China has launched the first 12 satellites in a bold plan to build a 2,800-satellite orbital supercomputer network, according to Space News.

The satellites are part of the “Star Compute” program by ADA Space, in collaboration with Zhijiang Laboratory and Neijang High-Tech Zone.

Unlike traditional satellites that depend on ground stations, these satellites are equipped to process data onboard using advanced AI models. Each satellite carries an 8-billion parameter AI model capable of performing 744 tera operations per second (TOPS), collectively reaching 5 peta operations per second (POPS). ADA Space’s long-term goal is to scale this to 1,000 POPS across thousands of satellites.


These AI-powered satellites can communicate via high-speed laser links of up to 100Gbps and share 30 terabytes of storage. Scientific instruments onboard include an X-ray polarization detector for capturing cosmic events like gamma-ray bursts. The satellites can also generate 3D digital twin data for applications in emergency response, tourism, and gaming.

Experts note that processing data in orbit offers significant advantages. According to the South China Morning Post, less than 10% of satellite data currently reaches Earth due to bandwidth and infrastructure limits. Harvard astronomer Jonathan McDowell highlighted that space-based data centers can efficiently use solar power and naturally radiate heat into space, reducing energy consumption and carbon emissions.

READ
Nvidia, Humain to Build AI Factories in Saudi Arabia