Satlyt, a software startup focused on enabling artificial intelligence operations aboard satellites, has closed an $8 million seed funding round. The company was co-founded by Rama Afullo, a Kenyan American engineer with prior experience in cloud computing at Google and a brief tenure at SpaceX’s Starlink division in 2024.
Afullo identified a significant gap in the aerospace market early in his career. “When I was at SpaceX, I tried to pitch this internally. They said no. When I was at Google, I tried to pitch this internally. They said no,” Afullo told TechCrunch. “So he left SpaceX and co-founded Satlyt,” a firm with headquarters in Sunnyvale, California, and Nairobi, Kenya.
Unlike competitors such as SpaceX, Google, or other startups like Starcloud and Cowboy Space Company, Satlyt is not manufacturing hardware. Instead, it is developing an operating system that allows multiple satellites to share heavy computational workloads. Afullo compares the model to VMware and Snowflake, noting, “If [SpaceX] are the iPhone, we’ll build Android as a horizontally integrated, open ecosystem.”
The timing of the raise coincides with a broader industry shift. SpaceX has recently committed to orbital data centers, and Satlyt’s software is scheduled to launch this Thursday on a SpaceX rocket alongside Google’s Project Suncatcher prototype.
Initial applications focus on operational efficiency. Satellites currently rely on ground controllers to resolve issues, but downlinking data is costly and slow. By processing sensor data and anomalies onboard using AI, operators can achieve substantial savings. Earlier this year, Satlyt deployed Google DeepMind’s Gemma model on a Momentus spacecraft, reducing transmission sizes for onboard software errors by over 60%—a metric Afullo estimates could save hundreds of thousands of dollars per satellite annually.
The upcoming mission will carry Satlyt’s software on a spacecraft built by Indian startup TakeMe2Space. The payload includes tests for three clients: NASA, which is evaluating cloud computing protocols in space; Stellerian, a space surveillance firm testing image processing; and TakeMe2Space itself.
While high-powered GPUs are still rare in orbit, Satlyt plans to demonstrate a shared computing cloud spanning two different satellites next year. If successful, this would establish a true third-party compute cloud service for spacecraft builders.
The seed round was led by Non Sibi Ventures, a Houston-based firm where partner Bernard Harris, a former NASA astronaut, helped validate the business case. Partner Kent Lucas noted that the investment does not rely on the uncertain economics of massive orbital data centers, but rather on the growing volume of satellites being launched.
Afullo aims to have his software operational on 20% of satellites by the end of the decade, predicting that future spacecraft builders will consider GPUs essential. “If you’re putting up a satellite without putting up a GPU on it, at the very least, you’re doing yourself a disservice,” he said.
Satlyt is also participating in TechCrunch’s Battlefield competition at Disrupt in San Francisco, taking place October 13-15.
An open-source Android for satellites? That’s a clever comparison. Here’s hoping it doesn’t become another fragmented mess.
Hyperscaling cloud workloads in orbit is a tough sell. Hardware failures up there are no joke. Let’s see the reliability numbers.
Wait, so Google and SpaceX both said no? I guess that explains why they aren’t doing it themselves!
Finally, someone is solving the latency bottleneck in space data processing. This feels inevitable.