Israeli technology company Maris-Tech has announced Venus-Space, a new version of its Venus-Pro platform designed for nanosatellites operating in low Earth orbit (LEO).
Get edge AI off the ground
Traditionally, small satellites have served as data collection devices, capturing images and relaying them to ground infrastructure for processing. Although effective, this approach consumes valuable bandwidth and can introduce delays between data collection and analysis.
Venus-Space is moving much of its processing into orbit. The platform is built on a durable FPGA-based architecture and can process ultra-high resolution video streams at speeds of up to 25 Gbps while storing both raw and compressed data onboard.
Process data before it reaches Earth
The platform includes a dedicated AI accelerator that can run neural networks directly on the satellite. This will allow spacecraft operators to analyze images in near real-time and identify specific targets or events, such as ships at sea, vehicle movement, wildfires, or infrastructure damage. Instead of transmitting every captured frame, the system can send compressed video clips, selected images, or event alerts to Earth, prioritizing relevant information.
According to Maris-Tech, this approach could reduce bandwidth requirements, optimize satellite communications, and improve the responsiveness of Earth observation satellite constellations.
From military vehicles to spaceships
Venus-Space is derived from the company’s Venus-Pro platform, which was originally developed for military vehicles operating in harsh environments. By adapting this technology to in-orbit missions, Maris-Tech is positioning the system as a compact, power-efficient payload for small satellites.
The company says the platform is designed to withstand difficult operating conditions while providing reliable onboard computing performance.
Growing trend towards orbital computing
The launch of Venus-Space reflects a broader shift to edge computing architectures across the space industry. Satellites are increasingly being designed to process information where it is generated, rather than acting solely as data collection platforms.
As the Earth observation satellite fleet continues to expand, onboard AI systems could help carriers alleviate data bottlenecks. It also aims to improve response times and make more efficient use of limited communication bandwidth.
If successful, platforms like Venus-Space could help transform future satellite networks from simple sensing systems to distributed computing infrastructures operating directly in orbit.
