At least 48 data center projects worth approximately $156 billion have been blocked or stalled due to local opposition, according to tracking by Data Center Watch and Carbon Direct. The resistance is accelerating rapidly: project cancellations jumped from 6 in 2024 to 25 in 2025, with more than 20 additional cancellations already logged in early 2026.

The pushback comes as hyperscale operators race to build capacity for AI workloads, which require far more power per facility than traditional cloud computing. Data centers are already among the fastest-growing sources of electricity demand in the United States, placing new strain on regional grids and utility planning.

The NIMBY wave hits silicon

The number of active local opposition groups has surged to 833 across 49 states as of March 2026, united by a common set of complaints: massive electricity consumption, excessive water usage, noise pollution, air quality concerns, and a permitting process that many communities describe as opaque.

Between January 2024 and May 2026, 46 AI data center projects across 20 states, representing investments of between $137 billion and $172 billion, were stalled, blocked, or withdrawn entirely. Two cases stand out for their sheer scale. In Arizona, a $14 billion project by Tract was withdrawn outright. In Virginia, a $24.7 billion QTS/Compass project became the target of lawsuits from local residents — a notable shift in a state that hosts the world's largest concentration of data centers in Northern Virginia's "Data Center Alley," where the industry has long enjoyed strong local support.

Regulators respond with moratoriums

In July 2026, New York announced a halt on large new data center facility proposals, one of the most aggressive regulatory moves to date. Other municipalities have implemented their own moratoriums and stricter zoning regulations, creating a patchwork of local rules that makes site selection considerably more complex for developers.

Google, Microsoft, and Amazon, the three largest hyperscale cloud providers, are all navigating this landscape. Each has announced aggressive AI infrastructure expansion plans, and each now faces the reality that willing communities with adequate power and water resources are harder to find than their growth projections assumed.

What this means for AI infrastructure costs

With $156 billion in projects currently stalled, tech companies face a strategic dilemma. They can spend more time and money on community engagement, offering better compensation packages and environmental mitigation. They can seek out locations with fewer residents and less political friction, though those sites often lack the necessary power grid infrastructure. Or they can invest in technologies that reduce the footprint of data centers, such as more efficient cooling systems or on-site power generation.

The accelerating cancellation rate suggests the problem is getting worse, not better. The jump from 6 cancellations in 2024 to 25 in 2025, with 20 more in early 2026, indicates that community awareness is spreading faster than the industry's ability to address concerns. How developers, utilities, and state regulators reconcile AI-driven demand growth with local acceptance will shape where the next generation of US computing infrastructure ultimately gets built.