Rama Afullo, a Kenyan American engineer with a background in Google’s cloud computing division and a brief tenure at SpaceX’s Starlink, harbored a conviction that was initially met with skepticism: the immense untapped potential of computing power in orbit. This foresight, he claims, was repeatedly dismissed by both Google and SpaceX when he attempted to champion the concept internally. Undeterred, Afullo departed SpaceX to co-found Satlyt, a company headquartered in Sunnyvale, California, and Nairobi, dedicated to developing software that enables multiple satellites to collaboratively handle significant computational tasks in space. Satlyt has now secured an $8 million seed funding round, an exclusive revelation to TechCrunch, signaling a significant validation of Afullo’s long-held vision.
The timing of Satlyt’s emergence appears opportune. SpaceX, which once rejected Afullo’s proposals, has since embraced the concept of orbital data centers with significant investment. Coincidentally, Satlyt’s software is slated for deployment on Thursday aboard a SpaceX rocket, coinciding with the launch of Google’s Project Suncatcher, its own ambitious space data center initiative. This convergence of industry giants and emerging startups underscores a burgeoning trend towards leveraging the unique environment of space for advanced computation.
A key differentiator for Satlyt lies in its strategic approach. Unlike competitors such as SpaceX, Google, or other emerging players like Starcloud and Cowboy Space Company, Satlyt is not focused on manufacturing its own spacecraft. Instead, Afullo’s team is concentrating on building the software infrastructure necessary to operate sophisticated AI models directly on satellites. He draws a parallel between Satlyt’s mission and the foundational work of companies like VMWare and Snowflake, whose platforms abstract complex software operations from the underlying hardware, allowing users to focus on functionality. "If the folks like SpaceX who are doing orbital data centers are the iPhone, we’ll build Android as a horizontally integrated, open ecosystem," Afullo stated, articulating Satlyt’s ambition to provide a universal software solution compatible across a diverse range of satellite platforms.
The Genesis of Orbital Computing
The concept of orbital data centers, or space-based computing, is rooted in the growing demands of space-based operations and the inherent limitations of terrestrial infrastructure. As the number of satellites in orbit rapidly increases, driven by commercial interests in communication, Earth observation, and scientific research, the need for on-board processing capabilities becomes paramount. Traditional satellite operations rely heavily on ground control for data transmission, analysis, and decision-making. However, this model presents significant challenges: the cost and latency associated with downlinking vast amounts of data from space are substantial, and the time delay can be critical in time-sensitive applications.
The idea of offloading computational tasks to space itself is not entirely new, but its practical implementation has been hindered by technological hurdles and economic viability concerns. Early forays into space computing were often limited to specialized, high-cost missions. The advent of more powerful, miniaturized processors, coupled with advancements in artificial intelligence and machine learning, has now made the prospect of widespread orbital computing more tangible.
Rama Afullo’s journey reflects this evolution. His experience at Google, a leader in cloud infrastructure, likely provided him with deep insights into the architecture and scalability of distributed computing. His subsequent involvement with Starlink, SpaceX’s ambitious satellite internet constellation, exposed him to the operational realities of large-scale satellite deployments and the immense data flows involved. It was during this period that Afullo’s conviction solidified: the future of computing would extend beyond Earth’s atmosphere, and a software layer was crucial to unlock this potential.
Satlyt’s Software: Enabling Intelligent Satellites
Satlyt’s core offering is its software platform designed to facilitate the deployment and execution of AI models on satellites. This approach offers several immediate benefits, particularly in enhancing operational efficiencies for spacecraft. Currently, satellites often require constant intervention from ground controllers to address anomalies or perform complex analyses. By enabling on-board decision-making, Satlyt’s software can reduce reliance on expensive and time-consuming data downlinks.
This capability translates into significant cost savings. The expense of transmitting data from orbit, known as downlink, can accumulate rapidly, especially for missions generating high volumes of sensor data. By processing this data locally on the satellite, or by using AI to compress it before transmission, Satlyt’s solution can drastically reduce downlink costs. Furthermore, the ability to resolve issues in real-time, without waiting for ground control intervention, can prevent mission-critical failures and extend the operational lifespan of satellites.
Satlyt has already demonstrated the efficacy of its technology through two successful demonstration missions. A notable early achievement was the deployment of Google DeepMind’s Gemma AI model on a spacecraft operated by Momentus earlier this year. This integration resulted in a remarkable reduction of over 60% in the size of transmissions detailing onboard software errors. Afullo estimates that such efficiencies could translate into annual savings of hundreds of thousands of dollars per satellite.
