Why Data Center Water Use Needs Better Operational Visibility

Data centers have become one of the most important infrastructure stories of our time.

They power cloud computing, enterprise applications, artificial intelligence, financial systems, healthcare platforms, logistics networks, public services, and the everyday digital tools modern life depends on. They are also increasingly part of broader public conversations around electricity demand, water use, land development, utility capacity, and local community impact.

That conversation is not going away. Nor should it.

Regardless of where someone personally falls on data center growth, one point should be easy to agree on: once these facilities exist, they should be operated as intelligently, transparently, and responsibly as possible.

At SitePro, that is where we see the real opportunity.

Our core is industrial automation. We help operators monitor, control, and optimize complex infrastructure. But our vision is broader than automation alone. We believe the same systems that improve uptime, safety, and asset performance can also support sustainability by making resource use visible, measurable, controllable, and continuously improvable.

For data centers, one of the clearest examples is water.

That is the gap we need to close.

Data center water use needs better operational visibility.

And the path forward starts with building a real water balance.

Water is not just a sustainability topic. It is an operations topic.

Water efficiency in data centers is often overlooked compared with energy efficiency, but water is directly tied to cooling, heat rejection, equipment performance, reliability, and local resource stewardship. Lawrence Berkeley National Laboratory notes that data centers consume water directly for cooling and indirectly through electricity generation, and recommends metering and monitoring systems as a best practice for tracking water consumption.

That point matters because data center water use is not a single number. It is a system.

Water may enter a facility from a municipal supply, a reclaimed water source, an onsite well, a storage system, or another supply arrangement. It may be used in cooling towers, evaporative systems, humidification, closed-loop fill, water treatment, domestic uses, cleaning, testing, or other support functions. It may leave the facility through evaporation, blowdown, drift, discharge, treatment reject, leaks, sanitary flow, or other losses.

In other words, water moves.

If operators only see a monthly bill or a single meter reading, they do not have operational visibility. They have hindsight.

Operational visibility means knowing where water is entering, where it is being used, where it is being lost, whether usage is expected, whether equipment is performing correctly, and whether the current operating mode makes sense for the facility’s load, weather, and reliability requirements.

That is a very different level of understanding.

The industry is moving beyond simple metrics

Water Usage Effectiveness, or WUE, is one of the most common metrics used to understand data center water efficiency. Microsoft describes WUE as water use relative to electricity consumed, measured in liters per kilowatt-hour, and calculated by dividing annual liters of water used for humidification and cooling by annual kilowatt-hours used to power IT equipment.

WUE is useful. But WUE alone is not enough.

A single annual WUE number does not tell operators whether a cooling tower is cycling properly, whether blowdown is excessive, whether a valve is stuck, whether a meter is drifting, whether a treatment system is underperforming, or whether two facilities with similar load are behaving differently because of climate, equipment design, maintenance history, or operating strategy.

ASHRAE’s AI Data Center Energy Performance Framework points in the right direction by encouraging operators to track and report broader performance metrics, including PUE, WUE, WUI, CUE, DCRE, and ITWC, and to use intelligent controls, real-time monitoring, digital twins, and continuous commissioning to keep systems operating efficiently.

That is the shift we believe matters most.

The future is not just annual reporting. The future is continuous operational intelligence.

One of the mistakes in the data center water conversation is treating all water use the same.

It is not.

That means operators need more than corporate-level averages.

They need site-level insight.

A facility in a water-stressed region may need to prioritize dry cooling, closed-loop systems, reclaimed water, peak-shaving strategies, or operating modes that minimize freshwater withdrawal. A facility in a water-abundant region may make different tradeoffs between energy efficiency and water use. A legacy site with open cooling towers may have a completely different profile than a new high-density AI facility using direct liquid cooling.

There is no universal answer.

But there is a universal requirement: operators need accurate, contextual, real-time data.

Building a data center water balance

A water balance is a practical operating model that shows how water enters, moves through, and leaves a facility.

At its simplest, the equation is:

Water In = Water Used + Water Lost + Water Discharged + Change in Storage

But in a data center, that equation must be broken down into the actual systems and assets that determine water performance.

A useful data center water balance should measure at least seven categories.

1. Water supply sources

Operators need to know where water is coming from and how much is entering from each source.

