At GTC 2026 on Monday, NVIDIA CEO Jensen Huang announced that accelerated computing is leaving Earth. The company unveiled the Space-1 Vera Rubin Module, a computing system designed for satellites and orbital data centers that delivers up to 25x more AI compute than the NVIDIA H100 GPU for space-based inferencing.

Intelligence in Orbit

The Space-1 module is engineered for the size-, weight-, and power-constrained environments found on spacecraft. It combines the Rubin GPU with NVIDIA's IGX Thor and Jetson Orin platforms to support three main use cases: orbital data centers (ODCs), geospatial intelligence processing, and autonomous space operations.

"Space computing, the final frontier, has arrived," Huang said. "As we deploy satellite constellations and explore deeper into space, intelligence must live wherever data is generated."

The module is not yet available but is expected to reach customers in the near future.

Partners and Applications

Six companies are already working with NVIDIA's space computing platforms: Aetherflux, Axiom Space, Kepler Communications, Planet Labs, Sophia Space, and Starcloud. Planet Labs announced a new collaboration targeting the ability to process Earth-imagery in seconds rather than hours — directly from orbit.

The Cooling Problem

Huang acknowledged the engineering challenges still ahead. Unlike ground-based data centers, space offers no convective cooling — heat can only dissipate through radiation.

"We have to figure out how to cool these systems out in space," he said. "But we've got lots of great engineers working on it."

NVIDIA's ground-side hardware also got a space-focused upgrade: the RTX PRO 6000 Blackwell Server Edition delivers up to 100x faster performance than legacy CPU systems for large-scale geospatial imagery analysis.