Why AI Data Centers Became a Bipartisan Backlash Issue in Texas

AI data centers are becoming politically difficult not because candidates have abandoned economic growth or the AI race, but because the costs of building the infrastructure are intensely local. Residents can see the construction, hear the noise and worry about water supplies and electricity bills; the promised gains—investment, construction activity and tax revenue—can feel more distant, particularly when projects receive public incentives. 12

Texas is the clearest example of that collision. A state that promoted itself as an AI-development hub is now pausing parts of the development pipeline and debating tougher rules. The shift gives Republicans a way to embrace AI competitiveness in principle while demanding that companies, rather than ratepayers and host communities, carry more of the infrastructure burden.

What is driving the backlash?

The opposition combines several concerns that cut across party and ideological lines:

• Electricity prices and grid reliability. Data centers are large, concentrated power users. Opponents argue that rapid new demand can strain a grid and leave households exposed to higher costs or reliability risks. 18
• Water, noise and light. Communities have raised concerns about the water needed for cooling, along with noise and night-sky light pollution. 123
• Land use and local control. Facilities can arrive quickly in rural and suburban areas, where residents and local officials may feel they had little role in planning. Abbott has argued that developers helped create the backlash by not collaborating sufficiently with state and local governments. 21
• Tax incentives versus permanent jobs. Critics question whether the long-term tax and employment benefits justify incentives offered to capital-intensive facilities. That argument is especially resonant when households are worried about utility bills. 12

These concerns do not prove that every facility will raise rates, deplete local water or fail to deliver economic benefits. They explain, however, why the political burden has shifted: developers increasingly have to demonstrate who pays for power, water and mitigation—and what a host community receives in return.

Why Texas is the central test case

Texas sits at the intersection of fast AI expansion and an already sensitive debate over electric-grid resilience. Requests from data centers and other large energy users to connect to the Texas grid rose from roughly 48 gigawatts in 2023 to more than 474 gigawatts, according to documents cited by Reuters. The scale of those requests prompted questions about which projects are real, financeable and ready to proceed. 17

Governor Greg Abbott responded by ordering a pause on approvals for new data-center projects in the grid-interconnection process pending an audit and verification process. His directive said projects that do not comply with the process meant to protect grid reliability and resilience must be denied. 18

Abbott's position is a notable change in emphasis. Less than a year after describing Texas as the “epicenter of AI development,” he was seeking broader restrictions, including scrutiny of incentives and stronger safeguards around electricity and water use. 120 Reporting has described subsequent Texas directives as temporarily affecting roughly 1,800 projects while pursuing limits on energy and water consumption, changes to some incentives and measures intended to lower customer electricity costs. 619

The policy logic is increasingly clear: verify demand, protect the grid, require transparency from developers and avoid asking ordinary customers to subsidize new infrastructure.

How Republican candidates are repositioning

The issue has opened an unusual divide within a party generally associated with deregulation, business recruitment and energy expansion. Republican candidates are not uniformly anti-data-center or anti-AI. Instead, many are moving toward conditional development: pauses, audits, local input, water limits and requirements that companies fund the infrastructure needed to serve them.

In Texas, Abbott's shift places data centers at the center of both the governor's race and the Senate contest. Democratic challenger Gina Hinojosa has attacked Abbott's earlier support for the projects, while Attorney General Ken Paxton has advanced his own proposal, including provisions concerning Chinese technology and AI-related child safety. 513 Senator Ted Cruz has also said developers should limit water use, avoid adding to local electric bills and make benefits to host communities clear. 22

That creates a powerful campaign incentive: candidates can criticize corporate giveaways or weak oversight without rejecting AI investment altogether. The most politically durable message is often not “stop AI,” but “protect ratepayers and make developers pay their share.”

The shift extends beyond Texas

Texas is not an isolated case. Republican candidates in several battleground states have taken more skeptical positions on data-center expansion. Michigan Senate candidate Mike Rogers, Pennsylvania gubernatorial candidate Stacy Garrity and Wisconsin Republican gubernatorial candidate Tom Tiffany have called for pauses or campaigned against permissive development, according to Politico. 12

The reaction is bipartisan, not solely Republican. Pennsylvania Democratic Governor Josh Shapiro imposed stricter standards on data-center development, illustrating how local infrastructure concerns can override the usual party divide over industrial development. 5

Campaign advertising reflects that convergence. NPR reported that campaigns and outside groups had spent more than $45 million since January on ads mentioning data centers in gubernatorial, House and Senate races. The major parties' ads frequently make similar arguments: facilities receive too many tax breaks, raise electricity costs and need tighter oversight. 2

National AI strategy versus local politics

President Donald Trump has framed data-center construction as part of U.S. competition with China and the broader effort to secure AI computing capacity. Yet Republican candidates have still found room to favor pauses or restrictions on individual projects; Politico described the moves by Abbott and other battleground Republicans as a break from Trump's pro-buildout approach. 12

This is less a settled disagreement over AI than a disagreement over who bears its physical costs. National advocates emphasize computing capacity, investment and strategic advantage. State and local officials increasingly ask whether projects will bring dedicated generation, pay for grid interconnection, manage their water use and provide meaningful local benefits.

What Big Tech and pro-AI groups are doing

Technology companies and allied political groups are trying to make a local economic case for the buildout, emphasizing investment, construction activity, tax revenue and U.S. competitiveness. Build American AI, a group affiliated with the pro-AI super PAC Leading the Future, said it would spend millions on advertising, grassroots engagement and public education in battleground states including Kansas, Ohio and Wisconsin. 1032

The spending numbers need careful interpretation. More than $45 million refers specifically to campaign and outside-group ads that mention data centers, according to NPR's AdImpact-based analysis. 2 Separate reports describe much larger AI-sector political fundraising, pledges or broader election spending; for example, CNBC reported that two major AI super PACs had raised more than $200 million by July. 37 Those totals should not be treated as data-center-ad spending alone.

The likely political outcome: stricter terms, not an AI freeze

The backlash is pushing candidates toward a new bargain for AI infrastructure. The likely demands are greater transparency over project ownership and power needs, tighter grid-connection reviews, water and pollution safeguards, local participation, and rules requiring developers to cover more of their own infrastructure costs. 171829

That approach lets officials maintain a pro-investment and pro-competitiveness posture while responding to the immediate concerns of communities hosting the facilities. For data-center developers, the lesson is increasingly practical: local acceptance may depend as much on credible commitments to ratepayers, water management and community benefits as on the national case for winning the AI race.