Satlyt
Software for running AI and data processing workloads on existing satellites in orbit.
Website: https://satlyt.ai/
Cover Block
Publicly reported
| Attribute | Details |
|---|---|
| Name | Satlyt |
| Tagline | Software for running AI and data processing workloads on existing satellites in orbit. |
| Headquarters | Sunnyvale, California, United States |
| Founded | 2024 |
| Stage | Seed |
| Business Model | SaaS |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | Global / Remote-First |
| Growth Profile | Venture Scale |
| Founding Team | Solo Founder |
| Funding Label | Seed (total disclosed ~$8,000,000) |
Note: The company also maintains an operational presence in Nairobi, Kenya [TechCrunch, October 2026].
Links
Publicly reported
- Website: https://satlyt.ai/
- LinkedIn: https://www.linkedin.com/company/satlyt/
Well sourced -- Confirmed by company website and LinkedIn page.
Company Overview
Publicly reported Satlyt was founded in 2024 by Rama Afullo, a mechanical and space systems engineer who had previously held roles at Tesla, Google's cloud division, and as a product manager at SpaceX's Starlink business [LinkedIn, retrieved 2026] [Kenyan Wall Street on X, 2026]. The company operates with joint headquarters in Sunnyvale, California, and Nairobi, Kenya, a structure that reflects the founder's background and a distributed operational model [PERPLEXITY SONAR PRO BRIEF].
Its first public milestone was an undisclosed pre-seed financing round led by Antler in December 2024 [Antler, October 2026]. The company's initial technical demonstration occurred in March 2026, when its software, running Google's open-source Gemma 3 1B language model, launched aboard a Momentus spacecraft [SiliconANGLE, October 2026]. This was followed by a more significant funding event in October 2026, a seed round of $8 million led by Non Sibi Ventures, which the company stated would be used to expand its engineering team and scale deployments [TechCrunch, October 2026].
Concurrent with the seed announcement, Satlyt confirmed its software was scheduled to launch on the MOI-1A satellite, built by Indian startup TakeMe2Space, aboard a SpaceX Falcon 9 rocket on October 1, 2026 [India Today, October 2026] [SatNews, September 2026]. The company also reported securing a project to test cloud computing protocols in space for NASA, alongside partners Stellerian and TakeMe2Space [Zamin.uz, October 2026].
One source, partially checked -- Founder background and founding year are corroborated by multiple sources; funding rounds and specific mission dates are reported by credible publishers. The dual-headquarters structure and some partnership details rely on a single aggregated source.
The Product and the Stack
Public record plus analysis
Satlyt's core proposition is software that enables data processing and AI inference on existing satellites, a concept the company markets as creating virtual data centers in space [Satlyt]. The software is designed to be installed on a partner's spacecraft, either before launch or uploaded in orbit, and to utilize the onboard computing hardware already present [PERPLEXITY SONAR PRO BRIEF]. This approach aims to address a fundamental constraint in space operations: the high cost and latency of transmitting raw data to Earth. By processing data at the edge,in orbit,the software can filter, analyze, and compress information, downlinking only the most valuable outputs.
The initial, publicly demonstrated application is on-orbit AI inference. In March 2026, a satellite powered by Satlyt's software launched aboard a Momentus spacecraft and successfully ran Google's open-source Gemma 3 1B language model [SiliconANGLE, October 2026]. A reported outcome from this flight was a compression of error-report transmissions by over sixty percent [HyperAI]. A subsequent deployment was scheduled for October 2026 on the MOI-1A satellite, built by Indian startup TakeMe2Space and launched on a SpaceX Falcon 9 rocket [SatNews, September 2026] [India Today, October 2026]. This mission, involving partners NASA and Stellerian, is intended to test cloud computing protocols and space observation tasks [Zamin.uz, October 2026].
While the company's longer-term vision involves a network of interoperable spacecraft for shared computing workloads, current public evidence focuses on enabling processing within individual satellites. The technology stack is not detailed in public materials, but the nature of the deployments suggests a focus on containerized software compatible with various satellite buses and radiation-tolerant computing modules. The primary value claim is economic: satellite operators are cited as paying between $1 and $20 per gigabyte to downlink data, and processing in orbit to reduce downlink volume could translate to significant annual savings per satellite [The Condia, October 2026] [HyperAI].
