Satlyt Inc., a startup that develops artificial intelligence software for satellites, has raised $8 million in funding.
TechCrunch reported today that the seed round was led by early-stage fund Non Sibi Ventures.
Satlyt was founded in 2024 by Chief Executive Rama Afullo (pictured, second from the left). He was previously a product manager at SpaceX Corp.’s Starlink business. Afullo earlier worked as a cloud architect at Google LLC, which launched its first AI-powered satellite into orbit last week.
Most satellites include a relatively limited amount of onboard computing equipment. As a result, they usually don’t process the data they collect locally but rather send it to ground-based servers. The process of beaming information down to Earth involves numerous technical challenges.
The first hurdle is that the average satellite only flies within range of its operator’s terrestrial antenna for a few minutes each day. That limits the amount of data that can be sent to Earth. Furthermore, the data loses some of its accuracy on its way to the antenna.
Satly’s namesake AI platform is designed to address those challenges. It enables satellites that include graphics processing units to process some of the data they collect internally. That reduces the amount of information they must send to Earth, which eases the associated networking challenges.
The first satellite to feature Satlyt’s platform was launched into low-Earth orbit in March. The system was built by a venture-backed startup called Momentus Inc. It uses Satlyt’s software to run Google’s open-source Gemma 3 1B language model.
The companies evaluated the system by simulating a set of common satellite software errors. Satlyt’s platform used Gemini to collect error logs, identify the root cause and suggest a fix. The platform reduced the amount of telemetry that had to be beamed down to Earth for analysis by more than 64%.
As of April, Satlyt was working to upgrade its stack to a newer, more capable AI model called Gemma 4 E2B. The company is also exploring ways of reducing AI models’ hardware footprint. Satly has set a goal of lowering the Gemma 4 E2B’s memory usage by more than 90%.
Another Satlyt-powered satellite is scheduled to take off later today aboard a SpaceX rocket. NASA will use the company’s software to test new orbital computing methods. Stellerian Inc., a space observation startup, plans to run image processing algorithms aboard the system.
Satlyt is reportedly working on a new software tool that will make it possible to distribute workloads across multiple satellites. The first system to feature the technology is set to launch into orbit next year.