Americans Turning Against Data Centers: Why Your Town Might Be Next

71% of U.S. adults now oppose data center construction in their local area, according to a recent Gallup poll. That number exceeds opposition to nuclear power plants — a threshold nobody in the tech industry saw coming two years ago.
This isn’t a fringe movement. It’s a structural shift in how communities across the country weigh the promises of AI infrastructure against tangible local costs. And if you’re in tech, you need to understand the mechanics of this backlash — because it’s already reshaping where data centers get built, how projects get approved, and what the next five years of AI infrastructure investment actually look like.
The core tension is straightforward: financial benefits flow to investors and hyperscalers, while rural communities absorb the noise, water consumption, energy price hikes, and land use disruption. That asymmetry is what’s driving the opposition — and the question of whether your town is next deserves a serious answer.
Key Takeaways
- According to a Gallup poll cited by Forbes, 71% of U.S. adults oppose data center construction in their local area, surpassing opposition to nuclear power plants.
- A Washington Post/Schar School poll found resident comfort with local data center construction in Virginia dropped 34 percentage points between 2023 and 2026, falling to just 35%.
- According to a Lawrence Berkeley National Laboratory report, data centers consumed 4.4% of U.S. electricity in 2023, potentially rising to 6.7%–12% by 2028.
- Over 1,500 data centers are in various stages of development nationally, with the majority proposed in rural communities, according to Pew Research Center.
- No federal uniform standards currently govern data center siting or environmental impact, leaving rural communities with minimal regulatory protection.
The Build-Out That Outran Public Trust
Three years ago, data centers were an infrastructure category most Americans couldn’t identify. Today they’re zoning fights, county commission battles, and state legislative debates.
The acceleration is real. Pew Research Center data puts over 1,500 data centers in active development nationwide, with the majority targeted at rural communities — often in economically distressed areas that previously absorbed manufacturing job losses. The pitch to local officials is consistent: jobs, tax revenue, and economic modernization.
That pitch worked until recently. The problem is the gap between promised and delivered outcomes. Brookings Institution research indicates data centers typically create far fewer permanent jobs than industry projections claim. The Box Elder County, Utah case from May 2026 illustrates exactly how this breaks down: a unanimously approved 40,000-acre facility projected 2,000 permanent jobs — but the majority of residents who showed up to the public meeting opposed the project outright. The facility will consume more than twice the energy currently used by the entire state of Utah.
Virginia’s trajectory tells the longer story. It’s the state with the most data centers nationally, and a Washington Post/Schar School poll found resident comfort with local construction dropped 34 percentage points between 2023 and 2026, landing at just 35%. That’s what overexposure looks like. Communities that initially accepted the trade-off are now watching utility bills climb, water tables drop, and promised economic benefits stay abstract.
The Energy and Water Math Is Getting Harder to Ignore
Lawrence Berkeley National Laboratory estimated data centers consumed 4.4% of U.S. electricity in 2023. That share could reach 6.7%–12% by 2028. Rural communities host the facilities but often pay elevated utility rates as grid strain increases.
Water consumption compounds the issue. Modern hyperscale facilities use evaporative cooling systems that drain thousands of gallons daily. In drought-affected regions like Box Elder County — already under extreme water stress — that’s not an abstract environmental concern. It’s a direct threat to agricultural operations and municipal water supply.
Pew Research data shows 39% of U.S. adults consider data centers “mostly bad” for the environment, 38% say mostly bad for home energy costs, and 30% say mostly bad for quality of life near facilities. Those numbers suggest the energy and water narrative has moved from activist talking point to mainstream perception.
This approach isn’t without solutions — some developers are piloting dry cooling and waste heat recovery systems that cut water use significantly. But adoption remains limited, and communities have no way to require it without regulatory leverage they currently don’t have.
The Regulatory Vacuum That Enables Fast Approvals
No federal uniform standards govern data center siting or environmental impact assessment. That’s not an oversight — it’s a structural advantage for developers. State and county-level processes vary wildly, and economically distressed communities often lack the legal and technical resources to mount effective opposition.
The Box Elder County vote makes this concrete. Commissioners approved a project that the majority of residents at the public meeting opposed. Investor Kevin O’Leary, one of the project’s backers, publicly dismissed opposition as “professional protesters,” according to USA Today’s reporting. That framing — opposition as astroturfed noise rather than legitimate community concern — is a recurring pattern in approval processes.
