China Officially Launches Space AI Data Center Initiative
The Chinese government has approved the establishment of a Space Computing Industry Innovation Center to build AI data centers in orbit. This move, made just a week before SpaceX's announcement of AI1, is intensifying the US-China space AI competition.
The Chinese government quietly approved the establishment of the “Space Computing Industry Innovation Center” in early June 2026. According to a report by Tom’s Hardware, this center aims to bring together companies specializing in rocket manufacturing, satellite production, semiconductor development, and AI technologies to construct an AI data center network operating in space. The Beijing municipal government explained that the objective is to “connect the entire industrial chain of space computing and promote the development of the satellite Internet of Things (IoT) sector.”
Notably, this approval came approximately one week before Elon Musk’s announcement of the satellite “AI1,” designed to perform AI processing in orbit. Research firm SemiAnalysis noted on its X (formerly Twitter) account, “Almost no one noticed. China moved on space-based AI computation a week ahead of Musk.”
Six Research Areas
The center is expected to officially launch later this month and will focus on six key research areas:
- Space-native computing chips with high reliability and heat resistance
- High-performance, ultra-interconnected space computing payloads
- Space-computing satellite platforms and standardized systems
- Large-scale language models for space-based computation under power constraints
- Cloud-based measurement and control networks integrating space and ground systems
- Service-oriented and tokenized operation of space computing power
These research efforts aim to develop orbital AI data centers that are independent of terrestrial energy sources, addressing the bottlenecks currently faced by ground-based data centers. Specifically, the project seeks to overcome power supply limitations and cooling challenges by leveraging solar power generation and the natural cooling properties of outer space.
US-China Space AI Competition
Musk has been discussing space-based computational processing since November 2025. By February 2026, SpaceX had submitted a proposal to the U.S. Federal Communications Commission (FCC) for its “Orbital Data Center System,” a network of one million satellites. Additionally, Jeff Bezos’ Blue Origin is advancing its “Project Sunrise,” a sun-synchronous orbit initiative involving 51,600 satellites.
China’s approach differs significantly from that of the U.S. in that the government is spearheading the collaboration of multiple companies. In contrast, Tom’s Hardware notes that SpaceX and Blue Origin are competitors with no evident collaboration in technology development. SpaceX has adopted a vertically integrated strategy, pursuing projects like the massive “Gigasat” satellite factory, which spans approximately 1.1 million square meters (equivalent to about 200 soccer fields), and Musk’s “TeraFab” mega-project.
While it remains uncertain which approach will ultimately prove more effective in the long term, the contrasting strategies—China’s state-led centralized investment versus the U.S.’s competition-driven, vertically integrated private initiatives—are set to intersect in the rapidly advancing field of space AI.
Editorial Opinion
In the short term, China’s move is likely to accelerate state-level competition in space AI computing. Unlike SpaceX and Blue Origin, which are pursuing independent projects, China’s ability to consolidate research resources under state direction could lead to faster progress in foundational technologies such as space-grade semiconductors and heat-dissipation systems. It is anticipated that 2026 will see a wave of satellite test launches from various companies and nations, making the feasibility of AI inference in space a significant topic of discussion.
From a long-term perspective, the realization of space data centers could address the energy constraints and environmental impact of terrestrial data centers. The ability to directly harness solar power in space and perform computational processing without atmospheric attenuation presents a significant advantage. However, technical challenges such as satellite communication delays, semiconductor degradation from space radiation, and maintenance in orbit remain formidable hurdles, suggesting that commercialization may be unlikely until the 2030s at the earliest.
The editorial team believes that the key point of interest is determining which model—China’s state-led, collaborative approach or the U.S.’s competitive, vertically integrated strategy—will ultimately gain the upper hand in the space AI domain.
References
- Tom’s Hardware: China unifies tech sector to build grid-free orbiting satellite AI data centers — Published June 20, 2026
Frequently Asked Questions
- When will the Space Computing Industry Innovation Center officially launch?
- The center is expected to officially launch later this month (late June 2026), following its initial approval in early June by the Beijing municipal government.
- How do space-based AI data centers compare to ground-based ones?
- Space-based AI data centers can directly utilize solar energy, minimizing power supply constraints, and leverage the cold temperatures of outer space for cooling. Additionally, they are expected to enable low-latency AI inference globally.
- What are the differences between China’s and the U.S.’s approaches?
- China is pursuing a government-led model that unites rocket, satellite, semiconductor, and AI companies for collaborative development. In contrast, the U.S. relies on competition between private companies like SpaceX and Blue Origin, which follow a vertically integrated development strategy.
Comments