Data Centers Are Coming to Lake Country. Start the Water Record Before the Concrete Is Poured.

As of mid-September, at least 138 Ohio municipalities and townships had paused new data centers, and voters in 18 Ohio communities will decide on data center measures this November. Several sit in the Lake Erie watershed. Cleveland, which draws its drinking water from the lake, is asking residents directly about water consumption and effects on natural resources.

The Lake Erie Foundation put it precisely: most of the ballot measures are about electricity and utility bills, not water. But water is a real and separate question. Depending on how a facility is cooled and where its wastewater goes, a large data center can place meaningful demands on local supply, wastewater treatment, stormwater systems and the grid.

Ohio is early, not alone. The same conversation is starting around inland lakes across the country, and it is there, one lake and one community at a time, that the answers will matter most.

The wrong yardstick

An industry campaign now compares all U.S. data center water use in 2023, about 17.7 billion gallons, with Lake Erie's roughly 127 trillion gallons. The math is right. The comparison answers nothing a lake community needs to know.

A lake's total volume says nothing about what one facility pulls from one intake or one aquifer. It says nothing about how much is consumed rather than returned, what the wastewater carries, or how that demand stacks on every other user in the watershed. Impacts are local and cumulative. That is the scale people live at.

It matters even more for inland lakes. Many are fed in part by groundwater, and many data centers will draw from wells, not the lake. When a nearby aquifer is drawn down, the effect can show up first in private wells, then in the lake's level and inflow. Surface water and groundwater are one system. Watching only one half tells you half the story.

You can't measure "before" after the fact

Every argument about impact comes down to one question: compared with what? Lakes rise and fall with rain, drought, evaporation and season. Water tables move with weather and pumping. Without a record from before a facility comes online, nobody can separate its effect from normal variation.

That cuts both ways. A community that suspects harm has no baseline to prove it. An operator that causes none has no baseline to show it. Both end up arguing from anecdote.

The window to start that record is now, while proposals are still under review and rules are still being drafted. Once construction starts, the baseline is gone for good.

Surface and groundwater, one record, under $4,000

Lake Pulse pairs two real-time sensors from the Toolbox so a community can watch both halves of the system, side by side, in its lake portal.

Together, the pair costs under $4,000 in hardware. Water Table Watch starts at $1,550.

One sensor answers a narrow question. Two answer a better one. If the lake drops during a dry August and the water table drops with it, that is weather. If the water table falls while rainfall is normal and the lake holds, then begins to lag, that is a pattern worth a closer look. Add a Lakewide Weather Station and the record carries its own rainfall context.

No single season proves cause. A continuous, dated record that starts before a facility does is the strongest evidence a community can bring to a hearing, a permit review or a negotiation.

For lake communities: start the record this season

If a data center is proposed anywhere in your watershed, the useful moves are simple.

• Install before the vote, not after the build. A few months of baseline is worth more than years of data that start too late.
• Pick the right well. A shoreline or near-shore household well, closest to the proposed site, tells you the most.
• Ask for the specifics. Peak and annual withdrawal, source, consumptive use, wastewater volume and chemistry, discharge point, stormwater plan, and whether lower-water cooling is an option. These are the questions the Lake Erie Foundation and Cleveland are already asking.
• Keep it in one place. Boards and consultants change. The lake's record should outlast them both.

Water level is the first signal, not the only one. If wastewater or stormwater will reach the lake, add temperature, nutrient and chloride testing from the Toolbox to the same record.

For data center operators: show it, don't just say it

Many operators expect their water footprint to be modest and want to prove it. A press release won't settle that. A credible, continuous record will.

Lake Pulse works with operators on the same terms it works with everyone else:

• The record belongs to the lake. Data is held in the lake's own Lake Pulse record, not on a corporate dashboard.
• Methods are fixed up front. What is measured, where and how often is set before the first reading.
• Nothing is suppressed. Unfavorable results stay in the record alongside favorable ones.
• Lake Pulse has no stake in the answer. No referral fees, no treatment sales, no preferred vendors.

That neutrality is what gives a good result its value. An operator that funds monitoring it doesn't control can point to evidence the community already trusts. When the data shows no adverse effect, that is good news worth sharing, and Lake Pulse is built to share it.

What is your lake actually doing?

Responsible growth and a healthy lake are not competing goals. But no one can claim either without the data to back it. The communities that start measuring now will be the ones whose questions get specific answers.

Lake Level and Water Table Watch are in the Real-Time Monitoring aisle of the Lake Pulse Toolbox. If you are a lake community facing a proposal, or an operator who wants a neutral record, a 30-minute Boathouse call is the fastest way to scope where the sensors should go.

Source: Lake Erie Foundation commentary on Ohio data center moratoriums, ballot measures and Cleveland's public-input process (2026).