When Rama Afullo looks up at the night sky, he doesn’t just see constellations; he sees a fragmented, underutilized network of hardware waiting for a unified operating system. As a veteran of both Google’s cloud computing division and SpaceX’s Starlink network, Afullo spent years witnessing the same technological bottleneck: the inability of satellites to process data autonomously.
Today, Afullo is the co-founder and CEO of Satlyt, a startup headquartered in Sunnyvale and Nairobi that has just secured $8 million in seed funding. His mission is audacious: to build the software layer that turns disparate satellites into a cohesive, high-performance computing grid. By providing an open, horizontally integrated ecosystem for orbital AI, Satlyt aims to do for space what VMware and Snowflake did for terrestrial data centers—abstracting away the hardware to let software developers focus on the mission.
The Genesis of an Orbital Vision
The story of Satlyt began as a series of frustrated internal pitches. During his tenure at SpaceX in 2024 and his previous years at Google, Afullo became convinced that the industry was missing a massive opportunity. As satellites grew more sophisticated, they were increasingly hindered by the "downlink bottleneck"—the slow, expensive, and latency-heavy process of sending massive amounts of raw sensor data back to Earth for analysis.
"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 tells TechCrunch.
Realizing that the internal roadmaps of tech giants were too rigid to accommodate his vision of a decentralized, multi-tenant orbital cloud, Afullo left to build it himself. He recognized that while companies like SpaceX and Google were racing to build proprietary "orbital data centers," the industry desperately needed an interoperable platform—an "Android" for space that could run on any hardware, rather than being locked into a single provider’s "iPhone-like" walled garden.
A Chronology of Orbital Innovation
The rise of Satlyt is characterized by a rapid transition from theory to flight-proven reality.
- 2024: Afullo departs from major tech firms to co-found Satlyt, focusing on the development of a software abstraction layer for satellite processing.
- Early 2024: Satlyt conducts its first series of demonstration missions, proving that AI models can be offloaded to spacecraft to handle anomaly detection.
- Mid-2024: The company achieves a milestone by deploying Google DeepMind’s Gemma AI model onto a Momentus-operated spacecraft. The experiment is a success, reducing the size of data transmissions regarding software errors by over 60%.
- Late 2024: Satlyt closes an $8 million seed round led by Non Sibi Ventures.
- The Present: Satlyt prepares for a pivotal launch alongside Google’s "Project Suncatcher" prototype, utilizing a TakeMe2Space platform to demonstrate multi-customer, cloud-based computing protocols.
The Architecture of Space-Based Intelligence
Satlyt’s value proposition lies in its ability to move computation from the ground to the edge. In traditional satellite operations, spacecraft are "dumb" sensors that transmit data to Earth, where ground-based flight controllers must process it, identify anomalies, and send commands back up. This process is time-consuming and costly.
By deploying AI models like Google DeepMind’s Gemma directly onto the satellite, Satlyt allows the spacecraft to perform its own "on-board triage." By processing images and sensor readings in orbit, the satellite only needs to transmit the most relevant, actionable insights back to Earth. This operational efficiency is not just a technical convenience; it is a massive cost-saver, potentially saving operators hundreds of thousands of dollars per satellite annually.
The current mission, launching this week, tests the limits of this architecture. Working with TakeMe2Space, Satlyt is managing workloads for three distinct entities:
- NASA: Testing new protocols for cloud computing in orbit.
- Stellerian: Utilizing the hardware for space surveillance and image processing.
- TakeMe2Space: Demonstrating the viability of hosting third-party software on its proprietary computing hardware.
Economic Implications: Beyond the Data Center
One of the most compelling aspects of Satlyt’s business model is its refusal to build its own physical satellites. While competitors like Starcloud and Cowboy Space Company are attempting to manufacture the massive, power-hungry orbital data centers of the future, Satlyt is betting on the ubiquity of processors.
Kent Lucas, a partner at Non Sibi Ventures, notes that the investment was driven by the realization that Satlyt doesn’t need a future of massive, stationary orbital data centers to be successful. "We don’t need data centers in space for Rama to be wildly successful," Lucas explains. "It can just be driven by the number of satellites going up."
By focusing on the software layer, Satlyt makes itself indispensable to the entire space economy. If Afullo’s prediction holds—that within the decade, every commercial satellite will carry GPUs or advanced AI processors—then Satlyt positions itself as the standard operating system for the next generation of space infrastructure.
Official Responses and Strategic Outlook
The investment from Non Sibi Ventures, led by former NASA astronaut Bernard Harris, adds a layer of deep technical validation to the startup. Having a 20-year veteran of space operations backing the firm suggests that the industry is ready for a shift toward "managed services" in orbit.
Afullo’s goal is to be live on 20% of all commercial satellites by 2030. He maintains a firm stance on the necessity of this shift: "If you’re putting up a satellite without putting up a GPU on it, at the very least, you’re doing yourself a disservice."
The company is now setting its sights on its next major hurdle: inter-satellite cloud computing. By next year, Satlyt plans to demonstrate a shared computing system that spans two different satellites, effectively creating a distributed cloud network in orbit. If successful, this will mark the first time multiple spacecraft have functioned as a single, unified computing resource, fundamentally changing the economics of space missions.
The Future of the "Space Cloud"
As the launch window approaches, the stakes for Satlyt are high, but the potential rewards are tectonic. By commoditizing space computing, Afullo is turning satellites from static, mission-specific assets into dynamic, revenue-generating managed services.
The transition from "spacecraft as a tool" to "spacecraft as a node in a global network" is the next frontier of the digital economy. While the hardware builders provide the physical reach into the stars, Satlyt is providing the intelligence that makes those stars accessible, efficient, and, ultimately, profitable. As the company prepares to showcase its technology at the TechCrunch Disrupt Battlefield this week, the industry will be watching to see if this "Android of space" can indeed unlock the latent computing power currently orbiting above our heads.
