Infrastructure, Political Challenges Impede US AI Data Center Growth; Component Glut Possible by 2027
The rapid expansion of AI data centers in the US is facing significant hurdles due to infrastructure limitations, particularly concerning power and water availability, and growing community opposition. These challenges, coupled with a potential oversupply of components, are expected to slow construction and may lead to a market glut by 2027. The article highlights these issues as impacting nationwide data center development, with Virginia cited as an example of increased opposition.
The rapid expansion of AI data centers in the United States is confronting substantial challenges stemming from both infrastructure constraints and escalating political opposition. Projections indicate that the power demands of planned US data center projects could reach 780GW by 2030, surpassing the nation's current peak load of 759GW, making such an expansion difficult to achieve. Concurrently, water consumption for cooling these facilities is set to double or quadruple 2023 levels annually, with a majority of new data centers being built in high water-stress regions, further straining vital resources.
Beyond resource limitations, a significant obstacle is growing community resistance. An estimated 70% of Americans oppose data center construction in their communities, citing concerns over rising utility bills, excessive energy and water use, and noise and air pollution. This opposition has already led to the blocking or stalling of projects worth at least $156 billion across 48 publicly disclosed projects in 2025. This resistance is evolving into legislative hurdles, as exemplified in Virginia, where efforts are underway to end state tax breaks for data center construction, which amounted to $1.6 billion in 2025. Brett Forster, VP of Renewables at McCarthy Building Companies, underscored the severity of these challenges at the Latitude Media’s Transition-AI 2026 Conference, stating that only one in four data center projects is truly viable.
These combined infrastructure and political headwinds are anticipated to slow the pace of AI data center construction, which could subsequently reduce demand for critical AI components like GPUs, memory, and digital storage. Paradoxically, many suppliers of these components have initiated production expansion projects in anticipation of earlier demand, with new capacity expected online by 2027. This scenario of decreased demand coinciding with increased supply could lead to lower prices and an industry glut, potentially ushering in a "bust" period following the recent "boom."