CHENNAI: Bengaluru-based spacetech company GalaxEye is likely to launch its AI-powered “Mission Drishti” satellite by March-end this year, with the company exploring feasibility of setting up orbital data centre using NVIDIA Jetson Orin GPU.
GalaxEye said the Drishti will feature world’s first “SyncFused OptoSar” payload, which uniquely integrates Electro-Optical (EO) and Synthetic Aperture Radar (SAR) sensors on a single satellite platform.
Usually, having both EO and SAR sensors in one vehicle is difficult for a payload to carry because of contradictory operational, power, and structural requirements. Powered by the NVIDIA Jetson Orin, the Mission Drishti satellite runs advanced AI workloads directly in orbit, dramatically accelerating how Earth observation data is processed, interpreted, and delivered to customers.
NVIDIA Jetson Orin is a high-powered GPU architecture which will be placed inside the satellite vehicle to store data. GalaxEye will also use Mission Drishti to explore the feasibility of Orbital Data Centres (ODC) where multiple satellites operate as interconnected compute nodes.
Insights from Drishti are expected to influence to design of future constellations, increase scalability, reliability, and operational economics.
It is also planning to use NVIDIA Nemotron open models and NVIDIA Cosmos world foundation models for development. Suyash Singh, CEO of GalaxEye, told TNIE, “We are likely to launch the satellite vehicle tentatively by May or June. We are targeting to launch the payloads within four months.”
Speaking about the business model and future growth, Singh said, “For us, it is demonstration of not just our sensors going into the space but also the assessing the feasibility of data centres in orbit.
We started this journey five years when are satellite design was frozen. Eventually, this has become a feasibility study to think about orbital data centre constellation for GalaxEye in the future.”
The primary reason for considering orbital data centres is to save electricity and energy on Earth, specifically by moving AI workloads into space. “Since AI workloads are increasing, and GPUs are essential for managing them, the decision was made to send a computational unit with Nvidia's GPUs to space to test its capability to handle AI workloads.
The successful performance of this unit will support the decision to proceed with orbital data centres,” he said. Singh further said, “Once the satellite goes into the space, we will be sending AI workloads to this computational unit and later we will see the performance.”