AI data centers are net positive for local communities

What's this about?

People disagree about whether AI data centers help the people who live near them.

These sites can bring money and jobs, but they can also use lots of power and water.

What supporters say

• Better tools, smart work plans, clean power, and heat reuse can cut harm to the local area.
• Data centers can help build power lines, roads, and fiber links (fast web lines) for wider use.
• New sites can bring big money, building work, and tax cash to local town halls.

What critics say

• The gains may go to firms and some towns, while local people pay more or face more harm.
• Data centers can use huge amounts of power and water, which may raise costs for nearby homes.
• These sites need many workers to build them, but they may need far fewer workers once built.
• Tax cuts for firms can shrink the cash that towns gain from hosting these sites.

How to read this

The number of points on each side does not show who is right; the strength of the proof matters more.

The bottom line

AI data centers can help some towns, but they do not always bring more good than harm.

The result rests on local needs, water and power supply, and rules that make firms pay their fair share.

AI data centers can bring major investment and infrastructure to host communities, but they can also consume large amounts of electricity and water while shifting costs onto residents and other customers. The available evidence does not show that these facilities are generally a net positive; instead, their impact depends heavily on local conditions and the rules governing each project.

The case for

Data centers can deliver substantial capital investment, construction activity and property development. Local governments often support them because they may generate property-tax revenue and help finance infrastructure. The final benefit, however, depends on whether agreements protect the public from excessive tax breaks and require companies to contribute to roads, utilities and other services. 1

Large facilities can also speed up improvements to local systems, including transmission lines, power generation, electricity distribution and fiber networks. Those investments may serve the wider community, not just the data center. But that happens only if the costs are properly recovered, capacity is shared and reliability improves beyond the facility’s own needs. 2 Growing data-center demand is already prompting investment in electricity systems in some regions (see Figure 1).

Technology offers another way to reduce local burdens. Better cooling, more efficient computing, flexible management of workloads and renewable power can lower energy use. Recovering waste heat and putting it to productive use could also reduce environmental impacts. These are credible ways to improve the balance, but the evidence does not show that they are routinely deployed at a scale large enough to guarantee community-wide benefits. 3

The case against

The strongest concerns involve resource use and who ultimately pays for it. AI data centers are increasing demand for electricity, while facilities using evaporative cooling can require significant amounts of water. Infrastructure costs may also be passed on to other electricity customers when connection fees and tariffs do not cover the full expense. The scale of these pressures varies with the local grid, climate, cooling system and water scarcity (see Figures 2 and 3). 4

The employment gains may be smaller than the investment figures suggest. Building a facility can create many temporary construction jobs, but permanent operations often require relatively few workers, many of them highly specialized. Some communities may still gain from construction and service work, but data centers do not necessarily produce broad, lasting local employment. 5

The distribution of benefits and burdens is a more serious challenge. A project can increase government revenue while nearby residents face noise, land-use changes, heavier demand for power and water, or environmental harm. Developers, utilities and governments may receive concentrated gains, while particular neighborhoods bear the day-to-day costs. Evidence is stronger for showing that these unequal impacts are plausible than for measuring their total effect in every location. 6

Tax revenue is also not the same as net public benefit. Tax abatements and exemptions can reduce the payoff, while publicly funded roads, utility upgrades and grid expansion can add long-term costs. A proper calculation must compare actual revenue with foregone taxes, public service expenses and resource commitments. Promotional estimates may not meet that standard. 7

The bottom line

The evidence does not establish that AI data centers are generally net positive for local communities. It supports a conditional conclusion: some projects may produce more benefits than costs, but others may leave communities worse off.

The case for data centers is credible, especially on investment, infrastructure and possible efficiency improvements. But the opposing evidence is stronger on the existence of resource pressures, limited permanent staffing, fiscal risks and unequal burdens. There is high confidence that outcomes vary by location and governance, but less confidence about the size or direction of the net effect in any particular community.

The decisive factors include water availability, grid conditions, facility design, tax agreements, cost recovery and the community’s bargaining power. Strong safeguards—such as infrastructure payments, limits on tax breaks, water protections and rules preventing other ratepayers from subsidizing the facility—can improve the balance. Without them, large investment figures may conceal costs borne by residents and public systems.