Space-Based Computing Race Intensifies with China’s On-Orbit Assembly

China has launched the first 12 satellites of its “Three-Body Computing Constellation,” marking a significant step toward establishing a space-based AI supercomputer network. The launch took place on May 14, 2025, using a Long March-2D rocket from the Jiuquan Satellite Launch Center. Led by Zhejiang Lab and ADA Space, the project aims to eventually deploy around 2,800 satellites with a combined computing power of 1,000 peta operations per second—equivalent to one quintillion operations per second—potentially rivaling Earth’s most powerful supercomputers.

How Does the Constellation Work?

The Three-Body Computing Constellation operates as a network of satellites equipped with AI models capable of processing data directly in orbit. Each satellite in the initial batch features an AI system with eight billion parameters, eliminating the need to transmit raw data back to Earth for analysis. This approach reduces latency, avoids terrestrial bottlenecks, and leverages space’s natural cooling properties—sidestepping the massive energy and water demands of ground-based supercomputers.

Wang Jian, director of Zhejiang Lab, emphasized that this initiative expands AI integration into space, calling it a “new frontier for technological development.” The satellites also use solar power and radiate heat directly into space, potentially lowering the carbon footprint of high-performance computing.

Comparing Space and Earth-Based Supercomputers

When fully deployed, the constellation’s computing capacity (1,000 peta operations per second) would compete with top-tier supercomputers like El Capitan at Lawrence Livermore National Laboratory, which currently achieves 1.72 peta operations per second. However, the initial 12-satellite batch offers a more modest 5 peta operations per second combined, along with 30 terabytes of storage.

Zhejiang Lab plans to accelerate deployment, aiming for more than 50 satellites by the 202 2025. The long-term vision involves scaling to 2,800 interconnected satellites, creating a distributed supercomputing platform in orbit.

Practical Applications

The constellation is designed for real-time, in-orbit processing, enabling applications such as:

• Digital twins of Earth’s terrain for disaster response, providing instant mapping without Earth-based processing delays.
• Gamma-ray burst detection via an onboard X-ray polarization detector developed by Guangxi University and the Chinese Academy of Sciences.
• Climate monitoring, communications, and navigation, with potential military uses not yet detailed.

Wang Jian stated the project will “greatly expand the boundaries of space applications,” signaling broader implications for Earth observation and scientific research.

Global Context: Who Else Is Competing?

China’s initiative places it ahead in the emerging field of space-based computing. Jonathan McDowell, a Harvard astronomer and space historian, noted that orbital cloud computing is “very fashionable” but added that the U.S. and Europe may follow with similar efforts. While other nations have explored the concept, China’s large-scale deployment—with thousands of satellites planned—suggests a strategic push to dominate this technological domain.

The constellation’s name, “Three-Body,” nods to Liu Cixin’s sci-fi novel The Three-Body Problem, which explores advanced extraterrestrial computing—an apt metaphor for this real-world leap into orbital supercomputing.

For now, China’s progress signals a new phase in the space race—one where computing power, not just exploration, takes center stage.