Company Spotlight: Satlyt Builds an Orbital AI Network

Satlyt is building a shared control plane for satellites: software that connects tasking, onboard compute, and data processing across spacecraft owned by different operators. Founder and CEO Rama Afullo and CTO Nelson Kigen Psenjen are pursuing a future where satellites process more information in orbit and coordinate like a network instead of behaving as isolated machines.

Founded in 2024 and operating across Sunnyvale, California, and Nairobi, Kenya, Satlyt enters the space economy from the software side. The company does not manufacture satellites or depend on a dedicated orbital data center. Its platform is designed to make the computing hardware already in space easier to program, monitor, and eventually share across missions.

That approach matters because the satellite market has become good at launching capable machines and much less consistent at making those machines work together. Satlyt’s opportunity sits between hardware owners, application developers, ground infrastructure, and customers who want answers without paying to move every raw file back to Earth.

About Satlyt

Rama Afullo founded Satlyt after roles spanning Tesla, Google’s cloud organization, and SpaceX’s Starlink business. TechCrunch reported that Afullo had proposed versions of the orbital-computing idea inside Google and SpaceX before leaving to build it independently. Nelson Kigen Psenjen, Satlyt’s CTO, brings a distributed-computing background that fits the company’s cross-spacecraft ambition.

Satlyt describes its product as a coordination layer for orbit. The Satlyt Network is intended to distribute data processing across satellites, run edge workloads near the source, and use inter-satellite communications to reduce delays created by routing every decision through the ground.

The distinction between current capability and destination is important. Satlyt can point to software operating onboard individual spacecraft and to a growing set of technical collaborations. A federation of satellites from different owners working as one programmable network is the next system to prove.

Onboard AI Is the First Wedge

Onboard processing gives Satlyt an immediate customer problem to solve before the full shared network exists. Satellites operate with constrained power, memory, thermal capacity, bandwidth, and communication windows. Moving analysis onto the spacecraft can help operators filter imagery, interpret telemetry, and diagnose faults before deciding what must be transmitted.

A Google DeepMind case study documented Satlyt running a quantized Gemma 3 1B model onboard a satellite. The model analyzed logs, software errors, and stack traces from image-processing workloads. In two representative fault-injection tests, it reduced diagnostic payloads by 64.4% and 64.8% while producing root-cause diagnoses and recommended actions.

Those results are narrow, useful evidence. They do not prove that every mission can cut downlink by the same amount, and they do not establish a multi-satellite cloud. They show that small models can produce actionable summaries under real spacecraft constraints, which gives Satlyt a credible entry point into broader orchestration.

Partnerships Turn the Thesis Into Hardware Access

A software company in space still needs places to run. Satlyt has been building that access through spacecraft operators, research programs, and hardware partnerships rather than financing its own constellation.

Satlyt has announced work with The Aerospace Corporation on DiskSat, a compact satellite architecture being explored for onboard computing and distributed coordination. It has also described demonstrations with Rogue Space Systems, a planned radiation-resilience experiment with Melagen Labs on the International Space Station, and a partnership with Star Catcher Industries around orbital energy and edge computing.

The company’s October 2026 deployment rode on a TakeMe2Space spacecraft carried by SpaceX’s Transporter-18 mission. The launch succeeded on October 1. Satlyt said its software package supported research connected to NASA-funded work and a commercial imagery-processing application. Launch success delivered the platform to orbit; operational results from the applications remain a separate milestone.

Leadership Spans Two Space Economies

Satlyt’s Sunnyvale and Nairobi footprint is more than a staffing detail. The company is selling into a global infrastructure market while building with a substantial African engineering presence and relationships that include the Kenya Space Agency and Angola’s GGPEN.

Afullo brings commercial experience across Tesla, Google, and SpaceX. Psenjen brings the technical depth required for distributed systems that must operate with delay, scarce resources, and limited intervention. Business Development Manager Sharon Njeri adds spacecraft-systems experience from NASA-supported programs and a market-facing role connecting the platform to operators and partners.

The company’s structure gives Satlyt access to Silicon Valley customers and capital while drawing technical and institutional relationships from African space programs. That can become a practical advantage if customers increasingly care about data sovereignty, neutral infrastructure, and technology that works across operators rather than locking every workload into one constellation.

What Satlyt’s Hiring Reveals

Satlyt’s current careers page lists three Bay Area roles: Network Engineer for Distributed and Space Systems, Senior Forward-Deployed Engineer, and Customer Success and Mission Deployment Lead. The mix says more than a generic hiring announcement.

A network engineer points to the cross-satellite coordination problem. A forward-deployed engineer suggests integration work close to customer missions and hardware. A deployment lead indicates that Satlyt is building the operating muscle required to move from research demonstrations to repeatable delivery. The company is not only hiring people to design the platform; it is hiring around the handoff from software into missions.

Satlyt describes its culture through innovation, collaboration, and ownership. Those values are common words on a careers page. The harder evidence is the shape of the roles, which sit where young infrastructure companies usually feel the most pressure: distributed systems, customer integration, and the operational work required to make advanced software survive contact with physical hardware.

Why Satlyt Matters Now

The orbital-computing market is attracting very large visions, from dedicated data centers to AI accelerators launched for specialized workloads. Satlyt is taking a narrower route that can create value earlier. It wants to improve the economics and usefulness of spacecraft already being launched, then connect more of that capacity over time.

The company’s recent $8M Seed round, led by non sibi ventures, gives that strategy more engineering and customer-delivery capacity. The capital follows visible technical work, but it also increases the burden of repeatability. Satlyt must support different spacecraft architectures, secure remote software delivery, preserve operator command authority, and prove that shared compute is worth the new trust relationships it requires.

Satlyt’s broader opportunity is not simply to put AI in space. It is to make orbital computing legible to operators and application developers as infrastructure they can use without owning a satellite. Each deployment adds another machine, customer, and operational boundary to that test. If the software can keep crossing those boundaries, the network begins to exist before Satlyt ever owns a spacecraft.