The recent United States midterm elections produced a wave of local initiatives that restrict the construction of new data-centre facilities. Dozens of ballot measures passed across the country, directly targeting the rapid expansion of power-hungry AI infrastructure. Analysts see the outcome as a clear signal that the electorate is wary of the environmental and financial burdens associated with large-scale AI deployment.
In Texas, newly elected state senator James Talarico recounted a conversation with a mother whose child believed it was daytime because the lights from a nearby data-centre were so bright. The anecdote highlighted how the physical presence of these facilities is affecting everyday life, adding a human dimension to policy debates that have largely focused on economic and technical considerations.
The electoral backlash arrives as the AI market narrative has been built on expectations of massive, debt-financed data-centre construction to meet soaring demand. Critics argue that promoting hundreds of humming warehouses while warning workers about replacement by autonomous systems creates a contradictory public image. The resulting skepticism is beginning to erode the optimism that has underpinned recent AI-related stock rallies.
Tech writer Cory Doctorow has described the current trend of forcing AI features into every app and device as "enshittification," a term that captures public frustration with perceived overreach. Voters have also observed a widening gap between the soaring fortunes of big-tech executives and persistent inflation pressures on ordinary households, fueling a broader sentiment that the industry has failed to win public trust.
Elon Musk’s recent spending in Texas, aimed at influencing local voters, was seized upon by Democratic candidates as evidence of tech-industry interference in the democratic process. Observers compare the emerging regulatory mood to the early 20th-century antitrust campaigns led by Teddy Roosevelt, suggesting a possible shift toward more aggressive oversight of dominant technology firms.
Economist Dario Perkins of TS Lombard warned that "people outside the ecosystem, you know, regular people, just weren’t that interested in AI," noting that the sector’s $200 billion in current revenue underpins more than $1 trillion in future commitments from two major firms. The International Monetary Fund recently cautioned that the anticipated economic gains from AI may be overstated and that massive hyperscaler investments might not be necessary, especially as lower-cost Chinese models improve. Together, these viewpoints challenge the AI lobby to demonstrate tangible societal benefits before further investment proceeds.