Technology
Space-Based Supercomputing: China’s £1.6 Billion Leap
- China intends to launch 2,800 satellites for its space-based supercomputer, with a projected processing power of 1,000 POPS.
- The expected cost surpasses ¥15 billion (£1.6 billion), making it one of the most expensive commercial orbital AI projects to date.
Data powers your daily life. It runs through your phone, your car, and your city. Usually, that data is processed in servers on the ground. Now, China is working to shift some of that processing to space.
In May 2025, China launched the first 12 satellites of its “Three-Body Computing Constellation�. This is the beginning of a much larger system planned by ADA Space, in partnership with and Neijang High-Tech Zone. Over the coming years, they aim to deploy 2,800 satellites. These will work together as a supercomputer in orbit.
This network is designed to take on complex computing tasks. The satellites don’t just collect data—they also analyse it right there in space. It’s a major step forward for global tech infrastructure.
What These Satellites Can Do
The full Three-Body Computing Constellation is expected to cost over ¥15 billion (£1.6 billion) according to statements from ADA Space and media reports. Each satellite is estimated to cost between ¥5 million and ¥8 million to manufacture and launch, depending on its configuration.
Each satellite is equipped with an AI model containing 8 billion parameters. It can handle 744 trillion operations per second. The satellites talk to each other using laser links that can reach speeds of 100 gigabits per second. Together, the first 12 satellites already offer a processing speed of 5 peta operations per second.
The plan is to eventually reach 1,000 peta operations per second. That’s one quintillion calculations every second—more than most of today’s Earth-based supercomputers.
This network is not just powerful. It’s fast, solar-powered, and built to work in a clean environment above Earth’s atmosphere.
Why Space-Based Supercomputing Matters
Traditional data centres are large, expensive, and dependent on energy grids. They require cooling systems and stable internet connections. The more we use digital services, the more strain we place on these centres.
Processing data in orbit helps reduce this load. Since the satellites are solar-powered and operate in a cold vacuum, they can stay active without the same energy needs as ground-based centres. They also cut down on transmission delays. Data is analysed closer to where it’s collected—sometimes instantly.
This means faster insights. Depending on how fast weather data, images from space, and climate readings are processed, the faster the disaster alert systems will be or, in fact, any scientific research.
From Ground to Orbit: Shifting the Model
We’re used to seeing satellites send data down to Earth, where it’s stored and analysed. But the volume of data being collected today is massive. Satellites observe everything from ocean patterns to city heat zones.
Sending all of that to the ground takes time and bandwidth. It also adds pressure to communication networks. With orbital AI infrastructure, the processing happens up there, reducing the need to send raw data down.
Think of this system as a filter in the sky—cleaning, sorting, and analysing information before it ever reaches us. That allows for faster responses and more efficient use of digital tools.
Practical Use Cases You’ll See First
There are already clear ways this technology could benefit daily life. As a real-time city and regional 3D map, one can contemplate an ever-changing twin meeting the essence of real-time updates, using satellite data for further updates; hence, better decision-making by city planners and emergency teams.
Another application is in transportation. Shipping, rail, and distribution networks rely on precise, up-to-date data. Space-based computing can assist with route planning and traffic prediction.
Scientific research also stands to benefit. Telescopes in orbit can transmit data directly to the supercomputer. That translates to less waiting and more real-time discovery. From solar activity to climatic patterns, the options are numerous.
Who’s Behind It?
The company leading the project is ADA Space. This firm focuses on commercial aerospace technology and is now working on one of the most ambitious AI projects to date.
It’s working closely with Zhejiang Lab and Neijang High-Tech Zone to develop the software and manage the satellite deployments. Together, these teams want to make space-based computing a core part of China’s national AI and data strategy.
The constellation’s name—Three-Body Computing—comes from a popular science fiction book, The Three-Body Problem. The name also reflects the complexity of orbital systems and the challenge of coordinating many moving parts in space.
Risks and Challenges
There are concerns, of course. One is space traffic. Thousands of satellites increase the risk of collisions. It takes careful planning to avoid disruption.
Another is cybersecurity. Like any network, space-based systems can be targets for interference. Developers will need to build strong protections into every layer of the system.
Then there are questions around data privacy. When satellites process personal or sensitive information in orbit, who controls the data? Clear rules and international agreements will be needed to manage access and responsibility.
What Other Countries Are Doing
China’s move into orbital AI has sparked global interest. In the US, NASA is looking at ways to integrate AI with Earth observation. In the UK, organisations like Satellite Applications Catapult are testing in-orbit servicing and AI modelling.
Europe is also taking steps. The IRIS² initiative is focused on building secure satellite networks that could, in the future, include computing functions.
These parallel efforts show that space computing is not a fringe idea. It’s becoming part of mainstream national strategies.
From Sci-Fi to Policy
Space computing used to be a fantasy. Today, it’s happening. And that means new policy questions.
Who owns the data that’s processed in space? Can countries share orbital computing infrastructure? What kind of oversight is needed to manage these systems?
Governments and industry leaders will need to work together. As the technology spreads, there will be growing pressure to set standards.
The Business Side
China’s space economy was valued at over ¥1.4 trillion (£150 billion) in 2024, with commercial satellite services accounting for more than ¥200 billion (£21 billion). Analysts project that the space-based AI computing segment could reach a market value of ¥100 billion (£10.5 billion) by 2030, driven by demand from logistics, climate modelling, and national defence contracts.
While pricing details haven’t been made public, analysts expect space-based computing to be offered as a service. Governments might use it for surveillance or disaster response. Private companies could use it to power large-scale AI models.
Interest is already coming in from logistics firms, financial services, and research groups.
This could open a new market, not just for China but for any nation or company willing to invest in orbital AI infrastructure.
Space-Based Computing and the UK
There’s an opportunity here for the UK.
British companies could join global efforts to manage satellite data standards. Universities might contribute to research or develop AI models suited for orbital conditions.
There’s also room for collaboration. UK aerospace firms and software developers could supply components or support services to international missions.
The key is to stay involved. As more satellites go up, the space above Earth will become part of your digital ecosystem.
What You Can Watch Next
In the coming months, you’ll likely see more launches. ADA Space is expanding fast, and new satellite batches are expected soon.
You might also hear about partnerships with foreign tech firms or early test results from data processed in space.
This is a fast-moving story, and the next 18 months will be important in deciding how widely space computing is adopted.
Long-Term Impact
By 2030, space-based computing could be as common as cloud storage is today. Many apps and services you use could draw on data that’s been processed above the Earth.
It’s not a replacement for existing systems. But it’s a strong compliment. It offers speed, reach, and resilience in a time when data demands are only increasing.
This is the next layer of the internet. And it’s orbiting overhead.
