China launches satellites to start building the world’s first supercomputer in orbit

China launches satellites to start building the world’s first supercomputer in orbit

China has launched the first batch of satellites for its space computing constellation, a system that could rival the most powerful ground-based supercomputers once fully deployed.

Twelve satellites, each equipped with intelligent computing systems and inter-satellite communication links, were sent into orbit aboard a Long March 2D rocket from the Jiuquan Satellite Launch Centre around noon on Wednesday, according to state-owned Guangming Daily.

They are part of the Three-Body Computing Constellation, space-based infrastructure being developed by Zhejiang Lab. Once complete, the constellation would support real-time, in-orbit data processing with a total computing capacity of 1,000 peta operations per second (POPS) – or one quintillion operations per second – the report said.

By comparison, the new El Capitan system at the Lawrence Livermore National Laboratory in California, which was the world’s most powerful supercomputer last year, achieves over 1.72 peta operations – or 1.72 quintillion operations per second.

Jonathan McDowell, a space historian and astronomer at Harvard University, said the idea of cloud computing in space was “very fashionable” right now.

“Orbital data centres can use solar power and radiate their heat to space, reducing the energy needs and carbon footprint,” he said.

China, the United States and Europe could be expected to deploy such orbital data centres in the future, McDowell said.

“Today’s Chinese launch is the first substantial flight test of the networking part of this concept.”

Data centres around the globe could consume more than 1,000 terawatt hours of electricity annually by 2026 – roughly equivalent to the entire electricity use of Japan – according to estimates by the International Energy Agency.

Cooling these facilities also requires vast amounts of water. In 2022 alone, Google used 19.7 billion litres (5.2 billion gallons) to cool its data centres.

Against this backdrop of rising resource costs, the limits of Earth-based data processing are becoming apparent.

Traditionally, satellites collect data in space but must send it back to Earth for processing – a method constrained by limited ground station availability and bandwidth. As a result, less than 10 per cent of the collected data typically makes it back to Earth, often with significant delays.