The recent controversy over new data centers in the United States centers on the explosive growth of AI-driven facilities and their massive electricity demands, which are straining the national power grid and contributing to sharp rises in residential electricity bills. This issue gained national prominence in 2025–2026 as hyperscale tech companies (Microsoft, Google, Amazon, Meta, and others) raced to build thousands of new data centers to support generative AI, cloud computing, and related technologies. While these centers promise economic growth and technological leadership, critics argue that ordinary households are unfairly subsidizing them through higher power costs, grid upgrades, and infrastructure burdens.
The Scale of Data Center Growth and Electricity Demand
Data centers have long existed, but the AI boom has supercharged their expansion. In 2023–2024, U.S. data centers consumed about 176 TWh of electricity annually—roughly 4–4.4% of total national usage. Projections vary, but the U.S. Energy Information Administration (EIA), Lawrence Berkeley National Laboratory, and Electric Power Research Institute (EPRI) forecast rapid growth: potentially 325–580 TWh (6.7–12%) by 2028, and up to 17% of U.S. electricity by 2030 in high-demand scenarios.
A single hyperscale AI data center can use as much power as 100,000 households (or more for the largest campuses). The EIA’s January 2026 outlook highlighted the “strongest four-year growth in U.S. electricity demand since 2000,” driven primarily by data centers, with national consumption rising from a 2025 record of 4,195 billion kWh to 4,244 billion kWh in 2026 and 4,381 billion kWh in 2027. Regions like PJM Interconnection (13 states from Illinois to North Carolina, serving ~65–67 million people) and ERCOT (Texas) are seeing the fastest growth.
Over 3,000 new or expanded projects are in development, with deals topping $61 billion in 2025 alone. This has created pipeline bottlenecks: nearly half of planned 2026 openings faced delays or cancellations by early 2026 due to insufficient power availability, transformers, and grid connections.
How This Translates to Higher Domestic Power Costs
Utilities must build new power plants (often natural gas, with some nuclear restarts and renewables), transmission lines, and distribution infrastructure to meet the demand. In regulated markets, these capital costs are recovered through rate cases approved by state regulators and spread across all customers via the “rate base.” Data centers pay high demand charges and often sign long-term contracts, but the upfront infrastructure costs (and capacity market obligations) are socialized.
Key impacts include:
National residential price surge: Prices rose more than 36% since 2020 (to ~17.44 cents/kWh by February 2026), outpacing inflation. The EIA expects further increases through at least 2027. Utilities filed a record $31 billion in rate hikes in 2025.
PJM region example: Data centers drove ~63% of the surge in capacity auction prices (from ~$29/MW-day in 2024–25 to $269–$329/MW-day for 2025–27 delivery years, hitting FERC caps). This added ~$9.3 billion in costs for 2025–26 alone, translating to $16–$18/month extra for average households in parts of Maryland and Ohio. Cumulative costs could reach $100–$163 billion through 2033 without reforms.
Virginia (“Data Center Alley”): Dominion Energy serves the world’s largest concentration of data centers. It proposed (and received approval for) rate hikes in 2025–2026, including a new GS-5 rate class for large users (>25 MW). Residents saw winter 2026 bill spikes, partly blamed on cold weather and inflation, but critics link it to data center-driven infrastructure needs. Proposed legislation (e.g., SB 253) would shift ~$5.50/month from residential bills to data centers.
Other hotspots: Similar dynamics in Georgia, Texas, California, and Pennsylvania, with localized hikes of 8–15% or more in data-center-heavy areas. Some states saw jumps as high as 267% over five years in extreme cases.
Goldman Sachs and others warned of broader “trickle-down” effects: higher production costs inflating food, transport, and goods prices.
Core Elements of the Controversy
The debate pits economic benefits against affordability and fairness:
Public and political backlash: Middle-class families feel they are subsidizing trillion-dollar tech firms. Bipartisan outrage spans President Trump (who has pushed tech to “pay their fair share” while fast-tracking permitting) to Democratic senators like Elizabeth Warren, who investigated utility/private equity profits. Proposals include moratoriums (e.g., New York, federal ideas), dedicated “large-load” tariffs, mandatory peak-load curtailment (data centers reducing usage during stress), and on-site generation requirements.
Utility and industry defense: Companies like Dominion argue data centers pay more via separate classes, upfront contributions (e.g., 85% of projected costs), and long contracts. Tech firms have pledged to cover incremental generation costs and explore flexibility (e.g., curtailing during peaks to avoid $40–150 billion in grid investments). They highlight jobs, tax revenue, and potential long-term bill reductions if they fund clean power.
Grid reliability and environment: Concerns include blackout risks, water use for cooling, backup diesel generators, and delayed clean energy transitions (some new demand met by gas). However, data centers can also anchor renewables or nuclear projects.
Regulatory responses (as of April 2026): States are debating cost-allocation reforms; PJM is considering rules for data centers to bid their own capacity or curtail load. The White House has highlighted tech commitments to self-fund. Project delays signal the grid’s limits.
Broader Context and Outlook
Electricity demand had been flat for decades before AI; now electrification (EVs, heat pumps) compounds it, but data centers are the dominant new driver. The controversy reflects deeper questions: How should society allocate the costs of the AI revolution? Is the current utility model outdated for “hyper-loads”?
As of mid-April 2026, the issue remains hotly debated in state legislatures, utility commissions, and Congress. Reforms could shift more costs to data centers, require greater flexibility, or slow builds—but unchecked growth risks sustained bill pressure. Tech’s $700 billion+ 2026 capex plans ensure the tension will persist. For households, the outcome will directly affect monthly bills for years. This is not just an energy story—it’s about who benefits from, and pays for, the AI era.
