Radiation is the heat-transfer route in a vacuum
On Earth, data-center heat is moved by fans, liquid loops, chillers, cooling towers, or combinations of those systems. In orbit, there is no surrounding air to carry heat away by convection. Equipment can still conduct heat into pipes, heat pipes, and radiator panels, but the final step is radiative: a radiator emits infrared energy toward space. NASA describes this as a standard spacecraft thermal-control method.
A computer cannot simply be exposed to space and expected to cool itself. Heat must be collected from processors and power electronics, transported to radiator panels, and released while the system controls its temperature. Radiator area, orientation, surface coatings, sunlight, Earth’s infrared emissions, and the spacecraft’s own heat all matter. High-performance processors generate concentrated heat loads, making heat spreading and radiator mass major design issues.
What the environmental case might be
An orbital system could avoid constructing a large facility on a particular parcel of land, drawing local freshwater for evaporative cooling, or placing new electrical demand on a constrained local grid—depending on the design and where it operates. Solar power in orbit can also be more continuous than at a ground site. These factors are why research programs are studying the concept as one possible way to expand computing infrastructure.
Those benefits are not automatic. A fair comparison must include the full lifecycle: manufacturing launch vehicles and satellites, propellant and launch emissions, solar arrays and radiator materials, replacement cycles, downlink and ground-station infrastructure, end-of-life disposal, and the energy and materials used to build equivalent Earth-based capacity. Avoiding a local water draw is a benefit to that location; it does not by itself establish lower total environmental impact.
Space changes the engineering and operational risks
Radiators have to reject the heat while avoiding unwanted solar and Earth heating. They add mass and area, and their performance can constrain the satellite’s shape and orientation. Electronics must also cope with radiation, thermal cycling, launch vibration, micrometeoroids, and limited opportunities for repair. Communications between satellites and users on Earth require high-capacity links and ground infrastructure.
Google’s Project Suncatcher moved from paper studies to an early prototype: Google reported in October 2026 that a small satellite with an AI accelerator had run an initial workload in orbit. The company said it plans further tests of thermal stress and radiation and a later test of two interconnected spacecraft. That is a meaningful experiment, but it is still an early research step rather than evidence of a commercial-scale orbital data center.
The European Commission-funded ASCEND project also examined the technical, economic, and environmental feasibility of orbital data centers. Its work is a system-level study: proposed environmental benefits depend on assumptions about launch costs, spacecraft architecture, useful lifetime, energy supply, and the terrestrial alternative.
What would demonstrate a real advantage?
Useful evidence would compare measured compute delivered over a satellite’s lifetime with the energy, emissions, water, materials, and waste of a matched ground system. It should include launch and replacement emissions, radiator and solar-array manufacturing, orbital debris management, and the emissions of the ground network needed to move data. The comparison should state geography and electricity mix because the impact of a ground data center varies considerably by location.
Radiative cooling makes space computing physically possible, but it is a difficult thermal-control system rather than a free cooling resource. Orbital data centers may eventually make sense for particular workloads or locations; the environmental case will depend on complete lifecycle evidence, not simply on the absence of a building or cooling tower on the ground.