Millions of households across the East Coast face higher electricity bills this summer, and a growing share of that cost pressure traces back to the explosive expansion of AI data centers. Federal forecasts from the U.S. Energy Information Administration project residential electricity prices rising in 2026, with East Coast regions expected to absorb some of the steepest increases in the country. At the same time, state officials and federal regulators are scrambling to decide who should pay for the grid upgrades these facilities demand.
How data center load growth is lifting residential rates in PJM territory
The connection between AI-driven electricity demand and household bills runs through the PJM Interconnection, the regional grid operator serving 13 states and the District of Columbia. PJM territory hosts the densest concentration of data centers in the United States, and the queue of new facilities seeking grid connections has ballooned in recent years. When utilities in this region invest in transmission lines, substations, and generation capacity to serve those large commercial customers, the costs often flow into rate bases that residential customers share.
The EIA’s latest short‑term outlook flags faster growth in commercial electricity use, including data centers, as one of several factors shaping near‑term price trends. East Coast regions are projected to see some of the largest residential price increases, a pattern consistent with the heavy data center buildout concentrated in northern Virginia, Maryland, and neighboring states. The agency’s detailed retail rate data show that cents‑per‑kilowatt‑hour charges in many PJM-heavy states have already been climbing for several years, even before the latest wave of AI-related projects reaches full operation.
The pricing dynamic works like this: data centers consume electricity around the clock at volumes that dwarf typical commercial buildings. Their demand pulls on the same transmission and distribution infrastructure that serves homes. When that infrastructure needs expansion or reinforcement, regulators have historically spread those costs across all ratepayer classes. The result is that a homeowner running air conditioning in suburban Maryland can end up subsidizing grid upgrades built primarily for a hyperscale computing facility miles away.
Utilities and grid planners argue that large new loads can also bring systemwide benefits, such as supporting investment in new generation and stabilizing revenues. But consumer advocates counter that the pace and scale of AI-driven development is unlike past industrial growth, and that existing cost-allocation rules were not designed for dozens of multi-hundred‑megawatt projects clustering around the same substations. In PJM’s case, the concentration of data centers near Washington, D.C., has already forced planners to revisit assumptions about how much spare capacity the system needs to maintain.
Federal and state officials push back on cost allocation
The Federal Energy Regulatory Commission has begun probing whether current rules unfairly shift the burden of this buildout onto households. In a recent order, FERC opened a proceeding on co‑location arrangements, where data centers running AI workloads connect directly to nearby power plants. By launching this co‑location inquiry, the commission signaled concern that bespoke deals could bypass normal grid-planning processes, undermine reliability, or leave other customers paying for shared facilities that primarily serve a few large users.
FERC’s move folds together earlier complaints from utilities, consumer advocates, and independent generators that had warned about “free‑riding” by data center operators. If a facility receives firm service from the bulk power system but avoids paying for the associated network upgrades, they argue, the remaining ratepayers must shoulder those capital costs. The show‑cause format gives the commission broad authority to demand changes to utility tariffs and interconnection procedures if it finds that current practices are unjust or unreasonable.
At the state level, Maryland Governor Wes Moore joined a coalition of regional governors formed to protect ratepayers from data‑center‑driven cost increases and to press for clearer planning from PJM. In a joint statement, the group cited “rising energy costs” linked to “explosive data center growth” and urged regulators to ensure that large-load customers “pay their fair share” of the grid investments they trigger. The governors also called for more transparency around where new projects are sited and how their impacts on reliability and local infrastructure are assessed.
Maryland’s Office of People’s Counsel has echoed those concerns in filings at FERC, arguing that residential and small-business customers should not be treated as a limitless backstop for speculative AI buildouts. The office has pushed for reforms that would more directly assign the costs of new high‑voltage lines, transformers, and reliability upgrades to the data center developers whose interconnection requests make those projects necessary. Similar arguments have surfaced from consumer advocates in other PJM states, who warn that unchecked cost shifting could erode public support for broader grid modernization and clean‑energy goals.
For now, households have little ability to disentangle data center impacts from other drivers of their bills, such as fuel prices, storm hardening, and general inflation in construction costs. But as AI’s electricity footprint grows, the debate over who pays for the underlying infrastructure is likely to intensify. The decisions FERC and state commissions make over the next few years will determine whether East Coast families simply absorb higher charges as the price of digital growth, or whether the companies building the next wave of AI capacity are required to carry more of the load.
