NVIDIA Space Computing was introduced at GTC 2026. Recently, NVIDIA officially released more information, aiming to move its accelerated computing platform from ground data centers to space orbits. The project focuses on AI infrastructure required for next-generation space missions, enabling satellites, orbital platforms, and ground stations to use NVIDIA GPUs and edge computing modules to speed up the processing of images, sensor data, and geospatial intelligence.
(NVIDIA GTC 2026｜NVIDIA sends Space-1 Vera Rubin into space to build a real “cloud computing” platform)
NVIDIA said that with the development of the commercial space industry, future missions will no longer be limited to sending data back to Earth, but will require real-time processing, analysis, and decision-making in orbit. This includes response to natural disasters, environmental monitoring, climate and weather forecasting, infrastructure management, and automated space operations.
From Earth to space: NVIDIA aims to solve satellite data latency and downlink costs
Traditional satellite missions often need to send vast amounts of raw data back to the ground, where ground data centers perform analysis. But in applications such as Earth observation, infrared imagery, SAR radar, and radio-frequency signal detection, data volumes can reach hundreds of TB. If the system relies entirely on downlink transmission, it is not only costly, but also extends response times.
NVIDIA’s Space Computing goal is to place part of the AI inference and data-fusion capability directly on the space side. Using Jetson Orin, IGX Thor, and the newly launched Space-1 Vera Rubin module, orbital platforms can process sensor data in real time in space, generate geospatial intelligence, and reduce reliance on ground transmission.
In other words, in the future, satellites will not just take pictures and transmit data—they will be able to “understand” the data directly in orbit.
Space-1 Vera Rubin: bring data-center-class AI compute into space
The most closely watched product this time is NVIDIA’s Space-1 Vera Rubin module. NVIDIA says the module can provide up to 25x AI compute capability per GPU for space inference and orbital data centers.
Space-1 Vera Rubin uses an integrated CPU-GPU architecture and high-bandwidth interconnects, targeting on-orbit processing of large data streams so that frontier models and foundation models can run on orbital platforms. This means some AI models will not need to rely entirely on ground data centers, and can complete real-time analysis directly within satellites or orbital data centers.
This also aligns with the direction NVIDIA CEO Jensen Huang presented at GTC: AI won’t stay only in the cloud and ground data centers—it will move into robots, factories, vehicles, and even space infrastructure.
Jetson Orin, IGX Thor: give satellites real-time edge AI capability
In addition to the Space-1 Vera Rubin, NVIDIA will also bring its existing edge AI platforms to space applications.
Jetson Orin focuses on small size, low power consumption, and high-performance AI inference, making it suitable for satellites, orbital maintenance vehicles, and space sensing platforms. It can process visual, navigation, and sensor data directly on the spaceborne platform, reducing latency and saving bandwidth.
IGX Thor is positioned as a more highly reliable, mission-critical edge platform, supporting real-time AI processing, functional safety, secure boot, and autonomous operation. For space missions that must operate in harsh environments, such platforms can provide a higher level of autonomous decision-making beyond ground control latency.
NVIDIA’s Space Computing is not only for the space side—it also includes data processing at ground stations. NVIDIA says that RTX PRO 6000 Blackwell Server Edition GPUs can be used for high-throughput ground-side data processing, including satellite image compositing, orthorectification, atmospheric compensation, and large-scale geospatial image analysis. Compared with traditional CPU batch processing systems, NVIDIA claims up to 100x performance improvement.
This means Space Computing is not a single chip product, but a complete AI computing architecture spanning the satellite end, orbital data centers, and ground stations.
NVIDIA Space Computing ecosystem unveiled: Axiom Space, Planet Labs join
NVIDIA said that space industry companies such as AetherFlux, Axiom Space, Kepler Communications, Planet Labs, Sophia Space, and Starcloud have already used NVIDIA’s accelerated computing platform to support both orbital and ground missions.
In terms of the ecosystem, NVIDIA also lists multiple hardware partners, including Aethero, Aitech, EDGX, Eizo, WOLF, and others, which are launching rugged, space and defense-grade edge computing devices built with Jetson Orin or Jetson Thor. Applications cover low-Earth-orbit missions, CubeSats, small satellites, unmanned systems, multi-sensor fusion, and real-time ISR missions.
Among them, Aethero’s NxA Edge Computing Module uses NVIDIA Jetson AGX Thor or Orin, supporting modular design, multi-layer redundancy, and even Kubernetes-enabled distributed on-orbit deployments. It is positioned as an edge AI computing node close to orbital mission operations.
Aitech launched the S-A2300 COTS AI supercomputer, using NVIDIA Jetson Orin. It targets low-Earth-orbit (LEO) missions and brings commercial off-the-shelf (COTS) products into space AI supercomputing applications.
EDGX’s Sterna is an edge computer with flight heritage, equipped with NVIDIA Jetson Orin NX, aimed at small satellite missions from CubeSat to micro-sat. It emphasizes delivering leading compute efficiency per watt-hour in five-year LEO sun-synchronous orbit missions.
Eizo’s Condor Thor 3U VPX Series is a rugged single-board computer powered by NVIDIA Jetson Thor. Use cases include real-time ISR, unmanned systems, and multi-sensor fusion, leaning toward defense, aerospace, and mission-critical scenarios.
WOLF’s WOLF-14T5 is a rugged 3U VPX single-board computer with NVIDIA Jetson AGX Orin and ConnectX-7, offering high-performance AI processing and up to 100GbE networking capability. It targets embedded missions requiring high security and high data throughput.
The article “NVIDIA Space Computing ecosystem unveiled, Space-1 Vera Rubin brings data-center-class AI compute into space” first appeared on ChainNews ABMedia.
