How a Campus Spends Water

Water is the least glamorous of the three levers and often the most decisive. To see why, you have to understand how a data center spends water in the first place, because the mechanism is not obvious and the quantities surprise people who have not looked closely.

A data center's servers generate heat, a great deal of it, and that heat must be removed continuously or the equipment fails. There are broadly two ways to reject heat: with water or with air, and most large facilities have historically leaned on water because it is efficient. The dominant water-based method uses evaporative cooling. Water is exposed to the warm air stream and allowed to evaporate; evaporation carries heat away extremely effectively, the same way sweat cools skin. The catch is that the evaporated water is gone — consumed, not merely borrowed. A large campus using evaporative cooling can consume water on the order of millions of gallons a day. That water is drawn from local sources: municipal supply, groundwater, or surface water, all of which belong to someone and serve other users.

There is a second, less visible way a data center consumes water: indirectly, through the electricity it uses. Many forms of power generation themselves consume water — thermal plants use it for cooling, and even some clean sources have a water footprint. So a campus that draws heavily on a water-intensive grid is spending water twice: once at the cooling tower and once, upstream and out of sight, at the power plant. This indirect water footprint is easy to ignore because it does not appear on the campus's own water bill, but it is real, and in a water-stressed region it deepens the total burden the campus places on the basin. Honest water accounting, like honest carbon accounting, insists on counting the upstream portion that convenient accounting omits.

The reason water rises to a first-order siting constraint is that, unlike electricity, water is intensely local and legally entangled. Electricity can be moved across a grid and generated far away; water largely cannot, at least not affordably. The water a campus consumes must come from the basin it sits in, and that basin has other claimants — farms, households, ecosystems — with rights that are often older and legally stronger than any newcomer's. In a wet region this contest is mild. In an arid or drought-prone region it is fierce, and it plays out in permits, public hearings, and politics. A community that shrugs at a large electricity draw may mobilize against a large water draw, because water scarcity is visible and visceral in a way that a distant power plant is not. Water, in short, is where the social license to build is most often granted or refused.

The cruel irony of water siting is that the regions richest in the clean power a carbon-serious campus wants are frequently the regions poorest in water. The sunny deserts with superb solar resources are dry. The windy plains can be arid. So the very map that clean-energy procurement draws — pulling campuses toward strong renewable resources — often pulls them toward water-scarce basins, where the water constraint then pushes back. This tension between the power map and the water map is one of the central problems of sustainable siting, and it cannot be wished away. It has to be managed, site by site, through the design choices we turn to in the next lesson.

There is also a question of what kind of water a campus uses, which softens the constraint where it can be applied. Not all water is equally scarce. Potable municipal water competing with household use is the most contested; non-potable, recycled, or reclaimed water is far less so. A campus designed to run on reclaimed wastewater, or on non-potable sources unsuitable for drinking or farming, places a much lighter burden on the community and meets correspondingly less resistance. This is why the source of a campus's water, not just the quantity, has become a siting and design question. A large water draw from a reclaimed source in a region that has invested in water recycling may be perfectly acceptable, while a smaller draw from a stressed potable aquifer may be impossible. The site-selection team must therefore ask not only how much water a site needs, but where that water comes from and whom it competes with.

The foundational lesson is to take water as seriously as power, and often more seriously, at the very start of a site search. A brilliant power-and-carbon site in a stressed, litigious water basin may be unbuildable, and no amount of clean electricity redeems it. Conversely, a site with ample or reclaimed water gains an advantage that does not show up on an energy spreadsheet at all. Water is the quiet veto, and the teams that learn to hear it early are the ones that do not waste years pursuing sites that were never going to clear the basin.