China has taken a groundbreaking step in space technology by launching the first cluster of its planned AI satellite constellation, a project designed to process artificial intelligence computations directly in orbit. This initial deployment of 12 satellites marks the beginning of a 2,800-satellite fleet known as the Three-Body Computing Constellation. Developed by ADA Space and Zhejiang Lab, the constellation aims to dramatically reduce China’s reliance on Earth-based data centers by harnessing the unique advantages of outer space.
The satellites launched aboard a Long March 2D rocket from the Jiuquan Satellite Launch Center on May 14. Their mission is to create a network capable of edge computing—processing data in real time aboard the satellite—before transmitting results back to Earth. This reduces transmission delays, bandwidth constraints, and data loss that typically occur when raw satellite data is sent to ground stations. The AI satellite constellation is part of a larger vision to decentralize computing infrastructure and move data processing closer to the source of data collection.
Each satellite contains an 8 billion-parameter AI model capable of performing 744 tera operations per second (TOPS). Collectively, the 12-satellite array achieves a processing speed of 5 peta operations per second. In comparison, high-end consumer devices like Microsoft’s AI-enabled Copilot+ laptops process around 40 TOPS. Eventually, the full constellation is expected to reach a computing capacity of 1,000 peta operations per second—a staggering milestone in artificial intelligence processing power.
Beyond computing speed, the AI satellite constellation uses the cold vacuum of space as a natural cooling mechanism, which dramatically reduces energy consumption and environmental impact. Powered by solar panels, the satellites radiate excess heat into space, allowing for efficient operation without the carbon footprint of traditional terrestrial data centers. This energy model aligns with growing concerns about the rising energy demands of AI data centers worldwide, which are projected to require an additional 29 gigawatts of power by 2027 and up to 67 gigawatts by 2030.
Another key innovation is the use of laser-based inter-satellite communication. This enables high-speed, high-capacity data exchange between the orbiting satellites. One of the satellites is even equipped with an X-ray polarization detector to study cosmic events such as gamma-ray bursts, combining AI infrastructure with astrophysical research potential.
The constellation takes its name from the three-body problem in classical mechanics and pays homage to Liu Cixin’s acclaimed science-fiction trilogy, “The Three-Body Problem.” This cultural and scientific inspiration reflects the project’s complexity and ambition. According to Zhejiang Lab’s director Wang Jian, the initiative is also intended to foster global collaboration, with the eventual goal of making the constellation available to international organizations.
While the U.S. and Europe have tested space-based computing systems, China’s launch represents the first operable deployment at scale. The project also echoes emerging concepts in the West, including former Google CEO Eric Schmidt’s proposal to launch AI-powered data centers into orbit. As terrestrial energy demands surge, the AI satellite constellation provides a visionary alternative—blending sustainability, high performance, and frontier technology into a single, transformative system.