AI software startup Satlyt raises $8M to power computing across satellite fleets

Rama Afullo pitched the idea of computing in space to two giants—and got rejected by both. After Google and SpaceX each declined his internal proposals, he decided to start his own company to make the vision real.

Afullo's startup, Satlyt, has just closed an $8 million seed round to build software that enables satellites to run AI models and share computational work while orbiting Earth. Rather than constructing its own spacecraft, Satlyt positions itself as the operating layer for satellite-based computing—similar to how VMware and Snowflake provide platforms that abstract away hardware complexity.

The pitch: Android for space#

Satlyt's core argument is that orbital data centers will proliferate, but no single company should own the software ecosystem. While SpaceX, Google, and other players are racing to build their own satellite infrastructure, Afullo aims to offer a horizontal platform that works across multiple operators' fleets.

"If they are the iPhone, we'll build Android as a horizontally integrated, open ecosystem," he explained to TechCrunch, capturing the company's strategy to position itself as the standard middleware for satellite computing.

Real-world cost savings already showing#

The company has already demonstrated measurable impact. Earlier this year, Satlyt deployed Google DeepMind's Gemma AI model aboard a Momentus spacecraft. The result: transmission data about software errors shrank by over 60%—a reduction that Afullo says can save hundreds of thousands of dollars annually per satellite.

The efficiency gain stems from a fundamental truth of space operations: sending data back to Earth (called downlink) is both slow and expensive. By processing information and making decisions directly on the satellite, operators avoid the bottleneck of ground-based troubleshooting and sensor analysis.

This week's milestone launch#

On Thursday, Satlyt's software launches aboard a TakeMe2Space satellite built by an Indian hardware startup. The mission will serve three purposes:

• NASA will test cloud computing protocols in space
• Stellerian, a space surveillance startup, will evaluate image-processing workloads in orbit
• TakeMe2Space itself will validate that the satellite can host third-party software