Reuters reports that China plans to launch space-based AI data centres over the next five years, citing a five-year development plan published by state broadcaster CCTV and attributed to China Aerospace Science and Technology Corporation (CASC). Reuters quotes the plan saying CASC will "construct gigawatt-class space digital-intelligence infrastructure" and that the centres will "integrate cloud, edge and terminal (device) capabilities" to process Earth-originating data in orbit. Reuters also reports Elon Musk told the World Economic Forum in Davos that SpaceX expects to launch solar-powered AI data center satellites within the next two to three years and that SpaceX is considering funds from a planned $25 billion IPO to develop orbital AI data centres. Tom's Hardware also covered the announcements and framed them as a competitive challenge between Beijing-backed space industry plans and SpaceX.
What happened
Reuters reports that China plans to deploy space-based AI data centres over the next five years, citing a five-year development plan referenced by state broadcaster CCTV and attributed to China Aerospace Science and Technology Corporation (CASC). Reuters quotes the plan promising to "construct **gigawatt-class space digital-intelligence infrastructure" and to "integrate cloud, edge and terminal (device) capabilities" so that data from Earth can be processed in orbit. Reuters also reports Elon Musk said at the World Economic Forum in Davos that SpaceX expects to launch solar-powered AI data center satellites "within the next two to three years" and mentioned using proceeds from a planned $25 billion IPO to support that effort. Tom's Hardware published parallel coverage and framed the announcements as a competitive rivalry between Chinese state-linked plans and SpaceX.
Technical details
Reuters attributes the Chinese plan's language to CCTV and CASC; the reported goals include creating high-throughput, high-power orbital infrastructure and integrating computing, storage, and communications capabilities in space. Reuters also reports Musk's claim that orbital solar generation can produce about five times the energy of ground panels, which he cited at Davos in support of SpaceX's timeline for orbital data centres.
Industry context
Editorial analysis: Companies and governments exploring orbital compute aim to address terrestrial energy and cooling limits for large-scale AI workloads. Comparable public statements from major space and cloud actors have emphasized three technical drivers: vastly higher solar insolation in orbit, the potential for direct line-of-sight communications to low-latency clients, and the engineering challenge of delivering sustained, high-density power and thermal management in space.
Implications for practitioners
Editorial analysis: For ML engineers and infrastructure teams, the reported initiatives signal a longer-term diversification of physical compute topology rather than an immediate change to current tooling. If realized, orbital data centres would affect long-term capacity planning, satellite-to-ground networking architectures, and resilience strategies for latency-sensitive or bandwidth-intensive workloads.
What to watch
Editorial analysis: Observers should track:
- •formal procurement or partnership announcements from CASC or other state entities
- •technical whitepapers or demonstrations detailing power-to-compute ratios and thermal designs
- •SpaceX operational milestones tied to the IPO proceeds and the timeline Musk cited at Davos. Public statements, regulatory filings, and demonstrator launches will be the clearest indicators of near-term feasibility
Scoring Rationale #
The story is notable because public plans from CASC and matching SpaceX statements imply major long-term infrastructure competition for AI compute capacity. It is not an immediate technical release for practitioners, and the reporting is several months old, reducing near-term urgency.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.