A recent voter poll indicates that opposition to data center construction played a significant role in the recent US midterm elections. According to a survey conducted by AP and a consortium of news organizations, 56% of voters considered data centers a crucial or major factor in their voting decision. Furthermore, 76% of those surveyed stated they would oppose the construction of a new data center in their community.
Voter Sentiment and Election Outcomes
The hostility toward these facilities was notably stronger among voters who supported Democratic candidates in this election cycle. This aligns with broader trends identified by Pew Research, which found that approximately half of Americans view data centers as mostly harmful to the environment, home energy costs, and quality of life. This represents a sharp decline in favorability since January.
While data center concerns were prominent, they trailed behind other major issues. Government corruption was cited as a crucial or major factor by 90% of voters, and the cost of living was cited by 89%. However, the salience of data centers is remarkable given that the issue likely would not have been on most Americans’ radar just five years ago.
Analysis of state-level polling suggests that states which swung to the Democrats either followed the national trend regarding data center concerns or appeared less concerned than the national average. It is difficult to determine exactly how many votes or Senate seats were directly delivered by this opposition. Nevertheless, it is notable that three key states in the US AI build-out, Texas, Georgia, and Ohio, all elected Democrats as senators and governors who have called for tighter regulation on data centers.
Regulatory and Economic Implications
The political shift comes as the US economy has been heavily supported by massive corporate investments in AI infrastructure. Official data shows that inflation-adjusted monthly spending on building data centers, excluding IT hardware, reached $5 billion in August 2026. This figure is five times higher than the spending levels recorded in 2023. Analysis by Columbia Business School suggests that AI infrastructure investment in the US is currently larger as a share of the overall economy than spending on new railways was at the height of that boom in the 19th century.
Three of the Senate seats flipped by the Democrats have significantly more data centers than the national median of 55. Texas leads with 595 data centers and 31 ongoing construction projects, followed by Ohio with 258 and Iowa with 117. The influx of these facilities has prompted political responses. Cori Crider, executive director of the Future of Tech Institute, told BBC Verify that the industry should anticipate that new state leaders will put conditions on or even stop runaway data center builds.
Seema Shah from Principal Asset Management noted that requirements for developers to pay for additional electricity grid investment, meet tougher water usage standards, or give up tax incentives could raise development costs and slow new supply. Heather West from the Eurasia Group consultancy added that while temporary bans and new regulations might delay the rollout, projects already permitted or under construction will largely proceed. Delays are expected to affect the next wave of building through local fights and slower permitting processes.
Despite the regulatory headwinds, analysts from Oxford Economics point out that strong economic incentives, including tax revenue and construction jobs, may prevent politicians from enacting policies that severely disrupt data center construction. Deepak Surana of the Futurum Group expects a negotiation rather than a moratorium from Texas state leaders to ensure no negative impact on residents’ energy and water bills.
In response to the political climate, OpenAI pledged on Thursday to ensure it fully pays for its own electricity. The company, which has a data center in Texas and under-construction projects in Ohio, made this commitment one day after the midterm results were finalized.
Source: BBC
