“This article builds on the diagnosis of water and energy consumption in data centers, developed here → [Energy and water consumption in data centers]”
Cooling without water: the physical limits of the digital cloud
For decades, the dominant narrative around the digital cloud has been one of dematerialization: intangible data, seemingly immaterial services, and processes that “happen on the internet.” However, this narrative increasingly clashes with the physical reality of the infrastructures that sustain the contemporary digital ecosystem. The cloud does not float: it occupies space, requires materials, consumes energy and, crucially, consumes water.
Several institutional and academic reports—such as those published by the Environmental and Energy Study Institute (EESI)—show that water consumption in data centers already represents a growing source of pressure in regions facing increasing water stress. In many territories, the primary constraint on expansion is no longer electricity availability alone, but access to sufficient water resources to ensure system cooling. This context forces a structural question: what happens when computation requires more water than a territory can sustainably provide?
The critical role of cooling in data centers
A substantial share of the energy consumed by a data center is inevitably transformed into heat. Dissipating this heat is essential to maintain hardware stability and prevent system failures. In current architectures, this function is primarily fulfilled through cooling technologies that depend directly or indirectly on water.
Cooling towers, evaporative systems and hybrid circuits represent the state of the art in data center cooling. Despite continuous improvements in energy efficiency, these systems remain bound by a fundamental physical principle: cooling entails a water cost. The water used is not an abstract resource; it comes from aquifers, river basins or urban supply networks that already face increasing pressure from agricultural, domestic and ecological uses.
In this scenario, strategies based solely on efficiency prove insufficient. While they reduce consumption per unit of computation, they do not eliminate the structural problem. As the scale of computation grows—driven by artificial intelligence, large-scale data processing and generative models—the physical limits of the system reassert themselves with increasing intensity.
The location of computation as a critical variable
Faced with these limits, a line of thinking has emerged that shifts the focus from how computation is performed to where it takes place. This involves considering the location of computation as a structural variable of the digital system. Within this framework, options once considered marginal begin to be explored, such as the partial relocation of processing beyond Earth.
From a physical standpoint, space offers two distinctive properties:
- Abundant and continuous solar energy, without the need for distribution grids or fuels.
- The presence of a vast natural thermal sink: the cold of deep space, which enables radiative cooling without water consumption.
Unlike terrestrial data centers, where cooling is intrinsically a territorial problem, in the space environment heat can be evacuated directly through radiation. There is no evaporation, no water intake, and no return flows. This difference, widely documented in the scientific literature, compels a reassessment of the assumptions underlying the current digital cloud model.
Orbital data centers: solution or mirror?
In this context, several research efforts—reported in recent articles in Nature Electronics and in industrial feasibility studies such as the ASCEND feasibility study by Thales Alenia Space—explore the viability of orbital data centers. These proposals rely on constellations of satellites with onboard processing capabilities, powered by solar energy and cooled through direct radiation into space.
It must be emphasized that these initiatives do not constitute a complete environmental solution. Satellite manufacturing, semiconductor production, launch processes and end-of-life management all entail significant environmental impacts. In this sense, orbital computing cannot be considered impact-free.
However, their primary value lies elsewhere: they function as a conceptual mirror. The fact that, under certain physical conditions, data processing can occur without consuming water reveals that water use in computation is not an inherent necessity, but the result of architectural, locational and scaling choices.
What does “a data center in space” actually mean?
Speaking of data centers in space does not imply a single architecture nor a direct orbital equivalent of terrestrial facilities. Current proposals—described both in academic literature and in specialized technology media as well as general reference sources such as Wikipedia—span multiple scales and functions, with widely varying levels of technological maturity.
In simplified terms, three main approaches can be identified:
- Satellites with onboard computing (orbital edge)
Existing satellites—used for observation, telecommunications or scientific research—equipped with local processing to analyze data at the source and reduce transmission volumes to Earth. - Constellations of computational satellites (distributed orbital cloud)
Systems specifically designed for computation, where each satellite acts as a node in a distributed cloud, powered by solar energy and cooled through radiation. - Monolithic or extraplanetary proposals
Still largely conceptual initiatives placing data centers in high orbits, on the Moon or at gravitational equilibrium points, to explore physical and operational limits.
Which “servers” can operate in space?
The main technical challenge of orbital computing is not energy access or cooling, but exposure to high-energy radiation. Outside Earth’s atmospheric protection, electronic components are subjected to particles that can degrade circuits, cause computational errors or drastically shorten hardware lifespan.
