that it is collaborating with NVIDIA to jointly develop the “Starmind AI1” satellite computing payload.
Technical configuration:
Each Starmind satellite will be equipped with NVIDIA’s Rubin GPU and Vera CPU
The goal is to achieve space computing capabilities at the level of a data center
Previously, NVIDIA CEO Jensen Huang made remarks in public about the landscape of the AI competition, basing his judgment primarily on the scale of computing power at the infrastructure level rather than the models themselves.
Three directions mentioned:
xAI, corresponding to foundational cognitive intelligence
Tesla, corresponding to autonomous driving and real-time data acquisition
Optimus, corresponding to humanoid robots
Tesla’s AI factories are equipped with large quantities of NVIDIA hardware, and its fleet is seen as the largest real-world data collection system currently in scale.
Notably, space computing is not a new concept; it has previously mostly been limited to research-level verification. This time marks the first publicly disclosed collaboration case in which a data-center-class GPU enters a commercial satellite payload. The key engineering challenges center on cooling and power consumption.
$SPCXB
#英伟达 #SpaceX #Aİ