In a groundbreaking advancement for space technology, Chinese commercial aerospace pioneer GuoXing Aerospace Technology has successfully deployed a general-purpose artificial intelligence model aboard its operational satellite constellation. The Chengdu-based startup announced this week that it has uplinked Alibaba’s Qwen3 large language model to its inaugural space-based computing center, enabling complete end-to-end reasoning tasks to be processed entirely in orbit.
Executive Vice-President Wang Yabo confirmed this represents the world’s first deployment of a general-purpose large-scale AI model from terrestrial control systems to an operational satellite network in space. The achievement follows China’s launch last May of 12 space computing satellites, forming the initial cluster of GuoXing Aerospace’s ambitious orbital computing project.
During recent trials, the Qwen3 model successfully completed multiple experiments with remarkable efficiency. Questions transmitted from Earth to the satellite were processed on-board, with results returned to ground stations within an impressive two-minute timeframe.
This development emerges as artificial intelligence continues driving unprecedented demand for computing resources, sparking a new technological frontier that pushes intelligent computing capabilities into space. The movement gained momentum when SpaceX recently orbited its Starcloud-1 satellite equipped with Nvidia GPUs.
Looking toward the future, Wang outlined the company’s visionary plan to construct an extensive network of 2,800 specialized computing satellites by 2035. This proposed constellation will comprise 2,400 inference satellites and 400 training satellites, deployed across sun-synchronous, dawn-dusk and low-inclination orbits at altitudes ranging from 500 to 1,000 kilometers.
The sophisticated satellite network will employ laser inter-satellite links to facilitate high-speed data transfer, with ambitious targets to deliver 100,000 petaflops of inference compute capacity and 1 million petaflops of training compute power worldwide. The company anticipates deploying its second and third satellite clusters within the year, with plans to complete a 1,000-satellite network by 2030.









