Asro
In the rapidly evolving landscape of the New Space economy, the dream of an "orbital data center" was once dismissed as a niche, even quixotic, ambition. For Rama Afullo, a Kenyan American engineer with deep roots in both Silicon Valley cloud infrastructure and the cutting-edge aerospace sector, the vision was not merely a dream—it was an inevitability. After stints at Google’s cloud division and SpaceX’s Starlink, Afullo found himself in the position of a prophet ignored by his own employers.
“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.
Today, that rejection has catalyzed a $8 million seed round and the launch of Satlyt, a company headquartered in Sunnyvale and Nairobi that aims to democratize computing in orbit. By providing a platform-agnostic software layer for satellites, Satlyt is positioning itself as the "Android" of space—a horizontally integrated ecosystem that allows developers to run complex AI workloads across a heterogeneous fleet of spacecraft.
The Chronology of a Vision
The journey from internal pitch decks to orbital reality has been swift. Afullo’s career trajectory provided the necessary bridge between two worlds that rarely communicate: high-scale software engineering and orbital mechanics.
His time at Google exposed him to the complexities of distributed computing at a global scale, while his tenure at SpaceX provided a front-row seat to the massive deployment of satellite constellations. He identified a glaring inefficiency: satellites are essentially "dumb" sensors, relying on costly, slow, and latency-heavy downlinks to send data to Earth for processing.
In 2024, Afullo pivoted to entrepreneurship. By mid-2024, Satlyt had secured $8 million in seed funding led by Non Sibi Ventures. The firm, led by former NASA astronaut Bernard Harris, saw in Satlyt a business model that didn’t rely on the capital-intensive gamble of building bespoke "space data centers," but rather on the software-defined infrastructure that turns any satellite into a revenue-generating node.
This week marks a significant milestone for the startup. Satlyt’s software is scheduled to launch on a SpaceX rocket, sharing the ride with Google’s own experimental "Project Suncatcher" data center. While the industry giants build the "iPhone" of space hardware, Satlyt is focusing on the operating system that makes that hardware useful to the broader market.
Supporting Data: Efficiency at Altitude
The primary value proposition of Satlyt lies in the economics of data transmission. Currently, the process of sending raw sensor data from space to a ground station is inefficient. It incurs significant costs and introduces latency that is unacceptable for time-sensitive applications like disaster response, space situational awareness, and real-time defense monitoring.
Satlyt’s software has already proven its worth in two demonstration missions. In one notable deployment, the company utilized Google DeepMind’s Gemma AI model onboard a Momentus-operated spacecraft. The results were stark: the model reduced the data transmission size related to onboard software errors by more than 60%.
Afullo estimates that such efficiencies can translate to savings of hundreds of thousands of dollars per satellite annually. These operational savings are critical for the burgeoning market of commercial small-sats, where every milliwatt of power and every byte of downlink bandwidth carries a premium price tag.
The current mission, involving an Indian startup, TakeMe2Space, serves as a stress test for Satlyt’s versatility. The mission hosts three distinct stakeholders:
- NASA: Testing protocols for cloud computing in a vacuum environment.
- Stellerian: A space surveillance firm requiring on-orbit image processing to track orbital debris.
- TakeMe2Space: Demonstrating the viability of a multi-tenant, hosted computing environment.
Official Responses and Strategic Pivot
The investment thesis from Non Sibi Ventures highlights the pragmatic nature of Satlyt’s approach. Unlike competitors who are betting the farm on the construction of massive, dedicated space data centers—a strategy plagued by high launch costs and regulatory bottlenecks—Satlyt is betting on the proliferation of processing power across the existing and future satellite population.
"We don’t need data centers in space for Rama to be wildly successful," says Kent Lucas, a partner at Non Sibi Ventures. "It can just be driven by the number of satellites going up."
Afullo’s philosophy is rooted in the "VMware/Snowflake" model. Just as those companies abstracted the complexity of data centers and cloud storage for enterprise users, Satlyt aims to abstract the complexities of satellite hardware. A user should not need to know the specific radiation-hardened processor architecture or the proprietary communication protocol of a specific satellite bus to run an AI model. They should simply be able to "deploy to space."
"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 stated.
Implications for the Future of the Space Economy
The long-term goal for Satlyt is the creation of a true "compute cloud" that spans multiple satellites. The company plans to attempt its first cross-satellite computing link next year. If successful, this would represent a paradigm shift in space operations. Instead of a single satellite acting as a standalone computer, the orbital environment would become a distributed network, capable of pooling resources for massive computational tasks.
This shift has profound implications:
1. The Rise of "Space-as-a-Service"
By turning satellites into managed services, owners of underutilized satellite hardware can lease their excess GPU and CPU cycles to third parties. This turns a high-CAPEX asset into an OPEX-generating tool, changing the math for satellite operators.
2. Democratization of Space AI
If a startup with a specialized algorithm for climate monitoring no longer needs to build its own satellite—and can instead simply lease "compute time" on an existing, Satlyt-enabled satellite—the barrier to entry for space-based innovation collapses.
3. Resilience and Redundancy
A distributed compute cloud is inherently more resilient. If a satellite suffers an anomaly or is compromised, the workload can be migrated to another node in the constellation. This is essential for critical infrastructure and defense applications.
4. Hardware Evolution
Afullo is bullish on the future of space-borne hardware, predicting that within the decade, every serious satellite builder will integrate advanced GPUs into their designs. "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 argues.
Conclusion: The Final Frontier of Software
As Satlyt prepares for its upcoming mission and its presentation at the TechCrunch Disrupt battlefield, the company sits at a unique intersection of aerospace and software engineering. While the "New Space" race was initially defined by the lowering of launch costs through reusable rockets, the next phase will be defined by what we do once we are in orbit.
By focusing on the software layer, Satlyt has effectively side-stepped the "launch bottleneck" and the inherent risks of hardware manufacturing. Instead, they are building the infrastructure that will allow humanity to process the vast amounts of data we collect in space, transforming the void from a place of mere observation to a place of active, intelligent computation.
Whether Satlyt succeeds in capturing 20% of the satellite market by 2030 remains to be seen. However, the industry’s shift toward on-orbit processing is undeniable. As satellites transition from relay stations to intelligent, cloud-connected nodes, the software that orchestrates this transition will be the most valuable piece of real estate in orbit. Rama Afullo may have been told "no" by the giants of the industry, but he is currently building the very system that those giants will likely depend on in the years to come.
Basiran
Azzam Bilal Chamdy
