More than 1,200 active candidate websites for House, Senate and governor’s races now mention artificial intelligence, appearing in close to 40 percent of those contests. The count comes from a Washington Post analysis that The Decoder covered on August 17, and it places AI ahead of several subjects that have anchored recent US election cycles. What is pushing AI onto so many campaign pages, per the analysis, is not chatbots. It is data centers.
The Decoder’s own headline names Israel, racism, and crypto as topics AI has now passed. The chart it published alongside the piece, credited to the Post, instead lists Israel, manufacturing, and racism as the comparison set. Neither The Decoder nor the excerpt available from the Post describes how individual website mentions were classified into topic categories or how the underlying ranking was built. Read the “AI overtook these issues” claim as the Post’s internal methodology, not as an independently reproducible count.
The substance behind the numbers is local, not abstract. Candidates are responding to data centers going up in their districts, specifically to power prices, water draw, and land use tied to those projects. That is a different kind of AI story than a model release or a funding round. It is a siting fight, decided by the same local and state officials now facing voters on the issue.
The Post’s analysis, as described by The Decoder, found an asymmetry in who raises AI and why. Democratic candidates bring it up roughly twice as often as their Republican counterparts, per the count, and center their message on risk, regulation, and safeguards for children. Republican candidates instead emphasize national security and rivalry with China. The Decoder also cites separate survey data pointing to broad public unease about AI, including concerns tied to employment, though neither the pollster nor the sample size is specified in the piece.
None of this is abstract for anyone financing AI infrastructure. A hyperscale data center is a multi-year commitment that runs through utility interconnection queues, water rights, and property tax abatements, all set by the same officials now under campaign pressure. AI Insiders has already covered what that exposure looks like in practice: a 10 gigawatt Ohio data center campus recently lost a financing guarantee it had been counting on, and separately, roughly $50 billion in debt has been raised across the industry to fund compute buildouts. Both are balance sheet stories. This analysis is a reminder that the physical assets behind them sit inside political jurisdictions that can now slow them down.
For anyone underwriting a data center project into the 2026 midterms, local political sentiment belongs in the model next to power purchase agreements and land options, not as a footnote to them. A siting timeline built only on utility and permitting risk is missing the variable that just became visible in close to 40 percent of American campaign races.
The Decoder reported this analysis, citing the Washington Post, on 17 August 2026.
