
A data center request is usually discussed in gallons, but the figure that governs a utility's exposure often sits in the cooling system's water chemistry. Evaporative cooling towers lose water as vapor and leave dissolved minerals behind, so the water remaining in the basin grows more concentrated over time. Operators hold that concentration down by bleeding off part of the circulating water as blowdown and replacing it with makeup water from the utility or another source. The amount of makeup a tower draws, and the amount it discharges, depends on how far the operator lets concentration climb before bleeding.
Engineers call the controlling ratio cycles of concentration, meaning the dissolved solids in the circulating water divided by the dissolved solids in the makeup water. In steady operation, blowdown volume is approximately the evaporation rate divided by the cycles minus one, with drift and other losses added on top. The arithmetic is simple, but the inputs are where errors enter, because each one is a claim from the requester. Raising the cycles reduces blowdown, and with it the makeup water needed for a given evaporation load. That is the efficiency argument, and it is why designers push cycles upward whenever the source water allows it.
Where the source water sets the ceiling
The limit on cycles is the makeup water itself. Hardness, alkalinity, silica, and total dissolved solids each reach a scaling or fouling threshold at a different concentration, and the first to reach its threshold ends the climb. A makeup source high in silica or hardness can cap cycles well below what the tower hardware could tolerate. Lower cycles then mean more blowdown, so the same facility both draws more water and discharges more of it than a headline gallon figure suggests. Operators often discover this limit only after startup, which is why the design basis should name it. Chemical treatment can extend the ceiling, but it adds its own supply, storage, handling, and discharge components. Source chemistry therefore shapes demand, the discharge stream, and the treatment program at once.
What the sector report frames
Water Finance and Management reported on a report from the American Water Works Association, the trade outlet's summary of the AWWA work on data center development, which carries the title "Considerations for Water Utilities Evaluating Data Center Development." The available summary does not list the report's specific recommendations, so this article does not attribute any particular finding to it. Utility planners should read the document directly before relying on any of its guidance in a local decision.
Questions to put to the requester
A request sized only on total gallons leaves the cooling design unexamined. Before the number is accepted, ask for the design cycles of concentration and the operating range expected in hot months, when evaporation is highest. Ask for the makeup source chemistry at its worst season rather than its annual average, because a silica or hardness peak can set the ceiling for the whole year. Ask for the blowdown volume, its chemistry, and its destination, since local pretreatment limits and the receiving collection system determine whether that stream can be accepted at all. Request the list of treatment chemicals along with their storage and handling plan. Finally, confirm whether the system is closed-loop, which recirculates with limited makeup, or once-through, which draws and discharges far more water. That single difference can change demand more than any efficiency measure in the design.
Where the uncertainty sits
The concentration ratio is a design assumption and an operating target, not a fixed property of the water. Operators can change cycles with the season, the treatment program, and the condition of the basin, so a figure in a request may describe design intent rather than what the facility runs at in August. The careful position for a utility is to record the cycles figure as a stated assumption, carry the peak-season chemistry alongside it, and treat any gap between design and operation as an open question until the customer provides operating data.