The Midterm Data Center Backlash Could Raise Your AI Costs

In Brief

U.S. midterm voters backed widespread limits on data centers and elected candidates who favor slowing their expansion.

Paul Roetzer sees a potential consequence for businesses using AI: higher subscription prices or tighter usage limits if computing capacity becomes harder to build.

What Happened

Voters favored restrictions on data centers in 40 of 41 local ballot measures during the U.S. midterm elections, either approving limits or rejecting proposals supporting development. The results, part of the broader backlash covered by Axios, put new political obstacles in the path of the infrastructure powering AI.

Incoming governors Gina Hinojosa in Texas, Keisha Lance Bottoms in Georgia, and Amy Acton in Ohio all supported pauses on new development, with different conditions for allowing projects to proceed. Texas is especially important: The Washington Post cited industry estimates placing a quarter of planned new data centers there.

The objections center on electricity bills, water use, and communities' ability to shape projects. SmarterX founder and CEO Paul Roetzer explored how those concerns could reach business AI budgets, on Episode 250 of The Artificial Intelligence Show.

The Key Numbers

40/41 - Local ballot outcomes favoring limits on data centers

25% - Share of planned new data centers in Texas, according to industry estimates cited by The Washington Post

25-35% - Share of next year's U.S. data center buildout potentially at risk, in investor Brad Gerstner's forecast reported by The Wall Street Journal

2028-2029 - Project completion years most likely to show the first substantial effects from the election, in Roetzer's assessment

Building Delays Could Reach Your AI Budget

The political environment has changed. Roetzer expects more regulatory activity, particularly at the state level, and greater scrutiny of AI leaders. "So it's just going to make it more challenging, definitely, for the AI labs to move forward at the pace that they were prior to this," he says.

Every AI task needs computing power. When a model answers a question, generates a video, or carries out a sequence of tasks, data center servers do the work. Roetzer's concern is that slower construction could constrain the computing supply available for those activities, putting pressure on the price of access.

"The cost of your licenses may go up real soon."

— Paul Roetzer, founder and CEO of SmarterX, Episode 250 of The Artificial Intelligence Show

Prices and timing are still uncertain, but Roetzer also sees the possibility of reduced usage allowances, especially for demanding tasks such as video generation and agents that perform multiple steps. An unchanged monthly fee could still buy less compute.

Negotiations could matter more than permanent bans. Roetzer expects Democrats to use their stronger position to secure benefits for local communities, while recognizing the economic importance of continued construction. "So the Democrats are going to hold a lot of negotiating power here, and I would imagine they're going to use it to an extreme," he says.

He expects projects already underway to keep moving, with the larger effects more likely on future capacity. That means the election is unlikely to stop current AI development.

SmarterX Take

Business AI planning now has another dependency to consider: the communities hosting the infrastructure. Leaders adopting more demanding AI workflows need to understand how changes in available computing power could affect the economics of those workflows. Today's subscription price is only part of that calculation.

The community concerns also deserve attention. Roetzer sees the election creating negotiating power for residents seeking protection from new costs, plus a share of the gains. How developers respond will help determine whether expansion moves with broader support or repeated delays.

What to Watch

OpenAI is proposing benefits for its host communities. Its National Community Compact makes commitments for future data centers it owns and operates, including covering energy and infrastructure costs, protecting local resources, and creating opportunities for local workers. Its Compute Campus Policy Blueprint proposes broader protections for host communities.

Pricing and usage limits will show how providers manage demand. Roetzer's scenario becomes more consequential if tighter access affects the tools businesses already depend on. For now, the election results and the eventual effects on an AI subscription remain separate.

How Dependent Is Your Business on AI Access?

Seventy-four percent of respondents rate AI as critically or very important to their success in the next twelve months, according to the 2026 State of AI for Business Report. That level of importance makes the practical terms of access worth understanding. A company building AI into daily work has a stake in both what the technology can do and the conditions under which employees can use it.

Based on more than 2,100 responses across roles, functions, and industries, the report examines adoption, readiness, and the barriers organizations face as they expand AI use. Read the full report →