A parcel can look perfect on a map and still be years away from becoming a working data center site, because the decisive question is no longer whether a location exists but whether the infrastructure around it can move at the same speed as the workload it is meant to support. The familiar sequence of fiber, customers, land and incentives once offered a practical shortcut for narrowing a market, yet that shortcut now breaks down when the electrical connection becomes the longest and least controllable part of the project. Recent research from CBRE, JLL and Cushman & Wakefield consistently points toward power availability, infrastructure delivery and resource constraints as increasingly decisive forces in data center location strategy. 

The emerging framework is therefore less a ranking of attractive real estate characteristics than a sequence of gates that a site must pass before design becomes meaningful. Power comes first because it establishes whether the project has a credible path to operation, land comes next because the powered site still needs physical room for buildings, yards, roads, substations and future expansion, water follows because cooling and local resource conditions can constrain an otherwise viable project, and people complete the stack because skilled labor and community acceptance determine whether the physical and infrastructure commitments can survive construction and operation. Fiber remains important throughout that process, particularly for latency-sensitive applications and network diversity, but its role changes from an automatic first filter to a requirement that can often be solved after a fundamentally viable site has emerged.

The Fiber-First Playbook Quietly Stopped Working

For much of the modern data center era, fiber provided a rational starting point because network density, customer proximity and interconnection ecosystems strongly influenced where operators could deliver useful services. Established hubs accumulated carriers, exchange points, cloud on-ramps, enterprise customers and technical talent, creating a reinforcing cycle in which another site inside the same geography could inherit much of the surrounding digital infrastructure. That model rewarded proximity because moving farther away from the established cluster often introduced additional network construction, latency considerations and commercial uncertainty, making the conventional site-selection exercise feel more like finding the best parcel inside a known ecosystem than discovering an entirely new geography.

Connectivity Was Once the Fastest Way to Narrow the Map

Paris illustrates the tension without requiring the market to abandon connectivity as a strategic asset, because the region remains deeply connected while new development increasingly follows areas where power and suitable land can support additional capacity. CBRE’s latest global research shows Paris continuing to expand while power constraints and development conditions are encouraging growth in emerging markets where scalable land and power are available. The lesson is not that Paris has become a poor data center location, because its network position, digital ecosystem and established market remain valuable, but that those strengths no longer guarantee that another project can be added wherever the network map looks most attractive.

The practical consequence for an end user is that connectivity should now be tested against the entire delivery path rather than treated as the first reason to choose a market. Network routes still need diversity, carrier access, low-latency paths and credible expansion options, but those questions become more productive after the project establishes that the site can receive the electrical capacity required for its intended workload and future growth. A fiber-rich location that cannot reach energization may create a beautifully connected project that never reaches production, while a less familiar market with strong transmission access can often attract network investment once a sufficiently large and credible demand base exists. The changing hierarchy also explains why the term “powered land” has become more important in market discussions, because land value increasingly reflects the infrastructure attached to it rather than the physical parcel alone.

The Market Map Is Moving Because Infrastructure Is Moving

The new geography becomes clearer when site selection is viewed through the relationship between existing hubs and the infrastructure required to extend them. A mature market can offer deep fiber, established suppliers and a familiar operating environment while simultaneously carrying a queue of proposed loads that compete for the same constrained grid resources, making another parcel inside that market less valuable than it appears from a property perspective. Secondary locations can reverse that equation when they combine available land with high-voltage infrastructure, credible utility expansion plans and a development environment that allows a project to move from control of the parcel toward construction without waiting indefinitely for the grid to catch up.

That distinction matters because the old map often treated fiber as a relatively fixed asset that determined the market boundary, whereas the new map treats infrastructure as a system that can evolve when the underlying economics support it. Connectivity, land, grid infrastructure and cooling can each adapt through different development strategies, but current market research shows that power availability increasingly determines where developers can add large-scale capacity. Power remains different because generation, transmission, substations, studies, easements, equipment and regulatory approvals can all shape the connection path, and the project sponsor often cannot control those dependencies directly. CBRE has noted that brownfield sites can attract interest when existing electrical infrastructure can shorten the path toward usable power, while JLL has highlighted transmission proximity and the long timelines associated with extending high-capacity lines to new development sites.

