Satlyt, a startup founded to transform satellites into virtual data centers, successfully tested software that distributes computing tasks among orbiting machines, marking a key milestone in its vision to process data onboard and reduce reliance on Earth-based analysis.
- Satlyt software now runs on three orbiting satellites after a successful test this month
- The test simulated a network sharing computing tasks among multiple machines in orbit
- Future steps include broader validation, secure links, and sharing work across separate spacecraft
What happened
Satlyt completed a significant test of its software designed to transform spacecraft into virtual data centers. Working through a NASA-funded project with the University of Houston, the company demonstrated how its software can dynamically assign computing tasks among satellites based on their available resources and communication links. Although only one physical satellite was involved, the test included simulated computers and showed that plans for workload distribution can be generated onboard.
The experiment involved a simulated network with multiple users, services, and computers to assess workload allocation in orbit. This step validated that Satlyt's software can operate in space hardware conditions, handling task scheduling and prioritization autonomously. The startup's software is already deployed on three satellites, bringing their ambitious goal of in-space distributed computing closer to reality.
Why it matters
Satlyt’s technology addresses key constraints faced by current satellite operations such as limited onboard computing power, intermittent communication opportunities with Earth, and the need to get relevant data back to users faster. Processing data closer to the source reduces dependency on ground stations, decreases latency, and accelerates decision-making—critical for applications like wildfire detection and disaster response.
By enabling satellites to share computing resources and intelligently distribute workloads, Satlyt aims to create a scalable and secure network of virtual AI data centers in orbit. This approach promises to make space missions more efficient and reduce costs linked to downlink bandwidth and human analysis. Ultimately, it can pave the way for new space applications that require real-time processing and lower reliance on Earth.
What to watch next
Satlyt plans further testing to validate its software under diverse workloads, hardware configurations, and connectivity scenarios. A key next step is demonstrating workload sharing across multiple physical spacecraft rather than simulations. The company also needs to establish secure communication protocols, operator permissions, robust scheduling, and failover mechanisms to handle spacecraft dropping from the network.
Longer term, Satlyt aims for developer ecosystems to build AI and other applications on their platform without reinventing infrastructure for each mission. As the technology matures, it could unlock new possibilities for satellite constellations and multi-mission cooperative processing in low Earth orbit and beyond.