Households and businesses across the mid-Atlantic grid are absorbing electricity costs that have climbed at a pace few analysts predicted, driven by a concentrated wave of hyperscale data-center construction. Wholesale power prices in one region of the PJM Interconnection, the largest wholesale electricity market in the United States, jumped by 833 percent in a single year. The surge traces directly to data-center demand that is growing faster than new generation or transmission lines can be built, creating a supply-demand mismatch with real consequences for tens of millions of ratepayers.
How data-center load growth is reshaping PJM’s capacity market
PJM coordinates the flow of electricity across 13 states and the District of Columbia. Its capacity market is designed to ensure enough generation exists to meet future demand. But the market’s forecasting tools were built for a world where load grew slowly and spread across broad geographic areas. Data centers break both assumptions. They arrive in clusters, often in a single county or utility zone, and they ramp to full consumption within months of energization rather than the multi-year timeline typical of industrial facilities.
Virginia sits at the center of this shift. Commercial electricity sales in the state have surged in recent years because of data centers, according to the U.S. Energy Information Administration, which used its own EIA-930 grid-monitoring dataset and PJM load forecast materials to document the trend. The EIA data shows that Virginia’s commercial sector electricity consumption has pulled sharply away from historical baselines, with data centers responsible for the bulk of the increase.
The problem is structural, not temporary. PJM’s capacity auctions set prices years in advance based on projected peak demand. When actual load arrives faster and in greater volume than the forecast anticipated, the market clears at higher prices because fewer megawatts of surplus generation remain available. Every ratepayer in the affected delivery area pays those higher capacity charges, whether or not they operate a data center.
This dynamic helps explain why wholesale prices in the most data-center-heavy zones have climbed so steeply. Generators that might otherwise retire stay online at higher cost, and new peaking plants command premium capacity payments. The 833 percent price increase in a single year reflects a market that was not designed to absorb this kind of demand shock in this short a timeframe.
Federal data and regulatory filings that document the price spike
Two federal agencies provide the primary evidence trail. The EIA, drawing on its real-time grid data collected through the EIA-930 reporting system, has tracked the volume of electricity flowing into Virginia’s commercial sector. That dataset captures hourly generation and demand across every balancing authority in the country, making it possible to isolate how much new load data centers have added to specific parts of the PJM footprint.
The EIA also references PJM’s own load forecast documents, which project demand growth by zone. Those forecasts have been revised upward repeatedly as hyperscale operators announce new campuses in northern Virginia, central Ohio, and other PJM territories. Each revision feeds into the next capacity auction, pushing clearing prices higher and signaling to developers that more generation will be needed to maintain reserve margins.
On the regulatory side, the Federal Energy Regulatory Commission maintains a public docket archive for PJM that covers capacity-market rules, reliability actions, and complaints tied to load growth. Those proceedings reveal an ongoing tension between PJM’s market design and the speed at which data-center demand is materializing. Filings in these cases show utilities and merchant generators arguing over how to allocate the cost of serving loads that did not exist in prior forecasts, and whether accelerated transmission investment should be recovered from all customers or targeted more directly at the beneficiaries.
Fuel costs add another layer. Natural gas remains the marginal fuel for much of PJM’s generation fleet, and when data centers run at near-continuous load factors, they keep gas-fired plants dispatched around the clock. The EIA’s weekly storage bulletins track inventory levels that reflect this sustained draw on gas supply. Higher gas consumption tightens storage balances, which in turn lifts the commodity price that sets wholesale electricity rates during most hours and amplifies the bill impact of capacity-market tightness.
Together, these datasets paint a clear picture: concentrated, fast-growing data-center demand is hitting a grid that lacks the generation headroom and transmission capacity to absorb it without significant price effects. The cost flows downstream to residential and small-business customers who have no direct relationship with the data centers driving the increase, but who pay for the same shared network of power plants and wires.
Unanswered questions about PJM’s forecast tools and cost allocation
Several critical gaps remain in the public record. No single FERC docket or PJM filing isolates data-center load as the sole driver of the 833 percent price spike. Wholesale prices reflect a mix of fuel costs, plant retirements, transmission constraints, and demand growth. Separating the data-center share from other factors requires nodal-level price analysis that neither the EIA nor PJM has published in a format that allows independent verification.
The EIA-930 dataset, while valuable, reports at the balancing-authority level rather than at individual substations or load zones. That means analysts can see Virginia’s total commercial demand climbing but cannot pinpoint exactly how much of the price increase in a specific pricing node is attributable to a single data-center campus versus other local loads. PJM, for its part, publishes auction results by zone rather than by customer class, making it difficult to trace how much capacity cost is being socialized to households compared with large industrial users.
There are also open questions about whether PJM’s forecasting tools are adequately capturing the next wave of data-center development tied to artificial intelligence workloads. Many of the largest announced campuses are still in early permitting stages, and their ultimate power requirements are uncertain. If those loads materialize at the upper end of current estimates, the region could face another round of unexpectedly tight capacity auctions and associated price spikes before new generation and transmission projects come online.
Cost allocation is likely to remain contentious. Some stakeholders argue that data centers, as the marginal driver of new infrastructure, should shoulder a greater share of the expense through targeted transmission charges or bespoke rate structures. Others counter that the regional grid is a shared resource and that carving out special tariffs for a single class of customer could undermine investment signals and complicate planning. FERC has not yet settled these debates definitively, leaving room for further litigation and rule changes as the impacts of data-center growth become more visible.
For now, the evidence from federal data and regulatory filings points in a consistent direction. Data centers are transforming PJM from a system that once enjoyed comfortable reserve margins and modest load growth into one grappling with rapid, localized demand surges. Unless planning tools, market rules, and cost-allocation frameworks evolve to match that reality, the 833 percent price spike seen in one PJM region may prove less an anomaly than an early warning of how the grid responds when digital infrastructure outpaces the power system built to serve it.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
