The rapid expansion of AI data centres in the United States has encountered a political backlash. On 3 November 2026, the date of the US midterm elections, voters across ten states overwhelmingly approved measures to restrict or block new data centre projects ([1]). A total of 41 local ballot initiatives involving data centres were on the ballot, and residents sided against the facilities in all but one case ([2]). This wave of community pushback reflects concerns that large AI computing sites can guzzle power and water while providing few local benefits ([3]) ([4]).
Nowhere was the backlash more consequential than Texas, a state on track to become the world’s largest data centre hub by 2030 ([5]). Democrat Gina Hinojosa won an upset victory in the Texas governor’s race after campaigning to halt new data centre projects and criticising her opponent for promoting a rapid, unaccountable buildout ([6]). The governor-elect has pledged an immediate moratorium on data centres until “real protections” for communities are in place ([7]). Her stance has rattled developers in Texas, which hosts roughly 27% of all planned US AI data-centre capacity ([8]), including major projects by Elon Musk’s SpaceX ([9]). Musk, whose company is constructing multiple gigawatts of AI computing facilities in Texas, reacted to the midterm results by quipping that “Data centres in space [are] looking pretty appealing right now” ([10]).
Leaders in other states face similar pressures to rein in AI infrastructure. In Ohio, another major data hub, the newly elected governor has backed a “conditional” ban on data centres until projects cover 100% of their own utility costs and use union labour ([11]). At the national level, technology stocks tumbled in the days after the midterms, as investors worried that rising data-centre costs might not translate into returns ([12]). The political mood has clearly shifted: in a Reuters/Ipsos survey, 42% of US voters said limiting the construction of AI-focused data centres was a very important or top issue for them ([13]).
Britain’s government is set to prolong its controversial partnership with Palantir, which operates the National Health Service’s new Federated Data Platform (FDP). According to reports in the Financial Times, officials have concluded that replacing Palantir’s system by the first 2027 break-point in the contract would be impractical, so ministers plan to grant a two-year extension to the £330 million FDP deal ([1]). The extra time will be used to design a fresh procurement process that could see a UK-based provider take over the platform in 2029 ([2]).
The decision has drawn fire from critics who cite ethical concerns and contested benefits. In July, a cross-party committee of MPs warned of “serious mistrust” of Palantir among the public, cautioning that the company’s involvement in NHS data could deter patients from sharing their information ([3]). Privacy campaigners and some health professionals have also questioned whether Palantir’s system is delivering on its promises. In a recent investigation, 29 NHS hospital trusts listed as “benefiting” from the Palantir platform said they were not actually using it, calling into question the value of its touted improvements ([4]). The UK’s statistics regulator has even instructed NHS England to ensure accuracy in any claims about the platform’s impact, after a dispute over unverified data on reduced waiting times and extra operations ([5]).
Palantir insists its software is already helping to improve patient care and hospital efficiency ([6]). Securing a contract extension is a short-term win for the US firm, but the writing is on the wall: unless it can demonstrate real benefits and uphold public trust in how NHS data is used, it risks losing the business when the platform is re-tendered.
The midterm results have also set the stage for greater federal oversight of data and AI in the US. With Democrats poised to take charge of Congress in January, party leaders are preparing a wave of legislation targeting both AI systems and data-centre development ([1]) ([2]). Senate majority-leader-in-waiting Chuck Schumer signalled that one priority will be protecting consumers from higher electricity bills linked to power-hungry AI data centres ([3]). Another proposal would require independent audits of advanced AI models to ensure they meet safety standards before deployment ([4]). Lawmakers are also exploring ways to clarify legal liability when AI systems cause harm, so companies that use personal data to train models can be held to account ([5]).
Although these measures face political hurdles, President Donald Trump has repeatedly opposed new AI regulations ([6]), the scrutiny alone may increase compliance costs for AI initiatives. Companies that rely on vast data infrastructure could be forced to invest more in sustainable power and community engagement to get projects approved. Those developing AI models will need stronger data governance and risk controls to satisfy regulators and earn public confidence. The message for executives is that mismanaging data practices, from poor quality or opaque use to lack of consent, can carry steep costs, from local opposition to multibillion-dollar lawsuits.