Monitoring Analytics did not hedge. The independent market monitor for the PJM Interconnection published its Q1 2026 state of the market report last Thursday and stated directly that data center load growth is the primary reason for recent and expected capacity market conditions, including total forecast load growth, the tight supply and demand balance, and high prices. Wholesale electricity prices across the 13-state grid rose 76% year on year to $136.53 per megawatt-hour. Capacity costs surged 398%. Data center load projections added $13 billion to customer bills across the grid. The price impacts on customers have been very large and are not reversible, the monitor said. Not probably attributable to data centers. Not partially related to data center growth. The primary reason. That is the language the independent watchdog for America’s largest power grid used to describe what the AI infrastructure buildout has done to the electricity bills of 67 million people across 13 eastern states. 

The industry’s response to this kind of finding has been consistent and inadequate. Data center operators and their trade associations acknowledge that demand is growing, note that data centers pay their electricity bills, point to the economic development and tax revenue they generate, and argue that the grid needs investment to keep up with a modern economy. These arguments are not false. They are also not responsive to the specific problem that Monitoring Analytics identified. The problem is not that data centers use electricity. The problem is that the infrastructure costs required to deliver that electricity to data centers at the scale of current AI buildout are being distributed across all ratepayers rather than being borne by the facilities whose load necessitated those costs. That is a cost allocation problem. The economic development arguments do not address it. 

What the Monitor Actually Recommended 

Monitoring Analytics did not simply diagnose the problem. It prescribed a remedy: data centers should be required to bring their own power generation. The bring-your-own-power requirement is not a novel concept. It mirrors the ratepayer protection pledge that the White House and major technology companies endorsed earlier this year. It reflects the policy direction that Florida’s SB 484 established in May, requiring large data centers to bear their full infrastructure costs rather than distributing them across residential and small business ratepayers. It is consistent with the Ohio regulators’ order requiring utilities to design pricing rules that make data centers pay more for the grid upgrades they require, and with AEP Ohio’s creation of a special rate class requiring large data centers to cover 85% of their connection costs. 

The monitor went further. Beyond the bring-your-own-power requirement, Monitoring Analytics recommended that PJM create a dedicated interconnection queue that would only connect new data centers when adequate generation capacity exists to serve them. That recommendation directly addresses the structural mechanism driving the crisis: data centers are being connected to the grid faster than new generation can be built to serve them, creating a supply-demand imbalance that raises prices for every customer on the system while the new load that caused the imbalance enjoys grid access it has not yet paid for. Other PJM customers, whether residential, commercial, or industrial, should not be treated as a free source of insurance, or collateral, or financing for data centers, the monitor stated. Yet that is what most of the proposals related to a backstop auction actually do. The monitor’s recommended solution is to stop using other customers as financial backstops for data center load additions and to require data centers to demonstrate that adequate supply exists before they access the grid. 

The $13 Billion Number That Changes the Conversation 

The $13 billion figure in the Monitoring Analytics report deserves specific attention because it is not a projection or a model output. It is a measurement of what has already happened in the two most recent PJM capacity auctions as a direct result of data center load projections being incorporated into those auctions. The $13 billion represents the incremental cost to PJM customers, spread across all ratepayers in 13 states, attributable to data center demand projections that drove up clearing prices in those auctions. Residential customers in parts of western Maryland and Ohio are seeing monthly bills rise by $16 to $18 just from capacity market effects. Businesses face commercial rate increases above 29% in some markets. The $13 billion is not a future liability. It has already been committed and will be collected from ratepayers through their utility bills through May 31, 2028. 

The irreversibility language in the Monitoring Analytics report is the detail that should concern the data center industry most. The price impacts will be even larger in the near term, the monitor said, unless the issues associated with data center load are addressed in a timely manner. The word irreversible in the context of grid economics means that the capacity auction results that generated the $13 billion cost increase cannot be undone regardless of what the industry does now. The costs are locked in through 2028. The only question is whether the next capacity auction cycle, which will incorporate data center load projections for 2027 and 2028, adds another $13 billion on top of the existing commitment or whether structural reform changes the cost allocation mechanism before the next auction. The window for influencing the next auction is not indefinite. PJM’s auction calendar and the regulatory timelines for implementing the bring-your-own-power requirement and dedicated interconnection queue that Monitoring Analytics recommended will determine whether the next auction runs under the current framework or a reformed one. 

The Political Calculation That Is Changing 

The data center industry’s ability to resist cost allocation reform has historically rested on two pillars: bipartisan political support for data center investment as economic development, and the diffusion of opposition across many small communities without a coordinating mechanism. Both pillars are weakening simultaneously. Monitoring Analytics found that more than 70% of Americans now oppose data center construction in their communities, with 48% strongly opposed, according to a Gallup poll cited in the report. More Americans say they would rather live next to a nuclear power plant than a data center. The political coalition that supported generous data center tax incentives and utility cooperation in Virginia, Ohio, Georgia, and Texas was built on the premise that data centers were broadly popular economic development assets. That premise no longer reflects public opinion. 

The bipartisan character of the pushback is what makes the current political moment different from previous waves of community opposition. Pennsylvania Governor Josh Shapiro, a Democrat, threatened to withdraw Pennsylvania from PJM if prices kept climbing. The CEO of American Electric Power, whose service territory covers Republican-leaning Ohio, opened the door to leaving PJM entirely. FERC Chair Laura Swett has pushed PJM to implement emergency market reforms. Florida’s Republican Governor DeSantis signed legislation requiring data centers to pay full infrastructure costs. The political coalition against data center cost-shifting now spans ideology, geography, and party in ways that the economic development coalition supporting data center investment cannot match with the same arguments that worked when data centers were less visible and their grid impacts were less measurable. 

What the Industry Must Accept 

The industry has a straightforward choice that the PJM report makes more urgent. It can continue resisting cost allocation reform through regulatory engagement and litigation, winning individual proceedings while the political environment for data center development deteriorates further with every quarterly electricity bill that arrives in the homes of PJM’s 67 million customers. Or it can accept the principle that Monitoring Analytics, Florida’s legislature, Ohio’s regulators, and the White House have all independently reached: that data center load additions should be paid for by data centers, not by the households and businesses that share the grid with them. 

Acceptance of that principle is not capitulation. It is the precondition for rebuilding the community consent that the industry needs to sustain its development pace through the end of the decade. As documented in our earlier analysis of the data center industry losing the public consent battle, the cost distribution problem is the structural source of community opposition. The PJM report puts a number on it: $13 billion in added costs distributed across 67 million customers in the most recent capacity auctions alone. That number will grow with every new gigawatt of data center load connected to the PJM grid under the current cost allocation framework. The industry that accepts the principle first and shapes the implementation will be better positioned than the one that fights the principle until it is imposed on unfavourable terms. The monitor has made the diagnosis. The industry now has to decide whether it will prescribe its own remedy or wait for regulators to prescribe one for it.