Desalination works. A reverse osmosis plant takes seawater, pushes it through a membrane at high pressure, and hands a city drinking water that was not freshwater before it entered the building.
That is created supply — not accounting, not a transfer, not somebody else's river rebranded. Potable reuse does the same job from the other end. So the useful question isn't whether a plant makes water. It's what kind of water it makes, and where that water shows up. Rain, where the region allows it. Late-season flow, almost everywhere you restore the land. Gallons at a tap, wherever you build the plant.
what desalination actually creates
Desalination is the removal of salt from seawater or brackish water to produce usable freshwater, most often by forcing seawater through reverse-osmosis membranes under pressure. A 2019 UN University–led synthesis counted roughly 15,900 operating plants producing about 95 million cubic meters of freshwater per day, with about half of it in the Middle East and North Africa. Israel now covers a large share of its municipal demand this way. That is not a rounding error, and it is not greenwash.
Potable reuse works the same way conceptually: treat municipal wastewater back to drinking standard and put it into supply instead of the ocean. Both are what utilities mean when they say new water — supply that exists because you built something, rather than supply you took from a neighbor or squeezed out of a leak-detection program.
A desalination plant is one of the few water projects that genuinely adds freshwater to a service area rather than reallocating someone else's.
the honest limits
None of these are reasons not to build. They are reasons to know what you bought.
Energy. Efficient large seawater plants run on the order of 3–4 kWh per cubic meter of product water; the best modern installations sit nearer 2, and a 2025 demonstration plant in the Canary Islands set a verified record at 1.794 kWh/m³ on its RO loop. The thermodynamic floor is around 1 kWh/m³, so there is no order-of-magnitude breakthrough waiting. Brackish sources are far cheaper energetically — roughly 0.5–1.5 kWh/m³. Electricity is the largest operating line item, which means desalinated water is priced to your power market, and its carbon intensity tracks your grid.
Brine. Seawater RO typically recovers 35–50% of what it takes in. The rest leaves as hypersaline concentrate. Globally that's an estimated ~142 million cubic meters of brine a day, about 1.5 liters per liter of freshwater produced. Well-engineered outfalls with diffusers manage it; it is still a disposal cost and an ecological question, and it is genuinely hard for inland plants with no ocean to discharge to.
Geography. Desalination is a coastal technology, and water is heavy. Lifting one cubic meter 1,000 vertical meters takes about 2.7 kWh of ideal work — roughly what it costs to desalinate that cubic meter in the first place, before friction and pump losses. That physics is why coastal plants serve coastal demand. A basin on the far side of a mountain range is not in the service area at any sane price.
Cost and lead time. Capital-heavy, years to permit and build, and each unit of water carries a running energy bill that never goes away. We already ran the full buyer's menu — dam, desal, reuse, imports, hauling, efficiency, source protection — with cost and lead time side by side in the cheapest water security isn't a pipeline. No point repeating it here.
If you came expecting a swipe at desalination, you'll be disappointed. Coastal cities with a hard supply gap should build plants, and the operators running them are solving a real problem competently. The mistake isn't the plant. It's the substitution — treating manufactured gallons as if they were a repaired basin.
two kinds of new water
| what it creates | where it shows up | what it cannot do | |
|---|---|---|---|
| plant new water (desal, potable reuse) | freshwater from seawater or wastewater — gallons that weren't potable before | at a tap, on a delivery schedule you control and can contract | restore baseflow in an inland river or change what falls upstream. A coastal desal plant does not raise a water table; potable reuse built for recharge can |
| land new water (restored soils, wetlands, floodplains) | infiltration, soil moisture, groundwater recharge, late-season streamflow — and in some regions, more locally recycled rainfall | across the basin: wells, springs, meadows, the dry end of the hydrograph | guarantee a fixed volume on a fixed date, or serve a city with no functioning watershed in reach |
A plant creates freshwater at a tap. A landscape creates soil moisture, recharge, and late-season flow. Neither one substitutes for the other, and only one of them works on the river itself.
what a plant does not fix
A coastal desal plant does not change the condition of the watershed upstream of your other intake. It does not slow runoff or return water to a stream in September. Potable reuse built for recharge — Orange County has done it for decades — does raise a basin's water table; it still does not touch the watershed upstream. The reach that goes dry stays dry. The wells that dropped stay down. The wildfire ash and sediment that spike your treatment costs after a burn keep arriving on the same schedule.
Rainfall recycling — the share of local precipitation that comes from moisture evaporated off nearby land — is real, well documented, and regionally variable. Restoring vegetation and wetlands can increase it in some landscapes and barely move it in others. It is not a lever you can promise on a parcel. We wrote the honest version of that in you can't seed a cloud that isn't there.
What is dependable is the near-field half: land that catches, holds, and slowly releases water produces more usable water in the same basin — later in the season, when the shortage actually bites.
the foundation is the place
The inland factory is a natural asset — the wetland, floodplain, or headwaters that produces and holds water. Nature-based solutions are the work on it. Ensurance is the mechanism that funds that underlying — not a desal offtake, and not a grant you re-justify every budget cycle. We do not sell plant capacity. Ensurance does not create water. A certificate attaches the funding to one named place. It is not a water right, not a delivery contract, and not a substitute for the energy your plant will use tomorrow.
what to do with this
If you're a coastal utility with a supply gap — build or contract the plant, then put source-water restoration on the same capital plan and compare cost per reliable acre-foot. Talk through the utility case →
If you're weighing options — start with the honest cost and lead-time menu in the cheapest water security isn't a pipeline.
If you're inland — a plant was never on your menu. The land is. Read next: rewater the land.