Even when data centers create jobs, their local economic benefits appear limited
Employment is moderate and housing prices slump when data centers come to town.
With technology companies expected to , many local governments are attempting to woo such investments with tax breaks and other incentives, in the hopes of creating jobs, raising fiscal revenue and fueling regional economic development.
But backlash against the projects is increasingly fierce 鈥 and politicized. Faced with the prospect of farmland and other natural spaces being converted into vast tracts of windowless warehouse-like structures, opponents warn of massive energy and water consumption, strains on other infrastructure and industrial-strength noise pollution.
滨贰厂贰鈥檚 Carles Vergara and Christoph Loch contribute to the debate over the net benefits of data centers to local communities with two pieces of research measuring the impact on employment and housing markets, with complementary information on indicators such as electricity prices and municipal bond yields.
Evaluating data centers presents unique challenges. Though hugely capital-intensive investments, they require relatively few direct employees once short-term construction workers have finished building them. The places where they choose to locate, normally with comparatively low-cost power and available land, may already be on a path to growth. In addition, facilities are drastically different in size, from hyperscale campuses to micro sites.
Still, what emerges from the research will likely disappoint the most gung-ho politicians. , though not immediately and only in a narrow range of sectors, according to the Vergara research. , the Loch study shows, while electricity prices and consumption climb. And neither study detects widespread economic benefits to local economies.
How data centers create employment
In a working paper published by the MIT Center for Real Estate Research, Vergara and coauthors examine 1,584 U.S. data centers between 2000 and 2023, decades in which developments, once concentrated in California and the Northeast, spread across the country.
Since data centers require few workers to operate, tracing their effect on the job market involves examining local suppliers of services such as maintenance and security. Strong demand for those kinds of jobs may then have a knock-on effect on local restaurants, retail and other services. The centers, connected to fiber networks and power infrastructure, could also potentially serve as magnets for other companies to set up shop nearby, which could conceivably generate jobs.
But do they? In general, the employment benefits are limited to the industries that supply the centers, and only accrue after two years. The research finds that:
- For the first two years after a data center becomes operational, local non-construction employment changes little.
- Around the third year, the response becomes more noticeable. From three to five years after opening, counties with data centers have, on average, about 1,000-1,300 more non-construction jobs than control counties. The delayed response may be due to the time it takes for local companies to respond to new demand, for workers to relocate gradually and for more capacity to be added to initial investments.
- By year five, the employment impact is approximately 2.74 jobs created per megawatt of installed capacity.
The employment gains are concentrated in two sectors:
- transportation and warehousing.
- professional and technical services.
In contrast, consumer sectors, such as local retail or restaurants, see little new hiring. Data centers generate jobs closely linked to their specific needs, rather than serving as broad-based consumption multipliers.
Turning to whether data centers draw new businesses to the area, the research finds that follow-on investment comes mostly from incumbents expanding their own campuses, rather than from new and varied entrants arriving on the scene.
jobs created
per megawatt of installed capacity, five years after a data center becomes operational.
Limited employment benefits
The job gains resulting from a data center being built in a locality are not immediate and they are not broad-based but are concentrated in a few sectors, mainly related to warehousing and technical services.
Why data centers lower housing prices
Widening the lens, Vergara鈥檚 research looks at local business activity, average monthly earnings, industrial electricity rates and housing prices. There is little variation in these metrics, one way or the other, five years after plants opened, underscoring how circumscribed the impact of data centers is.
Other studies point to broader and more negative economic consequences for local communities.
The working paper by Loch and coauthors from San Diego State and George Mason University finds data center development leads to slower growth in local housing prices, a bellwether for economic health, compared to places without the centers.
The weakening begins at the time when data centers are announced, indicating that the market immediately foresees a bleaker future. And the decline gathers pace over time, suggesting that lower valuations are not transitory but rather compound as the implications of projects become more apparent.
Using a dataset on U.S. data centers from 2015 to 2023, they find:
- One year after the announcement, housing price growth is 1.3 percentage points lower than in areas without centers.
- Two years post-announcement, growth is 2.1 percentage points lower.
- At three years, it is 3.1 percentage points lower.
Housing markets seem to be factoring in a future of higher utility bills and strains on existing electricity networks, among other indicators. In complementary analysis, the research looks at electricity consumption and residential electricity prices once actual construction begins. Consumption in counties where a center was being built rose by 1.9% compared to other counties; residential electricity prices were 3.1% higher.
That may, in turn, feed through to municipal bond markets, which reflect expectations regarding future infrastructure investment requirements and financing burdens. Spreads on bonds issued to fund utility-related projects rose by 22 basis points following a data center announcement; borrowing costs for cities hosting these facilities are higher.
At the same time, the research finds no substantial increase in property taxes collected after construction, and no growth in the establishment of new businesses.
percentage points
The drop recorded in housing price growth three years after it was announced a data center would be constructed, relative to areas without centers.
Housing down, electricity up
On top of lower housing price growth, electricity consumption in counties where a center was being built rose by 1.9% compared to other counties, and residential electricity prices were 3.1% higher.
Important questions on net benefits to communities
In approaching data centers, many government officials are returning to policy playbooks used to attract car makers and other large manufacturers in previous decades, when massive factories could serve as anchors for regional economic development.
But that may be the wrong playbook. Measured by economic impact, data centers may more closely resemble infrastructure projects, generating a burst of short-term construction jobs and longer-term maintenance work but little else. And whereas most infrastructure serves local communities, at least in part, these projects may serve technology giants headquartered hundreds if not thousands of kilometers away.
Given that their environmental impact may be substantial and damaging, government officials will need to analyze whether the new investments create enough economic activity around them to justify the resources used to attract and accommodate them.
In a telling indication of long-term impact, the Loch research examined income tax filings through 2022 to gauge shifts in income levels. Following data center construction, the share of low-income households increased by approximately 0.5-0.7 percentage points: The centers tend to impoverish rather than enrich the local communities that house them.
MORE INFO:
鈥溾 by Ernesto Aldana, Simon B眉chler, Alexander Carlo, Lyndsey Rolheiser, Albert Saiz and Carles Vergara. MIT Center for Real Estate Research (2026).
鈥溾 by Mingming Ao, Yi Cao, Christoph Loch, Chi Wan and Yaoyi Xi (2026).
This article is included in 天美传媒 Insight online magazine No. 174 (Jan.-April 2027).
READ ALSO:
