Data centers: How to make better location decisions for long-term viability

Data center location decisions set your operating limits and shape your uptime, operating cost and long-term asset value.

Your shortlist may look workable at first, with sites offering available power and water, but the real constraints may only emerge later as grid capacity tightens and water stress increases as temperatures rise just as chip technology demands greater cooling. By this stage, you could be paying for workarounds for risks you can have screened out early in the location planning stage.

In this insight, based on our multidisciplinary Outsmarting Uncertainty webinar on data centers, we break down exactly how you can make better, more informed decisions on data center location, thinking about the outlook for power, water and more.

How do you test data center power availability and grid resilience?

A grid connection alone tells you little about how a site will perform under stress.

To make more informed location decisions for long-term viability, ask the following questions:

• Where does the power come from? Identify the substations and transmission routes serving the site and test how they perform under credible hazard scenarios.
• How will grid stress change? Demand growth and rising temperatures often hit together. Hotter weather increases cooling load, which raises power demand when the grid may already be under pressure. That increases both cost and outage risk.
• How independent is the power supply across a cluster? Sites may look separate on a map but still share a substation, transmission corridor or constrained utility region, creating correlated outage risk.

How should cooling and water availability shape site selection for data centers?

Cooling. This is a direct test of long-term site viability because it links local climate, power constraints, water availability and community tolerance.

That means you need to match the cooling approach to local conditions rather than defaulting to a standard design, then test performance under hotter scenarios.

Water. This is often the harder constraint. Your site may secure power and still struggle to obtain water at the scale and reliability required. It may also face tighter permits, restrictions, higher tariffs or stronger community scrutiny over time, all of which will affect your operations and your chances of expanding them.

We would recommend you define operating thresholds, assess how many hours a year the site already runs close to them and project that forward using climate data over the expected ownership period.

You should also track measures such as power usage effectiveness and water usage effectiveness to see when performance starts to drift from your original assumptions.

How do heat and drought affect data center uptime, cost and asset value?

Heat increases cooling demand, energy use and, in many configurations, water consumption, raising operating costs and reducing margins.

Drought can lead to restrictions, tougher permitting, higher costs and reputational pressure in already stressed communities. Such factors can limit operations and expansion even without direct physical damage.

This means you need to screen heat and drought as rigorously as power capacity.

What does ‘shared fate’ mean for data center clusters and redundancy planning?

Shared fate is where two or more sites fail together because they depend on the same infrastructure or sit within the same event footprint. Shared fate is common to clusters of data centers.

Operators may risk building redundancy across locations, then finding the sites share upstream power, access routes, water infrastructure or telecom corridors, meaning a single disruption can hit more than one site at once.

To avoid this outcome, you should run scenarios across the cluster, not just site by site. Include substations, transmission routes, access roads, telecom routes, water treatment and fuel supply.

How do you evaluate data center dependencies beyond the fence line?

Map the dependencies required to keep the site running and recover operations, then test them under realistic scenarios.

That includes external power supply and upstream infrastructure, water supply and treatment, telecom routes, access roads and logistics routes, vendor support and spare parts pathways.

Next, test how those dependencies behave during regional disruption. Identify what fails first, what follows and what slows recovery.

A dependency map can give you more options early, such as changing the site, revising design assumptions or investing in targeted mitigations before committing capital.

How do permitting, planning and community acceptance affect data center schedule and viability?

Permitting and community acceptance often determine delivery timelines and future expansion. Power demand, water demand, land use, emissions and noise can all attract scrutiny that can manifest as tighter conditions, longer approvals and limits on growth.

By bringing in local insight early and assessing them consistently across regions, you give your project a higher chance of avoiding delays, redesigns and long-term operating ceilings you didn’t plan for.

How do you factor decommissioning and end-of-life obligations into site selection?

Decommissioning requirements, remediation obligations and exit constraints can create liabilities that rarely feature in the initial build discussion, so include these checks in early due diligence. That way, you can understand exit requirements and any barriers to redevelopment and ensure you’re making decisions informed by a view of the full asset lifecycle.

To get a more connected view of the risks to your data center’s long-term performance, get in touch with our team of specialists.

Find out more, watch the Outsmarting Uncertainty data center webinar.