That may include potable municipal water, reclaimed water, non-potable industrial water, onsite wells, rainwater capture, storage tanks, backup water supplies, or temporary supply sources.

The water balance should answer:

• How much water is entering the site?
• Which source is supplying it?
• What is the quality of that water?
• Is the source potable, reclaimed, non-potable, or mixed?
• Are there source-specific limits, costs, permits, or reporting requirements?
• Is supply changing over time?

2. Cooling system makeup water

Cooling is often the largest direct water use in facilities that rely on evaporative or hybrid heat rejection.

Operators should measure makeup water into cooling towers, evaporative coolers, adiabatic systems, humidification systems, and any other cooling-related water users.

This is where real-time visibility becomes essential. Makeup water can change because of IT load, outside air temperature, humidity, cooling mode, setpoints, tower cycles, equipment condition, water treatment performance, or leaks.

The water balance should answer:

• How much makeup water is being added?
• Is usage aligned with IT load and weather conditions?
• Is one tower, loop, or system using more than expected?
• Are abnormal makeup events tied to specific assets?
• Are cooling systems operating in the intended mode?
  

3. Blowdown, bleed-off, and discharge

In cooling tower systems, water is lost not only through evaporation but also through blowdown or bleed-off. Blowdown helps control mineral concentration and water quality, but excessive blowdown can waste water and indicate poor treatment, improper control, sensor issues, or low cycles of concentration.

Operators should measure blowdown flow, conductivity, total dissolved solids, pH, temperature, treatment system activity, and discharge events.

The water balance should answer:

• How much water is being discharged?
• Is blowdown controlled by actual water chemistry or fixed timing?
• Are cycles of concentration optimized?
• Are treatment systems performing correctly?
• Are discharge volumes increasing without a clear cause?

4. Evaporation and heat rejection performance

Evaporation is harder to meter directly, but it can be estimated using makeup water, blowdown, drift, basin levels, weather, load, and cooling system performance.

For facilities using evaporative cooling, evaporation is not necessarily a failure. It is part of the heat rejection process. The operational question is whether evaporation is expected, efficient, and appropriate for the current conditions.

The water balance should answer:

• How much water is likely being evaporated?
• How does evaporation compare with cooling load?
• Is the system rejecting heat efficiently?
• Are temperature differentials where they should be?
• Is the facility using water at the right times, or using it when dry cooling or free cooling would be sufficient?

5. Leaks, drift, overflow, and unaccounted losses

Every industrial water system has the potential for unaccounted losses.

In a data center, small leaks or control failures can become significant because the facility runs continuously. A valve passing water when it should be closed, a basin overflow, a failed level sensor, a stuck fill valve, or an unnoticed underground leak can quietly distort water performance.

The water balance should answer:

• Is there water entering the system that cannot be explained by known usage?
• Are basin levels changing unexpectedly?
• Are pumps cycling abnormally?
• Are valves reporting one state while flow data suggests another?
• Are nighttime, low-load, or low-temperature water patterns abnormal?
• Are there recurring losses after maintenance events?
  
  

6. Water quality and treatment performance

Water efficiency is not only about flow. It is also about quality.

Poor water quality can reduce heat transfer, increase scaling, create corrosion risk, increase biological growth, force more blowdown, reduce equipment life, and compromise performance. A good water balance should incorporate treatment data, not just meter data.

Operators should monitor conductivity, pH, oxidation-reduction potential, temperature, turbidity, hardness, corrosion indicators, chemical feed activity, filtration performance, and treatment system alarms where applicable.

The water balance should answer:

• Is water chemistry within target range?
• Is chemical feed aligned with actual system conditions?
• Are treatment issues causing higher water consumption?
• Are filters, softeners, side-stream systems, or other treatment assets performing correctly?
• Is water quality affecting cooling efficiency or asset life?

7. Context: load, weather, energy, and operating mode

Water use only makes sense in context.

A spike in water consumption may be normal during a hot afternoon at high load. The same spike may be abnormal during mild weather or low utilization. Without context, operators risk chasing false positives or missing real issues.

A useful water balance should correlate water data with:

• IT load
• Facility energy use
• Cooling load
• Outdoor dry-bulb temperature
• Outdoor wet-bulb temperature
• Humidity
• Cooling mode
• Chiller status
• Tower status
• Pump status
• Valve position
• Setpoints
• Alarms
• Maintenance events
• Water source
• Water quality
• Local restrictions or curtailment conditions

This is where basic monitoring becomes operational intelligence.