Why Energization Timeline Beat Location Score

A utility statement that a region has substantial generation capacity does not answer the question that matters most to a new data center user, because the project needs to know when its specific load can actually become serviceable. The distinction between available power and deliverable power has become central to site selection, particularly where large projects depend on transmission reinforcement, new substations or complex interconnection studies before the final connection can occur. The implication is straightforward but often missed in early site conversations: a theoretical power envelope is not the same thing as an energized connection, and a site with a smaller but credible near-term path can be more useful than a site promising much larger capacity at an uncertain future date.

Power Must Be Measured as a Date, Not a Number

This changes the first question a site team should ask because “How much power can this location support?” no longer captures the risk within the answer. The more revealing question becomes “What is the earliest credible date for the required load to reach service, what must happen before that date, and which dependencies remain outside the project’s control?” That inquiry forces the team to distinguish existing capacity from planned capacity, transmission proximity from transmission deliverability, utility interest from formal connection progress, and preliminary estimates from milestones that engineering and regulatory work support. It also brings equipment availability into the conversation because transformers, switchgear and other electrical components can influence construction sequences even after the grid connection receives approval.

The result is a site-selection process that increasingly resembles a schedule validation exercise rather than a conventional property comparison. A strong candidate should come with a documented understanding of the serving utility, transmission topology, substation position, interconnection pathway, required upgrades, study status, easement requirements and construction sequence, with each element tested against the project’s intended operating date. Developers that cannot explain those dependencies may still control attractive land, but they cannot yet demonstrate a credible path to live capacity, which means the site should not receive the same status as a project with verified utility milestones. CBRE’s current market research makes the hierarchy explicit by identifying power availability and infrastructure delivery timelines as major determinants of site selection, while current industry research similarly treats infrastructure availability, power access and delivery conditions as central elements of data center site selection rather than evaluating a location on land or connectivity alone.

The First Filter Is Speed to Live Capacity

Once energization becomes the first gate, the site-selection process shifts from a broad search followed by detailed diligence to a staged elimination model in which electrical evidence determines whether the team should invest further analysis. The strongest candidates are not necessarily those with the largest nominal power potential, because an enormous future capacity number offers little value when unresolved grid reinforcement still blocks access to that power. CBRE has observed developers moving toward markets where power can arrive faster, while JLL identifies speed to power as the primary site-selection criterion and notes that developers increasingly evaluate power availability and infrastructure constraints when deciding where to develop new capacity. Speed to live capacity therefore provides a more useful screening variable because it captures both physical power availability and the institutional process required to deliver service.

This framework also changes how the word “power” should appear in a site pitch, because a serious pitch needs to describe the infrastructure pathway rather than simply advertise proximity to a substation or transmission corridor. A credible power narrative should identify the serving utility, the relevant network assets, the expected connection architecture, the milestones already achieved, the remaining studies and approvals, the responsibility for network upgrades and the conditions that could move the energization date. The same narrative should distinguish between the initial operating requirement and the expansion path so that a site does not appear viable only because its first phase can connect while its later phases depend on an entirely different grid project. Power availability and delivery timing now play a central role in how prospective projects evaluate locations, making the distinction between theoretical capacity and a credible path to service increasingly important.

Cheap Land That Can’t Be Built On

Large parcels often look like the safest answer when data center projects need room for buildings, electrical infrastructure, cooling equipment, security setbacks, roads and future phases, yet gross acreage can conceal the physical conditions that determine how much of the property can actually support development. A site may contain wetlands, steep grades, flood exposure, protected areas, poor soils, fragmented ownership or access limitations that reduce the useful footprint long before the building design begins. The same problem appears when a parcel has enough room for the first building but not enough headroom for the electrical yards, cooling systems, utility corridors and expansion phases that make the project economically coherent over time. Current site-selection guidance emphasizes the interaction of available land, infrastructure access, power availability, planning conditions and the ability to support future development.