Expanding Capabilities and Market Validation
The upcoming launch on Thursday marks a significant milestone for Satlyt. Its software will be operational on a spacecraft built by TakeMe2Space, an Indian startup specializing in satellite computing hardware. This mission is a multi-faceted endeavor, involving three distinct customers, each with unique objectives:
- NASA: The space agency is funding the test of protocols for cloud computing in space, indicating a growing interest in distributed computational architectures beyond traditional terrestrial cloud services. This aligns with NASA’s broader objectives of fostering innovation and efficiency in space exploration and operations.
- Stellerian: This startup, focused on space surveillance, aims to leverage Satlyt’s software for image processing workloads. The ability to perform complex analysis of Earth imagery directly on orbit could enable faster threat detection, enhanced environmental monitoring, and more responsive intelligence gathering.
- TakeMe2Space: The satellite hardware provider intends to showcase its platform’s capability to host and run third-party software, demonstrating the versatility and open ecosystem approach that Satlyt advocates.
The Future of Orbital Computing: A Distributed Cloud
While the current landscape of orbital computing is characterized by a scarcity of high-powered Graphics Processing Units (GPUs) in orbit, Satlyt’s immediate focus on operational efficiencies and AI model deployment remains a highly valuable proposition. However, the company harbors a more ambitious long-term goal: the creation of a distributed computing system, or a true cloud, that spans multiple satellites.
Satlyt plans to undertake a pilot project next year to demonstrate its ability to create a shared computing environment across two different satellites. If successful, this would position Satlyt as a third-party provider capable of offering a comprehensive orbital compute cloud to spacecraft builders. This vision envisions satellites not just as data collectors, but as integral nodes in a distributed computational network, generating revenue through managed services. "The way to think about it is that we turn your satellite into a revenue-generating managed service," Afullo explained.
Investment and Strategic Partnerships
The recent $8 million seed round, led by Houston-based Non Sibi Ventures, provides crucial capital and strategic backing for Satlyt’s ambitious plans. Bernard Harris, a partner at Non Sibi Ventures and a veteran NASA astronaut with over two decades of experience, brings invaluable domain expertise and a deep understanding of the space industry’s challenges and opportunities. His involvement lends significant credibility to Satlyt’s vision.
Kent Lucas, another partner at Non Sibi Ventures, highlighted the pragmatic investment thesis behind their backing of Satlyt. He emphasized that the company’s success is not contingent on the immediate realization of full-scale orbital data centers, which face economic uncertainties and launch bottlenecks. Instead, Non Sibi Ventures sees a robust market driven by the sheer volume of satellites being launched. "We don’t need data centers in space for Rama to be wildly successful, right?" Lucas remarked to TechCrunch. "It can just be driven by the number of satellites going up." This perspective underscores the versatility of Satlyt’s software, capable of delivering value even in the current, less developed orbital computing infrastructure.
A Vision for the Next Decade
Rama Afullo has set an ambitious target for Satlyt: to have its software deployed on 20% of all satellites by the end of the decade. This projection is underpinned by his expectation that virtually all spacecraft builders will incorporate GPUs and other advanced processors into their satellites in the coming years. He views this trend as an inevitable progression, stating, "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?"
The proliferation of powerful processing hardware in orbit will create a fertile ground for Satlyt’s software solutions. As satellites become more capable, the need for intelligent software to manage and leverage these capabilities will grow exponentially. Satlyt’s focus on providing a standardized, interoperable software layer positions it to become a central player in this evolving ecosystem.
The broader implications of Satlyt’s success extend beyond cost savings and operational efficiencies. A thriving orbital computing ecosystem could revolutionize scientific research, enabling more complex experiments and real-time data analysis in space. It could enhance global security through more sophisticated surveillance and threat detection capabilities. Furthermore, it could pave the way for entirely new industries and applications that are currently unimaginable, transforming space into a truly computational frontier. Satlyt’s journey, from a dismissed idea to a well-funded startup poised for significant launches, represents a compelling narrative of innovation and foresight in the rapidly expanding domain of space technology.
Satlyt is among dozens of other startups vetted by TechCrunch and appearing at Disrupt as part of our broader Battlefield competition happening October 13-15 in downtown San Francisco. Learn more here.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
About the Author
