Satlyt, a startup based in Sunnyvale, California, and Nairobi, has raised $8 million in a seed funding round to develop software for satellite-based AI processing. The round was led by Non Sibi Ventures.
Satlyt raises $8M to bring AI processing to satellites
Founded by former Google and SpaceX product manager Rama Afullo, Satlyt focuses on building a software layer that allows satellites to process data in orbit. The company's approach aims to help satellites share computing workloads, functioning as a horizontally integrated ecosystem similar to how Android operates on various hardware.
Currently, satellite operators face high costs and latency when downlinking data to Earth, with some reports suggesting up to 90% of downlinked data may be redundant or corrupted before it reaches the ground. By running AI models directly on the spacecraft, operators can process sensor readings and resolve anomalies locally, significantly reducing the volume of data required for transmission.
Satlyt's software has already been deployed in demonstration missions. In one instance, the company used Google DeepMind's Gemma model on a Momentus-operated spacecraft, which reduced the transmission size for software error reports by more than 60%.
The company’s software is scheduled for launch this week on a SpaceX rocket alongside a prototype for Google's Project Suncatcher. The mission involves multiple customers, including NASA, which is testing protocols for space-based cloud computing. Satlyt plans to attempt a project next year to demonstrate a shared computing system that spans two different satellites.
Sources
- Satlyt, founded by a former Google and SpaceX product manager, raises $8M to run AI on satellites （TechCrunch AI, 2026-10-01）
- Satlyt