How atmospheric water generators work
An atmospheric water generator is a dehumidifier with a treatment train attached. A refrigeration circuit chills a coil below the dew point, moisture in the air passing over it condenses into liquid, and that condensate is collected, filtered, usually UV-treated and often remineralised before dispensing. The physics is entirely uncontroversial — it is the same process that puddles water under an air conditioner.
The appeal is obvious: water with no mains connection, no bore, no tank and no delivery. For genuinely water-scarce or off-grid situations, that is a real proposition. The question is not whether the technology works. It plainly does. The question is what it costs per litre, and how reliably it produces anything at all in a given climate.
What independent testing found
Because output depends on ambient conditions, manufacturer specifications are quoted at favourable temperature and humidity, and real-world performance diverges sharply. Published long-term field studies are the useful source here.
| Study conditions | Output | Energy per litre |
|---|---|---|
| 22°C, 63% RH (mild, humid) | 0.95 L/hr | 0.84 kWh/L |
| 36°C, 45% RH (hot, drier) | 0.36 L/hr | 2.1 kWh/L |
| 18.7°C, 48% RH (cool) | 0.13 L/hr | 1.98 kWh/L |
| 17.8°C, 37% RH (cool and dry) | Zero | — |
| Large commercial unit, favourable months | Up to ~1,500 L/day | 0.26–0.31 kWh/L |
| Same unit, cool/dry months | 71% production drop | Up to 0.53 kWh/L |
Three findings matter more than the individual figures. Energy per litre varies by roughly eightfold across conditions in the published range — the same machine is a different economic proposition in different weather. Efficiency and output move together in the wrong direction: the conditions that cut production also raise the energy cost of each litre produced. And output can reach zero: in the year-long field study, cool days at 37–47% relative humidity produced no water at all, while the machine still drew power.
The Australian problem
This is where the technology meets an uncomfortable local reality. AWGs perform best in warm, humid air — and much of populated Australia offers the opposite when demand for water is highest. Perth and Adelaide summers are hot and dry; the inland is drier still. The published zero-production conditions are not exotic edge cases here.
The pattern to be alert to is anti-correlation with need: the drought conditions that make an independent water source attractive are precisely the conditions in which an AWG produces least, at the highest energy cost per litre. A device marketed as drought resilience underperforms most during drought. Coastal humid regions — Darwin, Cairns, coastal Queensland — are a genuinely different and far more favourable case.
The cost comparison is the part no marketing page runs. At around 1 kWh per litre — a mid-range real-world figure — a litre of AWG water costs roughly one kilowatt-hour of retail electricity. Mains water in Australian capitals is billed per kilolitre in single-digit dollars, which puts the price of a thousand litres of tap water in the vicinity of what a handful of litres from an AWG costs to generate. The gap is not a percentage; it is orders of magnitude.
Is the water any good?
Condensate itself is close to distilled — very low in dissolved minerals — but it is not sterile or automatically clean, and that is the part worth scrutinising. The air it comes from carries dust, pollen, bacteria, mould spores and whatever else is local, including traffic pollution or bushfire smoke. Everything that lands on a cold wet coil ends up in the collection tank.
This makes the treatment train, not the condensing coil, the actual water quality determinant: filtration to remove particulates, UV or equivalent to handle microbiological load, and a sealed, regularly cleaned storage tank. A warm, damp, stagnant tank is an excellent environment for biofilm, so maintenance is not optional. Ask what treatment stages are fitted, what certification they carry, and what the cleaning schedule is — the same questions you would ask of a rainwater system, for the same reasons.
Consider what your air quality actually is at the intake. An AWG in a garage, near a road, or during a smoke event is drawing from that air. Siting matters more than it does for any other technology we review, because the source is the atmosphere at that spot.
Our verdict on the evidence
The technology is real and the physics is sound; the economics and the climate dependency are what decide it. Published field studies put specific energy consumption between roughly 0.26 and 2.1 kWh per litre depending on conditions, with documented zero-output days in cool, moderately dry air. That is an enormous performance range for a purchase decision, and it is why manufacturer figures quoted at optimal conditions should be treated as a ceiling, not an expectation.
Where an AWG genuinely earns its place: humid regions, sites with no mains, bore or reliable tank supply, emergency and remote applications, and situations where the alternative is carting or trucking water. Where it does not: as a general replacement for mains supply in an Australian capital, or as drought insurance in a dry-summer climate, where it will produce least exactly when it is needed most.
Before buying, the two numbers to demand are kWh per litre and output, both stated at your climate’s actual average temperature and humidity — not at the manufacturer’s rated conditions. If a supplier cannot or will not provide a generation table across conditions, that omission is the answer.
Primary sources: year-long field performance analysis of a commercial AWG under hot humid climate conditions (0.84–2.1 kWh/L; zero production recorded at 17.8°C/37.2% RH), published in Case Studies in Chemical and Environmental Engineering, 2022; long-term evaluation of VCRS-based atmospheric water generation in arid climates, Masdar City 2023–24 (0.26–0.53 kWh/L; 71% productivity decline in cooler, drier months); modelling of AWG specific energy consumption against historical weather data, Environmental Research: Infrastructure and Sustainability, 2023. Reviewed August 2026. FilterOut sells nothing and takes no commissions — see who to trust for filter advice.
FAQ
How much electricity does an atmospheric water generator use?
Published field studies report roughly 0.26 to 2.1 kWh per litre depending on temperature and humidity. Efficiency falls as air gets cooler or drier — the same machine can cost eight times as much per litre in unfavourable conditions.
Do atmospheric water generators work in dry climates?
Poorly. Output falls sharply as humidity drops, and one year-long study recorded zero production on cool days at 37–47% relative humidity. They produce least in exactly the dry conditions that make an independent water source appealing.
Is atmospheric water generator water safe to drink?
It depends entirely on the treatment train, not the condensing process. Condensate collects whatever the air carries — dust, pollen, bacteria, mould spores, pollution — so filtration, UV treatment and a clean, sealed, regularly maintained tank determine the result.