Satlyt Raises $8M Seed For Satellite AI Software

The space industry faces a massive data bottleneck. Satellites orbiting the planet collect enormous amounts of raw information every single day. Sending that information back to Earth requires expensive ground station connections. Operators frequently pay up to $20 for every gigabyte they download. By the time the files finally reach a server on the ground, the heavy majority of the data is redundant or completely useless. A new software company wants to fix this broken pipeline by moving the actual processing power directly into orbit.

Satlyt just secured an $8M seed funding round to build exactly that infrastructure. Non Sibi Ventures directed the financial transaction, supplying the capital required to scale the software engineering team. The company operates out of Sunnyvale and Nairobi, completely avoiding the traditional path of building custom physical spacecraft. Instead of manufacturing expensive metal boxes, the team focuses strictly on writing the code that operates inside them. The goal is to turn existing satellites into active data centers capable of running artificial intelligence models without any human intervention.

Rama Afullo launched the company after spending time inside two of the most demanding engineering environments on the planet. He previously managed products for Google and worked briefly on the Starlink communication network at SpaceX during 2024. During his time at these massive corporations, Afullo realized that building vertically integrated hardware was no longer the most urgent problem in space. The true barrier involved coordination. Thousands of machines currently orbit the Earth, but they operate completely isolated from one another. We tracked the extreme capital demands of launching isolated hardware when Loft Orbital and Marlan Space planned $1B for AI satellites earlier this year. Afullo wants to avoid those massive hardware costs entirely.

Processing Data In Orbit

The main product acts as a shared control plane for space operations. When a satellite takes a photograph of a weather pattern, it normally sends the entire raw image file down to Earth for analysis. The new software allows the satellite to analyze the image instantly using onboard processors. The machine identifies the exact weather pattern, deletes the unnecessary background pixels, and transmits only the final text conclusion to the human operator. This method reduces the required download capacity by massive percentages.

The company recently proved this capability using compressed language models. The engineering team deployed Google Gemma directly onto orbital hardware. They built an automated diagnostic tool that monitors the health of the spacecraft. If a sensor fails or a battery overheats, the software identifies the exact fault and attempts to fix it instantly. Google DeepMind recently published a case study highlighting how this exact application reduced diagnostic payload sizes by nearly 64 percent. By shrinking the amount of data that must return to Earth, the startup saves satellite operators millions of dollars in transmission fees. This push toward remote automated reasoning mirrors the structural logic we observed when Helogen raised $7.1M in seed funding to test in-orbit medical manufacturing. Operators want machines that think for themselves.

Building The Space Network

Convincing competing manufacturers to install third party software requires intense technical diplomacy. The startup pitches its product as a universal translation layer. Afullo frequently compares the software to the early days of internet protocols, the communication standard that allowed completely different computer networks to form the modern web. He wants to apply that exact same connective logic to orbit. If a machine built by one company can speak directly to a machine built by a rival company, they can share computing resources instantly.

This strategy is already yielding formal partnerships. Satlyt recently signed an agreement with Symphony Space and xOrbita to test modular infrastructure. Symphony Space designs physical stations that host and release free flying pods. When a pod detaches from the main station, it requires immediate instructions regarding navigation and docking procedures. The new software provides the communication layer that tells the pod exactly where to go and records the logical reasoning behind the decision. Securing early testing partners is an absolute requirement for unproven space technology. We documented similar early stage testing agreements when Kepler Aerospace secured $8M for autonomous military satellites.

Securing Federal Support

Federal space agencies are actively funding these exact communication tools. Satlyt recently confirmed it secured $4M in support contracts across multiple NASA centers. The federal government desperately wants to reduce its reliance on slow ground communication stations. If a deep space probe encounters an unexpected obstacle, it cannot wait twenty minutes for a human operator on Earth to send a new directional command. The probe must analyze the danger and adjust its course immediately. The software models designed by the Sunnyvale startup provide exactly that type of immediate logical processing.

The federal contracts also connect the startup to the broader defense sector. Military agencies want orbital networks that can survive severe communication blackouts. If an attacker destroys a ground station, the satellites above must continue operating independently. We noted the severe consequences of broken physical networks when SpaceX issued a public apology following a major infrastructure outage. Satellites that rely entirely on Earth for instructions become useless the moment the connection drops. The startup solves this vulnerability by making the machine self sufficient.

The Economics of Software in Space

Raising an $8M seed round gives the company the financial runway required to finish its mathematical algorithms and secure space on upcoming rocket launches. The business model is highly attractive to early stage investors because it requires very little physical manufacturing. Building rockets and casting titanium parts burns through capital incredibly fast. Writing code and licensing it to satellite operators generates high profit margins with minimal physical overhead. The startup acts as an operating system vendor for a rapidly expanding hardware market.

The competition to control orbital data is intensifying. Massive telecommunications companies are pouring billions of dollars into their own proprietary networks. We tracked the extreme scale of legacy investments recently when Viasat and Space42 committed $1B to a direct to device venture. Satlyt must prove that its open software platform can outperform the closed systems built by these massive industry giants.

The successful closing of this funding round proves that financial markets recognize the immediate need for better orbital software. Space is no longer just a place for passive observation cameras. It is rapidly becoming an active computing environment. The startup has the cash, the federal partnerships, and the technical pedigree required to build the first true orbital data network. The next phase involves proving the code works flawlessly while travelling thousands of miles an hour through a freezing vacuum.

Read More on TechRobust:

• Loft Orbital and Marlan Space Plan 1B AI Satellites
• Helogen Raises 7.1M Seed Funding In Orbit Medical Manufacturing
• Viasat Space42 Commit 1B Direct To Device Venture