Nvidia CEO Jensen Huang revealed the Space-1 Vera Rubin Module during his GTC 2026 keynote, announcing plans to put data-center-class AI computing into orbit.
The module delivers up to 25x more AI compute than the H100 GPU for space-based inferencing, with a tightly integrated CPU-GPU architecture and high-bandwidth interconnect designed for size-, weight-, and power-constrained environments like satellites and orbital data centers.
The Space-1 module enables large language models and advanced foundation models to operate directly in space, processing massive data streams from space-based instruments in real time — eliminating the latency of sending raw data back to Earth for processing.
Partners already building on Nvidia's space computing platforms include Aetherflux, Axiom Space, Kepler Communications, Planet Labs, Sophia Space, and Starcloud.
Engineers are still solving significant challenges, particularly heat dissipation in the vacuum of space where traditional cooling methods don't work. The IGX Thor and Jetson Orin platforms for edge space computing are currently available, while the Vera Rubin Space Module will be available at a later date.
The announcement represents perhaps the most ambitious frontier for AI infrastructure — extending GPU computing beyond Earth's surface into orbital operations.