This is no longer science fiction — it’s the dawn of a new tech race in space. Chinese company ADA Space has announced the world’s first orbital AI supercomputing network. The project, titled Star Compute and Three-Body Computing Constellation, is a direct nod to China’s famous sci-fi novel.
Last week, the first 12 of 2,800 planned satellites were launched, each equipped with an AI model containing over 8 billion parameters. Their processing power reaches 744 TOPS per satellite, with the entire network already achieving up to 5 POPS (peta-operations per second) — 100 times more powerful than a standard Microsoft Copilot-PC.
The main innovation: full orbital autonomy. The satellites communicate via laser at 100 Gbps, process data directly in space, and share a distributed 30 TB buffer. This isn’t just satellites — it’s the first distributed cloud above the atmosphere.
Applications range from disaster modeling and 3D digital twins for tourism and gaming to gamma-ray burst analysis using X-ray polarization detectors.
A major highlight is ecological efficiency: solar energy use and direct heat dissipation into space reduce carbon footprint and solve cooling issues common in Earth-based data centers.
Experts at Harvard call this a new chapter in computing history. While Western nations test AI in the cloud, China is literally building it above our heads.
