The startup Satlyt raised a seed round of US$ 8 million to develop software capable of running artificial intelligence models directly onboard satellites. The company was founded by Kenyan-American engineer Rama Afullo, who previously worked at Google and SpaceX, and this week will have its first commercial mission through a launch by Elon Musk's space company.
The round was led by Non Sibi Ventures, a Houston-based firm that decided to bet on a still nascent market: data processing directly in orbit. “We don’t need data centers in space for Rama to do very well,” explained Kent Lucas, a partner at the fund, in a conversation with TechCrunch.
Process data before sending it to Earth
Satlyt's proposal aims to introduce greater computing capacity within the satellites themselves. “It could simply be driven by the number of satellites launched,” Lucas stated while explaining why Non Sibi Ventures believes the business can grow even without the need for more ambitious projects for space data centers.
The goal is for satellites to be able to run artificial intelligence models directly in orbit, rather than relying solely on sending information to ground facilities for processing. The company seeks to develop the necessary software to leverage the computers installed onboard the new generations of satellites.
One of the drivers of the investment was Bernard Harris, a partner at Non Sibi Ventures and former NASA astronaut. His experience in the space sector helped the fund evaluate an idea that is still in an early stage of commercial development.
An idea rejected at Google and SpaceX
Before creating Satlyt, Afullo worked in the cloud computing area at Google and later went to Starlink, SpaceX's satellite network, during 2024. There, he became convinced that there was a business opportunity around incorporating more powerful computers within satellites.
“When I was at SpaceX, I tried to push this internally. When I was at Google, I tried the same. I was told no,” Afullo recounted about the origin of his project.
