US EditionPublished • loading... • Updated
China Races to Build 2,800-Satellite AI Network in Orbit
The plan would use 2,400 inference satellites and 400 training satellites, with commercial operations targeted by 2030, company executives said.
- ADAspace Technology Co., Ltd. plans to build a 'star compute' network of 2,800 AI computing satellites to form a global training and inference computing system by 2035, with initial satellites launched in May 2025 and further launches planned for 2026.
- The 'star compute' network will provide inference computing power at the hundred-thousand-petaflop level and training power at the million-petaflop level, aiming to begin commercial operations by 2030, primarily using inference satellites, and will utilize satellite-to-ground and inter-satellite laser communications to enable real-time AI operations even in network blind zones.
- Zhejiang Lab's 'Three-Body Computing Constellation' project has launched 12 AI computing satellites to process data in orbit and plans to expand to 100 satellites by 2027 to improve in-orbit AI data processing.
Insights by Ground AI
10 Articles
10 Articles
Global Times+3 Reposted by 3 other sources
From ground to orbit: China eyes computing in space
In a futuristic office resembling a space station, Zhao Hongjie, executive vice president of Adaspace Technology Co., Ltd. (ADAspace), outlined a grand vision: to bring intelligent services to every corner of the Earth.
·Beijing, China
Read Full Article
Chinese research institutes and technology companies are accelerating the development of orbital computing infrastructure. The project's strategic goal is to shift some of the workloads associated with data processing and neural network operations directly to low-Earth orbit. A key initiative in this area is the "Star Compute" program, implemented by Adaspace Technology. As explained by the corporation's Executive Vice President, Zhao Hongjie, t…
Coverage Details
Total News Sources10
Leaning Left2Leaning Right1Center2Last UpdatedBias Distribution40% Left, 40% Center
Bias Distribution
- 40% of the sources lean Left, 40% of the sources are Center
40% Center
L 40%
C 40%
R 20%
Factuality
To view factuality data please Upgrade to Premium
