Space becomes the new data center frontier as China launches its first 12 satellites for interstellar AI processing

Asia is becoming a hotbed for space-based AI data centers and computation networks as China leaps ahead with the launch of 12 new satellites that will create a supercomputer edge computing constellation of up to 2,800 linked satellites in space. However, they are not alone in chasing this opportunity – global interest is growing.

China has just launched a cloud of 12 AI-enabled satellites into orbit, which places it in a strong position for space-based computing capabilities. Each satellite – designed, built, and launched by Chinese company Guoxing Aerospace (also known as ADA Space) alongside AI specialists at Zhejiang Lab – will offer 744 tera operations per second (TOPS or 744 trillion operations per second) of computing power.

The so-called “Three-Body Computing Constellation” is a space-based infrastructure of supercomputers that would easily rival the most powerful ground-based supercomputers once fully deployed.

Creating space-based AI data centers

The constellation aims to harness unlimited solar power to power the satellites and to bypass the cooling effects demanded by earth-based data processing to perform astronomical observations. In this manner, it would rival the inefficiencies of advanced ground-based AI capabilities for applications such as super-fast edge computing, advanced AI processing, drone positioning, military and defense analytics, emergency situation control, and self-driving vehicles.

The constellation would enable data to be processed on the satellites rather than on the ground, thereby saving real estate, energy, and cooling requirements while reducing environmental impact. In addition, it could speed responses to emergency events that are impacted by land-based data bottlenecks. Moreover, the constellation could enhance the speed and efficiency of data processing without the need for expensive and unsustainable cooling methods for AI data processing centers.

This first dozen Chinese satellites, which were launched on board a Long March 2D rocket from the Jiuquan Satellite Launch Centre on May 14, 2025, will be connected through 100 Gbps laser inter-satellite links and run an 8-billion parameter AI model. They also provide remote sensing payloads – with data processed onboard to reduce space-to-earth data transmission requirements – and will offer 30 terabytes of onboard storage. 

A cosmic supercomputer

One of the satellites also carries a space X-ray polarimeter, which was developed by Guangxi University and the National Astronomical Observatories of the Chinese Academy of Sciences (NAOC). The polarimeter has the capacity to detect, record, and analyze cosmic entities such as black holes, neutron stars, and supernovas.

These first 12 satellites are the foundation of a planned constellation of 2,800 satellites, which will be built and launched over the coming years. This constellation, once operational, will reportedly provide a combined real-time processing power of 1,000 Peta operations per second (POPS) – equivalent to one quintillion (10¹⁸) operations per second. The fully functional constellation would put China in a leading position in orbital computer processing.

While there are many technical hurdles to overcome, such as radiation-proof equipment, and ensuring rapid in-orbit and space-to-earth communication, China’s launch confirms space as the new data processing frontier. China, though, is by no means the only actor aiming to create an AI hub in space.

For example, Florida-based Lonestar Data Holdings has proposed placing data centers on the moon, while Washington-based Starcloud plans to launch a satellite data center in 2025 with commercial operations expected to start in mid-2026.

Asia Pacific opportunities in space-based computation

Companies in the Asia-Pacific region are also getting into the act. For example, in April, Tokyo-based Space Compass Corporation performed in-orbit ship detection processing using an AI application on Microsoft Azure that reduced the volume of data that needed to be downlinked to earth by 98%. This project was undertaken as a part of the strategic alliance between NTT and Microsoft.

Other examples of the growing interest and emerging expertise in AI satellites are the Asian partnerships established by Rivada Space Networks – a Germany-based subsidiary of US Rivada Networks. Rivada recently announced partnerships with iO3 from Singapore, Acutec Systems from Australia, Tricom of India, Rikei Corporation from Japan, Taiwan’s Transcontinental Enterprise, and One Commerce from the Philippines. Rivada has launched what it terms the “Outernet” initiative, which aims to provide a secure space-based data network for governments and business, with the company securing more than $16 billion investment.

Other U.S companies, including Microsoft, IBM, and Space Compass, are also actively exploring the opportunities that Asia offers for space-based AI computation. To find out how we can help you minimize the risks and maximize the opportunities for working with such specialists in Asia contact us today.