Our grid, PJM, just finished up its next capacity auction for 2026/27, and to no surprise, the bill is high. We will break down the context of these evolving issues and why there will be little price relief for consumers in the next 5 years.
What happened in the auction?
The bottom line is that the auction cleared higher than last time, which, if you recall, was around 900% higher than the previous auction. These auction prices will result in roughly another 5% increase in consumers’ energy bills. That is approximately a 25% increase in the consumer’s electricity bill in 2 auctions.
What has changed since “last year’s” auction?
PJM recently lowered the estimated winter availability (ELCC rating) for gas-fired power plants from about 79% to 74%. That adjustment knocked roughly 2.8 GW out of the capacity cleared in the auction. This drop reflects PJM’s response to cold-weather plant failure, like widespread outages during Winter Storm Elliott in 2022, where over 60% of failures were linked to equipment issues and about one-third to fuel disruptions. Natural gas struggled to keep the lights on despite its slogan, and as such, PJM changed its ELCC rating.
Energy experts argue that improving gas plant winterization and maintenance can raise the average reliability from 74% to around 90%. This simple reliability measure could recover up to 12.2 GW of capacity almost immediately because of PJM’s ELCC rating and the equation they use to calculate reliable energy generation.That temporary boost would help PJM manage current demand pressure, while longer-term energy projects filter through the interconnection queue, which often takes 5–7 years to complete.
Why This Shift Already Happened
PJM recalculated ELCC ratings after the failures of 2014 and 2022, which made it clear that gas plants aren’t as dependable during extreme weather as assumed. Before the changes, PJM planned for summer peaks and assumed consistent gas availability—even when winter demand spiked.
Faced with these risks, PJM adopted a revised capacity accreditation model in mid-2024, dropping many gas plants from a >90% availability rating down to 79% and then further to 74% by the latest auction cycle.
What can be done to fix the problem in the short term?
Many experts are offering solutions to the capacity auction problem. PJM’s system has several issues, some of which are its fault and others not. The short-term solutions are few, but one of them is technically working right now.
The governor’s lawsuit and collar agreement have prevented the last capacity auction from continuing to rise to astronomical prices. The latest auction reached the collar agreement’s upper limit, effectively shutting down any higher bids. This agreement has saved Pennsylvanians a little less than 3 billion dollars on their electricity bills for the next cycle.
Other solutions involve weatherization and reliability upgrades to gas plants, which would immediately increase the energy represented in the ELCC rating. This work would “increase” the amount of energy PJM looks at/ acknowledges in the grid, as opposed to actually creating more energy, as described at the beginning of the blog. The final short-term solution would be to increase the number of small/ renewable projects bottlenecked in PJM’s interconnection queue. These projects would create real energy for the grid in a much shorter time frame than the larger generating projects PJM has planned. This is not the first time we have discussed the interconnection issues and PJM’s failure to address them fairly and protect consumers.
Who is really at fault for my higher energy bills?
Costs were elevated due to substantial projected increases in load forecasts driven by data center development, reduced reliability of gas-fired power plants stemming from winter-related risks, and PJM’s inability to clear enough resources from its interconnection queue.We have addressed these issues in various blog posts, some linked throughout this article. The rapid expansion of data center development across the PJM region, particularly in Pennsylvania, is poised to exacerbate existing electricity challenges. While data center developers continue to advocate for their interests, they often do so under the broad justification of urgent demand, without offering full transparency.
The substantial energy consumption of these facilities, coupled with the rising electricity rates paid by Pennsylvanians, effectively acts as a public subsidy to attract and support development. PJM recognizes the elevated demand driven by these facilities and adjusts its load forecasting models accordingly, resulting in higher electricity prices across the board. These increasing costs, borne by all residents of the Commonwealth, primarily serve to facilitate the data center boom, yet tangible benefits to Pennsylvania remain limited.
Data Center Numbers Game
Many reports and economic analyses highlight the dangers and overestimations of data center growth. Despite their massive physical footprint and billion-dollar construction costs, data centers create few long-term local jobs. Even the largest centers typically employ only a few dozen workers, many of whom are highly specialized and brought in from elsewhere. Local residents see little gain in employment, and communities see little increase in population or household income. Because these facilities are built for efficiency and automation—not integration into the local economy-–, they contribute almost nothing to small businesses or local supply chains.To take a quote from the Ohio River Valley Institute report linked above: “Beyond creating construction jobs, the potential economic benefits of an AI industry are questionable. Hype brings risks, and history shows that an AI bandwagon can run roughshod over people and the environment. A balanced policy approach to growing AI that considers communities, ratepayers, taxpayers, public health, climate, and natural resources is urgently needed,” said John Quigley, Senior Fellow with the Kleinman Center for Energy Policy at the University of Pennsylvania.
Another report found that at least ten U.S. states are losing over $100 million annually in tax revenue due to generous state-level tax exemptions for data centers—an issue exacerbated by the rapid expansion of cloud computing and artificial intelligence. These exemptions, primarily for building materials, equipment, and server replacements, make data centers some of the costliest subsidy recipients, despite being run by immensely profitable tech giants like Amazon and Google. Many states struggle to accurately project costs due to the automatic nature of these tax breaks, with some, like Texas, seeing projected losses skyrocket from $130 million to $1 billion in under two years. Transparency is also lacking, as 12 of the 32 states offering such incentives do not report even aggregate revenue losses. Some states appear to ignore accounting standards that require disclosing tax abatement losses, prompting calls to end what critics describe as unnecessary and out-of-control subsidies for a wealthy industry.
The End Result
Ultimately, data centers are structurally incapable of producing broad-based economic growth. Their capital-intensive, labor-scarce nature means they do not generate the economic ripple effects communities need. Data centers fall short compared to industries that bring sustainable jobs, long-term investment, and community integration. For local leaders seeking inclusive, resilient development, they are a poor choice—and potentially a costly mistake. All things must be considered when we make such gigantic shifts on the state level. The touting of large sums of investment in the state is meant to appeal to broad groups, hide the issues that can grow and ultimately hurt the state without proper guardrails and thoughtful, well-crafted development. Pennsylvania has seen enough boom and bust industries in its lifetime.