China launches 'Supercomputing-1' AI satellite as orbital data centers gather momentum

China has launched nine satellites — including its first integrated “rocket and satellite” AI computing project — as part of its push to put AI computing in orbit. According to a Digitimes report, the Kinetica 1 Y18 rocket, operated by Chinese commercial launch provider CAS Space, lifted off on September 20, carrying the nine satellites to their planned orbits.

The payload of greatest interest is the “Supercomputing-1” satellite, developed by Chinese AI computing infrastructure company S-AIDC. Also known as the S-AIDC-1, the satellite is designed to capture and process Earth-observation data in orbit instead of sending data back to terrestrial data centers. Keeping the processing local is meant to cut cross-regional data processing times from hours to minutes. The satellite features both a high-res optical payload and the image-processing AI computer.

According to Digitimes, the launch highlights China's efforts to develop orbital computing as part of a “broader integrated computing network.” In early June, the Chinese government approved the Space Computing Industry Innovation Center, which aims to bring together rocket and satellite manufacturers, semiconductor fabs, and AI tech companies to build a space computing network. A couple of factors are driving this spike in the development of off-planet computing, foremost among them being the AI boom.

The demand for artificial intelligence has spurred the construction of numerous data centers to house AI accelerators. These data centers, some of which can house hundreds of thousands of accelerators, consume unprecedented amounts of electricity, placing significant strain on the grid. This, along with land use, water consumption for cooling, and noise, has fueled growing anti-data sentiment, with local residents blocking $68 billion worth of new data center projects in Q2 2026.

Despite numerous economic and practical challenges, space might offer an alternative with unlimited area, access to solar power, constant cold conditions, and a dead-quiet vacuum. These advantages have made orbital computing a serious topic of discussion. SpaceX is leading the charge with the AI1 satellite, a proposed orbital data center that will deliver 150 kW of peak compute power. The company aims to achieve 1 GW per year of space AI compute by 2027 and has begun constructing the 11-million-square-foot Gigasat facility that would manufacture the 6,000 satellites required to meet this goal.

Beyond the large-scale serving of AI models, putting edge compute capacity in space is already useful for a range of applications. Earth-observation satellites can process images onboard to identify wildfires, storms, floods, ships, or other objects of interest, sending only the useful results back to Earth instead of dumping enormous amounts of raw data to ground stations. The same approach can speed up weather monitoring, disaster response, communications, and even the management of satellite constellations, where spacecraft could make more decisions independently rather than waiting for instructions from Earth.

The launch of a single satellite with edge AI capabilities on board is a far cry from an entire orbital data center, but China's approach could prove to be a more practical and immediate application for advanced computing capabilities in orbit.

Follow Tom's Hardware on Google News, oradd us as a preferred source, to get our latest news, analysis, & reviews in your feeds.