China is moving to put artificial intelligence infrastructure in orbit, betting that space-based data centres can leapfrog terrestrial rivals and reshape the balance of power in the commercial space race. By tying solar power, cloud computing and advanced chips together above the atmosphere, Beijing is setting up a direct challenge to Elon Musk’s SpaceX vision of satellite constellations that dominate both connectivity and AI.
The plan is not a distant thought experiment. Chinese agencies and research institutes are sketching concrete constellations, power systems and timelines that stretch to 2030, positioning the country to compete head-on with private US players just as AI workloads explode and governments look for strategic control over the hardware that runs them.
China’s orbital cloud gambit
Beijing’s strategy starts with a simple premise: if AI is becoming the core of economic and military power, then the infrastructure that runs it cannot be left solely on the ground. Officials have signalled that China is planning a space-based cloud computing network by 2030, designed to process data in orbit and beam results back to Earth. The goal is to create a resilient, globally accessible platform that is less vulnerable to terrestrial outages and can serve both civilian and defence users.
State media have gone further, reporting that China plans to launch solar powered AI data centres into orbit over the next five years, effectively turning satellites into floating server farms. These platforms are meant to handle intensive machine learning tasks, from real time image recognition to predictive analytics, without sending raw data back through congested ground networks. For a government that already treats digital infrastructure as a strategic asset, extending that logic into space is a natural, if ambitious, next step.
Inside the Beijing space lab blueprint
The most detailed glimpse of this architecture comes from a research institute in the capital that has laid out a specific orbital design. According to one Beijing space lab, the plan calls for a constellation of 16 centralized space data centres flying in a dawn dusk orbit at 700 to 80, a configuration chosen to maximise exposure to sunlight and simplify power management. By clustering processing nodes in a coordinated orbit, engineers hope to create a virtual supercomputer that can be dynamically allocated to different users and tasks.
That same blueprint emphasises that the constellation is meant to meet future large scale computing demand, not just today’s satellite imagery workloads. In practice, that means designing for high bandwidth inter satellite links, radiation hardened chips and modular platforms that can be upgraded as AI models grow more complex. The focus on a dawn dusk orbit at 700 to 80 also underlines how central solar power is to the concept, since a near constant angle to the Sun reduces the need for heavy batteries and allows more mass to be devoted to processors and cooling systems.
Solar power as a strategic weapon
Energy is the hinge that connects these orbital ambitions to a broader geopolitical contest. Space based solar power has long been discussed as a way to beam clean electricity back to Earth, but Chinese planners are now treating it as a dual use technology that can also feed power hungry AI hardware in orbit. In this emerging landscape, Beijing and Elon are portrayed as direct competitors, each racing to control a new frontier where solar energy and high performance computing converge.
For China, the attraction is clear. By putting solar arrays and AI accelerators together in orbit, it can sidestep some of the land, water and grid constraints that plague terrestrial data centres, while also gaining a platform that is harder for rivals to disrupt. Reports on the country’s broader space strategy describe how China Eyes Space as part of a New Space Race with the United States, explicitly linking orbital power systems to the relocation of AI processing into space. In that framing, solar energy is not just a clean power source, it is a strategic enabler for a new class of military and commercial services.
SpaceX, xAI and the private sector response
On the other side of this contest sits SpaceX, which has already built the world’s largest satellite internet constellation and is now looking to fuse that network with AI. Elon Musk has floated concepts for solar powered AI satellites that would ride on top of the existing launch and manufacturing infrastructure, effectively turning Starlink into a distributed computing platform. That vision is now intersecting with corporate strategy, as NEW YORK reporting describes SpaceX and xAI in merger talks ahead of a planned public offering, a move that would more tightly integrate launch, satellites and AI research under Elon Musk’s control.
For investors and regulators, that potential merger underscores how central orbital infrastructure has become to the AI economy. A combined SpaceX xAI entity could use its rockets and satellite factories to deploy proprietary AI accelerators in space, mirroring the kind of vertically integrated approach that Chinese planners are pursuing through state backed programmes. The difference is that while Beijing leans on national strategy and industrial policy, Musk’s ecosystem is driven by private capital and the promise of future revenue from services that do not yet exist at scale.
A crowded race for modular orbital computing
China and SpaceX are not operating in a vacuum. A growing set of companies and agencies around the world are experimenting with orbital data centres, treating them as the next logical step in the evolution of cloud computing. Industry analysis points to a wave of projects in 2025 and early 2026 that frame space based AI infrastructure as both The Future and inherently Modular and Orbital, with designs that favour small, upgradable units over monolithic platforms. That modularity is crucial, because it allows operators to refresh hardware as AI models change, without waiting for a full satellite replacement cycle.
Within this broader field, China is trying to lock in an early lead by committing to specific constellations and timelines, while also leveraging its existing dominance in terrestrial manufacturing. Official references to China plans to build space based AI data centres highlight how closely the orbital push is tied to domestic chipmaking, satellite production and launch capacity. At the same time, the country’s broader technological rise, captured in global profiles of China, gives it a deep bench of engineers and suppliers who can adapt terrestrial data centre know how to the harsher environment of space.
For now, much of this remains on paper or in early demonstration hardware, and some technical hurdles, such as in orbit cooling and secure data links, are still unresolved. Unverified based on available sources. Yet the direction of travel is unmistakable. As national governments and private giants like SpaceX race to put AI into orbit, the line between space policy and digital policy is blurring, and the outcome will shape who controls the most valuable computing resources of the coming decade.