The problem with water data today

Many facilities have some of this data already.

They may have utility meters, building automation systems, PLCs, cooling system controllers, treatment vendor reports, manual logs, spreadsheets, maintenance systems, DCIM tools, and sustainability reporting platforms.

The problem is that these systems are often disconnected.

One team sees water bills. Another sees cooling alarms. Another manages treatment chemistry. Another owns sustainability reporting. Another manages mechanical assets. Another manages compliance. Another manages enterprise dashboards.

When water data is fragmented, no one has the full picture.

That creates five common problems:

First, operators find issues too late. A monthly bill may show abnormal consumption, but it does not show when the problem started or which asset caused it.

Second, sustainability reporting lacks operational depth. The organization may know annual water use, but not the asset-level drivers behind it.

Third, teams cannot compare facilities consistently. Each site may use different naming, different meters, different calculations, different alarm thresholds, and different reporting methods.

Fourth, maintenance becomes reactive. Teams respond to failures instead of detecting early indicators like abnormal cycling, drift, chemistry issues, or unexplained flow.

Fifth, optimization does not scale. Even when one site develops a strong water-management process, it may be difficult to replicate across a portfolio.

This is exactly the kind of challenge industrial automation is built to solve.

Why SitePro is built for this problem

SitePro was built for dynamic infrastructure: distributed assets, real-time operations, remote control, equipment-level visibility, secure access, alarms, reporting, and operational decision-making.

That experience matters.

Data centers may be part of the technology economy, but mechanically and operationally, they are industrial environments. They depend on pumps, motors, valves, meters, sensors, treatment systems, cooling equipment, electrical systems, controls, communications, alarms, and field service workflows.

Those are not abstract software problems. They are real-world automation problems.

SitePro’s public platform is already positioned around real-time visibility, remote control, no-code configuration, reporting, APIs, integrations, and scalable operations. The company describes its platform as giving operators visibility and control across critical equipment, powering thousands of assets nationwide with 99.9% uptime. SitePro also describes configuration tools that allow users to view and manage data points from sensors, meters, power supplies, pumps, motors, cameras, and other devices, while creating logic, thresholds, and equipment changes without traditional programming dependencies.

That foundation is directly relevant to data center water management.

Because solving this problem is not just about installing another dashboard. It is about connecting field data to control logic, asset context, alarm response, reporting, and continuous improvement.

Where EdgeOS comes in

EdgeOS represents the next step in that evolution.

The challenge for data centers is not simply collecting water data. The challenge is making that data usable at scale.

A single facility may have thousands of relevant points across water, cooling, power, treatment, and environmental systems. A portfolio may have dozens or hundreds of facilities, each with different equipment, local constraints, naming conventions, control systems, and operating strategies.

EdgeOS is designed to solve that scale problem by moving intelligence closer to the asset while still connecting the full operation into a common platform.

In practical terms, that means EdgeOS can become the operational layer that helps data center teams:

• Connect field devices, meters, sensors, PLCs, controllers, and third-party systems.
• Normalize water, cooling, and asset data into a consistent model.
• Compute site-level and asset-level water balance in near real time.
• Detect abnormal water use before it becomes a monthly reporting surprise.
• Correlate water consumption with load, weather, cooling mode, and equipment status.
• Apply standard templates across facilities while still allowing site-specific configuration.
• Trigger alerts, workflows, reports, and control actions based on real operating conditions.
• Support portfolio-level visibility without losing asset-level detail.

That last point is important.

Executives need portfolio visibility. Operators need asset-level detail. Sustainability teams need credible reporting. Engineers need performance data. Maintenance teams need actionable alarms. Community and utility stakeholders need confidence that the facility is being managed responsibly.

EdgeOS is meant to connect those needs.

From water monitoring to water intelligence

The difference between monitoring and intelligence is action.

Monitoring tells you that water use increased.

Intelligence tells you that water use increased at Cooling Tower 3, during a low-load period, while wet-bulb conditions suggested the system should have remained in dry economization mode, and that the increase started after a valve position changed during a maintenance event.

That is the level of visibility operators need.

