China has launched 12 satellites, which experts describe as the world’s first operational space-based computing network, applying edge computing principles to orbital operations in a development that could reshape how enterprises manage global data. The satellites, launched by Guoxing Aerospace of China, are equipped with AI systems, advanced inter-satellite communication capabilities, and onboard computing power. The network, formally named the “Three-Body Computing Constellation” but also referred to as the “Star Computing Constellation 021” mission, represents China’s push to create computing infrastructure beyond Earth that could transform data processing while potentially reducing environmental impacts, Guoxing Aerospace said in a statement. The constellation is a project by China to build a network of 2,800 satellites to empower real-time in-orbit computing and data processing. “China’s orbital AI constellation is more than a technological feat—it’s a proof of concept for distributed processing, autonomy at the edge, and context-driven compute as core tenets of modern architecture,” said Sanchit Vir Gogia, chief analyst and CEO at Greyhound Research. “What makes this constellation distinctive is not just its scale, but its shift in control logic: inference and orchestration happen in orbit, across a high-speed inter-satellite mesh, without needing constant cloud fallback.” Edge computing in space For enterprise decision-makers, the constellation represents perhaps the most ambitious application yet of edge computing principles — processing data directly where it’s generated rather than sending everything to centralized facilities. This orbital implementation showcases how these principles can address even the most extreme bandwidth constraints. “China’s Three-Body Computing Constellation marks a radical evolution in edge computing — demonstrating a ‘hyper-edge’ model: autonomous, localized processing under extreme latency and bandwidth constraints,” explained Deepti Sekhri, practice director at Everest Group. “This leap forces enterprise edge strategies to move beyond basic edge nodes toward resilient, AI-infused micro-infrastructures. We can expect the biggest impact in industries like manufacturing, defense, and logistics, where decisions must happen instantly and locally.” Traditional satellites face a significant data transmission bottleneck, with a significant amount of data loss during transmission to Earth due to bandwidth constraints. This parallels challenges many enterprises encounter in remote operations with bandwidth-intensive applications. “We are now entering a post-centralisation era — where compute is pulled toward the edge not by ideology, but by necessity,” noted Gogia. “Whether it’s orbital satellites, smart grids, or in-field robotics, AI workloads are becoming heavier, more inference-driven, and intolerant to network-induced drag. Centralised clouds won’t disappear — but for many classes of applications, they will no longer be the first stop.” “As we enter an AI-native era shaped by data-heavy, latency-sensitive, and bandwidth-limited environments, distributed architectures are increasingly becoming relevant,” added Sekhri. “The appeal of processing data closer to where it’s generated is gaining ground fast, and space-based compute reinforces this directional shift.” Technical specifications showcase distributed potential As per Guoxing Aerospace, each of the 12 satellites contains specialized computing hardware capable of handling up to 744 trillion operations per second. With 12 satellites working together, the array delivers a combined computing power of 5 peta operations per second (POPS) — equivalent to 5 quadrillion calculations per second. To put this in perspective, when fully implemented, the constellation could reach 1,000 POPS of processing capacity—potentially surpassing Earth’s most powerful systems, according to the Chinese government. The El Capitan supercomputer at Lawrence Livermore National Laboratory in California, ranked as the world’s most powerful last year, achieves approximately 1.72 POPS! To function as a unified processing network, the constellation employs laser communication technology, achieving data transfer speeds up to 100 gigabits per second, comparable to some of the fastest terrestrial fiber optic networks, the statement added. The constellation also incorporates 30 terabytes of on-board storage and runs an AI model with 8 billion parameters, demonstrating how sophisticated artificial intelligence applications can be deployed at the network edge. For context, this means the system can run relatively sophisticated artificial intelligence applications similar to some large language models, though smaller than the most advanced AI systems used on Earth today. Environmental and economic advantages The space-based distributed approach also addresses growing environmental concerns about data centers — a priority for enterprises with sustainability targets. The International Energy Agency projects that global data centers could consume more than 945 terawatt hours of electricity annually by 2030—roughly equivalent to Japan’s entire electricity usage. “The economics of enterprise compute are undergoing a structural inversion — and China’s Three-Body Computing Constellation makes that undeniable,” observed Gogia. “As AI workloads are increasingly executed at the point of collection — in orbit or on Earth — the cost of centralisation becomes a liability.” By processing data closer to collection points, organizations can potentially reduce the energy footprint associated with moving massive datasets across global networks. Water consumption presents another environmental challenge for traditional computing facilities that distributed approaches can help address, as major technology companies use billions of liters of water annually to cool their data centers. Implications for global operations For multinational enterprises, the constellation demonstrates how distributed processing systems might eventually support truly global operations. Chinese officials have positioned this initiative as establishing globally accessible, mobile, and low-carbon space-based infrastructure. “This development presents not just an engineering milestone, but a geopolitical one — casting data governance into uncharted territory,” warned Gogia. “Orbital compute is redefining the boundaries of data sovereignty. The Three-Body system decentralises not only AI processing but also geopolitical accountability — placing inference and decision-making infrastructure into orbits that may not fall cleanly under existing legal regimes.” “However, as compute infrastructure stretches beyond sovereign borders, enterprises face a growing ambiguity around governance and jurisdiction,” cautioned Sekhri. “Enterprises will need to evolve their risk frameworks and operational policies to address a new class of sovereignty and control challenges.” It is expected that similar networks will likely be deployed by multiple nations in the coming years, potentially creating new infrastructure that enterprises could leverage for global operations, particularly in remote areas where traditional connectivity remains challenging. “Perhaps the more pressing question isn’t how far edge computing can go — but who ultimately orchestrates it,” Sekhri pointed out. “As the compute fabric fragments across clouds, geographies, and orbits, control may shift from centralized platforms to a more federated, contested, and geopolitically charged landscape. That’s a future worth preparing for.”