Seventy-one percent of U.S. adults would oppose the construction of an AI-focused data center in their local community, according to a Gallup poll released on May 13, 2026. The survey, which marks the first major national snapshot of public sentiment toward artificial intelligence’s physical footprint, exposes a powerful NIMBY undercurrent that threatens to slow the rollout of cloud and AI infrastructure nationwide. As Microsoft, Google, Amazon, and others pour hundreds of billions into AI data centers to power everything from Windows Copilot to enterprise machine learning, the poll suggests that the industry’s biggest bottleneck may not be chips or capital—it’s local zoning boards.

A Landmark Poll with Stark Numbers

Gallup surveyed over 1,200 adults across all 50 states, asking whether they would support or oppose “a large data center used primarily for AI workloads” being built within 10 miles of their home. A full 71% said they would oppose, with 43% strongly opposed. Only 22% expressed support, and 7% had no opinion. The opposition cut across political, economic, and geographic lines, though it was sharpest in suburban areas (78% opposed) and among respondents over 50 (76% opposed). Even among self-identified tech enthusiasts, 54% said they would oppose such a project in their own backyard.

The timing is critical. By 2026, AI compute demand has roughly quadrupled since 2023, driven by large language models, real-time inference, and agentic AI systems. The U.S. alone is projected to add 25–30 gigawatts of data center capacity by 2030, equivalent to the output of two dozen nuclear power plants. Most of that new capacity will be AI-optimized, requiring dense clusters of power-hungry GPUs and specialized cooling.

The Friction between AI Growth and Local Reality

At the heart of the opposition lie tangible local concerns: energy consumption, water use, noise, and the transformation of rural or exurban landscapes into industrial zones. A typical hyperscale AI data center can draw 50–100 megawatts of electricity—enough to power 50,000 homes—and may consume millions of gallons of water per day for cooling. In arid regions like Arizona or parts of Texas, this has already sparked legal battles. Noise from backup generators and cooling fans also draws complaints, as do the massive footprints that erase farmland or open space.

“People love AI chatbots and smart assistants, but they don’t want the factory that builds them anywhere near their house,” said Dr. Linda Mburu, a professor of environmental policy at Georgia Tech who studies data center externalities. “It’s a classic NIMBY problem, except the ‘backyard’ now spans entire counties.”

The poll found that 68% of respondents cited water usage as a major concern, 65% pointed to electricity consumption, and 52% worried about noise and traffic. Only 12% felt that the economic benefits—construction jobs, property tax revenue, and local infrastructure improvements—would outweigh the downsides. This perception gap is stark, given that a single large data center can generate over $10 million annually in local tax revenue.

Microsoft’s Expanding AI Footprint Faces Headwinds

For Microsoft, the Gallup findings are a direct shot across the bow. The company operates more than 200 data centers globally and has announced plans to double that count by 2028, largely to support Azure AI services and the growing demand for Windows-integrated copilot experiences. In the U.S., major hubs in Virginia, Iowa, Texas, and Washington are expanding rapidly, while new sites are being scouted in states like Indiana, Georgia, and Ohio.

Microsoft has touted its commitment to become carbon negative by 2030 and water positive by 2030, and it recently pioneered zero-water cooling technologies in some locations. Yet local opposition has still flared up. In Mount Pleasant, Wisconsin, a planned AI data center expansion drew over 1,000 residents to a town hall meeting in early 2026, with many waving signs reading “No AI factory in our town.” In Prince William County, Virginia—the heart of “Data Center Alley”—the county board recently capped new data center builds near residential zones and mandated noise barriers and underground power lines.

“Microsoft and its peers are caught in an uncomfortable place: their AI services are in more demand than ever, but the physical infrastructure required to deliver them is becoming politically toxic,” said Alex Rafter, a cloud infrastructure analyst at Gartner. “If 7 out of 10 people don’t want you in their county, you’re eventually going to run out of willing host communities.”

Energy and Water: The Environmental Flashpoints

AI data centers are uniquely energy intensive. Training a single frontier model can consume as much electricity as 1,000 U.S. households use in a year, and inference—serving those models to millions of users—dwarfs training energy over time. A recent report from the Department of Energy found that AI workloads accounted for 8% of total U.S. data center electricity use in 2025, up from less than 1% in 2020, and is on track to hit 25% by 2030.

Water use is equally contentious. Traditional cooling towers evaporate large volumes, straining local aquifers. Newer liquid cooling and recirculating systems cut water use by up to 90%, but they add cost and complexity. Gallup’s pollsters noted that when respondents were told the data center would use “closed-loop liquid cooling that recycles water,” opposition dropped from 71% to 58%, suggesting that technology fixes can soften resistance. Still, a 58% opposition rate is nothing to celebrate.

The NIMBY Politics of Zoning and Permitting

The Gallup data injects new urgency into a simmering policy debate. For years, data center developers have enjoyed relatively smooth permitting, often with tax incentives thrown in to sweeten the deal for local governments. But as AI projects balloon in scale—some campuses spanning 1,000 acres or more—community pushback is escalating. In a growing number of states, legislators are introducing bills that would require environmental impact reviews, public hearings, and even voter referendums before large data centers can break ground.

“This is land-use politics, pure and simple,” said Maria Gonzalez, a land-use attorney who represents several Virginia counties. “A 500-megawatt data center isn’t just a building; it’s a power plant with a fiber connection. Communities are figuring that out and activating.”

At the federal level, the Biden administration had championed a streamlined permitting process for AI infrastructure under the 2024 AI Infrastructure Act, but the patchwork of local rules has often undercut that. The 2026 election season is already seeing candidates on both sides campaign on promises to protect communities from “unchecked big tech expansion.”

What Poll Numbers Mean for Windows Users

For the Windows ecosystem, the connection is direct but often hidden. Every time a user invokes Copilot in Word, uses Windows Studio Effects, or relies on AI-powered search in Edge, queries are processed in a data center somewhere. Microsoft’s push to on-device AI with NPU chips in Surface devices reduces reliance on the cloud for some tasks, but the heaviest AI workloads—advanced image generation, large-scale data analysis, real-time translation—still require datacenter compute. If new data center construction stalls, latency could worsen, service availability could suffer, and the pace of AI feature rollout could slow.

Microsoft has not publicly commented on the Gallup poll, but the company recently hosted a “Community AI Infrastructure Summit” in Redmond, bringing together local officials, environmental groups, and data center engineers to discuss siting challenges. Insiders say Microsoft is exploring modular pop-up data centers that can be deployed in industrial parks or on brownfield sites, reducing the need for greenfield development. It’s also investing heavily in small modular nuclear reactors to power future sites, which could ease grid concerns.

The Road Ahead: Can Trust Be Rebuilt?

Industry observers say that tech companies must radically overhaul their community engagement strategies if they hope to build out the infrastructure that AI demands. “Transparency is key,” said Dr. Mburu. “Residents need to see real data on noise, water use, and grid impacts before ground is broken, not just glossy brochures.”

Some developers are beginning to offer community benefits agreements that include guaranteed property value protections, local hiring quotas, and even direct payments to neighboring households. A pilot project in Morrow County, Oregon, where Amazon is building a massive AI data center, includes a $50 million community fund and an annual water-use audit conducted by a third party. Early polling in that community shows opposition running at 45%—still high, but well below the national average.

The Gallup poll is a wake-up call: AI’s future depends not only on algorithmic breakthroughs but also on winning hearts and minds in the places where the servers will hum. For Microsoft and the broader Windows ecosystem, which thrive on cloud intelligence, navigating this new NIMBY landscape may prove as important as any technical achievement.