Voters reject data center expansion in most midterm measures
Voters overwhelmingly reject data center expansion, enacting bans and moratoriums in 38 jurisdictions nationwide.
(Photo via CNBC)
Voters across the United States expressed widespread opposition to the growth of artificial intelligence infrastructure during the Nov. 3, 2026, midterm elections. Out of 41 local ballot measures concerning the development of data centers, residents in 38 jurisdictions voted to enact bans, moratoriums, or other restrictive policies.
Analysts estimate that these results could affect approximately 45% of planned data center capacity expansion across the country in the near term. According to data from Ballotpedia, at least 50 measures were scheduled for 2026, with 41 reaching the ballot on Election Day.
The only jurisdiction to reject proposed restrictions on industrial development was Harwood, North Dakota, where voters opted against the measure. In other regions, residents cited concerns regarding the high consumption of electricity and water, potential noise pollution, and the impact of these facilities on local utility bills.
The election results follow a period of public resistance toward the rapid buildout of AI-related facilities. Prior to the vote, a CBS News poll indicated that Americans opposed the construction of new data centers in their areas by a margin of more than 2-to-1. Elizabeth Bast, executive director at Oil Change Action, described the results as a clear message from voters regarding their concerns over costs and infrastructure expansion.
Nick Owens, an attorney and co-founder of Conserve Ohio, stated that developers must address community concerns regarding health and land use to move forward with future projects. Meanwhile, SpaceX CEO Elon Musk commented on the election outcomes via X, suggesting that data centers in space might become an appealing alternative.
The industry faces further challenges as three additional ballot measures related to data centers are scheduled for December 2026. Analysts expect these developments may lead to a slowdown in the construction of the 800 to 1,000 data centers previously planned to support the AI sector.