Prabowo Widodo
The burgeoning commercial space economy is rapidly moving beyond simple observation and communication, setting its sights on a much more complex frontier: data processing in orbit. Rama Afullo, a Kenyan-American engineer with pedigree at both Google and SpaceX, foresaw this paradigm shift long before aerospace giants embraced the concept. Today, his startup, Satlyt—headquartered jointly in Sunnyvale, California, and Nairobi—stands at the vanguard of space-based software development. Having successfully closed an $8 million seed funding round led by Houston-based Non Sibi Ventures, Satlyt is positioning itself to become the universal operating layer for the next generation of smart satellites.
Unlike industry behemoths or direct hardware competitors like Starcloud and Cowboy Space Company, Satlyt is purposefully hardware-agnostic. Rather than manufacturing physical spacecraft, Afullo’s team is building horizontally integrated software designed to operate artificial intelligence models and distributed workloads across disparate satellites. By drawing parallels to enterprise software giants like VMware and Snowflake, Satlyt aims to provide an open ecosystem that allows satellite operators to run complex computational tasks without wrestling with underlying hardware infrastructure. As the aerospace sector prepares for an influx of graphical processing units (GPUs) and advanced edge-computing hardware in orbit, Satlyt’s software architecture arrives at a critical juncture for commercial space operations.
From Internal Rejection to Entrepreneurial Validation
Afullo’s journey toward founding Satlyt stems from institutional resistance he encountered while working within the tech and aerospace sectors. During his tenure in Google’s cloud computing division and a subsequent stint at SpaceX’s Starlink communications network in 2024, Afullo repeatedly pitched the concept of decentralized orbital computing.
"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," Afullo recalls.
Recognizing that the traditional gatekeepers of tech and aerospace were moving too slowly to capture the nascent market of space-based edge computing, Afullo departed SpaceX to establish Satlyt. The timing of his departure has proven fortuitous. Since Afullo’s exit, SpaceX has aggressively pursued orbital data center initiatives. Furthermore, Satlyt’s software is scheduled to launch on a SpaceX rocket alongside Project Suncatcher, Google’s inaugural space data center prototype.
The validation of Satlyt’s vision was cemented by its $8 million seed round. The investment was led by Non Sibi Ventures, a firm where partner Bernard Harris—a veteran NASA astronaut who spent more than two decades in the space program—provided critical insight into the viability of a software-first space business model. Non Sibi partner Kent Lucas noted that Satlyt’s long-term success is decoupled from the immense capital expenditures and launch bottlenecks associated with building physical space data centers. Instead, Satlyt’s financial model scales directly with the rapidly accelerating volume of commercial satellites launching into low-Earth orbit.
Operational Efficiency and Early Proofs of Concept
Historically, operating satellites has been an expensive, labor-intensive endeavor heavily reliant on ground-based flight controllers. When spacecraft experience technical anomalies, raw telemetry data must be transmitted back to Earth—a process known as downlink—for analysis. Because downlinking data is often slow, bandwidth-constrained, and costly, spacecraft frequently suffer from high latency in anomaly resolution. Satlyt addresses this bottleneck by deploying artificial intelligence directly onto the spacecraft, enabling edge computing that automates decision-making and sensor processing in real time.
The startup has already successfully validated its software across two separate demonstration missions. Earlier this year, Satlyt partnered with spacecraft operator Momentus to deploy Google DeepMind’s Gemma AI model directly onboard an orbiting vehicle. The implementation yielded immediate, quantifiable results: the onboard AI model successfully reduced the transmission size of error logs regarding spacecraft software by more than 60 percent. According to Afullo, efficiencies of this magnitude can generate hundreds of thousands of dollars in operational savings per satellite annually.
Building upon these early milestones, a pivotal mission highlights Satlyt’s software versatility. A SpaceX launch deploys Satlyt’s software onto a spacecraft engineered by TakeMe2Space, an Indian startup specializing in satellite computing hardware. The mission serves a tripartite customer base with distinct operational objectives:
- NASA: Contracting the mission to evaluate foundational protocols for cloud computing in space environments.
- Stellerian: An orbital surveillance startup utilizing the mission to test advanced image-processing workloads for tracking space debris and satellites.
- TakeMe2Space: Leveraging the flight to demonstrate that its hardware infrastructure can seamlessly host multi-tenant software from third-party providers.
Toward a Distributed Orbital Cloud
While high-powered GPUs remain relatively scarce in current orbital deployments, the industry is rapidly pivoting toward advanced edge-processing capabilities. Satlyt’s immediate revenue model centers on optimizing existing spacecraft operations, transforming standard satellites into revenue-generating managed services through intelligent automation.
However, the company’s roadmap extends far beyond localized anomaly detection. Satlyt’s next major engineering milestone, slated for next year, involves establishing a shared, distributed computing system—a true orbital cloud—spanning two distinct satellites in cooperation. If successful, this demonstration will establish Satlyt as the premier third-party infrastructure provider capable of orchestrating cross-satellite computational workloads.
"The way to think about it is that we turn your satellite into a revenue-generating managed service," Afullo explains, framing the software as an indispensable utility for satellite manufacturers looking to monetize idle computational capacity.
Market Implications and Future Outlook
The broader implications of Satlyt’s market entry point toward a profound structural shift in how space infrastructure is monetized and managed. As launch costs continue to decline and the commercialization of low-Earth orbit accelerates, the concentration of data generated by Earth-observation, communication, and surveillance satellites is growing exponentially. Transmitting all raw data back to terrestrial servers is no longer a sustainable economic or bandwidth-efficient model.
Industry analysts project that within the decade, the integration of GPUs and advanced enterprise processors into satellite architecture will transition from a luxury to an industry standard. Afullo is aggressively positioning Satlyt to capture this transformation, setting an ambitious corporate goal to have its software embedded on 20 percent of all operational satellites by the end of the 2020s.
"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?" Afullo asserts.
By bypassing the capital-intensive hurdles of hardware manufacturing and focusing exclusively on an open, scalable software ecosystem, Satlyt has charted a pragmatic course through the complex economics of the new space race. As this week’s rocket launches carry its code into orbit alongside tech giants and space agencies alike, Satlyt moves closer to cementing its role as the foundational operating system for the next era of human technological infrastructure.
