Hundreds of kilometers above Earth, a constellation of low-Earth-orbit (LEO) satellites is revolutionizing maritime operations by calculating fish movements for fishermen. Rather than flooding ground stations with raw imagery, these satellites process data in orbit, transmitting only actionable insights about fish populations. This shift—from passive imaging to intelligent data processing—is powered by a groundbreaking innovation: the space computing cloud.

Recently, the world’s first space computing cloud, spearheaded by Beijing University of Posts and Telecommunications (BUPT), transitioned from experimental trials to full-scale operational services. By offering satellite computing resources as a cloud-based solution, it eliminates the need for organizations to deploy their own satellites. Instead, users can deploy algorithms to orbit with a single click, transforming satellites from mere "eyes in the sky" into autonomous "brains" capable of real-time decision-making.

01. Why Move Computing Power to Space?

Traditionally, satellites acted as data relays: capturing images and transmitting them to Earth for processing. This model, however, faces critical bottlenecks. Ground stations, constrained by limited bandwidth and reception windows, struggle to handle the deluge of orbital data. "Marine fish migration patterns hold immense economic value, but conventional satellite systems can’t process them fast enough for time-sensitive applications," explains Xing Ruolin, founder of Yiwei Aerospace, a space operating system developer. The result? Valuable data languishes in orbit, unutilized.

The solution? Take computing power to the data. The space computing cloud integrates three platforms—in-orbit service, ground station management, and operational support—to automate task orchestration, resource allocation, and service delivery across satellites, ground stations, and data centers. Users interact with the system as if accessing a remote sky-based computer, submitting tasks and receiving results seamlessly. Professor Wang Shangguang of BUPT highlights its dual role: "It complements terrestrial computing for Earth-based users while serving as an independent resource for deep-space exploration."

This isn’t theoretical. BUPT’s "Cosmic Computation Constellation," launched in 2021 as the first of its kind, now operates 16 LEO satellites and seven ground stations. Over 60 organizations have conducted orbital experiments, with participation from researchers in 50+ countries. China’s space computing paradigm has shifted from isolated satellite tests to a scalable ecosystem, prioritizing operational efficiency over mere technical feasibility.

02. From "Eyes" to "Brains": A Paradigm Shift

The operational impact is transformative. Tested applications include orbital big data analytics, 6G communication protocols, and distributed storage, with over 100 users served and 10% market coverage achieved. For remote sensing, the change is revolutionary: satellites now process images mid-orbit, conducting disaster assessments and transmitting only actionable conclusions to Earth. This slashes response times during emergencies, where every minute counts.

Consider AI inference efficiency: the system achieves 10 tokens per joule, a remarkable feat given space’s harsh constraints—limited power, payload capacity, and radiation exposure. "This efficiency proves that constrained computing can support lightweight, long-duration space missions," says Wang, positioning China at the forefront of global in-orbit computing research.

The broader implications are profound. Space computing unlocks new avenues for satellite internet, 6G networks, and disaster communications. During ground station outages, orbital nodes double as communication relays and computing hubs, ensuring uninterrupted service. This sustainability model is expanding satellite systems’ service boundaries and applications.

The goals are threefold: build a space operating system ecosystem, transition from one-off experiments to reusable platform services, and enable coordinated satellite payload scheduling for real-time orbital processing. Satellites are evolving from isolated tools to interconnected smart nodes, while the spatial information industry shifts from hardware sales to ongoing service delivery. The gap between observation and intelligence defines the industry’s future.

03. Deploying an "Operating System" in Orbit

The true challenge lies not in launching satellites but in making them as accessible as cloud services on Earth. Central to this is an autonomous, open-source space operating system that lets users deploy programs instantly, tapping into satellite resources without reinventing infrastructure. Developers can now focus on coding, not satellite engineering—a democratization reminiscent of the internet’s early days.

The ecosystem is rapidly maturing. In June, Beijing’s Satellite Town inaugurated the Space Computing Innovation Center, unveiled China’s first open-source space OS standards, and launched a beta version of space computing services. These milestones—"platform construction," "standardization," and "service rollout"—occurred nearly simultaneously. "Commercial aerospace firms are building silos; we aim to consolidate them into a shared infrastructure," says Xing, reflecting a industry-wide push to commercialize space computing.

Beijing’s aerospace sector supports this vision, with 300+ companies and a RMB 100 billion+ industrial scale. From 2025 to Q1 2024, domestic commercial aerospace funding reached RMB 21.97 billion, with Beijing contributing RMB 12.15 billion. Amid rocket launches and satellite deployments, "computing power as a service" is emerging as a lucrative new frontier.

04. A Global Race with Equal Footing

Space computing is a global race, and China is neck-and-neck with leaders like the U.S., asserts Wang. "The regularization of space cloud services is a strategic milestone." As humanity ventures deeper into space—to the Moon, Mars, and beyond—computing power will lead the way.

Return to the fishermen. They don’t care about satellite names or technical jargon; they care about faster, more accurate fish data. When computing migrates to orbit, satellites transform from passive cameras to dynamic workstations, ready to serve humanity’s expanding frontier.

The path to "a space cloud for all" is long, but the door is now ajar.

- End -