Skyrocketing electricity prices fuel political backlash against tech sector’s AI data centers

Skyrocketing electricity prices fuel political backlash against tech sector’s AI data centers

Data centers are being partly blamed for rising electricity bills on the PJM Interconnection grid, which serves over 65 million people across 13 states and the District of Columbia. This increase in energy costs is leading to a political backlash against the tech sector’s rapid development of AI-focused data centers. The issue is gaining prominence in states like New Jersey and Virginia, becoming a significant factor in political discourse and regulatory scrutiny.

STÆR | ANALYTICS

Context & What Changed

The rapid proliferation of generative artificial intelligence has catalyzed an unprecedented demand for computational power, leading to a global boom in data center construction and energy consumption. This is not a linear continuation of past trends; it represents a paradigm shift. Unlike traditional data centers supporting cloud storage and enterprise applications, AI-focused facilities require densely packed racks of high-performance Graphics Processing Units (GPUs) that consume significantly more power per square foot and require more intensive cooling (source: International Energy Agency). This has transformed data center electricity demand from a predictable, manageable growth pattern into a sudden, exponential one.

The PJM Interconnection, the largest grid operator in the United States, is at the epicenter of this shift. Its territory includes Northern Virginia, the world’s largest and most concentrated data center market, often referred to as “Data Center Alley” (source: pjm.com). Historically, utilities and grid planners could forecast load growth with reasonable accuracy, allowing for methodical infrastructure planning. The AI boom has shattered this model. The sudden, massive increase in projected electricity demand from data centers is outstripping planned grid capacity and generation resources. This creates a direct conflict between the technology sector’s growth ambitions, the physical limitations of the energy grid, and the public’s expectation of affordable, reliable power. The issue has moved from technical planning documents into the public sphere, as utility companies file for significant rate hikes and grid operators warn of potential reliability issues, directly linking these challenges to the new data center load.

Stakeholders

Technology Companies: Major cloud providers—Amazon Web Services (AWS), Microsoft Azure, and Google Cloud—are the primary drivers of demand. They are engaged in an intense competitive race to build out AI capabilities, which requires enormous, reliable power sources. These companies face a tripartite challenge: securing sufficient energy to fuel growth, meeting corporate sustainability goals (often centered on 100% renewable energy matching), and managing the reputational and regulatory risk from their growing impact on public infrastructure.

Electric Utilities: Companies like Dominion Energy in Virginia, PPL in Pennsylvania, and Georgia Power are on the front lines. They face a surge in connection requests for loads that can equal a small city’s demand. This requires them to undertake massive, accelerated capital investments in new generation (often natural gas peaker plants for reliability), high-voltage transmission lines, and substations. The costs for these investments are typically socialized across their entire ratepayer base, creating significant public and regulatory friction.

Grid Operators: PJM Interconnection is responsible for maintaining grid reliability and administering the wholesale electricity market for 65 million people. It must balance the new, concentrated demand with existing generation and transmission capacity. This task is complicated by the concurrent retirement of dispatchable fossil fuel plants (coal and nuclear) and the challenges of integrating intermittent renewable resources, making the sudden addition of massive, inflexible data center loads a critical stability concern.

Regulators: State Public Utility Commissions (PUCs) and the Federal Energy Regulatory Commission (FERC) are the arbiters. PUCs must approve utility investment plans and decide how to allocate costs among residential, commercial, and industrial customers. FERC oversees interstate transmission planning and wholesale market rules. These bodies are now tasked with balancing economic development goals against consumer protection and grid reliability mandates in a rapidly changing environment.

Governments (State & Local): For years, state and local governments have actively courted data centers with generous tax incentives, valuing them as a stable, high-value source of commercial tax revenue without significant demand on public services like schools. They now face a potent political backlash from constituents over rising electricity bills, noise from cooling systems, and environmental concerns, forcing a re-evaluation of incentive policies and zoning regulations.

Consumers (Residential & Industrial): All other electricity users on the grid face higher bills to cover the costs of new infrastructure built primarily to serve a single industry. Non-tech industrial users, in particular, may see their global competitiveness erode due to higher energy input costs, creating tension between the digital and physical economies.

Evidence & Data

The scale of the energy demand shift is well-documented by energy authorities and industry analysts:

Global Demand Growth: The International Energy Agency (IEA) projects that global electricity consumption from data centers, cryptocurrencies, and AI could double by 2026, reaching over 1,000 terawatt-hours (TWh). This is roughly equivalent to the entire current electricity consumption of Japan (source: IEA, Electricity 2024 report).

