Satlyt, a startup founded in 2024 by Rama Afullo, completed a successful test this month for software designed to distribute computing tasks across in-orbit satellites—reducing dependence on downlinks and ground-based analysis.
The test, which followed 13 months of development, demonstrated the software's ability to dynamically assign tasks among networked satellites based on communication availability and processing capacity. Afullo, previously with Google and SpaceX, said the company's objective is to develop software that connects multiple satellites to facilitate shared computing assignments while users define permissions, priorities and operating limits.
Satlyt's software is already active on three satellites in orbit. The company worked with the University of Houston on a research project funded by NASA. While NASA funded the foundational research, the agency has not yet selected Satlyt for any specific mission or deployment.
For the test, Satlyt created a simulated network comprising four computers, four users and eight services. The software then evaluated how to assign and execute tasks across these simulated hosts, taking into account both compute and communication availability. The experiment ran on a single physical computer aboard a spacecraft, which evaluated the simulated network and transmitted results back to Earth—proving the software can generate operational plans on hardware in orbit.
The initial test revealed a performance trade-off, with one approach proving most effective though it demanded more computing power than alternatives. Afullo noted that NASA missions face similar constraints that Satlyt aims to address: limited computing resources, intermittent communication, and inefficiency in transmitting raw data to Earth.
Further testing is planned to observe how information transfers between multiple physical computers in practice. The company was recognized as a top five participant at TechCrunch Disrupt Battlefield this year.
Satlyt's software aims to enable operators to run artificial intelligence and other applications directly on spacecraft. On-board processing reduces the cost and latency of downlinks—particularly valuable for time-sensitive applications like rapid wildfire detection and analysis. NASA currently operates a separate technology network for moving data through space during communication delays, technology the agency began using earlier this year and has already tested on aircraft.
