The infrastructure bill for America’s AI ambitions is arriving at the power grid — and it is large. U.S. investor-owned utilities have increased their five-year capital spending projections by more than 27%, to at least $1.4 trillion through 2030, according to an analysis of spending plans from 51 major utilities. The jump, up from $1.1 trillion projected just a year ago, is driven primarily by surging electricity demand from AI data centers and the aging grid infrastructure that must be rebuilt to deliver it.
The scale of the buildout represents the largest planned expansion of U.S. electrical infrastructure in modern history — and it comes with a price tag that is increasingly being passed to consumers.
The Demand Driver: AI Data Centers
Utilities are candid about what is behind the numbers. Training and running large AI models is an energy-intensive operation. A single large-scale AI training run can consume as much electricity as thousands of homes use in a year. As cloud providers, AI companies, and enterprises race to build or lease data center capacity, the power draw on regional grids has grown faster than existing projections anticipated.
The pipeline is not slowing. Hyperscaler capex commitments for data center construction in 2025 and 2026 have been running at record levels, with Microsoft, Google, Amazon, and Meta collectively committing hundreds of billions to new facilities. Each facility needs reliable, high-capacity power connections — and utilities are now scrambling to provide them.
New transmission lines, upgraded substations, and expanded generation capacity are all part of the $1.4 trillion plan. For many utilities, the AI-driven demand surge has pushed investment timelines forward by years, compressing projects that were previously scheduled for the late 2020s or early 2030s.
The Consumer Equation
Grid investment does not happen in isolation from rate structures. U.S. electric and gas utilities have filed requests to raise customer bills by $31 billion in 2025 alone — more than double the requests filed in 2024. The trajectory suggests that residential customers could ultimately absorb close to $700 billion of the $1.4 trillion in planned infrastructure costs.
The political tension is emerging. Consumer advocates and some state regulators have begun questioning why households should subsidize grid upgrades that primarily benefit large commercial data center operators. Several state utility commissions are now reviewing whether cost-allocation frameworks need to be updated to assign more of the grid upgrade burden to the industrial customers driving the demand.
For now, utilities argue that broader grid modernization benefits all customers — through improved reliability, reduced outage durations, and integration of renewable energy that accompanies the buildout. The counterargument, gaining traction in regulatory proceedings, is that the timeline and scale of the current buildout are primarily AI-driven, and the beneficiaries should pay accordingly.
Supply Chain and Workforce Pressures
The $1.4 trillion commitment is not just a financial figure — it is a procurement challenge. Grid-scale transformers, high-voltage cable, and specialized switchgear face multi-year lead times from manufacturers who were not prepared for demand at this scale. Several utilities have flagged equipment delivery timelines stretching to 2028 and 2029 as a binding constraint on how quickly they can execute.
Skilled labor is an equally tight bottleneck. The electrical construction and engineering workforce has not grown proportionally to the planned spending, and competition for linemen, substation engineers, and project managers is intense across all U.S. regions.
A Structural Shift in Energy Economics
What the $1.4 trillion figure signals, beyond any single project, is a structural realignment of the U.S. energy economy. For decades, electricity demand growth in the U.S. was flat to declining, driven by efficiency improvements in buildings, appliances, and industrial processes. AI has reversed that trend abruptly.
The International Energy Agency and the Department of Energy both revised their U.S. electricity demand forecasts upward significantly in 2025, with data centers expected to account for a growing share of total national consumption through the end of the decade.
For utilities, the AI era is an unexpected growth mandate after years of flat revenues. For AI companies, it is a reminder that the most fundamental constraint on scaling intelligence may not be compute, data, or algorithms — it may be electrons.
