We’re leading an all-out national mobilization to defeat the climate crisis.

Join our work today to help us build a thriving and just clean energy future.

PJM’s Bandaid Proposal to Address Data Centers

As data centers threaten the grid and increase electric bills, customers shouldn’t be forced to subsidize the costs

TAGS:

100% Clean Power

Thursday, October 02, 2025

Key Takeaways

PJM Interconnection warns of potential energy shortages as early as 2027, due to data center growth and PJM’s mismanagement of the interconnection queue.
PJM's proposed plan, which will determine who loses power first, was weakened to make curtailment voluntary for large loads like data centers.
This change could lead to a less reliable grid and higher energy costs for consumers unless PJM requires large loads to be flexible, bring their own capacity, and pay their fair share.
PJM should protect ratepayers and prioritize affordability before the proposal is finalized and filed with FERC in December, making now the critical time to engage.

Introduction

PJM Interconnection, the nation’s largest grid operator that provides electricity to 67 million people in the mid-Atlantic and Midwest, is sounding the alarm about potential blackouts. This comes after PJM failed to unclog its interconnection queue and approve much-needed projects, which are mostly renewables. Now, in what can only be described as an own-goal, PJM is hurtling toward a resource adequacy crisis.

What’s making matters worse? The region’s data center boom is projected to drive an unprecedented surge in energy demand, meaning we’ll have to make unconscionable decisions about who gets power when there’s not enough to go around.

In September, PJM released its initial “Critical Issue Fast Path” (CIFP) proposal (PDF) to address these resource adequacy concerns. The initial proposal essentially said that during energy emergencies, new energy-hungry large loads, mostly data centers, should bring their own electricity sources to the table or be subject to power shutoffs. (And that makes sense, right? Wealthy tech giants that run these data centers should reduce their power consumption before everyday customers like your grandmother, who is trying to keep cool in a dangerous heat wave.)

But just weeks after sharing its initial proposal, PJM published an updated proposal (PDF) on September 26 that capitulated to industry pressure. PJM walked back its initial proposal, scrapping the mandatory components, along with any semblance of ambition.

Here’s what we think: PJM’s initial and updated proposals have taken a bandaid approach that ignores consumer costs. This crisis isn’t just about resource adequacy—it’s also an energy unaffordability crisis. Households in the PJM region are currently on the hook to subsidize over $100 billion in data center costs to the grid between 2028 and 2032—hitting working families hard when they’re already struggling with sky-high energy bills.

It’s time to ensure that data centers pay their fair share. We need a holistic approach that will protect ratepayers from data centers’ massive costs and risks. At the same time, PJM needs to fix its interconnection queue and accelerate the deployment of new renewables to reliably serve new and existing loads without greatly increasing energy costs to electricity customers.

Data Center 101

Data center development is exploding across the country due to new demands from artificial intelligence (A.I.). This boom is presenting very real and interrelated problems for household ratepayers, grid operators, and our climate:

Problem #1: Reliability and Resource Adequacy Risks. New data centers are expected to drive a surge in electricity demand. This large load growth is projected to strain existing grids, especially during peak summer and winter times. That’s leading to concerns about electricity shortages and prompting grid operators to invest in new energy infrastructure to meet this new demand.

Problem #2: High Energy Bills for Residents. Even though this electricity is needed just for data centers, the cost of new energy infrastructure and higher scarcity pricing is being shared by all customers, including residential ratepayers who are already experiencing rising household energy bills. Making matters more complicated, major uncertainty surrounds the scale of the data center boom and its future electricity demand. Some analysts have argued that data center electricity demand projections are likely overblown, meaning there’s a real risk of inflating capacity needs and overbuilding new energy infrastructure. That means residential energy bills could skyrocket to meet a level of electricity demand that may not even materialize.

