The concept of space data centers, which process data on satellites or in space rather than on Earth, is becoming more realistic. Google has launched a test satellite, SpaceX is advancing an ambitious plan, and Japanese companies are also beginning to enter the field.
Efforts to establish space data centers, which place data-processing infrastructure in space against the backdrop of power and land constraints at terrestrial data centers, are becoming more realistic. On 1 October 2026, Google launched a test satellite for Project Suncatcher, a research project to build an AI computing platform in space, and established communications and checked its operation after launch. The company is reportedly planning to test whether AI-specific semiconductors known as TPUs can withstand the space environment and then move toward realizing a computing platform comprising multiple satellites.
Although definitions and standard configurations have not yet been unified, a space data center refers to a system that places computing equipment on artificial satellites or space stations and processes data in space. Plans range from analysis on individual small satellites to large-scale concepts linking numerous satellites.
Power consumption at terrestrial data centers is expected to rise substantially in the future. An IEA report projects that consumption will nearly double from 485 terawatt-hours in 2025 to 950 terawatt-hours in 2030. For AI applications, it is expected to increase by about 3 times over the same period.
The advantages of space data centers are said to include extended use of solar power and reduced strain on terrestrial power grids, while significant challenges include rocket transportation costs, heat dissipation, communications and maintenance.
This article includes commentary from an SB Creative-related article and also introduces the view that practical implementation faces 5 barriers.