Satlyt raises $8M to run AI software on satellites
Satlyt, a startup founded by former Google and SpaceX product manager Rama Afullo, has raised an $8 million seed round. The company, based in Sunnyvale, California, and Nairobi, is developing software that can run AI workloads across satellites built by different companies rather than manufacturing spacecraft itself.
Afullo compares the approach with an open Android ecosystem, contrasting it with SpaceX’s integrated model. Satlyt has already tested its software on two demonstration missions. In one deployment, Google DeepMind’s Gemma model processed satellite data aboard a Momentus spacecraft and reduced the size of a transmission about onboard software errors by more than 60%. The company says such processing can reduce the cost of sending data to Earth and help spacecraft respond to problems locally.
The next mission will use a spacecraft built by India’s TakeMe2Space. NASA will test protocols for cloud computing in space, Stellerian will evaluate image-processing workloads, and TakeMe2Space will demonstrate hosting external software. Satlyt plans to attempt a shared computing system spanning two satellites next year. The round was led by Houston-based Non Sibi Ventures.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.