Acreage Is Not the Same as Developable Area

Grading becomes particularly important because a data center site cannot rely on boundary lines alone when large electrical and mechanical systems must cross the property. Cut-and-fill work, drainage, soil improvement, retaining structures and construction access can alter both the schedule and the cost of a site, while future expansion corridors can reduce the land available for the first phase. Utility corridors create another hidden constraint because transmission, water, gas, communications and road access each require space that a basic land-use calculation may overlook, and relocating those corridors can become difficult once detailed engineering begins. A parcel that appears inexpensive per acre can therefore become expensive per usable acre when the project excludes land that cannot support foundations, infrastructure or future expansion.

The right land question therefore sounds less like “How many acres are available?” and more like “How much of the controlled property remains usable after the full infrastructure stack occupies it?” That question should drive early mapping of setbacks, flood conditions, drainage, road geometry, substation placement, generation or storage areas, cooling infrastructure, water systems, fiber routes and expansion reserves before the team gives the land a final ranking. It also makes aggregation strategy more important because multiple adjacent parcels can create a more coherent development platform than a single large tract with internal constraints, while fragmented ownership can create another form of schedule risk even when the total land area looks sufficient.

Secondary Markets Win When the Land Actually Works

Secondary markets are becoming more interesting because they can offer a combination that mature hubs increasingly struggle to assemble: meaningful parcels, infrastructure access and enough physical room for phased development. The attraction is not simply a lower purchase price, because cheap land without power, water, access or permitting certainty does not solve the user’s underlying problem. The equation changes when developers find land where major infrastructure corridors already exist or can extend them without turning the project into a multi-year exercise in assembling unrelated dependencies. JLL and CBRE both identify emerging and secondary markets as beneficiaries of constraints in established hubs, particularly where those markets offer available power and scalable land. The shift matters because it changes the definition of a “good market” from one with a strong existing reputation to one that can assemble the full physical platform required by the workload.

Brownfield opportunities show why the distinction matters, because a previously developed industrial property can sometimes provide roads, utility corridors, electrical infrastructure, established access and a planning history that a vacant parcel cannot offer. That does not make every brownfield site suitable, since contamination, legacy structures, title conditions and redevelopment constraints can create their own risks, but the existing infrastructure can reduce the number of unknowns that typically accompany a greenfield project. Brownfield and previously developed sites can offer advantages when existing infrastructure, utility access or established development conditions reduce some of the barriers associated with new development. The important point is that secondary-market value often comes from infrastructure readiness rather than isolation, and the best sites may sit near existing industrial or energy infrastructure instead of inside a conventional technology cluster. 

Water Was a Footnote. Now It Vetoes Deals.

Water used to appear in many site discussions as a sustainability consideration that teams could address after major real estate and electrical decisions took shape, but higher-density computing has made cooling architecture inseparable from site feasibility. A data center’s water requirement depends on its cooling system, climate, operating profile and local resource conditions, so the site team cannot assess water needs accurately without understanding how the technical design will reject heat. Uptime Institute’s research finds that local climate, cooling conditions and heat-rejection technology shape a data center’s water-use profile, making water assessment inherently site-specific. That finding changes the timing of diligence because the team cannot treat water as a sustainability detail after a parcel reaches the shortlist.

Cooling Has Turned Water Into a Front-End Constraint

The technical evaluation begins with the source, but it cannot stop there because the project also needs to understand treatment, reliability, discharge, seasonal conditions and the relationship between water demand and local resource constraints. A municipal connection may provide a straightforward source in one location while another project may require reclaimed water, alternative supply arrangements or a design that minimizes or eliminates routine cooling water consumption, and each pathway carries different infrastructure and permitting implications. Uptime Institute notes that cooling choices can materially change a data center’s water profile, with dry or low-water approaches becoming more relevant in water-stressed locations. The same principle applies to discharge because removing heat from a system does not eliminate the need to understand where process water goes, what quality requirements apply and whether the receiving system can accept the proposed discharge.

