• President Trump calls data centers a 'Golden Goose,' urging communities to welcome them for jobs and tax revenue.
  • Local resistance grows over electricity costs, water use, and transparency, becoming a 2026 election issue.
  • Federal push accelerates AI infrastructure, but local siting battles and grid constraints could define where capacity gets built.

A Federal Endorsement Amid Local Pushback

President Donald Trump made his strongest case yet for data-center development, telling communities that rejecting these projects would leave them 'backwards and poor' and hand China a strategic advantage. 'Let Data Reign,' he said in an interview aired August 23, framing data centers as economic engines that bring jobs, investment, and lower taxes. He also suggested developers could build their own power plants to avoid straining the grid.

But Trump's endorsement comes as local resistance is hardening. Candidates in Texas, Ohio, Michigan, Wisconsin, and Pennsylvania are campaigning on concerns about rising power bills, water use, land conversion, and the modest number of permanent jobs. Texas even paused new data-center development on August 3, prompting the EIA to slash its 2027 electricity-load growth projection for the state from 14% to 6%.

The Economics of the 'Golden Goose'

Trump's case rests on genuine benefits: multibillion-dollar capital projects can broaden tax bases and support construction jobs. But construction work is temporary, and operational workforces are far smaller than traditional factories. That mismatch is central to community disputes.

'The key economic question is not simply whether a data center produces investment, but who pays for the supporting infrastructure,' said one energy analyst. If developers fund dedicated generation and transmission upgrades, the local bargain looks different than if costs are recovered through household rates. Utilities are already responding: Xcel Energy (XEL)'s Colorado proposal would impose higher rates and stricter requirements on data-center customers needing more than 50 MW.

Political and Global Context

Trump's push aligns with Executive Order 14318, issued July 23, which fast-tracks federal permitting for qualifying AI data centers. But Washington can only do so much: local zoning, property taxes, water access, and grid interconnection remain in state and local hands. That creates tension, as states unevenly weigh moratoria against incentives.

Internationally, the stakes are strategic. The IEA says the U.S. accounted for 45% of global data-center electricity use in 2024, versus China's 25%, and expects the U.S. to lead growth through 2030. But supply-chain vulnerabilities remain: China controls about 99% of refined gallium, critical for advanced chips.

Public Sentiment and the Road Ahead

Public sentiment is increasingly wary. A Reuters/Ipsos survey cited by Brookings found 64% of respondents viewed rapid construction negatively, and 77% worried about higher electricity rates. Only 14% were willing to live near a facility.

This makes data centers a flashpoint in the broader AI dilemma: national ambitions impose concentrated local costs. As one community organizer put it, 'We're not against progress, but we want to know who pays and what we get in return.'

Near-Term Implications

  • More state utility commissions are likely to adopt dedicated large-load tariffs, requiring developers to pay for grid upgrades upfront.
  • Developers may increasingly pair projects with dedicated generation, batteries, or renewable procurement.
  • Local approvals will likely lengthen unless developers disclose water needs, grid plans, and permanent-job estimates.

The issue is poised to loom large in the 2026 midterms, putting Trump's pro-build stance at odds with candidates in both parties responding to voter affordability concerns.

Long-Term Outlook

The most likely outcome is a tougher 'community bargain': faster approvals for projects that transparently fund infrastructure and mitigate impacts. Industry growth remains strong, but bottlenecks in grids, transformers, and public acceptance could determine where capacity gets built. As the IEA warns, transmission can take four to eight years, and half of U.S. data centers under development sit in existing clusters, amplifying reliability risks.

In the end, available power—not just cheap land or tax breaks—may become the decisive factor in site selection. As Trump urges communities to embrace the 'Golden Goose,' the real test will be whether that goose delivers more than it consumes.