Watch the InnovateUS workshop 10 Things to Know About Data Centers

AI data centers are moving into communities long before most public institutions have figured out how to talk about them. Today, 38% of Americans live within 5 miles of an operational data center, according to the Pew Research Center.  And that number is increasing.

In many places, public officials are first encountering these projects through rezoning hearings, utility discussions, or public meetings.  And many of the new AI data centers are now in rural areas new to industrial development.

Source

Residents are asking whether electricity prices will rise and whether water availability will fall. Environmental groups are raising concerns about impacts on air, water, and climate. Economic development teams want to know if new jobs and tax revenue will increase. Utilities are trying to plan for major new demand on the grid.

Underneath all of it is the looming question of what it means for a community to host the infrastructure behind AI. Public opinion is constantly changing.  And that opinion is often influenced by whether AI data centers are built in their backyard.

A recent poll found that 7 in 10 Americans would oppose data centers being built near them.  So much so that some would prefer to live near a nuclear power plant than a data center.

Source

Lessons from the Field

For the past few years, I’ve been analyzing AI data centers and the systems that support them.  I’ve learned a great deal about the opportunities and challenges these centers present. I hope what I've learned will help public officials respond to this difficult situation.

My recently published work for ReImagine Appalachia, for example, examines whether waste heat from data centers could be reused to support nearby communities and industry. For News for the States, I shared my perspective on using AI data center waste heat to reduce energy costs and benefit communities. 

I also analyzed AI data center infrastructure, supply chains, and regional economic impacts for the Energy & Manufacturing in Appalachia Initiative managed by Catalyst Connection. There, I learned from manufacturers the importance of state policies to “buy local” so that communities hosting data centers can benefit from the jobs rather than those outside the region.

Source

What has emerged across this body of work is that AI data centers are rarely just technology projects. They quickly become debates about electricity, water, land use, industrial development, workforce needs, and public trust.

I recently explored these issues in an InnovateUS workshop 10 Things to Know About Data Centers, focused on the questions public officials are increasingly being asked as AI infrastructure expands into more communities.

This blog examines the issues public officials should understand, investigate, and debate before making decisions about AI data centers. While the rate of data center construction spending is slightly declining, data centers remain a major topic of debate across the country.

Source

And this debate is not partisan, as was recently described in a recent New York Times headline that called the situation “The Most Bipartisan Issue Since Beer: Opposition to Data Centers.”

What Should Public Officials Know About AI Data Centers?

At many public meetings, the debate is often reduced to simple arguments for or against AI data centers. What should public officials understand about AI data centers to have honest, practical conversations with the communities being asked to host them?

Here’s my response – focused on the ten things I believe public officials should know. 

As background, it is important to note that AI data centers differ from traditional data centers. Data centers have long supported websites, cloud services, hospitals, universities, businesses, and government systems.

AI data centers differ from traditional data centers because they are built for high-intensity computing. The specialized equipment used to train and run AI systems consumes far more electricity and produces far more heat than many traditional facilities.

That difference matters locally. AI data centers may require larger buildings, more cooling, and stronger connections to energy and digital infrastructure.

Public officials should not assume a proposed AI data center will have the same local effects as a smaller or more traditional facility already operating nearby. Residents may not distinguish between the two, so officials need to explain the difference clearly.

1. Electricity infrastructure may require upgrades to respond to AI data center energy demands

Electricity infrastructure is different from electricity demand.

Demand is the amount of power the facility needs. Infrastructure is whether the grid can reliably deliver that power, where upgrades are needed, how long they will take, and who pays for them.

Large AI data centers may require new substations, transmission lines, distribution upgrades, backup power systems, battery storage, or renewable-energy connections.

This is especially important in rural communities, where many proposed facilities are now being considered. Rural areas may have inexpensive land and access to energy or fiber infrastructure, but they may lack electrical systems designed for high-load industrial users.

Public officials should ask: What are the project’s total expected electricity demands at full buildout, and who pays for the required grid upgrades?

2. Electricity demand and affordability are central public concerns

AI data centers can use enormous amounts of electricity because they run energy-intensive computing equipment around the clock. A large facility may consume as much electricity as many thousands of homes. 

As a result, residents and businesses may reasonably ask whether new demand will increase their electricity costs.  This can be true even for communities not located near a data center, as long as they are on the same electricity grid. They may also want to know whether the data center companies will pay the full cost of the infrastructure they require.

A balanced answer should acknowledge both possibilities. Large electricity customers can sometimes support grid investment and renewable energy development that might otherwise be paid for by ratepayers. But utilities may also seek rate increases to pay for new generation, substations, transmission lines, or other infrastructure upgrades. 

Public officials should ask: Will residential and small-business electricity customers face higher costs? Will data centers take actions to improve their energy efficiency, including the reuse of waste heat for community benefit?

3. Water use varies by location and technology

Water consumption is another major public concern, but broad claims can be misleading.

Some cooling systems use substantial amounts of water. Others use less water but require more electricity. Some facilities may use reclaimed water, closed-loop systems, or advanced cooling technologies that reduce freshwater demand.

Local context matters. A water demand that seems manageable in one region may become highly controversial in a drought-prone or water-stressed area.

Water questions also extend beyond consumption. Residents may raise concerns about wastewater, chemical treatment systems, stormwater runoff, and impacts on streams, wetlands, or groundwater. Current policies may already respond to these concerns, but increased monitoring and enforcement may be needed in regions that previously did not face this challenge.

Public officials should ask: How much water, and from what source, will the facility require under normal, peak, and drought conditions? 

4. Water pollution and wastewater impact may need monitoring

Water questions also extend beyond consumption. Residents may raise concerns about wastewater, chemical treatment systems, stormwater runoff, and impacts on streams, wetlands, or groundwater. Current policies may already respond to these concerns, but increased monitoring and enforcement may be needed in regions that previously did not face this challenge.