Community conditions add another layer because a water allocation can appear technically available while households, agriculture, industry or environmental requirements compete for the same resource and influence the project’s long-term acceptability. Uptime Institute’s work emphasizes that water use can attract community criticism and that local conditions determine the impact of a data center’s water profile, reinforcing the need to understand the resource in its local context. Recent industry research shows that governments and communities increasingly examine data centers through the combined lens of energy demand, water use and local infrastructure impacts. For an end user, this means the question should not stop at whether a utility can provide water on the opening date, because a durable site also needs a defensible long-term resource position that can withstand changing operating requirements and local scrutiny.

The Water Test Now Includes Source, Discharge and Consent

A modern water review should begin before the team locks the final cooling architecture because the technology choice can materially alter the project’s resource profile, capital requirements and relationship with the local water system. The site team needs to identify the actual water source, determine whether the utility can provide firm or conditional capacity, map the infrastructure needed to deliver it and assess whether the resource remains viable under the project’s operating conditions. The team should give discharge equal attention because treatment requirements, receiving systems and local environmental rules can impose constraints that a simple water-availability statement may overlook. Uptime Institute’s guidance makes clear that local conditions and the selected heat-rejection system shape water use, making cooling technology an important part of early site assessment.

The diligence also needs to distinguish between physical water availability and social permission to use that water, because a project can encounter resistance even when the engineering solution appears straightforward. Local communities increasingly examine how large computing projects affect shared resources, and that scrutiny can influence planning decisions, operating conditions and the broader willingness to support new development. Current industry research and reporting show that community concerns over energy, water, infrastructure and local impacts can influence the development environment for data centers. This means the water narrative should explain not only how the cooling system works but also why the selected approach fits the local resource environment, what safeguards exist and how the project will manage resource use over its operating life.

The People Layer No Model Accounted For

A site can clear the power, land and water gates and still struggle if the surrounding labor market cannot support the construction and operating demands that follow, because infrastructure does not become reliable simply because teams permit and install the physical systems. The workforce question extends beyond the technicians who eventually operate a completed environment, because data center development and operation require specialized technical skills across construction, engineering, commissioning and ongoing maintenance. Lightcast’s 2026 workforce research identifies workforce availability as a potential deciding factor in whether data center projects can succeed after teams address land, power and other development challenges.

Workforce Depth Is Now Part of Infrastructure Readiness

The workforce layer also changes how a site team should judge a location because a large labor pool in a nearby metropolitan area does not necessarily provide the specialist skills that high-density computing infrastructure requires. A project may need local contractors who can work around high-voltage systems, advanced cooling equipment, automation controls, commissioning procedures and tightly sequenced construction schedules, while the operating phase requires personnel who can maintain electrical, mechanical and network systems under demanding reliability requirements. Recent workforce research identifies labor availability as a potential constraint on data center growth, reinforcing the need to evaluate local workforce conditions alongside land, power and community factors.

Community acceptance creates the other half of the people layer because a technically feasible project still depends on whether residents, local authorities and surrounding businesses accept the physical changes that accompany large-scale infrastructure. Current industry research and reporting show that community concerns can center on water, electricity, land use, infrastructure and other local impacts, making community conditions an important part of data center development. The issue is becoming more visible as local governments introduce conditions around resource use and community benefits, while developers increasingly need to demonstrate how a project fits into the existing economy and infrastructure system. 

Acceptance Determines Whether the Other Three Hold

Acceptance does not mean that every resident must welcome a project, because infrastructure decisions naturally involve competing interests, but it does mean that the project needs a credible relationship with the place where it will operate. The strongest approach starts before final land commitment by identifying the issues most likely to shape local concern, including water demand, electricity infrastructure, traffic, construction disruption, land conversion, environmental effects and the relationship between private development and shared resources. Recent regulatory activity in several jurisdictions shows governments responding to concerns about data center electricity demand, water use and local infrastructure impacts.

The people layer also affects the credibility of a project’s expansion plan because workforce and community conditions can change the pace at which later phases become possible. A developer that can build an initial environment but cannot secure enough specialist contractors for subsequent construction phases may find that theoretical land capacity does not translate into usable capacity, while an operator that has no durable local labor strategy may face recurring recruitment and retention challenges after commissioning. Uptime Institute identifies staffing challenges alongside power constraints, supply-chain delays and the operational demands created by AI workloads, while Lightcast identifies workforce availability as a potential constraint on data center growth.

