Tested Intelligence — October 3, 2026
Satlyt, a startup building software to run AI workloads on satellites, has raised an $8 million seed round — and its software is already flying. The company’s tech launched October 1 aboard a SpaceX rocket on a satellite built by Indian startup TakeMe2Space, sharing the ride with the first prototype of Google’s Project Suncatcher. Why it matters: downlinking data from satellites is expensive and slow, so running AI onboard — to resolve anomalies or process sensor readings in space — can save real money.
What happened
Co-founder Rama Afullo, a Kenyan American engineer who worked at Google Cloud and then spent a stint at SpaceX’s Starlink network in 2024, told TechCrunch he pitched orbital computing internally at both companies. Both said no — so he left SpaceX and co-founded Satlyt, based in Sunnyvale, California, and Nairobi.
Unlike SpaceX, Google, or startups like Starcloud and Cowboy Space Company, Satlyt isn’t building spacecraft. It’s building the software layer that lets AI models run across fleets of satellites built by other companies — “If they are the iPhone, we’ll build Android as a horizontally integrated, open ecosystem,” Afullo said, comparing the approach to VMware and Snowflake.
The $8 million seed round was led by Houston-based Non Sibi Ventures, whose partner Bernard Harris — a former NASA astronaut — helped ground the firm’s conviction in the bet. Notably, the firm says Satlyt doesn’t depend on grand orbital data-center visions: “We don’t need data centers in space for Rama to be wildly successful,” partner Kent Lucas said. “It can just be driven by the number of satellites going up.”
The tech, in orbit
Satlyt has already flown its software on two demonstration missions. Earlier this year, it ran Google DeepMind’s Gemma model aboard a Momentus spacecraft, cutting the size of a transmission about onboard software errors by more than 60% — the kind of efficiency Afullo says can save hundreds of thousands of dollars per satellite each year.
This week’s launch carries three customers: NASA, which is paying the company to test protocols for cloud computing in space; space-surveillance startup Stellerian, testing image-processing workloads; and TakeMe2Space itself, demonstrating that a satellite can host other companies’ software. Next up, Satlyt plans to demonstrate a shared computing cloud spanning two different satellites next year, and Afullo hopes to be live on 20% of satellites by the end of the decade.
Why it matters
Orbital computing is getting crowded fast — SpaceX is going all-in on orbital data centers and Google just launched its first Suncatcher hardware into orbit. Satlyt’s bet is that the software platform wins, not the hardware. If satellite operators stop building bespoke software for every mission and instead run third-party AI workloads on shared infrastructure, that’s a genuinely new market. Our take: the real validation here isn’t the $8 million, it’s NASA as a paying customer testing cloud protocols in space. That’s how infrastructure markets start.
Frequently asked questions
What does Satlyt do?
Satlyt builds software that runs AI models and computing workloads on satellites, letting operators process data in orbit instead of sending it back to Earth — without building the spacecraft itself.
Who founded Satlyt and who funded the round?
Rama Afullo, formerly of Google Cloud and SpaceX’s Starlink, co-founded the company. The $8 million seed round was led by Houston-based Non Sibi Ventures.
Why run AI on satellites?
Downlinking data to Earth is expensive and often slow. Running AI onboard can cut transmission costs — Satlyt’s Gemma deployment cut one transmission type by more than 60% — and enable faster responses to anomalies.
How is this different from Google’s Project Suncatcher?
Project Suncatcher is Google’s own orbital-computing hardware effort. Satlyt builds software designed to run on many companies’ satellites, positioning itself as the neutral platform layer rather than another hardware competitor.
Sources: TechCrunch, The AI Insider, AI Daily Post