Beijing University of Posts and Telecommunications has confirmed that the space computing cloud—the first of its kind to provide external in-orbit testing services—has transitioned to normal, continuous operation. The platform has already supported in-orbit pilot programs across key scenarios including big-data processing in space, 6G communications, and distributed storage. It also supports mixed-criticality task deployment, and has cumulatively completed hundreds of orbital computing calls serving over 100 institutional users.
The significance of the space computing cloud lies in its ability to place real computing capacity into orbit. Built on a domestically developed native software and hardware ecosystem, the platform achieves an inference energy-efficiency ratio of 10 tokens per joule for on-orbit LLM workloads, with space computing service coverage exceeding 10 percent. It unifies task orchestration, computing scheduling, data processing, and service delivery across satellite platforms, spaceborne servers, ground stations, and terrestrial data centers—enabling fully automated, end-to-end execution of complex operations.
Behind this achievement is a structural shift in the space information industry. In the past, satellites functioned as single-mission carriers whose value was concentrated in delivered hardware. Today, the space computing cloud establishes a schedulable, reusable, and operable in-orbit digital infrastructure, transforming satellites into networked intelligent nodes and shifting the industry's center of gravity from hardware delivery toward continuous service provision.
From an industrial perspective, the formation of sustainable in-orbit computing services relies on several critical capabilities: cross-domain satellite-ground task orchestration, distributed computing power scheduling, real-time data processing, and low-latency decision support. These capabilities have already evolved into continuous service offerings in areas such as real-time data analytics and iterative satellite-network capability upgrades, creating new growth vectors for satellite internet, 6G, commercial aerospace, and emergency communications.
The deployment of the space computing cloud effectively extends the paradigm of "computing scheduling" from terrestrial data centers to Earth orbit. Whether on the ground or in space, the underlying logic is the same: computing resources are moving from fragmented to unified scheduling, and from exclusive ownership to shared reuse. StarWar Cloud's focus on GPU computing platforms and intelligent orchestration directly addresses this industry-wide trend toward large-scale, cross-domain compute resource dispatch.
Looking further ahead, coordination between orbital and terrestrial computing power will give rise to entirely new service architectures. Satellites will perform preliminary computation in orbit and transmit only critical results back to Earth, dramatically reducing space-ground communication costs. In the 6G era, the joint scheduling of edge computing and orbital computing will become an integral component of communications infrastructure itself.
The normalization of the space computing cloud marks only the beginning. The boundaries of the computing power industry are expanding beyond server rooms into orbital space—a trajectory that anyone tracking computing infrastructure should be watching closely.