The Broker Became a Power Scout

The traditional site broker earned value by understanding land markets, ownership structures, incentives, zoning, development patterns and connectivity, but the modern shortlist requires a much deeper understanding of infrastructure dependencies. A broker working on a serious data center search now needs to know which utility serves a parcel, where transmission infrastructure sits, whether a substation can support the intended development concept and what evidence exists behind claims about future electrical service. CBRE’s 2026 research places power availability and scalable infrastructure among the factors reshaping data center growth, increasing the importance of infrastructure diligence within site-selection work.

Real Estate Knowledge No Longer Stops at the Property Line

That change makes verification more important than presentation because a polished site package can conceal the difference between land that sits near a transmission corridor and land with a credible path to connection. The broker increasingly needs to trace utility filings, understand the status of interconnection work, identify land aggregation risks, review access routes and recognize whether water infrastructure can support the intended cooling strategy. Research published for Wisconsin’s data center market similarly identifies power, connectivity, water, buildable land and workforce as interconnected siting considerations, reinforcing the need for a broader diligence process than conventional commercial property analysis provides. 

The most useful broker therefore becomes the person who can identify infrastructure risk before the engineering team spends weeks proving it, because early elimination has become as valuable as early discovery. That requires enough technical fluency to challenge vague phrases such as “power nearby,” “water available” or “fiber at the property,” while still knowing when a question requires utility engineers, environmental specialists or legal diligence. The broker does not need to replace those specialists, but the broker increasingly needs to know which questions to ask before a parcel enters the serious shortlist. 

The New Shortlist Is Built From Evidence

A stronger brokerage process begins by building an infrastructure evidence file alongside the property file, with utility information, land control, water conditions, road access, network routes and development constraints linked to the same parcel. That file allows a prospective user to distinguish between infrastructure that exists today, infrastructure that has received meaningful commitments and infrastructure that remains dependent on future decisions. The distinction is especially important for power because speculative applications can create uncertainty in interconnection queues, while developers increasingly seek earlier clarity around grid access to protect project schedules. 

The broker also needs to understand land aggregation as an infrastructure problem rather than a property transaction, because the final site may require several parcels to create enough room for electrical yards, roads, cooling systems, water infrastructure and expansion. Fragmented ownership can become a schedule risk when one parcel controls a necessary access route or utility corridor, while an apparently large tract can lose value when environmental or physical constraints divide the usable area into disconnected pieces. CBRE’s current research identifies power availability, scalable land and broader development conditions as factors shaping where new data center capacity can be added.

Developers Thinking Like Utilities, Utilities Thinking Like Developers

The previous development model allowed the developer to control the site, advance the design and then coordinate with the utility as a largely separate infrastructure workstream, but that separation becomes difficult when electrical availability determines whether the project should exist in the first place. Developers increasingly need to understand grid conditions before land commitment, while utilities need better visibility into the scale, timing and operating characteristics of prospective loads before they can plan infrastructure efficiently. The current evidence shows this convergence clearly, with data center developers pursuing early grid interconnects, power agreements and alternative supply strategies while utilities and governments confront growing requests for new electrical capacity. 

The Old Handoff Model Cannot Carry the New Load

The developer’s role consequently expands from securing a parcel to constructing a credible infrastructure narrative around that parcel, including the relationship between the site’s electrical demand, cooling architecture, water requirements, expansion sequence and local development conditions. That narrative needs to survive scrutiny from utilities, regulators, financiers, communities and eventually the user who will depend on the completed capacity. Current industry research identifies power availability, land, water, workforce, permitting and construction conditions as factors that can constrain data center development.

Utilities face a corresponding change because large data center loads can no longer be treated as ordinary incremental demand that simply arrives after the property decision has already been made. The utility needs to understand load timing, ramp profiles, reliability expectations, redundancy requirements and the potential for phased development, while the developer needs enough visibility into the utility’s infrastructure plans to determine whether a site can support its commercial schedule. The relationship increasingly requires earlier coordination between developers and utilities so that electrical requirements and project-development schedules can be evaluated together.

