An Indian startup has switched on a surveillance network it says can watch low Earth orbit around the clock without relying on data from foreign countries.
Digantara, based in Bengaluru, unveiled MOSAIC on Aug. 15, calling it an AI-powered wide-area sensing architecture that searches for, detects and tracks hundreds of objects at once. The first deployment strings together five optical sensor nodes, each pairing a sky-imaging unit with electronics that handle onboard processing, timing and autonomous operation.
The system runs on solar power with battery backup and is built to survive punishing sites, from the heat of the Thar Desert to Himalayan winters. Machine learning separates stars from resident space objects, including faint satellites and debris, using lost-in-space attitude estimation to fix an object’s position without prior tracking data.
The timing matters. During Operation Sindoor, India’s war with Pakistan earlier this year, military planners leaned on ISRO’s Cartosat and RISAT constellations plus commercial imagery said to come from US company Maxar to confirm targets and evaluate strikes. The mix worked, yet it also exposed how dependent India remained on outside sources for time-critical orbital information.
Digantara frames MOSAIC as the answer: a fully indigenous capability for continuous orbital awareness that feeds space domain awareness and defense intelligence, built in India for the scale of modern orbital traffic.