The Pulse
Satlyt Raises $8 Million to Run AI Across Satellites
Satlyt raised an $8 million seed round to build software that runs AI workloads on satellites and coordinates computing across spacecraft. The company plans to test its software on a TakeMe2Space satellite, with NASA among the customers for
AI.info Team · Written with AI ·
“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.”
Rama Afullo, Satlyt co-founder and CEO, speaking to TechCrunch
Rama Afullo’s rejected pitches at Google and SpaceX became the starting point for Satlyt, a company developing software to run AI on satellites. The Sunnyvale, California- and Nairobi-based startup has raised an $8 million seed round led by Houston-based Non Sibi Ventures, Afullo told TechCrunch.
Satlyt is not building its own spacecraft. Its software is meant to run across satellites made by different companies, letting operators process information in orbit and, eventually, share computing work between spacecraft. That approach puts the company’s bet on software that can work across hardware providers rather than on a new fleet of satellites.
Satlyt wants satellites to process data before downlink
Satellites often send sensor readings and other information to Earth for analysis, a process that can be expensive and slow. Satlyt’s initial use cases put AI onboard to handle some tasks locally, including diagnosing spacecraft software problems and processing sensor data. The spacecraft can then transmit a smaller set of results rather than all the raw information.
Satlyt has already flown its software on two demonstration missions. Earlier this year, the company deployed Google DeepMind’s Gemma model on a spacecraft operated by Momentus. TechCrunch reported that the model reduced the size of a transmission about onboard software errors by more than 60%; Afullo said the resulting efficiencies could save hundreds of thousands of dollars per satellite each year.
That figure is Afullo’s estimate, not a reported price cut or a result established across satellite missions. The cited demonstration focused on transmitting information about software errors. Satlyt’s broader commercial case depends on showing similar benefits in other workloads and on satellites with different systems.
NASA and two startups are customers for the next test
At the time of the funding announcement on October 1, Satlyt’s software was scheduled to fly that day on a SpaceX rocket aboard a spacecraft built by Indian startup TakeMe2Space. The mission has three customers: NASA is paying Satlyt to test protocols for cloud computing in space, space-surveillance startup Stellerian plans to test image-processing workloads, and TakeMe2Space wants to demonstrate that its satellite can host software from other companies.
The mission tests whether Satlyt can deploy its software on another company’s spacecraft and support practical workloads. It is not yet the multi-satellite cloud the startup ultimately wants to offer. Satlyt’s next stated milestone is a test of shared computing across two different satellites, which Afullo expects the company to attempt in 2027.
The shared orbital cloud remains unproven
Satlyt’s ambitions extend beyond processing data on a single spacecraft. If it can coordinate workloads across satellites, the company could offer satellite operators access to computing capacity without requiring each operator to build a dedicated orbital data center. For now, that network is a plan: the two-satellite test has not yet taken place.
Non Sibi partner Kent Lucas told TechCrunch that Satlyt’s prospects do not depend on orbital data centers becoming a major market: “We don’t need data centers in space for Rama to be wildly successful, right? It can just be driven by the number of satellites going up.” Afullo says he hopes Satlyt will be operating on 20% of satellites by the end of the decade. Its nearer test is whether software running on two separate spacecraft can share computing work.