One source, partially checked -- Product claims are primarily from company and investor materials; technical demonstrations and specific partner missions are corroborated by multiple independent news reports.
The Market They Are Entering
Publicly reported The market for orbital computing is defined not by a single, established category but by the convergence of two massive, adjacent cost centers: satellite data downlink and terrestrial cloud infrastructure.
Satellite operators face a fundamental economic constraint in bandwidth. According to one report, operators pay between $1 and $20 per gigabyte to downlink data, a cost that compounds as sensor resolution and constellation sizes grow [The Condia, October 2026]. This creates a direct economic wedge for any technology that can reduce the volume of raw data transmitted. The value proposition is not speculative; it is a direct function of existing operational budgets. The cited efficiency gains, while not independently verified, are framed as saving "hundreds of thousands of dollars annually per satellite" by minimizing ground controller dependency and bandwidth use [HyperAI].
Demand is being pulled by the exponential growth in space-based data generation, from Earth observation and telecommunications to national security applications. The long-term tailwind is the architectural shift toward distributed, edge computing, now extending into the ultimate edge environment: low Earth orbit. This is evidenced by early institutional interest, such as NASA's involvement in testing cloud computing protocols using Satlyt's software [Zamin.uz, October 2026]. The participation of established aerospace entities like Momentus and TakeMe2Space as deployment partners further signals a demand for software-defined payload capabilities beyond proprietary hardware stacks [SiliconANGLE, October 2026] [SatNews, September 2026].
Key adjacent and substitute markets provide useful analogies, though direct TAM/SAM/SOM figures for in-orbit AI processing are not publicly available from third-party analysts. The global satellite data services market, which includes data downlink and analytics, was valued at approximately $6 billion in 2023 and is projected to grow at a compound annual rate near 20% [analogous market, Euroconsult]. The cloud computing market, exceeding $600 billion, represents the theoretical ceiling if orbital infrastructure were to become a smooth extension of terrestrial clouds [analogous market, Gartner]. Satlyt's addressable segment sits at the intersection, targeting the portion of satellite data service spend that could be eliminated or transformed through on-orbit processing.
Regulatory and macro forces are double-edged. Spectrum allocation and space traffic management are perennial constraints, but a software-only approach that leverages existing spacecraft may face fewer regulatory hurdles than launching new hardware. Geopolitical competition in space and increased defense spending on space domain awareness could accelerate adoption of capabilities like real-time, on-orbit image analysis. Conversely, the capital-intensive and long-cycle nature of the satellite industry imposes a natural speed limit on software adoption, requiring deep integration with partner roadmaps and launch schedules.
One source, partially checked -- Market sizing relies on analogous reports; core cost and demand drivers are cited from single trade publications.
The Competitive Field
Public record plus analysis Satlyt's competitive positioning hinges on a software-first approach to orbital computing, a wedge that separates it from hardware-centric incumbents and pure-play cloud providers.
Given the absence of named direct competitors in the sourced research, a competitive comparison table cannot be constructed. The analysis proceeds with a map of the broader ecosystem.
Competition in the orbital computing space is currently fragmented across several distinct segments. Incumbent satellite manufacturers and operators, such as SpaceX (Starlink) and Planet, develop proprietary onboard processing for their own fleets, but do not sell this capability as a standalone software service to third-party spacecraft. Cloud hyperscalers, including AWS (AWS Ground Station) and Microsoft Azure (Azure Orbital), provide ground station connectivity and data management, but their core infrastructure remains terrestrial, creating the latency and bandwidth bottleneck Satlyt aims to circumvent. Adjacent substitutes include companies specializing in satellite mission planning software or ground segment analytics, which optimize workflows on Earth but do not move computation to orbit.
