Musk confirms SpaceX-Nvidia alliance to build AI data centers in space

Elon Musk officially confirmed that SpaceX and Nvidia are partnering to build ground and orbital AI data infrastructure, centered around SpaceX's ambitious Starmind initiative.

The tech billionaire said SpaceX will exclusively use Nvidia technology to power its next generation of AI infrastructure, describing Nvidia's platforms as the industry's best.

The announcement was made during SpaceX’s inaugural earnings call, highlighting a massive strategic push to bypass Earth's mounting energy and grid limitations by taking artificial intelligence computing into space.

Musk stated that SpaceX has committed to using Nvidia systems exclusively for its AI infrastructure, calling Nvidia's technology the best on the market.

Specifically, SpaceX will deploy Nvidia's cutting-edge Vera Rubin architecture including the Vera Rubin NVL72 rack-scale system (codenamed Kyber) across both terrestrial data centers and space-based hardware.

The orbital component relies on specialized satellites designed to host localized computing workloads directly in low Earth orbit.

These satellites utilize sun-synchronous orbits to harvest continuous, uninterrupted solar energy while leveraging the vacuum of space for effortless radiative cooling, bypassing the massive water and chiller requirements of ground data centers.

Processed AI workloads and data will be beamed back to Earth seamlessly via high-speed optical laser links interconnected with the Starlink satellite network. 

SpaceX indicated that the first launches of these AI-capable orbital data centers are slated to begin next year.

The announcement comes as SpaceX announced plans to rapidly scale its total computing capacity from over 2 gigawatts up to nearly 10 gigawatts by the end of 2027.