Edge Computing Mapped for Earth Observation
A new analysis of edge computing for Earth Observation outlines the emerging landscape for onboard satellite data processing. The deep dive covers AI implementation paths and a survey of over 30 companies, highlighting future directions like foundation models running in orbital data centers.
The push for onboard processing is driven by the sheer volume of data generated by modern Earth Observation satellites, with a single satellite capable of capturing over 100 terabytes of imagery daily. Transmitting this volume of raw data to Earth is increasingly unsustainable due to bandwidth and latency constraints, making in-orbit data filtering and analysis a mission-critical necessity. This shift reduces the data downlink burden by up to 80%, accelerating the delivery of actionable insights from hours or days to just minutes. Standardization bodies are actively working to create the framework for this new orbital infrastructure. The 3rd Generation Partnership Project (3GPP) is integrating Non-Terrestrial Networks (NTNs) into 5G and future 6G standards, with Release 17 introducing the first specifications for satellite-to-device connectivity. The ITU-T is also developing recommendations for edge computing, including ITU-T Y.3223, which addresses the convergence of fixed, mobile, and satellite networks to support multi-access edge computing (MEC). The development of orbital data centers is becoming a focal point of geopolitical competition between the U.S. and China. China's ADA Space has already launched a 12-satellite network capable of running AI models in orbit and is planning a 2,800-satellite "Star Computing" project. In response, U.S. companies like StarCloud, backed by Nvidia, are also deploying GPU-powered satellites, framing the competition as a race to control the next layer of global digital infrastructure. This emerging market intersects with complex intellectual property dynamics, including Standard Essential Patents (SEPs) and FRAND (Fair, Reasonable, and Non-Discriminatory) licensing terms. Patent pools, which aggregate SEPs from multiple companies, are being encouraged by bodies like the European Commission to streamline licensing for complex technologies like IoT and wireless communications, reducing transaction costs and providing greater transparency on aggregate royalties. This model is critical for the satellite communications sector, where interoperability is key. Regulatory frameworks are also evolving to govern this new domain. The European Union's proposed Digital Networks Act aims to create a single, EU-level authorization framework for satellite communication services, simplifying cross-border operations and spectrum access. Concurrently, regulations like the GDPR extend to satellite-collected data, defining location data as personal data and imposing compliance obligations on operators who process information on EU citizens, a critical consideration as image resolution improves. The ultimate vision is a hybrid network where terrestrial and non-terrestrial systems are seamlessly integrated. This requires not only technical standards for interoperability but also new governance models. The rise of orbital data centers raises questions of data sovereignty and security, as these platforms could provide on-orbit threat detection and extend a nation's data governance beyond terrestrial borders.