Data centers are moving into orbit: why China is taking computing into space
The number of satellites in orbit is growing rapidly, and so is the amount of data they generate. However, the capacity to transmit all of that information back to Earth remains limited. China is developing a new approach in which part of the data is processed directly in space using computing systems installed aboard satellites.
This is giving rise to orbital computing — essentially space-based data centers capable of analyzing images and other information with artificial intelligence before transmitting results back to Earth. China currently operates more than 640 civil and commercial remote-sensing satellites. They monitor weather, agriculture, oceans, natural disasters and infrastructure.
According to CCTV+, less than 5 percent of the data generated by these satellites can currently be transmitted to Earth because of limited communication bandwidth.
Under the traditional model, a satellite collects information, sends it to a ground station and only then does processing begin. Orbital computing allows analysis to take place immediately after the data is captured. For example, if a satellite detects a wildfire, an onboard AI model could identify the fire, determine its location and transmit an alert instead of sending large volumes of raw imagery.
One of China's key projects in this field is the Three-Body Computing Constellation. The first 12 computing satellites were launched in May 2025. They are equipped with onboard processors, inter-satellite communications and the ability to run AI models in orbit. The first group provides around 5 POPS of computing power and about 30 terabytes of storage.
In the future, the satellites are expected to operate as a distributed computing network, sharing tasks between spacecraft. The constellation is planned to expand to around 1,000 satellites by 2032.
Orbital computing could be used for disaster monitoring, agriculture, ocean observation, infrastructure management, aviation, shipping and emergency communications.
In August 2026, China also began providing experimental space-computing cloud services, allowing users on Earth to send computing tasks to orbital infrastructure and receive processed results.
Artificial intelligence is expected to play a central role. Instead of sending all raw data to Earth, AI systems can determine which information is important enough to process and transmit immediately.
The technology also faces major challenges. Electronics must operate under radiation, vacuum and extreme temperature conditions. Cooling, launch costs and high-speed communications between satellites remain difficult engineering problems.
For this reason, orbital computing is unlikely to replace terrestrial data centers in the near future. Instead, it is expected to complement them in applications where rapid processing of satellite data provides a clear advantage.
Preliminary estimates from the China Academy of Information and Communications Technology suggest China's space-computing industry could exceed 250 billion yuan by 2030.
According to CCTV+, China is moving from individual experiments toward building a broader orbital computing infrastructure that could eventually change the way satellite data is processed.
