China aims to deploy 1,000 general-purpose computing satellites by 2032, enabling AI to reach every corner of the world like water and electricity, according to a Chinese expert.
Space computing now is one of the hottest topics in AI industry around the world.
A project put forward by Wang Jian, director of Zhejiang Lab, demonstrates the direction that China is heading towards.
"I often say that the future satellites will be general-purpose satellites. By applying AI models to satellites, we're making the satellites far more versatile. This will be a dramatic structural change. That's also why AI is so remarkable today -- it has evolved from single-task models to general foundation models," said Wang.
In his view, when the number of computing satellites reach a sufficient scale, they will no longer be just individual technological products, but a type of infrastructure.
"Why do we often talk about infrastructure? It's easier to understand when compare it to water or electricity. First, everyone needs it. Second, its coverage is wide enough. Third, it is really cheap enough. Only when you have scale can you serve more people; and only when you serve more people can you have the chance to make it cheaper," Wang said.
China to equip satellites with AI for general application: expert
The unique operating mechanism of Zhejiang Lab, characterized by extensive consultation, joint contribution, and shared benefits, has successfully attracted private investment to fuel its space computing development, according to Wang Jian, director of the lab.
Zhejiang Lab is a research institution in the space computing domain based in Hangzhou, east China's Zhejiang Province. It focuses on frontier technology and represents a formidable force.
The 2026 World Artificial Intelligence Conference has just concluded late July in Shanghai, with space computing power emerging as one of the hottest topics. How can artificial intelligence be sent from the ground into space?
Wang, also an academician of the Chinese Academy of Engineering, has proposed the "Three-Body Computing Constellation" project, with the goal of deploying 1,000 general-purpose computing satellites by 2032, making AI accessible to every corner of the globe as ubiquitously as water and electricity.
On May 14, 2025, the first batch of 12 computing satellites of the "Three-Body Computing Constellation" successfully entered orbit. However, back in 2023 when Wang first proposed the concept of a computing constellation, no one in the international community had publicly raised such an idea.
The computing constellation was named "Three-Body." Wang explained that it was not a reference to writer Liu Cixin's sci-fi novel, but rather to Newton's "three-body problem", where the addition of a third celestial body makes precise calculation impossible. From its very inception, the project was destined to be something no single person could accomplish alone.
Zhejiang Lab is supported by Zhejiang Province, yet launching over 1,000 satellites would require at least 20 to 30 billion yuan (around 2.9 to 4.4 billion US dollars). Instead of seeking government financial support, Wang chose to pioneer a unique "extensive consultation, joint contribution, shared benefits, and common development" model at the lab, binding together a group of private enterprises in what appeared to be a relatively "loose" partnership.
"In the [space] industry, the biggest fear is that everyone becomes too loosely organized. That is why the command structure is very strict, isn't it? If you talk nonsense, no one will follow you. That's why we emphasize extensive consultation, joint contribution, and shared benefits. I would say consultation is the most important. People often want to skip consultation and jump straight to contribution and sharing, some even want to skip everything and go directly to sharing. Without proper consultation, there should be no contribution; without proper contribution, there should be no talk of sharing. You know it's difficult, but if you're willing to go all in, there's still a chance to make it happen. But if you're hesitant, if you're opportunistic, thinking you'll jump in only if it seems feasible and profitable, then no one will follow you," said Wang.
Unique operating mechanism of Zhejiang Lab attracts private investment for space computing development
