The Data Center Backlash:
A Live Case Study
This page documents a movement that did not exist three years ago and is now reshaping the politics of AI infrastructure nationally. It is a perfect illustration of the site's central argument: communities objecting to data centers are responding rationally to real local harms — and are largely unaware of the broader structural story those harms are part of.
blocked outright
delayed
groups in 40 states
What Communities Are Objecting To
The opposition is not monolithic — different communities emphasize different harms. But the concerns appear in virtually every campaign in roughly the same order of frequency.
- >40% Water consumption. By 2028, AI data centers could use as much water as 18.5 million households just for cooling their servers. In regions already experiencing water stress, this is the primary flashpoint.
- 2nd Electricity costs and grid strain. In the PJM electricity market stretching across 13 states and 65 million people, data centers drove a $9.3 billion price increase in the 2025–26 capacity market. Carnegie Mellon estimates data centers could increase the average US electricity bill by 8% by 2030, with the figure potentially exceeding 25% in high-demand markets. Electricity costs rose nearly 7% last year — more than twice the overall rate of inflation — costing the average household $123 more in 2025.
- 3rd Noise pollution. Large cooling systems run 24 hours a day. Facilities located near residential areas generate continuous industrial-level noise. This concern appears in virtually every local opposition campaign regardless of geography.
- Also Loss of farmland, property values, and community character. Additional documented harms include massive amounts of e-waste, unrelenting noise pollution, loss of farmland, and lost revenue from data center tax incentives.
Case Study: Utah, May 2026
Utah's approval of a major data center project in May 2026 — over sustained community opposition — illustrates the tension at the heart of the backlash movement. Local residents raised concerns about water use, power consumption, and the adequacy of the public review process. State officials approved the project citing economic development and job creation.
The Utah case demonstrates a pattern playing out across the country: community opposition is real, organized, and well-documented — and is frequently overridden by state-level economic development priorities. The question of whether local concerns can prevail against state and federal economic pressure is unresolved.
Source: The Guardian, May 13, 2026. Full article at theguardian.com.
The Legislative Response
The backlash has moved from town halls to state legislatures to Congress faster than most observers anticipated. The political alignment has surprised commentators — opposition is genuinely bipartisan at the local level.
| Level | Action | Status |
|---|---|---|
| Federal | The AI Data Center Moratorium Act (Sanders/AOC, March 2026) — nationwide halt on constructing or upgrading data centers with power demand of 20 MW or more until "strong national safeguards" are in place. | Introduced · limited co-sponsors |
| State | Lawmakers in at least 11 states have proposed pauses on data center construction. Virginia's long-running data center tax break — worth roughly $1.6 billion annually — came under direct challenge. | Active in multiple states |
| Local | Communities in at least 14 states have enacted temporary pauses on data center development. Wisconsin passed the nation's first anti-data center referendum. North Carolina moved toward a one-year moratorium after project withdrawal. | Accelerating |
Communities objecting to data centers are right about their local harms. The water consumption is real. The electricity rate increases are documented. The noise is continuous. Their opposition is rational and informed.
What most communities don't yet know — because it's rarely part of the local conversation — is the broader structural context: the AI being built in these facilities is owned by some of the wealthiest people and institutions on earth, while the infrastructure costs, the rate increases, and the environmental burden are socialized to ordinary ratepayers and residents.
A moratorium shifts the geography of who bears the burden — AI development migrates to Canada, the EU, UAE, and Southeast Asia. It doesn't fix the underlying inequity. The people who lose most from a moratorium are not the owners of AI, who adapt. They are the communities that were promised jobs and tax revenue and instead got opposition campaigns. The backlash is the right instinct aimed at the wrong target. The target is the structure, not the building.
What a Moratorium Does — and Doesn't Do
- Pauses new construction pending regulation
- Creates time for communities to organize and respond
- Forces a policy conversation at state and federal level
- Protects specific communities from immediate burden
- Signals political accountability to data center developers
- Stop AI development — existing models and cloud infrastructure continue
- Fix the ownership and wealth distribution structure
- Prevent development migrating to other countries
- Resolve the annotator labor conditions
- Change who profits from AI or who bears its costs