Power and Control: AI Data Centres and the Struggle for Democratic Influence in the U.S.
Artificial Intelligence is often discussed as a digital technology, but in recent years the demand for physical infrastructure expansion is becoming increasingly apparent. Behind every generative AI model are data centres requiring large amounts of land, water and electricity. As governments and technology companies race to expand AI capacity, data centres are reshaping energy systems, local environments and public policy. However, this transformation is occurring under limited democratic oversight and growing concentration of corporate power.
The scale of AI infrastructure growth is unprecedented. Multinational firms including Microsoft, Meta, Google and Amazon are investing billions of dollars into hyperscale data centres designed to support advanced AI systems. According to the International Energy Agency (IEA), global data centre electricity consumption is projected to more than double to approximately 945 terawatt-hours by 2030, a figure that slightly exceeds Japan’s current annual electricity consumption (IEA, 2025). The United States is expected to account for the largest share of this increase, making it the focal point of the global race to build AI infrastructure.
America’s Digital Capital
Nowhere is this transformation more visible than in Northern Virginia. Home to the highest concentration of data centres in the world, the region has become central to America’s AI infrastructure boom. Concurrently, local Virginian communities have made their opposing views on data centre construction publicly known, with the Prince William Digital Gateway emerging as the clearest example of this tension. A proposal for 37 data centre facilities adjacent to Manassas National Battlefield Park was approved by the county’s Board of Supervisors in December 2023, despite a public hearing attended by hundreds of residents. Two community-led groups subsequently sued the county, which led the state court to halt the project after finding that officials had failed to comply with public notice requirements ahead of the vote (Prince William Times, 2026). In this instance, communities managed to disrupt the infrastructure construction, though the victory rested on a narrow procedural technicality rather than on any judgement surrounding the merits of the development itself.
Financial pressures also exacerbate these tensions. Dominion Energy, Virginia’s largest utility provider, reported that data centres accounted for 28% of its Virginia electricity sales in 2025, up from 26% the previous year (Dominion Energy, 2026). The increase in electricity demand led to an 833% rise in the regional grid operator’s capacity auction price for 2025-2026, prompting Virginia’s state regulator to approve a dedicated electricity rate class for large AI data centre customers (American Action Forum, 2026). Such interventions suggest that even when state institutions act, their purpose is to redistribute the new energy costs among existing ratepayers, rather than to constrain the expansion itself.
Environmental Impacts
Water consumption presents a parallel but arguably more concerning issue with the expansion of data centres. Modern data centres rely heavily on water-based cooling systems, drawing particular scrutiny in states such as Arizona, where prolonged drought and declining groundwater reserves have heightened anxieties over resource allocation choices. In 2023, U.S. data centres consumed approximately 66 billion litres of water, with hyperscale and colocation facilities accounting for 84% of that total (Shehabi et al., 2024). This figure is expected to continue rising substantially in the coming years. Communities in states such as Texas and Arizona are regularly encouraged to use water responsibly and support conservation efforts, even whilst data centre developments continue to receive approval nearby.
The City Council in Tucson, Arizona has been fighting against these data centre developments, highlighting the risk of worsening water shortages in Arizona. They successfully halted a proposed data centre plan, Project Blue, by voting against the annexation of the property in 2025 (Hyatt, 2025). Despite this, news reports from Southern Arizona now claim that construction has started in 2026 (AZ Luminaria, 2026). What the Tucson case makes clear is that environmental burden and economic benefit rarely land in the same place. Local residents face water shortages, while investors and technology firms collect the returns. And unlike in Virginia, where a procedural misstep gave communities a foothold in court, water governance across much of the western United States offers no equivalent lever. Once approvals are granted, the decisions tend to be final.
Corporate Power
Beyond environmental concerns, AI infrastructure expansion has intensified questions regarding corporate influence. While local communities often bear the costs associated with increased electricity demand, water consumption and land use, the economic benefits are consistently concentrated among a small number of large technology firms.
