Trump oblivious to voter fury about data centers, saying ‘the jobs are enormous and the money paid, the taxes paid, are just enormous’

Trump endorsed AI data centers as job and tax engines, but political backlash is escalating: multiple states’ candidates are using data-center fights to attack opponents, with voters opposing local construction per a Fox poll and polling showing ~4 in 10 cite AI data centers as a key election issue. In Nevada, Aaron Ford proposes halting ~$200M in new state tax breaks and auditing/auditing projects while requiring companies to pay for electricity and protect water; in Ohio, the NRSC calls data centers a “sleeper issue” as Sherrod Brown runs ads framing Sen. Jon Husted as the “face of data centers.” Near-term risk is increased regulatory and permitting friction (local votes, utility/water requirements, delayed power-plant timelines), which could pressure data-center expansion plans despite continued AI demand.

Analysis

This is not an AI demand shock; it is a permitting and cost-of-capital shock. The near-term loser is the marginal data-center developer model: the more a project depends on local incentives, cheap utility interconnects, and fast zoning, the more its IRR gets squeezed by political pushback, community veto risk, and higher mitigation spend. That argues for lower multiples on the more levered “build-on-spec” ecosystem and for a bigger premium on hyperscalers that can self-fund, self-power, and internalize siting risk.

Over 1-3 months, the important mechanism is not cancellation but friction: longer approval cycles, more local concessions, and more scrutiny on who pays for grid upgrades and water. That tends to shift bargaining power toward utilities, grid equipment, and power-generation vendors, while pressuring pure software/AI capex narratives if investors start marking down the speed of rollout. GOOGL is more insulated than smaller cloud peers because it can absorb front-end capex, but its AI multiple is still vulnerable if market perception shifts from “unlimited deployment” to “socially constrained buildout.”

The contrarian miss is that this may actually be bullish for the adjacent picks-and-shovels names, not a reason to short AI outright. If the political system forces more on-site generation, backup power, switchgear, and cooling redundancy, the winners are the infrastructure suppliers with pricing power and backlog visibility, not the land-bank owners. The thesis breaks if states standardize fast-track permitting or if tech firms sign binding cost-sharing deals that make local opposition financially irrelevant.

Key Decisions for Investors

• Over the next 2-6 weeks, buy pullbacks in AI power-infrastructure names (ETN, VRT, GEV, PWR) versus a basket of data-center-exposed REITs; the trade benefits if approvals slow but capex still has to be spent on-site and closer to the load.
• Keep GOOGL on a watchlist, not an immediate short: the risk is not demand destruction but rising AI deployment friction. Falsify the bearish view if the company reiterates AI capex growth with no change in delivery timelines or power procurement costs over the next earnings cycle.
• Short-duration hedge: buy 3-6 month put spreads on a data-center proxy ETF or the most valuation-sensitive cloud names if state-level incentive rollbacks accelerate; target a move lower only if local referendums or tax changes start passing, not on headlines alone.
• Use DJT only as a sentiment/event-volatility vehicle, not a fundamentals trade: any rally from the issue’s political salience is likely tactical and fades unless it translates into concrete federal permitting actions within 30-60 days.
• Alert level: if you see formal state utility-cost pass-through rules or binding moratoriums in key growth states, increase conviction on the infrastructure long/AI capex short pair; absent that, keep exposure modest.