Independent conceptual engineering project

SolarDune

SolarDune was created to explore whether seawater could be turned into freshwater in a hot, dry desert, with the longer-term aim of supporting desert greening.

How the idea works Technical paper ↗
Why SolarDune exists

The starting point is freshwater.

Greening a desert ultimately runs into a simple constraint: water. SolarDune started from the question of whether seawater could be brought into a hyper-arid environment and converted into freshwater by making use of the abundant sunlight and heat available there.

As the idea developed, photovoltaic power and a data centre became part of the system. A data centre needs electricity and cooling, but it also produces a steady stream of low-grade waste heat. SolarDune explores whether those needs and outputs can be connected to seawater evaporation and condensation instead of being treated as separate systems.

In engineering terms, the reference concept combines photovoltaic electricity generation, data centre cooling and waste-heat recovery with a large-scale, low-pressure humidification–dehumidification (HDH) desalination cycle. Warm seawater is spread across enclosed evaporation decks; humid air moves through the structure; and a large condenser recovers part of that moisture as freshwater.

SolarDune integrates established physical processes into a single system for electricity generation, seawater evaporation and condensation, heat management and freshwater production.

01

Electricity

Solar photovoltaic panels generate renewable electricity for computing, pumps, fans and other auxiliary systems.

02

Heat

Residual solar heat and low-grade data centre waste heat are treated as useful inputs for warming seawater and supporting evaporation.

03

Air

Air carries heat and water vapour from the evaporation area toward the condenser. The design aims to work at low pressure and with modest airflow speeds.

04

Water

Seawater is evaporated and the vapour is condensed again as freshwater, while the remaining concentrated brine has to be managed separately.

Reference configuration

A first design used to test the idea.

~25 mstructure height
~150 × 75 mapproximate footprint
~3.6 GWh/yrphotovoltaic electricity
~50 m³/dayfreshwater target

These values are preliminary first-order engineering estimates for a reference configuration. They are not experimentally validated performance claims.

How it is meant to work

Evaporate seawater, move the moisture, condense freshwater.

In simple terms: sunlight produces electricity and heat; the data centre produces additional low-grade heat; that heat helps seawater evaporate; and the humid air is then cooled so part of the water condenses as freshwater.

The SolarDune reference design treats electricity, water, air and heat as four linked flows. Changing one of them affects the others. The system is therefore closer to a regulated thermodynamic environment than to a conventional desalination plant with one fixed process line.

Three stacked evaporation decks create a large seawater surface inside the structure. The current concept uses broad, shallow channels and low-pressure airflow to bring warm humid air toward a condenser on the north side. The cooler desert night is intended to help reject heat and improve condensation.

The data centre provides an economic use for the photovoltaic electricity while supplying a continuous source of low-grade waste heat that can support freshwater production. By designing renewable electricity, computing, cooling and heat recovery as parts of the same system, freshwater becomes an additional output of infrastructure that serves several functions simultaneously. This integration is central to the economic logic of SolarDune.

Salt aerosol, corrosion, demisting, brine disposal, dust, pumps, fans and maintenance remain practical engineering problems to be solved during prototyping.

Using the water

The aim is gradual desert greening.

Producing freshwater only matters if the water can be used sensibly. The current concept therefore links freshwater production to a gradual restoration strategy around the structure rather than to an immediate attempt to transform a large desert area.

A possible first stage would use nurseries or greenhouses, drip irrigation, pioneer vegetation, soil stabilisation and windbreaks. If vegetation survives and soil conditions improve, later stages could introduce more complex planting and, where appropriate, productive species suited to dry climates.

This part of SolarDune is particularly site-specific. Water demand, species selection, soil quality, salinity, evapotranspiration, wind exposure and establishment losses would all need local testing. Ecological restoration is therefore an intended use of the freshwater, not a guaranteed outcome of the current design.

Robust design

Designed for harsh desert conditions.

SolarDune is intended for an environment where extreme heat, dust, saline water and remoteness make complex infrastructure difficult to maintain. The concept therefore relies on large, simple geometries, low-pressure flows and relatively low-complexity components that can be inspected, repaired or replaced without highly specialised equipment. Wherever possible, the design uses robust materials and natural processes such as convection, evaporation and day–night temperature cycles. Where pumps, fans or other mechanical equipment are required, the aim is to use conventional and readily replaceable technology. Maintainability and replaceability are treated as design requirements.

What is known — and what is not

This is still a concept, not a proven system.

SolarDune is an early-stage conceptual engineering project. The reference geometry has been worked through with first-order thermodynamic, geometric and techno-economic calculations, but the system has not been experimentally validated or peer reviewed.

The largest uncertainties are not hidden. They include real evaporation rates, coupled airflow and humidity behaviour, condenser effectiveness, heat rejection during hot weather, the usable fraction of solar and data centre waste heat, salt and aerosol control, corrosion, maintenance and the practical integration with data centre cooling.

The most useful next step would be a coupled dynamic model followed by experimental testing of the evaporation and condensation subsystems. The present first-order calculations have not identified a fundamental physical constraint that would rule out further investigation.

Reference geometry

A preliminary 3D model.

An interactive 3D CAD model in Onshape shows the current SolarDune reference geometry. It is useful for understanding the scale and wedge shape of the structure and how the main spaces fit together.

The CAD model is a conceptual reference, not a construction drawing. Dimensions, internal layouts, materials and subsystem interfaces may change substantially after more detailed modelling and testing.

Open the 3D model ↗
Technical detail

The calculations and assumptions are public.

SolarDune: Greening Desert Landscapes by Building AI Infrastructure — A Conceptual Techno-Economic Assessment of Integrated Photovoltaic, Datacentre and Solar Desalination Infrastructure for Hyper-Arid Regions

Stadhouders, N. (2026)
Technical concept paper. Preliminary; not peer reviewed.

The report describes the system architecture, first-order heat and water balances, airflow assumptions, reference geometry, data centre integration, maintenance considerations, preliminary costs and the main unresolved engineering questions.

DOI: 10.5281/zenodo.22170017

Read the report on Zenodo ↗
Contact

Questions and criticism are welcome.

SolarDune is an independent project, and technical feedback is very welcome. If you work on desalination, heat transfer, data centre cooling, photovoltaics, renewable energy, arid-land ecology or a related field, I would be glad to hear from you. Relevant papers, comparable projects and practical engineering experience are equally welcome.

I am particularly interested in finding the assumptions that may be wrong or missing. SolarDune currently appears plausible at a first-order engineering level, but that does not mean it will work in practice. If there is a physical constraint, scaling problem, overlooked interaction or incorrect assumption that fundamentally challenges the concept, I would very much like to find it. If the concept continues to hold up under scrutiny, I would like to explore opportunities to test and develop it further.

contact@solardune.org

Get in touch