AI companies pour $265 million into the U.S. midterms as opposition to data centers grows

The artificial intelligence industry has thrown itself fully into the battle over the 2026 US midterm elections, with a combined outlay estimated at $265 million in super PACs and political groups, according to a Wall Street Journal analysis cited in the article. Along with the cryptocurrency and betting industries, AI has become the sector injecting the most money into this election cycle, which by June was already recording 40% more corporate spending than the 2024 elections.

The trigger for this push is an unexpected phenomenon: the boom in data center construction, indispensable for training and running AI systems, has run into across-the-board public opposition. An Economist/YouGov poll cited in the text reveals that barely 20% of Americans view the construction of new data centers positively, with only 14% of Democrats and 31% of Republicans seeing them favorably. The reasons for the discontent are concrete: fears that they will drive up electricity prices, deplete the water resources used to cool the facilities and hurt property values. The article notes that some two-thirds of the data centers built since 2022 are located in water-stressed areas, many of them in the west of the country.

The volume of investment underpinning this boom is enormous: the five big hyperscalers —Amazon, Microsoft, Google, Meta and Oracle— expect to spend more than $560 billion on AI infrastructure this year, and the pace of data center construction in July grew 60% year on year, according to Census Bureau data cited in the article. Electricity demand from these centers is expected to double in the United States by 2027. Against that backdrop, states of both political stripes are weighing moratoriums: in July, New York approved the country's first statewide suspension of large data center construction, lasting one year.

President Donald Trump himself weighed into the debate this week, calling data centers a 'Golden Goose' and stating on Truth Social that communities opposing them 'want to end up backward and poor', as well as accusing critics of acting in collusion with China. According to the article, Trump has seen polls showing that most Americans object, but that has not curbed his public backing. His cabinet has followed the same line: Commerce Secretary Howard Lutnick went so far as to claim that data centers 'don't use water', a statement that contradicts both the available evidence and his own words the previous year, when he acknowledged that these facilities 'suck up' water.

On the political money front, the super PAC Leading the Future has raised $140 million for the midterms, with plans to fund pro-AI candidates from both parties. Its donors include $50 million from the venture capital firm Andreessen Horowitz and a combined $25 million from Greg Brockman, president of OpenAI, and his wife Anna, according to Federal Election Commission (FEC) filings. Anthropic, the creator of Claude, has reportedly donated at least $40 million through a 'dark money' organization called Public First Action, which has a $100 million war chest. Elon Musk's super PAC handles a similar sum, and Meta is linked to several PACs with tens of millions of dollars available. OpenAI, for its part, pointed to a June statement in which it says it has not donated to campaigns nor founded or funded any PAC, clarifying that the Brockmans' support for Leading the Future was given in a personal capacity.

One peculiarity noted by Brendan Glavin, director of analysis at the transparency group OpenSecrets, is that AI companies have chosen to intervene directly in specific races —even district-level ones— rather than donating generically to the parties, replicating the tactic the crypto industry used in 2024. Many of the television ads funded by these companies do not even mention AI or data centers, focusing instead on attacking or backing candidates according to their alignment with the sector's interests.

The case most cited in the article is New York's 12th district, where AI money poured into the Democratic primary against Alex Bores, a state assembly member and former Palantir employee who authored a state AI safety law. A PAC linked to Public First (tied to Anthropic) devoted more than $13 million to supporting him, while a super PAC linked to Leading the Future spent more than $8 million against him; in total, the race is estimated to have mobilized $27 million in AI-related money. Bores lost by a narrow margin to Micah Lasher, who, ironically, had co-sponsored Bores's law and backs a temporary national moratorium on data center construction; in his victory speech, Lasher said he would not follow the 'instructions' of the companies that intervened in the race.

The phenomenon is repeating across the country: according to a CNN analysis of AdImpact data, more than $31 million has been spent on political ads related to data centers this year, almost all of them negative in tone, compared with the complete absence of such ads in 2024. The article documents several cases that break the usual partisan script: in Texas, Republican Governor Greg Abbott —who last November was celebrating the state becoming the 'epicenter of AI development' after a $40 billion investment by Google— has ordered an audit of data center projects seeking to connect to the power grid, while his Democratic rival, Gina Hinojosa, is calling for a moratorium and accusing him of being too close to the industry. In Pennsylvania, Democratic Governor Josh Shapiro went from declaring that his state was 'all in on AI' to restricting new data center development and attacking his rival, state Treasurer Stacy Garrity, calling her 'Pennsylvania's number one data center fan'; Garrity, in turn, accuses Shapiro of 'misleading' public opinion given his record of prior support. In Ohio, Democrat Sherrod Brown accuses his rival, Senator Jon Husted, of being 'the face of data centers' because of his earlier work as lieutenant governor, while Husted has proposed a bill to protect consumers from electricity bill increases linked to these facilities.

Faced with this climate of distrust, the big tech companies have launched public relations and community initiatives: Meta unveiled a $1 billion fund called 'Future Is For Everyone' to invest in the communities where it operates data centers, and pledged to be 'water positive' by 2030 in those areas, as well as to create a job-training academy. Other companies have funded recreation centers, fire departments and municipal wifi networks. In West Virginia, where Microsoft and Nvidia are building one of the world's largest data centers, there has been talk of using these projects to eliminate the state income tax and increase school funding. The Trump Administration, for its part, has promoted a voluntary 'Ratepayer Protection Pledge' for large AI companies to commit to not passing energy costs on to consumers, while at the same time pushing deregulation to speed up the construction of these facilities.

The article gathers the voices of several experts to frame the phenomenon. Nicol Turner Lee, director of the Center for Technology Innovation at Brookings, sums up the political lesson of the moment: trying to sidestep local communities in the development of these projects 'is not working, and people are going to take it to the ballot box'. Journalist and author Paris Marx, who is preparing the book 'Hyperscale: The Ambition and Excess of Big Tech's Data Empires', argues that citizens' concerns are real and not a mere overreaction; the article adds that the industry often operates opaquely, resorting to shell companies and confidentiality agreements with lawmakers. Turner Lee adds that discontent with AI is intertwined with other economic tensions, such as the high fuel prices linked to the war in Iran, which makes the big tech companies —dominated by billionaires— an easy target for local frustration. Finally, Chris McGuire, a former Biden Administration official and now a researcher at the Council on Foreign Relations, warns that 'the window for thoughtful, proactive policy is open now, but it won't last forever', anticipating that AI will be an unavoidable issue in the 2028 presidential cycle.

Taken together, the article portrays a central paradox: while the AI industry multiplies its political spending to shield its infrastructure expansion, that very expansion —with its tangible impact on electricity bills, water and the landscape of specific communities— is generating a popular backlash that transcends traditional partisan lines and that neither money nor presidential backing appear to be reversing.