Article · AI translation
Beijing time on November 5, according to U.S. technology outlet The Verge, as the results of the 2026 U.S. midterm elections continued to come in, construction of the data centers on which the artificial intelligence industry depends encountered major obstacles.
Voters in many parts of the United States used local referendums to restrict data center expansion, and several candidates who supported pausing construction of such projects won their elections, meaning U.S. AI data center developers may face a stricter regulatory environment in the future.
An election-day poll conducted by NBC News showed that 76% of voters surveyed opposed building a data center in their own community. Meanwhile, local referendums on data center construction also showed a clear tendency to restrict development.
According to figures compiled by consultancy Redbud Consulting, as of the morning of November 4 local time, 33 of the 45 local ballot measures across the United States concerning data center development had produced results supporting restrictions on data center construction.
Some local governments adopted fairly strict measures. For example, voters in Wells, Maine, supported a 180-day pause on construction of large data centers; several cities in Ohio voted to prohibit data centers with power demand exceeding 25 megawatts; and El Dorado, Kansas, voted to prohibit construction of data centers with peak power demand exceeding 10 megawatts for the next 10 years.
This wave of opposition to unchecked data center expansion also affected the political landscape in several states.
In Texas, Democratic state Representative Gina Hinojosa defeated incumbent Republican Governor Greg Abbott and was elected the next governor. Texas has the second-largest number of planned data centers among U.S. states, and Hinojosa supports a statewide pause on data center development. By contrast, Abbott had long supported data center expansion, although he also introduced new review measures this summer to assess the impact of connecting such projects to the power grid.
Texas's U.S. Senate race was also affected by the data center issue. Democratic candidate James Talarico defeated Republican candidate Ken Paxton. Both advocated cutting some data center tax breaks and requiring developers to bear the costs of their own energy infrastructure, but Talarico placed greater emphasis on environmental impacts and allowing the public to share in the economic benefits brought by data centers.
In Ohio, Democrat Sherrod Brown regained a U.S. Senate seat, defeating incumbent Republican Senator Jon Husted. During the campaign, Brown called his opponent a “spokesman for data centers” and criticized his earlier efforts to advance related tax incentives. He argued that data center developers should not receive construction permits or tax breaks without the participation of local residents.
Michigan's newly elected Democratic U.S. Senator Abdul El-Sayed also supports pausing new data center projects until adequate safeguards and federal regulatory measures are in place.
In addition, Georgia's newly elected Governor Keisha Lance Bottoms and Maine's newly elected U.S. Senator Troy Jackson were among those who supported pausing data center construction during their campaigns to reassess its environmental and economic impacts.
Notably, opposition to unrestricted data center expansion was not a position held only by Democratic candidates.
In Florida, Republican Byron Donalds was elected governor. Although he opposes a blanket ban on data center construction, he previously introduced a bill in Congress requiring developers to bear the costs of their own energy and water needs.
Pennsylvania Governor Josh Shapiro won reelection. He previously signed an executive order prohibiting the use of nondisclosure agreements for data center development projects and ending expedited approval for AI data center projects. In Wisconsin, although neither party's gubernatorial candidate supported a complete pause on data center construction, both advocated ending related state tax subsidies.
The Verge noted that although some candidates who supported data center development still won their elections, almost all candidates connected to the issue had to address it during their campaigns.
Behind the data center controversy are U.S. residents' concerns about rising electricity prices, water consumption, environmental impacts and public subsidies. As demand for AI model training and inference grows rapidly, technology companies continue to increase investment in computing infrastructure, but the costs of related power grid upgrades and resource consumption are becoming a focus of concern for local communities.
However, it remains difficult to quantify how much the data center issue directly affected candidates' election results. Voters' choices were also influenced by the economy, immigration and other political issues.
These midterm elections sent at least one clear signal: the expansion of AI data centers in the United States is encountering increasingly strong local political resistance. For technology companies, building data centers in the future will require them to address not only chips, electricity and funding, but also more restrictions from voters, local governments and regulators.