Balancing AI Data Center Demand with Community and Environmental Realities

Data center development has become one of the most contested issues in local planning across the United States. A tech company announces a major facility, and not long after, community groups raise concerns about water use, electricity demand, environmental impact, and neighborhood disruption. Meanwhile, demand keeps growing.

The tension isn’t new, but it has intensified as AI-related computing needs have accelerated. What makes the current moment notable is how much of that tension traces back to a single variable: where a facility gets sited. Location decisions shape nearly every aspect of the debate—which concerns apply, how severe they are, and what tradeoffs actually exist.

The Scale of Demand

AI runs on computation at a scale most other technologies never required, and data centers supply the infrastructure. That places them in the same foundational category as power grids and telecommunications networks.

Data center build-out is a global phenomenon. China, India, and other countries are expanding data center capacity as a strategic priority, treating computing infrastructure as an economic and national asset. 

American technology companies need domestic infrastructure to remain competitive, and dependence on foreign facilities carries its own risks. The result is that development pressure is high and not especially sensitive to local opposition in any single market. When a project doesn't move forward in one location, demand tends to resurface somewhere else.

The relevant question is often less about whether development occurs and more about the specific conditions under which it does.

Valid Community Concerns Over Data Centers

Opposition to data center development isn’t uniform, and it isn’t uniformly unfounded. The concerns that come up consistently have substance.

Water is among the most cited. Modern data centers use roughly 0.5 to 1 liter of water per kilowatt-hour. In water-stressed regions such as parts of Arizona, New Mexico, and other arid states, that consumption competes with existing agricultural, municipal, and ecological demand.

Other well-grounded concerns include housing cost pressures and displacement, construction-phase impacts like traffic and noise, and energy consumption that accounts for 2–3% of total U.S. electricity use. Then there are questions about labor practices, wages, and corporate accountability. Even highly efficient facilities—power usage effectiveness ratios now hover around 1.1 to 1.2—represent significant demand on local resources. These are real tradeoffs that planning processes are designed to evaluate.

Concerns That Don’t Hold Up Under Scrutiny

Several concerns that circulate in public debate don’t hold up well under scrutiny.

Water

The consumption figures most commonly cited—roughly 0.5 to 1 liter per kilowatt-hour—reflect older evaporative cooling systems that continuously consume fresh water. Modern closed-loop cooling systems recirculate the same water through a sealed circuit rather than evaporating it, reducing consumption by 80–95% compared to traditional designs. Some advanced implementations approach near-zero freshwater consumption by utilizing recycled or nonpotable water sources. As closed-loop adoption has grown, the water intensity of new data center development has dropped substantially. As a result, water availability has become a far more manageable siting consideration than older figures suggest, particularly when facility design requirements are built into the approval process.

Electricity prices

The assumption that large data center loads raise utility costs runs counter to how electricity markets typically function. Data centers operate with steady, continuous demand rather than the variable peaks that characterize residential consumption. That predictability allows utilities to run more efficiently, and long-term power purchase agreements can fund infrastructure improvements. Analyses of regions with significant data center presence have generally found electricity prices stable or lower relative to comparable regions without it.

Grid capacity

Concerns about data centers exhausting available grid capacity reflect a misunderstanding of utility planning. Formal capacity planning accounts for all anticipated demand, and large industrial customers typically go through interconnection processes that include infrastructure investment requirements.

Local heat effects

The claim that operations raise ambient community temperatures by several degrees doesn't match the mechanics of thermal discharge. Waste heat is concentrated at cooling exhaust points and dissipates quickly. Facility design and regulatory requirements can manage thermal impacts to highly localized areas.

Global displacement

When a project doesn’t move forward in the U.S., the computing demand doesn't disappear. It migrates elsewhere. China, for example, is building hundreds of coal-fired plants to support data center expansion along with dozens of nuclear plants. Facilities built under weaker environmental standards can produce considerably higher emissions than comparable American facilities operating under renewable or clean energy requirements.

