News
Mar 9, 2026
baCta Raises €7M to Build AI-Powered Molecular Factories for Sustainable Industrial Ingredients
News
Enterprise
Artificial Intelligence
Americas
NewDecoded
4 min read

Image by Nvidia
NVIDIA announced new accelerated computing platforms today that bring data-center-class AI performance to orbital data centers and autonomous space missions. Unveiled at the annual GTC event, these tools enable real-time inference and decision-making directly on satellites. This move reduces the dependency on massive data downlinks to Earth by processing information at the source.
The lineup is headlined by the NVIDIA Space-1 Vera Rubin Module, which provides up to 25 times more AI compute than previous hardware for orbital inferencing. Alongside it, the IGX Thor and Jetson Orin modules offer energy-efficient processing for size and power-constrained environments common in spacecraft. These platforms allow large language models to operate directly in orbit.
CEO Jensen Huang emphasized that as we explore deeper into space, intelligence must live where data is generated. This transition transforms spacecraft into self-navigating systems capable of real-time sensing and autonomy. By processing data at the source, missions can now prioritize actionable insights over raw and unrefined telemetry.
Industry leaders such as Planet Labs PBC and Kepler Communications are already integrating these technologies. Planet is using the platform to move from raw pixels to near real-time planetary intelligence. Meanwhile, Kepler has deployed a scalable cloud infrastructure across its satellite network to manage data across its constellation.
While edge AI dominates the orbit, NVIDIA also introduced the RTX PRO 6000 Blackwell Server Edition for ground stations. This hardware accelerates the analysis of massive historical imagery archives by up to 100 times compared to legacy CPU-based systems. It ensures that data collected in space can be combined with terrestrial archives for deep trend analysis.
The practical applications range from immediate disaster response, such as identifying wildfires or oil spills, to precise climate monitoring. These platforms also automate complex object detection for managing global energy grids and transportation networks. The result is a more responsive and intelligent global infrastructure.
The IGX Thor, Jetson Orin, and Blackwell systems are available immediately, while the Space-1 Vera Rubin Module is scheduled for a future release. This rollout establishes a unified computing architecture that spans from the Earth surface to the final frontier. It marks a significant milestone in making space a seamless extension of the global cloud.
This move marks a pivot from satellites acting as simple cameras in the sky to becoming intelligent, autonomous edge nodes. Historically, space data faced a massive downlink bottleneck where huge amounts of raw imagery sat waiting for limited transmission windows to Earth. By moving AI compute to the source, NVIDIA is effectively extending the cloud into low Earth orbit. This enables satellites to make split-second decisions for disaster response or defense without ground intervention, fundamentally changing the economics and utility of orbital infrastructure.
Related Articles