For someone who saw orbital data centers coming, Rama Afullo has an unusual résumé: he's the guy both Google and SpaceX told "no." Now he's building the idea they passed on — and it launches this week.

Satlyt, the startup Afullo co-founded after leaving SpaceX, has raised an $8 million seed round, which he exclusively told TechCrunch on Thursday. The company, headquartered in Sunnyvale, California, and Nairobi, Kenya, builds software that runs AI models on satellites built by other companies — a kind of VMware for orbit, splitting large computing jobs across fleets of spacecraft the way cloud platforms distribute workloads across servers on the ground.

The timing is hard to beat: Satlyt's software launches Thursday aboard a SpaceX rocket, on the same ride as the first prototype for Google's Project Suncatcher, the search giant's own push into orbital computing.

The pitch Google and SpaceX passed on

Afullo, a Kenyan American engineer, worked in Google's cloud computing business before a brief stint at SpaceX's Starlink communications network in 2024. Both stops convinced him that satellites could handle serious computing, not just relay signals — and both employers rejected the idea when he pitched it internally.

"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," he told TechCrunch.

So he left and built it himself. Unlike SpaceX, Google, or startups like Starcloud and Cowboy Space Company, Satlyt isn't building spacecraft at all. The framing Afullo uses: if the orbital data-center builders are the iPhone, Satlyt will be the Android — a horizontally integrated, open ecosystem that works across everybody's hardware.

Intelsat VI communications satellite drifting over Florida's Kennedy Space Center, seen from orbit
Intelsat VI over Florida's Kennedy Space Center. Computing in orbit is moving from concept to reality: Google's Project Suncatcher prototype flies Thursday. (Image: NASA)

Why compute in space at all

The economic case starts with the downlink. Sending data back to Earth is expensive and often slow, so running AI on the spacecraft — to resolve anomalies or process sensor readings where they're captured — can mean real savings. Satlyt has already flown its software on two demonstration missions, and earlier this year it deployed Google DeepMind's Gemma model aboard a spacecraft operated by Momentus. The model cut 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 per year.

The initial focus is operational: spacecraft today depend on flight controllers on the ground to solve any problem that arises, and onboard AI can shrink that dependence.

Three customers on Thursday's flight

This week's launch puts Satlyt's software to work aboard a spacecraft built by TakeMe2Space, an Indian startup that builds computing hardware for satellites. The mission carries three customers:

  • NASA, which is paying the company to test protocols for cloud computing in space;
  • Stellerian, a startup focused on space surveillance — tracking objects in orbit — that wants to test image-processing workloads; and
  • TakeMe2Space itself, to demonstrate that its satellite can host other companies' software.

Satlyt's next project is bolder: a shared computing system — a true cloud — spanning two different satellites, an attempt the company expects to make next year. "The way to think about it is that we turn your satellite into a revenue-generating managed service," Afullo said.

The SBS-4 communications satellite silhouetted above Earth's limb and cloud cover
The SBS-4 communications satellite above Earth. Afullo's goal: Satlyt's software running on 20% of satellites by the end of the decade. (Image: NASA)

Backed by a former astronaut's fund

The seed round was led by Houston-based Non Sibi Ventures, where partner Bernard Harris spent more than 20 years as a NASA astronaut. Partner Kent Lucas told TechCrunch the firm bet on Satlyt precisely because it doesn't need the full orbital-data-center vision to materialize: "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's target is Satlyt software live on 20% of satellites by the end of the decade — by which point, he expects virtually every spacecraft builder to be putting GPUs and similar advanced processors in their satellites. "If you're putting up a satellite without putting up a GPU on it, at the very least, you're doing yourself a disservice, right?" he said.

Sources: TechCrunch (original report, Oct 1, 2026) and AI Daily Post (Oct 1, 2026). Images: NASA (public domain).