Specialized literature identifies three main strategies to address this challenge:
- Radiation-hardened hardware specifically designed for space environments.
- Controlled use of commercial hardware, accepting shorter lifecycles.
- Fault-tolerant architectures that prioritize system resilience.
This approach implies a paradigm shift: orbital computing relies less on the durability of individual servers and more on the system’s ability to absorb failures and continue operating.
Lifecycle: manufacture, launch, operate, retire
Assessing the environmental footprint of these infrastructures requires a full lifecycle perspective. While the operational phase may be nearly emissions-free and water-neutral, impacts concentrate in other stages: manufacturing, launch, limited operation and end-of-life management, including space debris risks.
For this reason, the most rigorous proposals do not present orbital computing as a “clean” solution, but rather as a redistribution of impacts aimed at reducing pressure on critical resources in inhabited environments.
Data access and storage
Orbital data centers are not designed to replace large-scale terrestrial storage. Current proposals focus primarily on temporary processing of intensive tasks, intermediate storage and integration with terrestrial clouds through hybrid orchestration models.
In this way, orbital computing does not eliminate the terrestrial cloud, but may reduce its load in the most resource-intensive processes.
Pros and cons (current state, in theory)
Potential advantages
- Cooling without water consumption
- Direct access to solar energy
- Reduced pressure on terrestrial infrastructures
- No direct territorial impact
Limitations and risks
- High manufacturing and launch costs
- Limited hardware lifespan
- Complexity of space debris management
- Dependence on robust communications
- Uncertain economic viability
In the short term, orbital computing does not represent a general-purpose alternative to terrestrial data centers. However, as a limit case, it highlights a fundamental insight: water consumption associated with digital computation is not inevitable, but the outcome of architectural and locational decisions.
The future of the cloud will not be ethereal
Thinking about the future of the cloud requires acknowledging that digital systems are subject to physical laws. Scaling computation without accounting for energy, materials, heat and water is unsustainable. Space, with all its constraints, makes this reality explicit.
Even if orbital data centers never become commonplace, their existence as an idea serves an essential function:
to shift the debate from purely technical solutions toward the digital infrastructure model we are willing to sustain.
On a planet where water is becoming a critical resource, rethinking where computation takes place may be as important as rethinking how it is performed.
References and recommended reading
Academic literature
- Nature Electronics
Recent work on orbital data centers and space computing examines the potential of radiative cooling, continuous solar power and the material limits imposed by radiation. These publications provide the conceptual framework for understanding orbital computing as a limit case to rethink the current cloud model.- Aili, A., Choi, J., Ong, Y. S. et al. (2025). The development of carbon-neutral data centres in space. Nature Electronics, 8, 1016–1026. https://doi.org/10.1038/s41928-025-01476-1
Institutional reports and public policy analysis
- Environmental and Energy Study Institute (EESI)
Data Centers and Water Consumption.
A key report on data center water use, the territorial impacts of cooling and risks linked to growing water stress. It provides context for why water is becoming a limiting factor in digital infrastructure.
https://www.eesi.org/articles/view/data-centers-and-water-consumption
Aerospace industry and applied research
- Thales Alenia Space
ASCEND feasibility study – Space-based data centers.
Results from the ASCEND feasibility study, assessing space-based data centers as energy-autonomous infrastructure cooled by radiation and integrated into a hybrid Earth–orbit architecture. This source adds an industrial, applied perspective that complements academic literature.
https://www.thalesaleniaspace.com/en/press-releases/thales-alenia-space-reveals-results-ascend-feasibility-study-space-data-centers-0
Specialized technology coverage
- Singularity Hub
Pieces such as Future data centers could orbit Earth, powered by the Sun and cooled by the vacuum of space offer an accessible yet informed overview of emerging orbital computing proposals, connecting academic research, technology forecasting and industrial strategies.
https://singularityhub.com/2025/11/03/future-data-centers-could-orbit-earth-powered-by-the-sun-and-cooled-by-the-vacuum-of-space/
General reference sources
- Wikipedia
Entries on Data center, Space-based computing, Satellite constellation and Radiation hardening can be useful as a starting point for terminology and broad technical context—always complemented with academic and institutional sources.
Editorial note
The sources cited here are not used to present orbital computing as an immediate or definitive solution, but as a conceptual instrument to make the physical—especially water-related—limits of today’s terrestrial data center model more explicit. Their function in this article is to support an informed debate about sustainability, location and the architecture of digital computation.
