Data Center Siting and Community Opposition
Presented by:
Grant Jacobsen
Economics Program, School of Public Policy
Oregon State University
TODAY, September 11, 2026
12:00 pm-1:15 pm
Taylor-Hibbard Seminar Room (Rm103)
As AI demand for data centers expands a growing number of communities have enacted policies to slow or stop their construction. We present evidence on the determinants of data center siting and community opposition thus far to understand where future development is more likely. Using novel county-level datasets on data center development and opposition events we document four main findings. First, large data centers (i.e., ``hyperscalers'') are being located where power is currently inexpensive---residential electricity costs, household energy burdens, and marginal carbon emissions rates are all lower in host counties---but this advantage may be fleeting, because data centers are also locating where the marginal costs of energy and capacity are higher. Second, data centers are locating in comparatively prosperous counties, a pattern that runs counter to the tendency for industrial disamenities to be located near populations with fewer economic resources. Third, Democratic vote share predicts opposition well in excess of what siting patterns alone would imply, a result that also holds in a parallel analysis of renewable energy---despite Democrats' favorable views of renewables---suggesting that Democratic-leaning communities are effective at mounting opposition to large-scale development in general. Because Democratic vote share is negatively associated with marginal emissions rates, growing local opposition could push data center construction toward dirtier segments of the electric grid. Fourth, after accounting for other demographic variables, counties with higher educated populations are more likely to oppose data centers but less likely to oppose renewables: this may suggest that local advocacy against AI infrastructure is partly driven by threats AI presents to white-collar employment.