With EdgeOS, the goal is to turn water data into operational decisions:

• Is this water use expected?
• Which asset is driving it?
• Is the facility in the right operating mode?
• Is water quality forcing unnecessary blowdown?
• Is an equipment issue increasing consumption?
• Is a sensor providing bad data?
• Is a cooling strategy creating a water-energy tradeoff that leadership should understand?
• Is this site performing differently from similar sites?
• Are we trending better or worse over time?

These are the questions that matter inside the control room.

They are also the questions that matter in sustainability conversations.

A practical example: abnormal makeup water

Consider a facility with evaporative cooling.

The monthly water bill shows a 12% increase. Traditionally, the team may investigate after the fact. They may compare weather data, check tower performance, review maintenance logs, ask the treatment vendor, and look for obvious leaks.

By then, weeks have passed.

With a real-time water balance, the issue can be detected much earlier. EdgeOS can compare makeup water against expected consumption based on IT load, ambient conditions, cooling mode, blowdown, basin level, and tower operation. If the system sees makeup water increasing while load and weather remain stable, it can flag the anomaly.

Then it can add context.

Maybe blowdown increased because conductivity readings changed. Maybe conductivity readings changed because a sensor drifted. Maybe a fill valve is passing. Maybe a tower basin level is unstable. Maybe a treatment skid is behaving differently. Maybe one tower is carrying more load than the others.

The point is not just that the system sees the problem.

The point is that it helps narrow the problem.

That saves water. It saves time. It improves reliability. It gives operators confidence. And it creates a record of what happened, when it happened, and how the team responded.

Better water visibility also improves asset management

Water efficiency and asset management are deeply connected.

A pump operating outside its preferred range can waste energy and affect flow. A valve that does not fully close can waste water. A fouled heat exchanger can increase cooling demand. A failed sensor can trigger unnecessary blowdown. A poorly tuned control loop can cause unstable operation. A treatment issue can shorten equipment life and force more water use.

This is why SitePro views environmental optimization as an extension of operational excellence.

Sustainability is not separate from reliability. In many cases, it is the same discipline viewed through a different lens.

When assets are healthy, calibrated, visible, and controlled, they tend to operate more efficiently. When data is accurate, teams make better decisions. When alarms are meaningful, operators respond earlier. When configuration is standardized, best practices scale.

That is how industrial automation supports sustainability.

Not through slogans.

Through better operations.

The future is a portfolio-level water operating model

The data center industry is growing quickly. The Department of Energy reported that data centers consumed about 4.4% of total U.S. electricity in 2023 and could consume approximately 6.7% to 12% by 2028, with total data center electricity usage rising from 58 TWh in 2014 to 176 TWh in 2023 and an estimated 325 to 580 TWh by 2028.

As the industry scales, water visibility has to scale with it.

One-off dashboards will not be enough. Manual spreadsheets will not be enough. Annual sustainability reports will not be enough. Operators need a repeatable, site-aware, portfolio-ready water operating model.

That model should include:

• Standardized water-balance templates.
• Site-specific water source tracking.
• Real-time monitoring of key water assets.
• Cooling-system context.
• Water quality and treatment visibility.
• Asset health indicators.
• Anomaly detection.
• Alarm workflows.
• Automated reporting.
• Historical trending.
• Audit-ready data.
• Integration with existing automation, BAS, DCIM, maintenance, and enterprise systems.
• Secure remote access and control.
• Continuous improvement loops.

This is the type of operating model EdgeOS is being built to support.

Responsible infrastructure requires operational proof

The data center conversation will continue to be complex.

There will be debates about where facilities should be built, how utilities should plan for them, how water should be sourced, how communities should benefit, how operators should report impact, and how AI-driven infrastructure should be governed.

Those debates are important.

But inside the facility, the work is more direct.

Measure what matters.

Know where the water is going.

Understand the assets driving consumption.

Detect abnormal conditions early.

Give operators the tools to act.

Give leaders trustworthy data.

Give communities confidence that resource use is being managed with discipline.

That is the practical path forward.

At SitePro, we believe industrial automation has a central role to play in responsible infrastructure. Data centers need more than uptime. They need intelligent, transparent, scalable operations. They need systems that can connect physical assets to environmental performance. They need platforms that can turn water data into decisions.

That is why water visibility matters.

That is why the water balance matters.

And that is why we are building EdgeOS.

Because the next generation of infrastructure will not be judged only by how much capacity it creates.

It will be judged by how responsibly it operates.