By David Kim-Hak, Vice President, Wastewater, Energy Recovery
In Phoenix, a water resources director at a food and beverage plant is looking at two numbers that nobody, a decade ago, would have read together: her Central Arizona Project (CAP) allocation, operating under a Tier 1 shortage that has cut CAP’s supply by roughly 30 percent, and her utility bill, up more than 30 percent since 2020. In central Texas, a plant manager at a semiconductor fab is signing a new multi-year power contract in ERCOT, surrounded by a data center buildout adding gigawatts of competing demand. In Flanders, a process engineer is designing a direct potable reuse system whose economics now depend as much on kilowatt-hours per cubic meter as on treatment performance.
These three people have never met. But they are looking at the same problem. For most of the last century, water scarcity, energy price volatility, and the infrastructure demands of computing were three separate concerns in three separate industries. Those boundaries are gone.
I have spent twenty years in water treatment, most of it focused on doing more with less energy. The first decade of my career was about making reverse osmosis membranes better: more selective and less energy-hungry. The industry has always cared about energy efficiency, especially in seawater desalination, where the pressures involved made the math unavoidable. What has changed is the pace of what operators are up against, and how far down the pressure spectrum the math now reaches. This is no longer just a sustainability story. It is a risk management story, and it applies to every facility that treats water through a membrane.
That reverse osmosis is energy-intensive is not news. What has shifted is the scope. The same math that made energy recovery essential in seawater desalination is now reaching into brackish water reuse, industrial wastewater treatment, and municipal drinking water reuse. Pressures once considered too low to bother with (100 to 200 psi rather than 1,000) are now large enough to matter. Pulling treatment into these new applications is water stress: the World Resources Institute estimates that roughly four billion people already live under high water stress at least one month of the year, with some 70 trillion dollars of global GDP exposed by mid-century. Stress forces reuse. Reuse means treatment. Treatment, at any pressure, means electricity.
For the better part of a decade, industrial electricity prices behaved like a slowly shifting floor. From 2016 through 2020, wholesale prices in most major markets were remarkably stable. That baseline is gone. According to U.S. Energy Information Administration (EIA) data, US retail prices rose 7.1 percent in 2021, 12 percent in 2022, 6.3 percent in 2023, and jumped another 7.6 percent through mid-2025. US wholesale prices in the first half of 2025 rose roughly 40 percent year over year. In top data center states, cumulative retail increases since 2020 range from 31 to 64 percent. The EIA expects US commercial electricity consumption to grow 5 percent in 2026, driven primarily by hyperscaler construction. The International Energy Agency (IEA) projects global data center electricity consumption doubling to roughly 945 terawatt hours by 2030. Every water-treating facility is now competing for electrons with a buyer whose willingness to pay is higher than theirs.
Public conversation about this risk is dominated by the supply side: more renewables, more gas, more nuclear, long-term PPAs. All of it matters. None is quick. The grid is constrained by permitting timelines, interconnection queues measured in years, and a competition for electrons that water infrastructure is not winning.
The demand side is the lever almost no one talks about, and it is the only one any facility fully controls. Every kilowatt-hour a plant does not consume is a kilowatt-hour it does not have to buy, forecast, hedge, or explain. The industry once treated energy recovery as a seawater desalination concern. That framing is outdated. The same physics that justifies energy recovery at 70 bar in a desalination plant justifies it at 10 bar in a municipal reuse system and at 120 bar in an industrial zero-liquid-discharge line. The pressures differ. The logic does not.
Consider two sites at opposite ends of the spectrum. In Hofstade, Belgium, one of Europe’s first direct potable reuse systems delivers 400 million liters of drinking water a year to 12,000 people using low-pressure reverse osmosis at 100 to 200 psi; integrating a low-pressure energy recovery device cut electricity consumption by about 23 percent, with a three-year payback. In central China’s Hubei province, a lithium iron phosphate cathode facility uses ultra-high-pressure reverse osmosis up to 120 bar to drive wastewater toward zero liquid discharge; integrating energy recovery cut consumption in the high-pressure stages by roughly 51 percent. One Belgian town at low pressure. One Chinese factory at ultra-high pressure. The physics is the same, and the economics now work across the full range.
The next decade will not be kinder to water-treating facilities than the last one. The ones that navigate it most successfully will be the ones that built or retrofitted toward using less energy per unit of water, across every pressure range their operations touch. Their reward will be quiet. It is the reward of having prepared for weather that has already begun.
