AI data centers deliver uneven economic gains as metro areas capture most benefits

AI Data Centers Drive Historic Investment Wave Across America

Artificial intelligence is fueling an unprecedented surge in data center construction across the United States, with more than 2,700 facilities either active or under construction1

1

. Individual hyperscale AI data centers often cost more than $1 billion to build and can span over 1,000 acres, consuming as much electricity as a small city2

2

. This represents one of the largest waves of capital investment in American history, yet the economic impact on host communities remains deeply uneven and largely misunderstood.

New research from Georgia Tech Scheller College of Business faculty member Daniel Yue and postdoctoral fellow Yiyang Zeng examines how data center openings affect local jobs, wages, business activity, and electricity prices using detailed facility-level data paired with county-level data1

1

. Their findings reveal that geography plays a decisive role in determining whether communities see meaningful local economic benefits.

Metropolitan Areas Capture Disproportionate Job Growth and Wage Increases

The research shows that when a data center opens, host counties experience measurable gains over time. Employment rises by approximately 3.5%, wages increase by 5%, business establishments grow by nearly 5%, and household income climbs by about 2%2

2

. Building permits also spike sharply, reflecting construction activity tied to new facilities. In the first three years after activation alone, employment rises by about 0.9%, wages by 1.1%, and business activity by 1.0%1

1

.

However, these gains are modest relative to the scale of investment and distributed unevenly. Metropolitan areas capture most economic benefits from data centers, while rural areas see far fewer spillover effects. Metro counties experienced increases in employment and new business growth, while non-metro counties saw no measurable gains2

2

. "Enormous amounts of capital are flowing into specific communities, much of it tied to new construction, while rigorous evidence on the local benefits of these facilities has been thin," Zeng explains1

1

.

Agglomeration Effects Explain Why Rural Areas Miss Out

The disparity between metropolitan areas and rural areas comes down to what economists call agglomeration effects—the clustering of economic activity. Data centers don't operate in isolation. They rely on construction contractors, engineers, equipment suppliers, professional services, and a skilled workforce1

1

. These connections are far easier to establish in places with deep labor markets and established business networks.

Metropolitan areas absorb the indirect spending that data centers generate more effectively. High-wage technical employees support restaurants, retail, and local services, while suppliers and contractors scale up quickly. These spillover effects amplify the initial investment's impact2

2

. In rural areas, this amplification proves much harder to achieve. Facilities typically employ fewer than 100 permanent workers, and many specialized services are imported from outside the county1

1

.

Rural communities do see some benefit. The research identifies a small but measurable decline in unemployment rates in non-metro counties, and local governments may gain tax revenue or infrastructure upgrades2

2

. Yet the sweeping job and wage growth often promised during local recruitment efforts is unlikely to materialize. "It's the characteristics of the host community, not just the facility opening there, that determine whether the local benefits show up," Zeng notes1

1

.

Electricity Prices Rise 5% as Infrastructure Demand Surges

The research uncovers a significant trade-off. A single large data center can consume as much power as roughly 80,000 homes1

1

. In areas where researchers could cleanly measure price effects, electricity prices rise by approximately 5% after a data center enters2

2

. "When the benefits to communities are small, downsides like higher electricity prices and added pressure on local infrastructure will be felt by locals," Zeng warns1

1

.

Who pays for increased electricity costs isn't straightforward. The research suggests communities should ask specific questions about who pays for new infrastructure and how those costs will be distributed through cost-sharing arrangements1

1

. Local utility companies divide costs differently among homeowners, businesses, and large industrial users, including data centers. Because cost-sharing systems vary by state, each development presents unique challenges2

2

.

As states and cities work to attract or resist data center construction, Yue and Zeng hope their research will encourage more evidence-based decision-making. For metro areas with strong labor markets and dense business ecosystems, data centers can deliver meaningful gains. For rural communities, the calculus is far less certain, and the infrastructure demand may outweigh limited economic returns.