Data centers lift urban jobs while rural gains lag, study finds

Data centers can raise local employment, wages and business activity, but the gains are uneven, according to research by Georgia Institute of Technology scholars Daniel Yue and Yiyang Zeng. Their study found the strongest benefits in metropolitan counties, while less populous areas saw little job or wage spillover.

The findings arrive as state and local officials face growing pressure over whether to approve, subsidize or restrict new data center projects. The Data Center Tracker, a public database of U.S. community responses to data center siting, has recorded more than 1,200 public actions since early 2024, including zoning disputes, campaigns and temporary moratoriums.

Opponents have frequently raised concerns about electric grid capacity, water use and transparency in site selection, according to the tracker and reports cited by the researchers. In Maine, lawmakers passed a contested bill in spring 2026 that would have created the first statewide moratorium on new data centers, but Gov. Janet Mills vetoed it, saying it could derail a $550 million plan to convert the closed Androscoggin Mill in Jay into a data center.

Mills also said she supported the idea of a development pause in principle, signed separate legislation blocking state tax incentives for data centers, and pledged to create a council to study the industry statewide. The researchers noted that as many as 10 other states are considering similar measures to limit data center growth.

Study finds gradual local gains

Yue and Zeng compared counties before and after their first data center opened with similar counties that did not have one. They combined operating records for data centers with local economic data to estimate changes in employment, wages, business formation, income and building activity.

Across all counties studied, the researchers found that employment rose about 0.9% on average in the first three years after a data center opened. Wages increased 1.1%, while the number of business establishments rose 1%.

The longer-term gains were larger, according to the study. Employment increased about 3.5%, wages rose 5%, and business establishments grew 4.7%. Household income climbed 1.9%, and building permits increased 16.1%.

The researchers said those gains developed over time rather than appearing immediately. They also found that local conditions shaped the results more than the data center project alone.

Urban counties benefit more

In metropolitan counties, employment rose about 4.1% and wages increased 5.5%, according to the study. Yue and Zeng attributed the stronger effect to the presence of nearby construction firms, equipment suppliers, professional services and skilled workers.

In less populous counties, the researchers found negligible job and wage spillovers. They said that suggests a community’s existing level of economic development can matter more than the scale of a proposed facility.

Other factors also affected outcomes. The study found that facilities run by major technology companies lifted local wages more than those operated by smaller firms, and counties that attracted multiple data centers within five years posted larger cumulative gains than those with a single facility.

Power costs remain a central dispute

The researchers also examined electricity prices in counties where utility service areas were localized enough to study the effect. They found retail electricity prices rose by about 5% after a data center began operating.

Yue and Zeng cautioned that the estimate is not universal because utility territories often cross county lines and rate-setting rules vary by state and utility. Weather, transmission systems and regulations also affect electricity prices, making the power-cost finding harder to isolate than other economic measures.

The study found opposition to data centers was more common in places where facilities already operate. The researchers said that pattern may reflect local experience, though they could not determine how much of the opposition comes from direct effects versus broader concern about artificial intelligence.

Yue and Zeng said policymakers should look beyond headline investment figures when weighing data center proposals. They pointed to subsidy terms, electricity tariffs, grid and water upgrades, and the use of new tax revenue for public services as factors that can shape whether residents share in the benefits.

This story draws on original reporting from Fortune.