Planning Data Center Growth Around Power, Water, and Supply Constraints
For years, data center planning has largely been an exercise in solving for capacity. Operators needed more compute, so they added servers, racks, buildings, and campuses. As rack densities increased, cooling systems changed to handle the additional heat. Each technical problem created an engineering response.
That process is becoming more complicated because some of the constraints on data center growth now sit outside the building. A site may have enough land but limited access to power. A cooling design may work as intended but place demands on a water supply shared with a growing community. Equipment specified for one project may be difficult to source in the quantities needed for the next five. A design that can be repeated quickly in one region may encounter different utility, environmental, or regulatory conditions somewhere else.
The result is a different kind of planning problem. Data center developers still have to solve the technical requirements inside the facility, but they also have to understand the limits of the place, supply network, and public infrastructure around it. As development accelerates, solving one constraint can simply move pressure somewhere else.
How Is Faster Data Center Growth Changing Infrastructure Design?
Faster data center growth is driving demand for modular, turnkey, pre-engineered systems that reduce on-site engineering and construction work and can be deployed across multiple locations.
Engineering each project from scratch becomes harder to sustain when similar capacity is needed across campuses and regions. Repeatable designs let teams carry established decisions into the next project rather than revisit them for every build.
But utility capacity, climate, water resources, regulations, and construction conditions vary by location. A system that works well in one region may need adjustments before it can be used in another.
Teams therefore need to distinguish between design choices they can carry forward and those that require a fresh site assessment. That lets them retain the speed of a repeatable approach without locking the next facility into assumptions that do not fit its location.
How Are Supply Chain Constraints Affecting Data Center Design?
Supply chain constraints are pushing data center teams to assess equipment availability, supplier stability, and regional sourcing options before designs are finalized.
Equipment must meet performance, cost, and reliability targets, but a large deployment program also needs confidence that enough units can be delivered across multiple projects. Shortages or long lead times can force changes after engineering decisions have been made, especially when a design depends on components with few qualified alternatives.
Supplier selection is part of that design decision. Partners with stable supply chains and global manufacturing and distribution footprints can offer more ways to keep equipment moving when geopolitical uncertainty disrupts trade or transportation. Access to qualified equipment from multiple regions can reduce dependence on a single country or shipping route.
A global footprint alone does not guarantee continuity. Teams need to verify which alternative locations can supply the specified equipment and whether those options meet project requirements. Establishing those alternatives before a disruption gives developers more room to respond without starting a new round of qualification and redesign.
That support needs to extend beyond construction. Operators will eventually need maintenance parts and replacement equipment in each market where they operate, along with workable alternatives when the original specification is no longer practical.
Evaluating those capabilities early helps teams choose specifications they can support across repeated builds and years of operation, rather than relying on availability at the time of the first order.
Why Should Data Centers Evaluate Water Use at the Watershed Level?
Data centers should evaluate water use at the watershed level because local supply, competing demand, and downstream water quality determine how much water a site can draw and how it should manage what leaves.
Facility efficiency figures alone do not capture seasonal shortages, municipal capacity, or demand from homes, agriculture, and other businesses. A site water plan should account for those conditions and identify where capture and reuse can reduce withdrawals. Reclaimed water and captured roof runoff may support non-potable uses, including irrigation or cooling tower makeup, where availability, storage, treatment, and local rules allow.
Retention ponds can help manage stormwater by slowing its release and allowing sediment and some pollutants to settle. Their design must fit local rainfall, available land, maintenance needs, and the receiving waterway, including potential temperature effects. Before runoff or used process water returns to the watershed, treatment and monitoring should address its quality and applicable discharge requirements; a pond alone may not be enough.
How Should Data Centers Balance Water, Energy, Cost, and Reliability?
Data centers should balance water, energy, cost, and reliability by comparing designs against the site’s limiting resources and assessing what each option requires to build, operate, and adapt.
Reducing water consumption may change energy use or cost. A customized system may suit local conditions but take longer to engineer and build, while a standardized design may deploy faster but offer less room for site-specific adjustments. Flexibility for future equipment can also raise upfront costs. Each option needs to be assessed against the same site conditions and operating requirements, rather than judged by its best-performing measure.
The practical test is whether a design relieves the site’s main constraint without creating another barrier to operation or expansion. Where water availability limits growth, lower consumption needs to be weighed against the resulting power demand, cost, and reliability requirements. The takeaway is to make that exchange explicit before committing to a design—not treat an improvement in one metric as an improvement in the whole facility.
How Can Data Centers Plan for Changing Cooling and Compute Requirements?
Data centers can prepare for changing cooling and compute requirements by avoiding design choices that make future equipment changes unnecessarily difficult or expensive.
Uncertainty also exists inside the data center. Facilities built today will operate through multiple generations of computing hardware, while the cooling systems used to support that equipment continue to change. Cold-plate and direct liquid cooling are growing alongside other approaches such as two-phase and immersion cooling.
Trying to predict which approach will dominate years from now is unlikely to be a useful design strategy. Neither is assuming today's requirements will remain unchanged for the life of the building.
A more practical approach is to consider where future changes would be expensive or disruptive. Can equipment be replaced without redesigning a much larger system? Can capacity be added as loads increase? Can specifications accommodate more than one source of supply? If future cooling requirements change, which parts of the supporting infrastructure would have to change with them?
There are limits to how much flexibility can reasonably be built into a project. Designing for every possible future would add its own cost and complexity. The objective is simpler: avoid making today's assumptions unnecessarily difficult to change.
How Should Data Centers Plan for Future Infrastructure Constraints?
Data centers should plan for future infrastructure constraints by looking beyond current technical requirements to the limits that could emerge from power, water, equipment supply, local infrastructure, or future expansion.
There probably will not be one bottleneck that defines the next decade of data center development. Different projects will run into different limits.
Power may determine what can be built in one market. Water conditions may influence cooling and site decisions in another. Equipment availability may slow a project even when land and utilities are available. Local infrastructure or permitting may determine how far a campus can expand. Technology changes may alter requirements before a facility reaches the end of its planned life.
Those conditions also change. A site selected under today's assumptions may operate for decades, while the infrastructure and community around it continue to develop.
That changes the job of planning for data center growth. Solving the current engineering problem is only part of it. Owners and the companies supporting them also need to understand what their decisions require from suppliers, utilities, watersheds, and communities, and how those requirements could change as capacity expands.
The next constraint will not necessarily be where the last one was. Designing with that in mind gives operators more room to respond when it moves.
About the Author
General Manager, Data Centers - Watts®
With over 30 years leadership experience in the data center and critical infrastructure industries, Kevin leads go-to-market strategy, strategic partnerships, and sales execution focused on enabling resilient, efficient, and scalable infrastructure for next-generation digital facilities. Throughout his career, he has led global sales and marketing organizations, built large-scale partner ecosystems, and helped launch managed services and remote monitoring solutions that transformed how critical infrastructure is operated and maintained.