The Energy Picture

Natural gas is the most commonly cited near-term solution for large baseload requirements. It produces roughly half the CO₂ per unit of energy compared to coal, with significantly lower particulate emissions. U.S. shale reserves provide a stable long-term supply, and natural gas plants deploy relatively quickly while supporting grid balancing functions.

Small modular reactors (SMRs) represent a longer-horizon option attracting serious attention. SMR technology can deliver carbon-free baseload power at scale in modular configurations co-located near large facilities, requiring substantially less water than conventional nuclear plants. Several companies are advancing commercial deployment programs—though cost trajectories are increasing, fuel supply chains are still maturing, licensing processes remain uncertain, and community acceptance isn’t guaranteed. SMRs are a credible part of the long-term energy mix, but timeline and scalability remain open questions.

An emerging model pairs dedicated generation—behind-the-meter and in-front-of-the-meter—with interconnection through independent system operators, allowing excess capacity to flow into the broader grid with stabilizing effects on regional supply.

What Negotiated Outcomes Look Like

Projects from Meta, Google, Apple, and Microsoft have produced a body of case studies in negotiated development. Common elements in projects that moved forward with community support include the following:

• Renewable or clean energy commitments
• Water conservation measures
• Environmental impact assessments with enforcement provisions
• Thermal monitoring
• Community benefit agreements
• Labor and wage standards with local hiring targets
• Ongoing transparency and reporting

Outcomes tend to be better when engagement begins early, and commitments are structured into the project rather than addressed after opposition has formed. Communities that engage in negotiation have generally secured more concrete protections than those whose primary posture was opposition. Underlying all of it is site selection. The choice of location determines which concerns are most acute, which mitigation measures are most relevant, and what the baseline tradeoffs look like before any negotiations begin.

The Economic Dimension: Tax Revenue and Jobs

Data center development generates substantial economic benefits through real and personal property tax revenue and significant job creation—impacts that can be transformational in regions facing economic transition or declining tax bases.

Tax Revenue Impact

Real property taxes on structures and personal property taxes on computing equipment create substantial revenue given the capital-intensive nature of operations. A typical large-scale facility represents hundreds of millions, if not billions, in investment, translating to significant annual contributions to funding schools, infrastructure, emergency services, and community programs.

A representative range of outcomes:

• Northern Virginia: Fairfax, Arlington, and Loudoun Counties have collected billions in cumulative tax revenue over two decades, with operational facilities providing roughly $300–500 million annually.
• Arizona: Multiple data center projects generated between 2021–2023 approximately $2.3 billion in combined local and state revenues, over 500 permanent jobs averaging $70,000–$90,000 in annual salaries, plus 3,000–4,000 construction jobs at peak.
• Columbus, Ohio: Recent projects contributed over $500 million in cumulative revenue and created approximately 1,200 permanent and 2,500 construction jobs, significantly benefiting local schools.
• Prineville, Oregon: Google and Facebook facilities contributed approximately $300 million since 2013, stabilizing a rural tax base with roughly 2,000 construction jobs and 600 permanent positions.
• Richland Parish, Louisiana (HyperionAI): Tax revenue will result in each district teacher receiving a one-time $50,935 bonus, transforming the fiscal capacity of a rural school system while creating approximately 300 permanent jobs.
• Mississippi (Amazon): With $25 billion in capital investment over two years, Amazon is providing high-tech, high-paying employment. Governor Tate Reeves stated: “Amazon isn't just reinvesting in Mississippi. The company is once again betting on our people. ... This is what transformational economic growth looks like.”

Construction Employment

Major builds typically create 2,000–5,000 temporary jobs over 2–3 years—construction trades, equipment operators, specialized installers and engineers—at average wages of $50,000–$80,000 annually, with local suppliers and service businesses benefiting as well.

Ongoing Operations

A mid-sized facility typically employs 150–300 permanent staff—facility managers, technicians, network engineers, security, and maintenance specialists—with competitive wages ($45,000–$85,000 for technicians, higher for management) and long-term career stability.

Multiplier Effects

Every permanent job generates an estimated 0.8–1.2 additional jobs, while each construction dollar generates $1.50–$2.00 in secondary economic activity.

