BY:SpaceEyeNews.
China wants to transform satellites from simple data collectors into intelligent computers orbiting Earth. Shanghai Xingshu Tiansuan Space Technology has introduced a project that could eventually connect 1,000 spacecraft.
These China space computing satellites would process remote-sensing information and run AI models in orbit. They could then return selected findings instead of transmitting every raw file.
The concept could reduce delays and ease pressure on communication bandwidth. However, China has not completed the planned network. The current development represents an early stage of a much larger commercial ambition.
China Space Computing Satellites Enter the Spotlight
Shanghai Xingshu Tiansuan presented the project during the 2026 World Artificial Intelligence Conference. The event ran from July 17 to July 20 in Shanghai.
Reuters reported that the company said it had launched the project’s first constellation. However, Chinese institutional reports used wording closer to “unveiled” or “officially released.”
Those reports describe the proposed architecture and its expansion stages. Yet they do not clearly identify a launch date, rocket, or orbital catalogue entry for this particular cluster.
Therefore, the safest conclusion is that Shanghai Xingshu Tiansuan has publicly introduced the first configuration. Further confirmation will clarify its exact orbital status.
Fudan University supports the project’s underlying space-based AI technology. Together, the participants want to create a global intelligent computing network above Earth.
This is different from a standard satellite internet constellation. The system’s main purpose involves processing information rather than only relaying communications.
Milestone: China Successfully Launches the First Satellites of 1,000-Satellite Space Supercomputer!
Inside the First Orbital Cluster
The first configuration uses what developers call a “one primary, two supporting” structure. It includes one central computing satellite and two specialized companion spacecraft.
A Computing Satellite as the “Space Brain”
The central platform acts as the cluster’s “space brain.” It would receive information from both companion satellites and combine several data formats.
A distributed scheduling platform would determine where each task should run. Onboard AI could then analyse observations before sending useful results to Earth.
Developers say the central satellite will carry a domestically produced space-grade GPU. It will also test remote-sensing models and an intelligent health-monitoring system.
High-efficiency solar cells would generate electricity. The design also includes experimental perovskite tandem power technology.
Intensive computing produces heat, even in orbit. A microchannel liquid-cooling module would help control processor temperatures. Laser communication equipment would connect the cluster and selected ground facilities.
Weather and Earth Observation
The Fuxi weather satellite forms one supporting element. Its microwave radiometer would measure atmospheric temperature, humidity, and cloud-water distribution.
Those measurements could enter the Fuxi forecasting model directly. This process may shorten the journey from atmospheric observation to a useful forecast.
The other companion is an ultra-low-orbit Earth-observation satellite. It carries visible-light and infrared instruments.
Its reported two-POPS edge-computing capability would support onboard image analysis. Potential applications include monitoring farmland, forests, cities, coastlines, and environmental changes.
How Orbital AI Processing Works
Traditional Earth-observation systems often transmit large raw datasets before extensive analysis begins. That process depends on available ground stations and communication capacity.
The proposed China space computing satellites would change that sequence. A spacecraft could collect an image, examine it with AI, and identify important features onboard.
Laser links could then move information between different computing nodes. The network would divide workloads according to available resources.
Finally, the constellation could send a smaller, focused result to Earth. Ground teams would not need every original file immediately.
This approach could reduce bandwidth pressure. It might also deliver time-sensitive information faster.
For example, an observation platform could examine changing cloud patterns before transmitting its findings. Another satellite could identify unusual changes across crops, forests, or coastal areas.
Ground-based facilities would remain essential. They would manage control, long-term storage, verification, and more demanding analysis. Space computing would complement those facilities rather than replace them.
The Road to 1,000 Satellites
The published roadmap divides development into three stages. Each phase would expand the network’s capacity and test more demanding services.
The engineering-validation stage calls for two computing satellites and 12 edge-computing satellites. This phase would examine AI inference, power management, cooling, laser links, and distributed scheduling.
Next, commercial deployment would introduce 50 computing satellites and 100 edge-processing spacecraft. Primary nodes would handle larger workloads. Edge platforms would analyse information near its source.
During commercial operation, the network could grow to approximately 1,000 satellites. That scale could provide wider coverage, shorter revisit times, and greater processing power.
Yet the final target remains a plan. Reaching it would require repeated launches and reliable spacecraft production. The operators would also need financing, regulatory coordination, replacement strategies, and paying customers.
China has other orbital-computing programs as well. The Zhejiang Lab-led Three-Body Computing Constellation follows a separate development path. The projects share broad ideas but involve different organizations and architectures.
Why Space Computing Matters
Orbital processing could prove especially valuable when satellites collect more information than they can quickly transmit. AI could prioritize the most relevant observations first.
Weather services might receive focused updates sooner. Agricultural groups could identify changing crop conditions across large regions.
Environmental organizations might monitor shorelines, forests, water systems, and urban expansion more efficiently. Faster analysis could also support assessments after major natural events.
Still, the technology faces significant engineering tests. Powerful processors need electricity and cooling. Radiation can affect electronics and stored information.
Laser links also require precise alignment between rapidly moving spacecraft. Clouds may disrupt optical connections with terrestrial stations.
Cost presents another major question. Companies must prove that faster results or unique coverage justify building and operating advanced orbital computers.
China Space Computing Satellites Face Their Real Test
The China space computing satellites project presents a compelling new model for orbital infrastructure. It combines remote sensing, AI analysis, workload sharing, and laser communication.
Its greatest promise lies in returning useful knowledge instead of only raw observations. That could make selected satellite services faster and more efficient.
However, the 1,000-spacecraft network remains a long-term objective. Verified launches, reliable onboard performance, successful laser connections, and commercial agreements will reveal its true potential.
If those elements come together, China space computing satellites could become an important extension of terrestrial AI infrastructure.
Main Sources:
Reuters – Shanghai Xingshu launches first constellation of its space-based computing project
ScienceNet – Star Hub Plan first constellation officially presented
Shanghai Songjiang District Government – Star Hub Plan and its 1,000-satellite objective
The News International – China launches first of 1,000 space-computing satellites