U.S. Grid Impact: In the United States, data centers are projected to consume 9% of the country’s total generated electricity by 2030, a significant increase from approximately 4% in 2022 (source: Boston Consulting Group). This growth is not evenly distributed, creating acute stress points in specific regions.

PJM Load Forecasts: PJM has dramatically revised its load growth forecasts. Its 2024 forecast projects a 7.5 GW increase in its winter peak load forecast over the next decade, nearly doubling the 4 GW increase projected just a year earlier. The report explicitly cites “accelerated growth of data centers” as the primary driver (source: PJM Load Forecast Report 2024).

Utility-Level Strain: Dominion Energy, the primary utility for Northern Virginia, stated in its 2023 integrated resource plan that it expects data center load in its service territory to more than double, from roughly 2.7 GW to a potential 10 GW by 2030. To meet this, the utility has proposed building new natural gas plants, sparking controversy (source: Dominion Energy regulatory filings). In 2022, the utility had to temporarily pause new data center connections in parts of Loudoun County due to a lack of transmission capacity, a clear signal of infrastructure falling behind demand (source: company press releases).

Cost & Price Impacts: The grid upgrades required are substantial. Across the U.S., estimates suggest over $500 billion in new transmission infrastructure investment is needed by 2035 to maintain reliability amid electrification and load growth (source: U.S. Department of Energy). These costs translate directly to consumer bills. In Georgia, regulators approved a plan for Georgia Power to build new gas turbines, citing data center demand, which is projected to add between $6 and $16 per month to residential bills over the coming years (source: Associated Press).

Scenarios (3) with probabilities

1. Muddled Response & Growing Friction (High Probability: 60%): In this scenario, action is fragmented and reactive. Individual states and localities enact a patchwork of regulations, including zoning moratoriums, special energy tariffs for high-density loads, and stricter efficiency standards. Utilities struggle to get regulatory approval for large-scale grid upgrades, leading to delays and cost overruns. Political friction remains high as the tech industry lobbies for its needs while consumer advocates and other industrial groups push back. Grid reliability becomes a more frequent concern, with localized capacity shortfalls. This path leads to suboptimal outcomes, slowing digital infrastructure growth while failing to fully protect consumers.
2. Coordinated Strategic Build-Out (Medium Probability: 25%): A more proactive approach emerges, driven by a combination of federal incentives (akin to the Inflation Reduction Act) and state-level planning. Governments, regulators, utilities, and tech companies collaborate on integrated energy and digital infrastructure plans. This involves fast-tracking permitting for critical transmission projects, reforming wholesale energy markets to better value reliability, and creating “green data center zones” that co-locate new data centers with new clean, firm power sources like advanced nuclear (SMRs), geothermal, or massive solar-plus-storage projects. This scenario requires significant political will and public-private partnership but offers a path to sustainable growth.
3. Tech-Led Decoupling & Balkanization (Low Probability: 15%): Frustrated by the slow pace of public grid upgrades and regulatory hurdles, major tech companies significantly accelerate their own energy infrastructure investments. They move beyond purchasing renewable energy credits to directly financing and operating their own power sources, such as dedicated SMRs or large-scale microgrids, effectively taking their load “behind the meter” and off the public grid. While this could alleviate strain on the public system, it raises profound policy questions about equity, the privatization of critical infrastructure, and the potential for a less reliable, higher-cost public grid serving remaining customers.

Timelines

Short-Term (1-2 Years): Expect an increase in contentious utility rate cases, local zoning battles over new data centers and substations, and the introduction of state-level legislation to study or pause data center development. Electricity prices in affected regions will continue to rise.

Medium-Term (3-5 Years): The first major grid upgrades and new gas peaker plants planned today will begin to come online. New state-level regulatory frameworks specifically for high-density loads may be implemented. Tech companies will face increasing pressure to fund grid infrastructure directly as a condition for connection.

Long-Term (5-10+ Years): The success or failure of medium-term strategies will become clear. If a coordinated build-out is pursued, new long-lead-time generation sources (e.g., SMRs, advanced geothermal) could begin to be deployed. The energy consumption of AI models may become a key factor in their design and deployment, driven by cost and regulation.

