From Visionary Rejection to Orbital Innovation: How Satlyt is Building the ‘Android’ of Space Computing

For years, Rama Afullo was the person in the room who saw the future of orbital computing—and for years, he was told that the future wasn’t ready. During his tenure at Google’s cloud computing division and a subsequent stint at SpaceX’s Starlink, Afullo pitched an ambitious concept: shifting the heavy lifting of data processing from Earth-bound servers to the satellites themselves. His internal proposals were met with skepticism, bureaucratic resistance, or outright rejection.

Today, that skepticism has evaporated. Afullo has emerged as the founder of Satlyt, a Sunnyvale- and Nairobi-based startup that has successfully raised an $8 million seed round to build the foundational software layer for space-based AI. As the space industry pivots toward orbital data centers, Afullo is betting that the real gold rush isn’t in building the "hardware boxes" in orbit, but in controlling the operating system that runs them.

The Genesis: A Founder’s Rejection-Fueled Ambition

Afullo’s journey from a disillusioned engineer to a venture-backed founder is a classic tale of Silicon Valley resilience. His frustration peaked while working at the intersection of two of the world’s most innovative tech giants. At SpaceX, he witnessed the sheer scale of the Starlink constellation; at Google, he saw the limitless potential of cloud infrastructure. Yet, when he suggested that these satellites could act as distributed nodes in a global, orbital compute network, the internal consensus was that the industry was too focused on connectivity to worry about processing.

"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 told TechCrunch.

Realizing that his vision was unlikely to gain traction within the rigid hierarchies of multi-billion-dollar corporations, Afullo took the leap into entrepreneurship. By establishing Satlyt, he sought to create a platform-agnostic solution. He didn’t want to build a better satellite; he wanted to build the "VMWare or Snowflake" of the final frontier—software that allows complex workloads to run seamlessly across disparate hardware, effectively decoupling the software from the satellite’s manufacturer.

The Chronology: From Concept to Orbit

The trajectory of Satlyt has been remarkably rapid, moving from a conceptual framework to live orbital demonstrations in a matter of months.

• 2024: Afullo exits his corporate roles to co-found Satlyt. The company secures a critical seed round led by Houston-based Non Sibi Ventures.
• Mid-2024: Satlyt conducts two successful demonstration missions, proving that AI models could operate effectively in the harsh, high-radiation environment of space.
• The "Gemma" Breakthrough: In a partnership with Google DeepMind, Satlyt deployed the Gemma AI model on a spacecraft operated by Momentus. The results were staggering: the model reduced the size of data transmissions related to onboard software errors by more than 60%.
• This Week: Satlyt is set to launch its latest software iteration on a SpaceX rocket. This mission includes a prototype for Google’s own "Project Suncatcher" and hardware from Indian startup TakeMe2Space.

Supporting Data: Why In-Orbit Computing Matters

The economic argument for Satlyt is built on the inefficiency of current space operations. Traditionally, satellites are essentially "dumb" sensors; they collect massive amounts of raw data and beam it back to Earth—a process known as downlink. This process is not only expensive but fraught with latency issues. If a satellite encounters an anomaly, it must wait for instructions from ground-based flight controllers, creating a dangerous delay in resolution.

Satlyt’s software changes this paradigm by enabling "edge computing" in space. By processing data on-orbit, satellites can make autonomous decisions, filter out irrelevant data before transmission, and resolve technical anomalies in real-time.

The financial implications are significant. Afullo notes that the efficiency gains demonstrated by the Gemma AI model alone could save operators hundreds of thousands of dollars per satellite annually. In an industry where launch costs and operational overhead are the primary barriers to entry, such savings are transformative.

The "Android" Strategy: A Horizontal Ecosystem

Satlyt distinguishes itself from industry behemoths like SpaceX and startups like Starcloud or Cowboy Space Company by remaining intentionally "hardware-light." While others are focused on the massive capital expenditure of building custom orbital data centers, Satlyt is building the horizontal ecosystem.

"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 explained.

This strategy is designed to survive the volatility of the space market. By focusing on software that works across various satellites, Satlyt isn’t betting on a single hardware manufacturer’s success. Whether a client uses a massive SpaceX Starlink-class satellite or a smaller nanosatellite from an emerging firm, Satlyt’s OS acts as the common denominator.

Implications: The Future of the Orbital Cloud

The implications of Satlyt’s work extend far beyond mere efficiency. The upcoming launch serves as a bellwether for the future of space infrastructure. The mission, involving partners like NASA, Stellerian, and TakeMe2Space, is a stress test for cloud computing protocols in space.

NASA’s interest is particularly telling; the agency is testing whether cloud protocols can be standardized for space missions. If successful, it could signal the transition from proprietary, siloed satellite operations to a collaborative, interconnected orbital cloud.

The Next Frontier: Shared Computing

Satlyt’s next major milestone, scheduled for next year, involves creating a shared computing system that spans two different satellites. If they can successfully orchestrate a task between two distinct spacecraft, it would mark the first true "compute cloud" in orbit. This would effectively turn a customer’s satellite into a revenue-generating managed service, allowing companies to rent out their onboard processing power when it isn’t being used for their primary mission.

Official Responses and Investor Conviction

The backing of Non Sibi Ventures, led by former NASA astronaut Bernard Harris, provides more than just capital—it provides a seal of technical and operational credibility.

Kent Lucas, a partner at Non Sibi, highlighted why Satlyt’s business model is particularly attractive in the current climate. "We don’t need data centers in space for Rama to be wildly successful," Lucas stated. "It can just be driven by the number of satellites going up."

This perspective shifts the risk profile of the investment. Because Satlyt’s software is valuable regardless of whether the industry moves toward massive data centers or stays decentralized, the company is insulated from the "launch bottleneck" and the unpredictable economics of building custom space hardware.

Conclusion: The GPU Revolution in Orbit

Afullo’s ultimate goal is bold: he wants his software running on 20% of all satellites by the end of the decade. He believes that the rapid integration of GPUs into spacecraft is an inevitability, not a luxury.

"If you’re putting up a satellite without putting up a GPU on it, at the very least, you’re doing yourself a disservice," Afullo said.

As the industry moves toward an era of ubiquitous sensing, high-resolution imaging, and AI-driven telemetry, the need for a robust, flexible, and scalable operating system in space has never been higher. Rama Afullo, once the engineer told "no," is now the architect of the platform that may well become the standard for the next generation of space exploration.

Satlyt will be presenting its vision as part of the TechCrunch Disrupt Battlefield competition, held October 13-15 in San Francisco.