A University of Illinois and Purdue survey of roughly 1,000 US adults found rural Americans are more worried than urban or suburban residents that AI data centers will raise their electricity bills, with 57.9% of rural respondents 'very worried' about the issue versus 46.9% of urban and 49.7% of suburban respondents. The Gardner Food and Agricultural Policy Survey, fielded in February 2026 and published June 25 by farmdoc daily, also found rural residents more concerned than others about data centers converting farmland. Separately, Pew Research Center reported in April that 87% of existing US data centers are in urban areas, but 67% of planned facilities are headed to rural communities, and 39% of planned centers are in counties with none today. The gap matters because rural grids and local governments often have less capacity to absorb new large loads or negotiate mitigation terms.
For AI infrastructure teams and site-selection planners, this survey confirms what permitting fights already suggest: the communities now hosting most new AI-scale data centers have measurably less tolerance for utility-cost increases than the urban markets that host most existing capacity, and fewer resources to negotiate mitigation terms.
University of Illinois and Purdue researchers (Mark White, Sarah Low, Maria Kalaitzandonakes, Jonathan Coppess, and Brenna Ellison) published results from the quarterly Gardner Food and Agricultural Policy Survey (GFAPS Wave 16, roughly 1,000 US adults, fielded February 2026) on June 25, 2026. Asked to rate their worry on a 1-7 scale about AI/data center effects, respondents rated electricity costs the top concern overall (5.17/7 average), ahead of water overuse (4.77) and farmland conversion (4.56). Rural respondents were more worried than urban or suburban respondents on both electricity costs (5.41 vs. roughly 5.11) and farmland use; 57.9% of rural respondents were 'very worried' (rating 6-7) about electricity costs, versus 46.9% of urban and 49.7% of suburban respondents. Urban respondents were the most concerned group about water usage (44.5% very worried).
The survey lands alongside a broader siting shift. Pew Research Center reported in April 2026 (Seets and Radde) that 87% of existing US data centers are in urban areas, but 67% of planned facilities are slated for rural communities, and 39% of planned centers are in counties with no existing data center today. Modern AI-scale data centers typically require 500-800 acres, pushing developers toward available rural land, often farmland. Separately, the University of Virginia's Weldon Cooper Center projects data center energy demand will more than double in Illinois and triple in Indiana, Michigan, Minnesota, and Wisconsin by 2030.
The survey authors note a practical asymmetry: large data center operators bring specialized lawyers, engineers, and financial analysts to site-selection and permitting negotiations, while many rural jurisdictions rely on part-time or volunteer officials without comparable capacity. That gap is a real risk factor for project timelines, community-benefit negotiations, and utility rate cases, independent of any individual project's technical merits.
State-level rate cases and grid-interconnection filings in counties with newly announced projects, and whether states with partially deregulated generation markets, like Illinois, shift more construction costs directly onto data-center customers rather than spreading them across all ratepayers.
Original-source survey data (University of Illinois/Purdue Gardner Food and Agricultural Policy Survey) plus Pew's verified rural/urban siting split give this concrete, citable numbers on a recurring practitioner-relevant AI-infrastructure friction point. Solid but not landmark, since it is descriptive survey data rather than new policy or litigation; modestly bumped from 6.1 given the stronger sourcing found this audit.
Public references used for this report.
A 5-minute Tuesday brief on AI & data science. Curated, no fluff.
No spam. Privacy.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
News on Let's Data Science is compiled from multiple public sources with editorial oversight. See our Editorial Standards and Corrections Policy.
Leave a Reply