Quantified Ranges

Demand Increase: Data center load on the U.S. grid is expected to grow by 15-25 GW between 2023 and 2028, an unprecedented 5-year increase (author’s synthesis of PJM, MISO, and utility forecasts).

Capital Investment: Meeting this demand will require an estimated $150-$250 billion in investment in new generation and grid infrastructure specifically attributable to data centers over the next 5-7 years (author’s estimate based on utility capital expenditure plans).

Ratepayer Impact: Residential electricity rates in high-growth regions like Virginia and Georgia could see cumulative increases of 15-30% above baseline inflation over the next five years, directly attributable to data center-driven grid expansion (author’s projection based on public rate case filings).

Risks & Mitigations

Risk: Grid Instability & Blackouts. The primary risk is that demand outpaces the construction of new generation and transmission, leading to reliability events, especially during extreme weather.

Mitigation: Accelerate permitting for transmission. Reform capacity markets to better incentivize dispatchable resources. Implement sophisticated demand-response programs where data centers can curtail non-essential processing during peak demand.

Risk: Stranded Assets. Utilities may build natural gas plants to meet near-term demand, only to see them become economically or regulatorily obsolete as decarbonization goals tighten.

Mitigation: Prioritize investment in flexible resources like long-duration energy storage and advanced grid management software. Require new data centers to be “grid-friendly,” with capabilities for demand flexibility.

Risk: Political Backlash Derails Digital Infrastructure. Public opposition to rising costs and local impacts could halt the development of data centers essential for U.S. competitiveness in AI.

Mitigation: Improve transparency in cost-benefit analysis of data center projects. Mandate community benefit agreements that direct a portion of tax revenue to affected localities. Require tech companies to engage in proactive community outreach.

Sector/Region Impacts

Technology Sector: Faces rising operating costs (power is a primary opex driver), significant regulatory hurdles, and reputational damage. The sector will be forced to evolve from being a passive consumer of electricity to an active partner, and potentially owner, of energy infrastructure.

Utilities & Energy Sector: This is a massive growth opportunity, but one fraught with execution risk. It requires unprecedented capital investment, a more agile regulatory approach, and a rapid expansion of the skilled labor workforce for construction and engineering.

Public Finance: The calculus for governments is changing. The substantial tax revenue from data centers may be partially or wholly offset by the political cost of higher electricity rates for all constituents and the need for public subsidies for grid upgrades.

Regions: The pressure is most acute in existing hubs like Northern Virginia, but the search for cheap power and welcoming regulation is driving development in new areas like Ohio, Pennsylvania, and Georgia, spreading the infrastructure and political challenges nationwide.

Recommendations & Outlook

For Government & Regulators:

1. Develop Integrated Planning: Create state-level integrated energy and digital infrastructure plans that align economic development goals with grid planning realities. Do not assess data center proposals in a vacuum.
2. Reform Rate Design & Cost Allocation: Explore new tariff structures, such as fixed-cost contributions for grid upgrades or dynamic pricing, that ensure data centers bear a more direct and equitable share of the infrastructure costs they necessitate.
3. Streamline Permitting: For critical infrastructure that supports broad public benefit, such as major transmission lines, federal and state governments must streamline and accelerate the permitting process.

For Utilities:

1. Proactive Scenario Planning: Move beyond traditional load forecasting to model more extreme growth scenarios and develop contingency plans for generation and transmission.
2. Engage as Partners: Work with tech companies early in their site selection process to guide them toward locations with existing or planned grid capacity, avoiding future bottlenecks.

For Technology Companies:

1. Fund the Fix: Move beyond Renewable Energy Credits and Power Purchase Agreements to making direct, upfront investments in the grid infrastructure required to serve their facilities.
2. Design for Efficiency: Make energy consumption a core design parameter for both hardware (chips, servers) and software (AI models). Promote research into more energy-efficient computing.

Outlook:

(Scenario-based assumption) The tension between exponential digital growth and the physical, analog constraints of the energy grid will be a defining policy and infrastructure challenge for the next decade. The Muddled Response scenario is the most probable path in the near term, given the complexities of our regulatory systems and political polarization. However, the sheer scale of the demand and the severity of potential grid failures may act as a catalyst, forcing stakeholders toward the more collaborative Coordinated Strategic Build-Out scenario over the medium term. The risk of inaction is not just higher electricity bills, but a potential constraint on the development of artificial intelligence and the future of the digital economy in the United States.