Power Narratives Are Becoming Part of Development Strategy

The new developer-utility relationship does not mean every developer should attempt to operate as a utility, because grid planning, generation, transmission and distribution remain specialized functions with distinct regulatory responsibilities. It means the developer needs enough understanding of those systems to build a project around realistic infrastructure conditions rather than assume that electrical service will appear after the building is designed. Current market activity includes projects pursuing alternative power arrangements when conventional grid access cannot align with the desired development schedule.

Utilities are increasingly central to development discussions because large data center loads can affect grid planning, infrastructure investment and the timing of new connections. The most productive relationship starts with transparent information about what the project needs, when it needs it and how the load could evolve, allowing the utility to assess network implications while the developer tests whether the connection path fits the project schedule. That approach can also surface alternatives earlier, including phased energization, different connection points, demand-management strategies or other infrastructure arrangements that might otherwise emerge only after the site has already been selected. 

Design Only Starts After The Stack Says Yes

The new site-selection stack works because it recognizes that not every site deserves to reach the design stage, and that the fastest way to protect a project is to eliminate weak candidates before engineering effort accumulates around them. Power comes first because without a credible path to energization there is no operating schedule, land comes next because the powered location must physically accommodate the complete infrastructure system, water follows because cooling and resource conditions can constrain the technical design, and people complete the sequence because construction, operation and community acceptance determine whether the infrastructure can remain viable. This is not a claim that fiber has become unimportant, since network diversity and latency still matter deeply for many workloads, but it does mean that connectivity should be evaluated inside a site that has already demonstrated basic physical and infrastructure feasibility.

Power, Land, Water and People Form a Sequential Gate

The sequence also creates a better decision structure for users because it replaces a broad scorecard with a series of questions that have clear consequences. If the power path fails, the site stops; if the land cannot support the infrastructure concept and expansion plan, the site stops; if water cannot support the cooling strategy and local resource conditions, the site stops; and if the workforce or community environment cannot support construction and operation, the site should not advance simply because the earlier gates looked attractive. This approach reduces the temptation to compensate for one weakness with another, because inexpensive land cannot compensate for an unusable grid connection and strong fiber cannot compensate for a cooling system that lacks a defensible water strategy.

The final test is whether all four layers can reach readiness on the same timeline, because a project does not become operational when its strongest component is ready but when its slowest essential dependency stops holding it back. That principle changes how the industry should talk about “site readiness,” because readiness must mean more than controlled acreage, a favorable utility conversation or a network route at the property boundary. The useful site is the one where the evidence supports a coherent path from infrastructure commitment through construction and commissioning to dependable operation, with enough flexibility to absorb changes in workload, cooling technology and local conditions. Uptime Institute identifies power constraints, staffing challenges and the operational demands associated with AI among the pressures facing data center operators, while Lightcast identifies workforce availability as a potential constraint on future data center growth.

The Winning Market Is the One That Moves Together

The next generation of data center markets will not be defined only by how much land they can advertise or how close they sit to an established network cluster, because users ultimately need capacity that arrives as a working system rather than as four separate infrastructure promises. A market becomes strategically useful when its electrical system can support a credible energization path, its land can accommodate the physical platform, its water conditions can support an appropriate cooling strategy and its people can construct, operate and accept the resulting infrastructure. That combination is harder to assemble than a conventional property shortlist, but it creates a much more meaningful basis for deciding where capital and computing demand should go.

Design begins only after the stack says yes, because architecture cannot solve a site that lacks the infrastructure required to make the architecture useful. The strongest development strategy therefore starts outside the building envelope, tracing the electrical connection, usable land, cooling resource and human environment before translating those conditions into a physical design. That approach does not eliminate uncertainty, but it moves uncertainty into the part of the process where it can still change the site rather than forcing the user to redesign the project around a site that should never have advanced. As AI and other high-density workloads reshape infrastructure requirements, markets that can combine power availability, scalable land, suitable cooling conditions and workforce capacity are better positioned to support new data center development.