Wed5

Data Centers and 2023 Home Sales in Northern Virginia Keith Waters and Terry Clower August 2025 The Center for Regional Analysis Schar School of Policy and Government George Mason University 

Introduction Over the past several decades, data centers have become an increasingly important asset to Northern Virginia’s regional economy. Positively, they generate substantial employment during construction as well as tax revenue after construction. However, their expansion has raised questions regarding energy usage and potential impacts on neighboring properties. This research note focuses narrowly on the impact of data centers on for-sale housing values in Northern Virginia.1 Descriptive Statistics To provide a broad overview of the impact of data centers on home values, Figure 1 maps the locations of home sales and data centers. [1]. Darker reds indicate high home sale prices . In general, the areas with the highest home sale prices are North McLean, Western Loudoun County, Clifton/Farrs Corner, and Belle View. These areas generally have larger properties, larger homes, and natural amenities such as the Potomac River or country views. Data centers in contrast are generally located around Dulles International Airport in Loudoun County and between I-66 and Manassas Regional Airport in Prince William County, despite a few being located in Tysons. Examining the direct relationship between distance to data centers and price reveals a negative relationship. That is, the farther a home was from a data center, the lower its sales price. The overall negative relationship holds for single-family detached homes, townhomes, and condos. Despite the very high end of the market being somewhat separate, such as North McLean, simple scatter plots reveal that proximity to data centers is correlated with higher home prices for the bulk of sold homes. 

Findings The model developed for this analysis explains almost 87% of the variance in homes observed in Northern Virginia in 2023. This is a strong model, though there are other factors impacting housing prices not accounted for in the model. For example, we did not include the presence of swimming pools or outbuildings that often impact housing prices. Also, the data used for this analysis has no measure for the condition of the home that is suitable for regression analysis. The regression analysis demonstrated the expected relationships with two exceptions. The distance to the Potomac River variable, which was meant to account primarily for exclusive neighborhoods in northern Fairfax County was not statistically significant. More interestingly, while the variable denoting distance to a data center was statistically significant, the coefficient carried the unexpected negative sign meaning that the closer a home was, holding all other variables equal, to a data center, the value was higher. We structured the test as a one-tailed statistical test, which leads to the conclusion: The analysis fails to demonstrate statistical evidence that proximity to a data center negatively impacts housing values. This suggests that any negative externalities associated with data centers, such as noise, do not have a systemic effect on housing values.