AI-Driven Power Boom Forces Decision on U.S. Electricity System Growth
The Energy Information Administration reported U.S. power consumption will rise further in 2026 and 2027, driven by AI-hungry data centers, forcing critical decisions on electricity system expansion financing and infrastructure investment.
Objective Facts
U.S. power consumption, which hit its second straight annual record in 2025, is projected to rise further in 2026 and 2027 to 4,271 and 4,397 billion kWh respectively, driven by AI-hungry data centers and electrification. Load growth that historically ran below 1% annually hit 4% at some grid operators last year, creating unprecedented pressure on transmission and capacity markets. The rapid expansion of artificial intelligence infrastructure is reshaping how U.S. utilities finance massive grid upgrades, with millions of consumers increasingly paying for projects before they are completed. In January 2026, the Trump administration and governors from 13 PJM states issued a statement urging PJM Interconnection to hold an emergency backstop auction to support new capacity entry and preserve regional reliability. AI data center demand contributed to an 833-percent increase in PJM's capacity market auction price for 2025–2026 compared to the previous year. Regional media reporting on data center impacts focuses heavily on water depletion and local infrastructure strain in Virginia, Arizona, and Texas—emphasizing community costs differently than national coverage.
Left-Leaning Perspective
Consumer Reports and Harvard's Howell Peskoe argued that tech companies' voluntary pledges do nothing to help consumers since utilities hold the pen on these issues and utility regulators make final decisions. Consumer advocates emphasize that CWIP financing shifts construction and financing risks onto households through higher electricity bills, and warn that if projected AI demand slows, households could remain tied to the costs of oversized infrastructure investments. Senators Sheldon Whitehouse, Martin Heinrich, and Chris Van Hollen warned that if all gas-powered data center projects currently under development are completed, they will add 12.1 billion tons of lifetime carbon emissions—double the annual emissions from all other U.S. sources—violating climate targets requiring 45% emissions reductions by 2030. Progressive critics argue the Trump administration's deregulatory approach and emphasis on fossil fuels actively undermines clean energy deployment. Julie McNamara, associate policy director at Union of Concerned Scientists, noted that permitting backlogs and efforts to sideline renewable energy are "creating a real near-term crunch on the system". Lori Bird, director of the U.S. Energy Program at the World Resources Institute, said "Companies are scrambling to try to get as much power as they can as quickly as possible. It's a mad rush and a lot of competition for resources", which limits options for sustainable deployment. Total emissions from Google, Amazon, Microsoft, and Meta have risen despite renewable energy purchases, with Google's emissions jumping nearly 50%, Amazon's rising 33%, Microsoft's more than 23%, and Meta's more than 60%. Left-leaning coverage emphasizes that tech companies have broken climate commitments and that ratepayers—particularly low-income households—bear hidden costs through CWIP charges. Progressive outlets downplay industry claims of self-regulation and highlight that existing regulatory frameworks already permit utilities to pass construction costs to consumers before projects operate, making the Ratepayer Protection Pledge essentially unenforceable.
Right-Leaning Perspective
President Trump has emphasized that companies must "pay their own way" on data center energy costs, arguing this approach will "make electricity affordable again but will also encourage these companies to innovate". On March 4, 2026, President Trump convened seven leading AI companies at the White House to sign the Ratepayer Protection Pledge, with signatories agreeing to build, bring, or buy new generation resources sufficient to meet data center electricity demands, and to cover the full cost of all power delivery infrastructure upgrades. Trump's AI Action Plan aims to accelerate U.S. leadership in artificial intelligence through deregulation and expanded energy infrastructure, prioritizing easing restrictions on data centers and energy projects, while the administration argues the plan will counter China's technological advancements. Right-leaning outlets and the Trump administration frame aggressive deregulation as essential for U.S. competitiveness. The biggest issue facing data center development is "time to power," and the Trump Administration has launched an ambitious federal infrastructure initiative to fundamentally reshape the regulatory landscape for data centers through the AI Action Plan and supporting executive orders. Policy advocates argue that "regulatory reforms at both the state and federal levels aimed at ensuring that price signals, not government intervention, determine the trajectory for data-center growth" are needed. In January 2026, President Trump's National Energy Dominance Council took unprecedented action to intervene in the failed PJM power market, with the intervention expected to drive what is expected to be the single largest development of power plants in U.S. history. Right-leaning commentary emphasizes that the current regulatory bottleneck is the real problem—not corporate responsibility. Conservative outlets argue that streamlined permitting, competitive bidding for long-term power contracts, and allowing companies to build their own generation capacity will solve both affordability and supply challenges more efficiently than government mandate or cost-shifting rules.
Deep Dive
The AI data center boom presents a genuine infrastructure inflection point requiring unprecedented investment speed. Bain and Company projects AI data centers could consume 9% of total U.S. electricity by 2030, adding 150+ terawatt-hours of demand the grid was never built to handle. PJM's capacity auction prices surged 833% in just one year due to AI demand, demonstrating how rapidly competitive markets can amplify costs when supply lags. PJM faces an unprecedented power shortfall of up to 60 GW, making the timing of infrastructure decisions genuinely critical. Both sides correctly identify real problems but propose conflicting solutions. The left emphasizes consumer protection and environmental goals—noting that consumer advocacy groups estimate nuclear plant construction costs overrun by an average of 100%, costing roughly twice initial estimates, making CWIP a mechanism for shifting risk to vulnerable ratepayers. The right emphasizes permitting speed and capital efficiency, arguing that market pricing and corporate accountability mechanisms work better than utility commission oversight. Neither side adequately addresses the core tension: the speed required to build 150+ gigawatts of generation by 2030 may be physically incompatible with either traditional utility rate regulation or purely voluntary corporate climate commitments. MIT researchers note that "the demand for new data centers cannot be met in a sustainable way" because "the pace at which companies are building new data centers means the bulk of the electricity to power them must come from fossil fuel-based power plants"—a dilemma that market deregulation and consumer cost controls alone cannot resolve. What to watch: Whether the PJM emergency auction model successfully attracts private capital for 15+ GW of new generation, whether tech companies' Ratepayer Protection Pledge survives first contact with actual utility commission rate cases, and whether the Trump administration's permitting reforms meaningfully accelerate nuclear and renewable deployment or merely enable faster fossil fuel expansion. The June 2027 PJM auction results will reveal whether market-driven procurement outperforms traditional utility planning. State-level rate commission decisions in Virginia, Texas, and California over the next 12 months will test whether voluntary corporate commitments have any teeth against structural utility incentives to shift costs to ratepayers.