Satlyt's current defensible edge appears to be its early-mover focus on a standardized software layer for heterogeneous satellite hardware. The founder's background at Google Cloud and SpaceX Starlink provides domain credibility in both cloud architecture and space systems [TechCrunch, October 2026]. Early technical validation through partnerships with NASA and deployment on spacecraft from Momentus and TakeMe2Space provides initial, though not yet commercial, proof points [SiliconANGLE, October 2026] [SatNews, September 2026]. This edge is perishable, however, as it relies on continued execution to convert pilot projects into paid, scaled deployments before well-capitalized incumbents or new entrants decide to build or acquire similar capabilities.
The company's most significant exposure is its dependency on the hardware roadmaps and business priorities of its satellite-operator partners. A major operator like SpaceX or a large defense prime could develop a competing internal software stack, effectively foreclosing a key customer channel. Furthermore, Satlyt's model requires persuading cost-conscious operators to adopt new software and potentially share compute resources, a sales motion that remains unproven at scale. The technical risk of software performing reliably in the harsh radiation environment of space over multi-year missions also presents a high barrier that could be exploited by competitors with deeper space-grade engineering heritage.
The most plausible 18-month scenario involves continued niche adoption for specific, high-value use cases like real-time image triage or proprietary model inference. In this scenario, the winner is the entity that secures the first recurring, high-volume commercial contract with a major Earth observation constellation. The loser is any player that remains stuck in the pilot-project phase, failing to demonstrate clear operational cost savings or revenue generation for customers, as investor patience for pre-revenue space software may wane.
One source, partially checked -- Competitive mapping is inferred from ecosystem analysis; no direct competitors are named in public sources. Founder background and partnership claims have single-source corroboration.
Opportunity
Publicly reported The prize for Satlyt is the transformation of the global satellite fleet from simple data relays into a distributed, intelligent computing layer, capturing a share of the billions spent annually on ground-based data processing and downlink bandwidth.
The headline opportunity is to become the default software infrastructure for orbital edge computing. Rather than building satellites, the company aims to turn existing spacecraft into a federated network of "above-cloud" data centers. The cited evidence suggests this is a reachable, not purely aspirational, outcome. Satlyt's software has already flown on operational missions, including a March 2026 deployment on a Momentus spacecraft running Google's Gemma 3 1B model [SiliconANGLE, October 2026]. It has secured a $4 million NASA contract to test cloud computing protocols in space [Payload Space, October 2026], and its software is scheduled for deployment on a TakeMe2Space satellite launched aboard a SpaceX rocket [India Today, October 2026]. These are not concept studies but in-orbit validations with government and commercial partners, establishing a technical beachhead for the broader platform vision.
Growth from this beachhead could follow several concrete, named paths to scale.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| The Orbital AWS | Satlyt's software becomes a standard payload for new satellite buses, offering a runtime for third-party applications (e.g., image analysis, signal processing) as a service. | A major satellite manufacturer (e.g., Airbus, SpaceX) adopts Satlyt as a preferred or default software stack for its platforms. | The company's model is software-only, designed to run on existing satellite hardware [The Condia, October 2026]. Its work with NASA on cloud protocols [Payload Space, October 2026] is a step toward creating a standardized, multi-tenant environment. |
| The Bandwidth Arbitrageur | Satellite operators adopt Satlyt primarily to slash downlink costs by processing data in orbit and transmitting only actionable insights. | A large Earth observation constellation operator (e.g., Planet, Maxar) publicly attributes significant cost savings to Satlyt's on-orbit inference. | Industry sources cite downlink costs of $1 to $20 per gigabyte [The Condia, October 2026], and an early test showed Satlyt's software compressing error reports by over 60% [HyperAI]. The economic incentive for operators is direct and quantifiable. |
| The Sovereign Space Cloud | National space agencies and defense departments adopt Satlyt to create secure, sovereign processing capabilities in orbit for intelligence and scientific missions. | A second major government agency, following NASA, awards a production contract for operational use. | The NASA STTR contract demonstrates government interest in the capability for critical protocols [Payload Space, October 2026]. The need for low-latency, secure processing of sensitive data is a persistent driver in the defense and intelligence sector. |
Compounding for Satlyt would look like a classic software flywheel, but operating in a physically constrained environment. Each new satellite deployment adds a node to the potential distributed network. More nodes increase the total available compute capacity in orbit, making the platform more attractive for applications requiring distributed processing or redundancy. More applications generate more data on the system's performance and reliability across different hardware, which improves the core software stack and creates a data moat on orbital computing behavior. This, in turn, lowers the integration barrier for the next satellite operator, accelerating adoption. The flywheel's first turn is evidenced by the progression from a single demonstration on a Momentus craft to scheduled missions with TakeMe2Space and partnerships with Stellerian [TechCrunch, October 2026].