Public officials should ask:  How transparent will the project be regarding operational impacts, environmental monitoring, and public reporting?

5. Economic costs and benefits are community-specific.

Economic development is often a key argument for attracting AI data centers.

Potential benefits can include tax revenue, utility revenue, construction activity, infrastructure investment, and business opportunities for suppliers and service providers. In small or rural communities with limited tax bases, those benefits can matter.

But economic gains do not automatically flow to the people or places most affected. And, some states offer tax incentives or abatements to attract data centers, which may reduce the net public benefit. 

The economic case should be examined carefully rather than treated as automatic.

Public officials should ask: What specific benefits will local residents receive?  What public incentives, tax abatements, or infrastructure subsidies are being requested?  Ultimately, what is the projected net public return on investment?

6.  Workforce impacts vary by phase and role

The job question deserves more precision than it often gets. AI data centers can create construction jobs and support work in electrical trades, HVAC, cybersecurity, engineering, maintenance, and related supplier industries. They may also create opportunities for local manufacturers and service firms.

Some equipment and supplier contracts, however, may come from outside the region. Permanent employment inside the facility may also be smaller than many residents expect, as some jobs may be temporary, specialized, indirect, or located outside the host community. A facility may require a large construction workforce during development, but a much smaller permanent workforce once operational.

A construction job, a permanent facility job, and a supplier contract are different outcomes. They should not be bundled into one impressive-sounding number.

Public officials should ask: How many short-term and long-term jobs will be created? How many of those jobs may go to community residents? Will the AI data center support educational and training programs for local residents, giving them the opportunity to secure those jobs now and in the future?

7. Data Centers may affect the character of nearby communities

For residents living near a proposed site, an AI data center may not feel like just another office building.

Concerns may include building size, traffic, lighting, noise from cooling systems and generators, truck movement, security fencing, road use, and changes to the look and feel of a rural or residential area.

Some of these concerns can be mitigated through design, setbacks, landscaping, traffic planning, sound controls, or site selection. Others may remain points of disagreement.

Public officials should ask: How close is the facility to homes, schools, farms, parks, or small businesses? Will it change the character of our community?  What actions are being taken to reduce those changes? How will complaints be handled after the facility opens?

8. Air pollution, backup generators, and public health concerns

Air quality questions often focus on what happens when data centers rely on backup power. Diesel emergency generators can emit nitrogen oxides, particulate matter, and other pollutants during testing, maintenance, or extended outages. 

These impacts may be limited under normal operating conditions, but they can become more significant in communities already facing air quality burdens or cumulative industrial emissions. Cleaner alternatives, such as battery backup systems, hydrogen fuel cells, or lower-emission generator technologies, may reduce these risks but may not yet be standard across projects.

Public officials should ask: How often will backup generators be tested or used, what emissions controls will be required, and how will local air quality and public health impacts be monitored and reported?

9. Transparency and public trust shape community response

Transparency may be the most underestimated issue.

Many data center decisions involve negotiations among developers, utilities, state agencies, local officials, and economic-development organizations. Some confidentiality may be common in business recruitment. But when communities hear about a project late, with limited information and few opportunities to ask questions, trust can erode quickly.

Once trust is lost, even accurate information may be treated with suspicion.

Communities need to know what is being proposed, what remains unknown, which decisions remain open, and how public input will be used.

A credible process can begin with a simple commitment: collect public questions, require developers to answer them, and make those answers public.

Public officials should ask: What information is available and what is still missing? When will the public be able to ask questions? Which commitments will be enforceable?

10. Cybersecurity, critical infrastructure, and community risk worries

Cybersecurity questions extend beyond the data center fence line. As AI data centers become part of energy, communications, public services, and regional economic infrastructure, cyberattacks or system failures can create risks for utilities, businesses, governments, and residents. 

Communities may also raise concerns about the concentration of digital infrastructure, who controls essential systems, and how vulnerabilities are disclosed and managed. Existing cybersecurity requirements may apply, but local officials may still need clarity on emergency response, incident reporting, coordination with utilities, and protections for public-sector partners.

Public officials should ask: What cybersecurity standards, incident response plans, and public-private coordination mechanisms will protect critical infrastructure and inform the community in the event of a major disruption?

Proposed Policy Response: Community benefit agreements to shape local outcomes

My conclusion from my analysis and field lessons is that community benefit agreements can help translate general promises into clearer expectations. These agreements may address workforce training, local hiring, supplier access, infrastructure contributions, environmental monitoring, broadband access, noise mitigation, or public reporting requirements.

But agreements vary widely. Some are specific and enforceable. Others are vague or difficult to monitor. A strong agreement should answer:

• Who helped define community priorities?

• What commitments are included?

• Who is responsible for delivering them?

• How will progress be reported?

• What happens if commitments are not met?

Community benefit agreements are not substitutes for regulation, permitting, utility oversight, or environmental review. As noted by the Brookings Institute, "transparency and cooperation between firms, local institutions, and residents are essential to facilitate community input into CBAs, and for proposals to support residents' digital access, well-being, employment, and more.

Matching Data Center Decisions to Community Priorities

AI data centers are becoming a test of public-sector capacity because different communities experience the trade-offs differently.

In a dry region, water may dominate the conversation. In a fast-growing energy market, electricity affordability may come first. In a rural community with limited jobs and a declining population, tax revenue and workforce opportunities may matter most. In neighborhoods near proposed sites, noise and lighting may shape public opinion.

When communities do decide to move forward, community benefit agreements can help clarify expectations, document commitments, and ensure that promised benefits, workforce opportunities, infrastructure investments, environmental protections, or local revenue are specific, enforceable, and responsive to local concerns.