The American cloud computing market is dominated by Amazon Web Services, Microsoft Azure and Google Cloud, which collectively control much of the infrastructure required to train and deploy advanced AI systems (Synergy Research Group, 2026). In terms of hardware, Nvidia has also established a commanding position in the market for AI chips.
Governments have facilitated this concentration through tax incentives and planning support. In 2025 at least 38 states offered financial incentives to attract data centre investment, ranging from sales tax exemptions to property tax abatements, often with minimal job creation requirements attached (National Conference of State Legislatures, 2026). Critics argue that such policies weaken public scrutiny by prioritising economic development over accountability, leaving local authorities with far fewer resources and less technical expertise than the corporations with which they negotiate.
This concentration places considerable influence in the hands of a small collection of firms. Decisions regarding where data centres are built, how resources are consumed and how computing power is distributed are increasingly shaped by corporate priorities, yet unlike traditional public utilities, these companies are not subject to equivalent democratic supervision.
Final Thoughts
The expansion of AI data centres is frequently presented as an inevitable consequence of technological progress. However, the infrastructure supporting artificial intelligence is neither invisible nor politically neutral. It consumes public resources, reshapes local communities and influences national energy policy. As AI becomes more central to economic and strategic competition, the debate must move beyond innovation alone. The question facing the United States is not only whether AI infrastructure should continue to expand, but whether democratic institutions can respond to the unprecedented concentration of power emerging behind the companies that increasingly shape modern society.
References
American Action Forum (2026) Virginia’s new data center electricity rate class. Washington, DC: American Action Forum. Available here: https://www.americanactionforum.org/insight/virginias-new-data-center-electricity-rate-class/
AZ Luminaria (2026) ‘“We said no!!”: Dozens protest as construction of Project Blue data center begins’, AZ Luminaria, 24 April. https://azluminaria.org/2026/04/24/we-said-no-dozens-protest-as-construction-of-project-blue-data-center-begins/
Dominion Energy, Inc. (2026) Form 10-K for fiscal year 2025. Filed with the U.S. Securities and Exchange Commission. Richmond, VA: Dominion Energy. https://www.sec.gov/Archives/edgar/data/715957/000095017025028387/0000950170-25-028387-index.htm
Hyatt, A. (2025) ‘Tucson council blocks Amazon data center after public outcry’, Tucson Spotlight, 7 August. Available here: https://www.tucsonspotlight.org/tucson-city-council-blocks-amazon-data-center-after-public-outcry/
International Energy Agency (2025) Energy and AI. Paris: International Energy Agency. https://www.iea.org/reports/energy-and-ai/
National Conference of State Legislatures (2026) Policy snapshot: Data center incentives. Updated 17 April 2026. Available here: https://www.ncsl.org/fiscal/policy-snapshot-data-center-incentives
Prince William Times (2026) ‘Judge overturns Prince William Digital Gateway rezoning’, Prince William Times, 23 February. https://www.princewilliamtimes.com/news/breaking-judge-overturns-prince-william-digital-gateway/article_951d6362-1bbd-416f-89c5-ff322bf49fa8.html
Shehabi, A., Smith, S.J., Hubbard, A., Newkirk, A., Lei, N., Siddik, M.A.B., Holecek, B., Koomey, J., Masanet, E. and Sartor, D. (2024) 2024 United States data center energy usage report. LBNL-2001637. Berkeley, CA: Lawrence Berkeley National Laboratory. https://eta-publications.lbl.gov/sites/default/files/2024-12/lbnl-2024-united-states-data-center-energy-usage-report_1.pdf
Synergy Research Group (2026) Cloud Infrastructure Services. Available at: https://www.srgresearch.com/research/cloud-infrastructure-services
World Resources Institute (2025) 7 ways data centers affect US communities. Washington, DC: World Resources Institute. https://www.wri.org/insights/us-data-center-growth-impacts