The absence of federal standards means the fight stays local, fragmented, and asymmetric. Developers arrive with experienced legal teams and pre-built approval playbooks. Rural county commissions don’t have equivalent resources. That gap is why projects keep clearing approval despite visible public resistance.
Why Opposition Now Exceeds Nuclear
Forbes analysis citing Gallup frames this as a visibility problem: technologies gain public acceptance when benefits become tangible and locally visible. Roads meant mobility. Power plants meant light. AI infrastructure’s value proposition remains abstract for most host communities. The servers running large language models don’t translate into a faster commute or cheaper electricity.
Nuclear opposition historically ran high because communities understood the downside risk even without direct experience of it. Data center opposition now exceeds that threshold because communities are experiencing the downsides — and the upsides remain invisible.
Who Bears the Cost vs. Who Gets the Benefit
| Factor | Local Community | Developer / Investor | Tech End-User |
|---|---|---|---|
| Energy cost impact | Higher utility bills | Fixed power purchase agreements | Indirect / minimal |
| Water consumption | Direct depletion | Operational cost | None |
| Noise / land use | Direct disruption | Site selection variable | None |
| Job creation | Fewer than projected (Brookings) | Construction boom | None |
| Tax revenue | Short-term, often negotiated down via incentives | Tax breaks common | None |
| AI capability benefit | Abstract / delayed | Revenue-generating | Immediate |
This asymmetry explains the political math. Communities aren’t reflexively anti-technology — a Politico poll found 37% of Americans would support a local data center versus 28% opposed. But support erodes fast when promises don’t materialize and costs arrive immediately. Virginia is the clearest case study in what that erosion looks like over time.
Three Groups, Three Different Problems
For infrastructure developers and hyperscalers, the approval pipeline is getting longer and more expensive. Virginia’s 34-point comfort drop should register as a leading indicator. Siting strategies built on rural community inexperience won’t hold for another investment cycle. Proactive community benefit agreements — real ones, not vague jobs projections — are shifting from nice-to-have to approval-critical.
This doesn’t always work either. Even well-structured benefit agreements can collapse when energy and water impacts exceed initial projections. Developers who treat community engagement as a one-time approval step rather than an ongoing commitment tend to face the sharpest backlash when problems emerge later.
For state and local policymakers, the regulatory vacuum is the immediate lever. Without federal standards, individual states that move first on environmental review requirements and community impact assessments gain both policy credibility and negotiating power with developers. Watch for state-level legislation in Virginia, Texas, and Utah within the next 12 months.
For tech professionals and investors, the backlash is now a material risk factor. Opposition compounding existing Wall Street skepticism about AI ROI — Forbes notes revenues outside a handful of firms remain modest relative to infrastructure costs — creates a double exposure. Projects that stall in approval add years to already-stretched timelines.
What to watch next:
- Federal siting legislation gaining traction in the 2026–2027 session
- Whether hyperscalers shift toward co-location in existing industrial zones to reduce community friction
- How Virginia’s continued comfort-level decline affects the most data-center-dense market in the country
Where This Goes in the Next 12 Months
Rural communities are next because that’s where approvals have historically been fastest and resistance weakest. But that calculus is changing.
A few things are now clear from the data:
- Opposition surpassing nuclear plant resistance represents a real threshold shift, not a polling artifact
- Virginia’s 34-point decline in three years shows the trajectory in mature markets
- No federal framework means the fight stays fragmented — but that can flip quickly with one piece of federal legislation
- The jobs-and-tax-revenue pitch is losing credibility as Brookings data on actual job creation spreads
Near-term, expect more Box Elder County-style fights in Utah, Texas, and upstate New York. Medium-term, expect the first serious federal siting standards proposal to gain committee traction.
The infrastructure buildout for AI isn’t slowing down. But the approval environment just got meaningfully harder. The companies that treat community engagement as a real constraint — not a PR problem — will move faster than those that don’t. The ones dismissing opposition as “professional protesters” are building a longer queue for themselves.
What’s your read on whether federal siting standards would actually slow AI infrastructure buildout, or just redirect it to urban industrial zones?
References
- Take This Data Center and Shove It - The American Prospect
- How Much of Data-Center Activism Is Really AI Slop?
- From Texas To New York, Resistance To Data Centers Is Growing - Newsweek
Photo by Etienne Girardet on Unsplash