Infrastructure Investment and Long-Term Stabilization

Developers frequently fund road upgrades, utility expansion, and workforce training partnerships. For communities facing economic decline, the combination of sustained tax revenue, direct employment, and induced activity can create a stable long-term base, as Prineville, Richland Parish, and Mississippi have each demonstrated.

The Incentive Landscape

State-level incentives are in flux, and the divergence between states is widening.

States reducing or eliminating incentives:

• New York scaled back programs amid fiscal pressure and skepticism about corporate subsidies. 
• Massachusetts shortened property tax abatement periods, questioning whether data centers generate sufficient permanent employment to justify them. 
• Minnesota formally reviewed its sales tax exemption’s cost-effectiveness. 
• Vermont eliminated several programs, redirecting investment toward other priorities.

States maintaining or expanding incentives: 

• Texas continues offering aggressive programs, including Chapter 313 successor agreements and property tax abatements, leveraging its deregulated electricity market and available land. 
• Virginia maintains a comprehensive framework of tax exemptions, workforce funding, and tax credits. 
• Ohio has strengthened its incentive package through Enterprise Zone and Community Reinvestment Area programs, attracting multiple large investments. 
• Iowa’s sales tax exemption on equipment has drawn Microsoft, Google, Meta and Apple.

This divergence affects project economics considerably. The incentive environment is now one of the more consequential variables in location analysis, requiring current, state-specific evaluation rather than general assumptions.

The Sources of Opposition

Opposition draws from several distinct perspectives worth understanding separately.

Environmental opposition is grounded in documented history: Corporate environmental wrongdoing is well-recorded, and regulations exist largely because voluntary compliance proved insufficient. This perspective tends to produce sustained, organized opposition that responds to concrete, enforceable commitments rather than general assurances.

Concerns about corporate power and labor practices reflect broader skepticism about how large technology companies treat workers and communities, showing up in debates about wages, working conditions, and the distribution of economic benefits.

Local opposition is often the most immediate and varied. It may encompass concerns about neighborhood character, infrastructure strain, housing costs, and pace of change. It’s highly site-specific and shaped by local political dynamics. These perspectives frequently overlap, and projects encountering opposition are often navigating all three simultaneously.

Where Site Selection Sits in All of This

Location is among the most determinative variables in how a data center project unfolds. It shapes which concerns are most relevant, what mitigation is feasible, what incentives apply, and what the long-term operating environment looks like.

Effective site selection involves a multi-variable analysis: power availability and proximity to existing transmission infrastructure; water resource availability relative to regional capacity; distance from environmentally sensitive areas; alignment with industrial zoning; workforce availability; and the current state of incentive programs. 

In today's market where available power is constrained, community scrutiny is high and the incentive landscape shifts constantly—that analysis has become considerably more complex than it was even five years ago.

Site Selection Group is built for this. 

With decades of data center experience and an extensive network spanning utilities, economic development organizations, grid operators, data center and infrastructure developers, and state and local governments, Site Selection Group carries a depth of market intelligence that’s difficult to replicate. In a market where the most viable opportunities are often identified before they’re publicly visible, those relationships matter.

The work spans the full lifecycle of power and grid analysis, water resource assessment, workforce mapping, zoning and environmental screening, and incentive navigation across a complex, shifting state-by-state landscape. 

Beyond technical analysis, Site Selection Group supports the stakeholder engagement that keeps projects moving: working with local governments, utilities and grid operators, and helping clients navigate the regulatory and community dynamics that determine whether a project advances or stalls.

The net result is reduced uncertainty early in the process by identifying locations where the fundamentals align, and the path forward is navigable before significant capital is committed.

Closing Observations

Data center development sits at the intersection of several legitimate and sometimes competing interests: the operational needs of technology companies, the infrastructure requirements of the broader AI sector, the environmental and economic concerns of host communities, and the regulatory environments of individual states.

None of those interests operates in isolation. A project's success—whether it gets built, operates efficiently, and maintains its social license over time—depends on how well the location and development process account for all of them. 

Site selection is the discipline that attempts to make that accounting systematic, identifying where tradeoffs are most manageable and positioning projects to navigate them with the least friction.