In the burgeoning ecosystem of the new space economy, the next great frontier isn’t just about putting more metal in orbit—it’s about what that metal can actually do. Rama Afullo, a Kenyan-American engineer with deep roots in both the cloud computing giants of Silicon Valley and the rapid-fire innovation of SpaceX, is betting that the future of space exploration lies not in hardware, but in the software layer that orchestrates it.
His company, Satlyt, is attempting to solve a fundamental bottleneck in the modern space race: the inefficiency of the "ground-based command" model. By building a platform that allows satellites to process data in situ—effectively creating a decentralized cloud in the sky—Afullo is positioning his startup to become the "Android of orbit." With a fresh $8 million seed round in hand and a mission set to launch this week, Satlyt is moving from theoretical pitch to operational reality.
The Genesis: From Rejection to Orbit
The story of Satlyt began in the boardrooms of the world’s most powerful tech and aerospace firms. During his tenure at Google’s cloud computing division and later at SpaceX’s Starlink, Afullo saw a glaring misalignment. He recognized that as satellites became more complex and sensor-heavy, the traditional method of downlinking massive amounts of raw data to Earth for processing was becoming unsustainable, both in terms of cost and latency.
"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.
The industry’s collective skepticism stemmed from a hardware-first mindset. For years, the space industry focused on launch vehicle performance and constellation deployment. Afullo, however, was looking at the problem through the lens of a software architect. He realized that the industry needed a platform-agnostic layer that could operate across various satellite architectures, enabling AI-driven autonomy. Undeterred by the corporate roadblocks, Afullo exited the big-tech fold to co-found Satlyt, establishing headquarters in both Sunnyvale, California, and Nairobi, Kenya.
Chronology of a Disruption
Satlyt’s rise has been marked by rapid iteration and strategic validation.
- Early 2024: Afullo departs SpaceX to formalize Satlyt, focusing on the development of a lightweight, scalable operating environment for satellite-based AI.
- Mid-2024: The company conducts two successful demonstration missions, proving that AI models could reliably execute onboard spacecraft.
- The DeepMind Milestone: Earlier this year, Satlyt made headlines by deploying Google DeepMind’s Gemma AI model onto a Momentus spacecraft. The experiment was a watershed moment; the model successfully reduced the size of error-reporting data transmissions by over 60%, showcasing the tangible, high-dollar savings that autonomous data processing could offer.
- The Current Milestone: This week, Satlyt is launching its software aboard a spacecraft built by Indian startup TakeMe2Space. This mission serves as a critical testbed for three distinct stakeholders: NASA, which is vetting cloud-computing protocols; Stellerian, which is testing AI-driven image processing for orbital surveillance; and the host, TakeMe2Space, which aims to demonstrate that its hardware can serve as a multi-tenant platform for third-party software.
Supporting Data: Why "On-Orbit" Intelligence Matters
The economics of space are currently dominated by the "downlink problem." Sending raw, uncompressed sensor data from orbit to Earth is expensive, slow, and prone to atmospheric interference.
Satlyt’s value proposition is centered on operational efficiency. By processing sensor data onboard, a satellite can filter out noise, identify anomalies, and transmit only actionable intelligence to the ground. This capability reduces the burden on ground-station infrastructure and flight controllers.
The metrics from the Momentus mission provide a compelling case study: by cutting transmission volume by 60%, Satlyt estimates it can save operators hundreds of thousands of dollars per satellite annually. In an industry where mission margins are notoriously thin and launch costs remain a primary barrier, this level of optimization is not just a feature—it is a business model.
The Strategy: Building the "Android" of Space
Perhaps the most distinctive aspect of Satlyt’s business model is its refusal to build hardware. In an era where companies like SpaceX and Google are launching their own "orbital data centers," Satlyt is choosing a different path: horizontal integration.
"The folks like SpaceX who are doing orbital data centers—if they are the iPhone, we’ll build Android as a horizontally integrated, open ecosystem," Afullo says.
By functioning as a software provider similar to VMWare or Snowflake, Satlyt aims to provide a unified platform that works regardless of whose satellite is carrying the payload. This modular approach is designed to outlive the specific hardware cycles of any single manufacturer. As Afullo notes, "We don’t need data centers in space for Rama to be wildly successful. It can just be driven by the number of satellites going up."
Official Perspectives: The Investor’s Conviction
The $8 million seed round led by Houston-based Non Sibi Ventures speaks to the shifting investor appetite in the space sector. For Bernard Harris, a partner at Non Sibi and a veteran NASA astronaut of 20 years, the investment was a logical extension of his experience in the field.
"We don’t need data centers in space for Rama to be wildly successful," says Non Sibi partner Kent Lucas. "It can just be driven by the number of satellites going up." The firm recognizes that the real bottleneck isn’t just power or compute, but the integration of that compute into the broader space ecosystem. By backing a software-layer company, they are hedging against the volatility of the hardware market while capturing the growth of the data-processing sector.
Implications for the Future of Space
Satlyt’s trajectory signals a broader shift in how humanity manages its assets in low Earth orbit (LEO). As the density of objects in orbit increases, the need for autonomous decision-making—whether for collision avoidance, traffic management, or real-time data analysis—will become a prerequisite for safety and efficiency.
The Road to the "Cloud"
The company’s next grand ambition is to link multiple satellites into a single, cohesive computing network. By creating a cloud environment that spans across different physical spacecraft, Satlyt hopes to offer a distributed compute utility. If successful, this would fundamentally change the satellite industry from a collection of isolated "islands" of intelligence into a global, interconnected grid.
A New Standard for Satellite Design
Afullo is bullish on the inclusion of advanced processors in every future spacecraft. "If you’re putting up a satellite without putting up a GPU on it, at the very least, you’re doing yourself a disservice," he asserts.
By 2030, Afullo aims for Satlyt’s software to be present on 20% of all orbiting satellites. This goal represents a massive cultural shift for the aerospace industry, which has historically been risk-averse and slow to adopt new software paradigms. However, as the competition for orbital real estate intensifies and the volume of data generated in space continues to explode, the transition to software-defined satellites seems not only inevitable but necessary.
For now, the eyes of the industry are on this week’s launch. If the TakeMe2Space mission performs as expected, Satlyt will have cleared its most significant hurdle yet: proving that its software can facilitate complex, multi-tenant cloud operations in the harsh, unpredictable environment of space. Should they succeed, the dream of an orbital "Android" will move from a visionary pitch to the backbone of the next generation of space exploration.