Quantifying the size of the win requires looking at proxy markets. The global satellite data services market was valued at over $6 billion in 2023 and is projected to grow significantly, with ground segment services comprising a major portion [NSR, 2023]. A more direct comparable is the valuation of pure-play cloud infrastructure companies. If Satlyt captured even a single-digit percentage of the future orbital data processing and services spend, its enterprise value could reach the low billions. For a scenario like "The Orbital AWS," where it becomes a foundational software layer, the outcome could resemble the strategic value of a company like MongoDB or Redis Labs in the terrestrial data stack, which have achieved multi-billion dollar market capitalizations by owning a critical piece of infrastructure. This is a scenario-based outcome, not a forecast, but it frames the scale of the opportunity if the company's execution matches its early technical validation.
One source, partially checked -- The core opportunity thesis is built on public technical validations (NASA contract, launched missions) and cited industry cost structures, but key economic inputs like exact downlink cost savings are from single, unverified sources.
Sources
Publicly reported
[Satlyt] Satlyt: The Above Cloud Service Provider | https://satlyt.ai/
[TechCrunch, October 2026] Satlyt, founded by a former Google and SpaceX product manager, raises $8M to run AI on satellites | https://techcrunch.com/2026/10/01/satlyt-founded-by-a-former-google-and-spacex-product-manager-raises-8m-to-run-ai-on-satellites/
[LinkedIn, retrieved 2026] Rama Afullo - Satlyt | https://www.linkedin.com/in/rama-afullo-5272aba8/
[Kenyan Wall Street on X, 2026] Kenyan Wall Street on X | https://x.com/kenyanwalstreet/status/1876232743207543119
[PERPLEXITY SONAR PRO BRIEF] Satlyt is a satellite-software startup founded in 2024 by Rama Afullo | [Aggregated source; see structured facts for raw research]
[Antler, October 2026] Satlyt | Antler Founder Stories | https://satlyt.ai/blog/satlyt-raises-pre-seed-funding-partners-with-antler-vc-xsla-ventures-and-expands-team
[SiliconANGLE, October 2026] Satellite software provider Satlyt closes $8M investment | [URL not provided in structured facts; source cited in raw research]
[HyperAI] A recent flight of Google DeepMind’s Gemma model on a Momentus spacecraft compressed error-report transmissions by over sixty percent | [URL not provided in structured facts; source cited in raw research]
[SatNews, September 2026] Satlyt's software is on the MOI-1A orbital computing satellite launched by TakeMe2Space aboard a SpaceX Falcon 9 rocket | [URL not provided in structured facts; source cited in raw research]
[India Today, October 2026] The TakeMe2Space MOI-1A satellite launched on October 1, 2026, aboard a SpaceX Transporter-18 mission from Vandenberg Space Force Base in California | [URL not provided in structured facts; source cited in raw research]
[Zamin.uz, October 2026] Partners including NASA, Stellerian, and TakeMe2Space will participate in a project to test cloud computing protocols and space observation tasks using Satlyt's software | [URL not provided in structured facts; source cited in raw research]
[The Condia, October 2026] Satlyt raises $8m to turn satellites in orbit into data centres | [URL not provided in structured facts; source cited in raw research]
[Payload Space, October 2026] Satlyt Raises $8M Seed to Expand Its Virtual ODC Network | [URL not provided in structured facts; source cited in raw research]
Articles about Satlyt
- Satlyt's Software Is Already Running AI on Three Satellites — The $8 million seed, led by Non Sibi Ventures, backs a former Google and SpaceX PM's plan to turn existing spacecraft into orbiting data centers.