Problem #3: New Rush to Build Dirty Fossil Gas Plants. Many developers are proposing to build new dirty and expensive gas plants to serve data center load growth. New gas plants emit climate pollution and health-harming air pollution that tangibly impact everyone, especially fenceline communities living near these facilities. New gas plants are also slow to build, with gas turbines facing widespread supply-chain shortages. That is to say, new gas plants may not even be built in time to meet data center demand. In contrast, renewables and battery storage can be built quickly and cheaply.

Here’s the bottom line: The data center boom poses a steep challenge for the power grid. Massive and rapid load growth from hyperscale data centers is an exogenous shock that PJM’s slow-moving processes are ill-equipped to handle. Big tech companies are some of the wealthiest companies in the history of the world, and they should pay their own grid costs. We shouldn’t expect residential ratepayers to shoulder the financial burden of rapidly building out new energy infrastructure via higher bills—or put up with new fossil gas pollution—to meet this projected electricity demand surge.

What’s Happening Across the PJM Region?

Households in the PJM region are facing eye-wateringly high energy bills. That’s because energy demand from data centers has skyrocketed while PJM has repeatedly failed to address the interconnection queue and approve new energy resources, primarily fast-to-build renewables.

And now, new AI data centers and other large loads are forecasted to drastically increase across the PJM region over the next decade. This projected spike in electricity demand will further inflict sky-high energy prices upon households that are already struggling with expensive bills—especially since PJM’s process to approve the connection of new power plants takes five years, on average.

Data center load growth is already costing PJM customers $10 billion per year, according to a report from NRDC. Without action from PJM, data centers are expected to further increase average household energy bills by hundreds of dollars a year.

What Was PJM’s Initial Proposal?

This August, PJM began a Critical Issue Fast Path (CIFP) process and released an initial proposal (PDF) for managing new large load additions, including data centers. Its proposal was significant because it was the first regional transmission organization (RTO) in the country attempting to address reliability issues created by data centers. Here’s some of what it initially proposed.

PJM proposed creating a new class of demand for new large loads: “Non-Capacity-Backed Load” (NCBL). Simply put? No guarantee of power during energy emergencies. New large loads of 50MW or higher—like data centers—would volunteer to be the first entities to have their power curtailed when we’re reaching peak demand during energy emergencies, before other actions would kick in. If not enough large loads volunteered to be part of this program, PJM proposed that it might require mandatory participation. This makes sense. When there’s not enough power to go around during an emergency, who should lose power first? Working families or the tech giants exacerbating the reliability and affordability crisis?

In return for participating in this program, these large loads wouldn’t have to participate in PJM’s capacity auctions or pay for capacity charges. Removing large load entities from the capacity market is expected to reduce rate hikes for consumers in future capacity auctions. That’s because new data centers wouldn’t be counted in the demand curve that sets the price for the capacity auction.

At the same time, PJM initially proposed exploring:

Providing incentives for large load additions to bring their own generation or participate in load flexibility to avoid power curtailment during capacity emergencies
Offering “primary [fast track] interconnection pathways for generation resources with offtake agreements with large loads. This could apply to already queued resources or brand new projects.
Considering “new demand response products to better match the operating capabilities and business models of the large loads that are now connecting to the grid.”

PJM’s initial proposal was a baby step in the right direction, but it had a massive shortcoming because it didn’t explicitly protect energy affordability. Unsurprisingly, Big Tech and many profit-driven industry groups are deeply opposed to PJM’s proposal. They staunchly asserted that data centers should be allowed to run during energy emergencies, even if it breaks the grid.

PJM’s Updated Proposal Is a Significant Backslide

In response to industry opposition, PJM released an updated proposal (PDF) that buckled to these interests. PJM removed the mandatory NCBL requirement and instead proposed that the NCBL should be voluntary. That means consumers could be fully exposed to data center-driven price hikes in the capacity market. It also proposes creating a new expedited interconnection track for sponsored generation, one more way for gas to cut in front of clean energy in the queue.

Simply put? This is a big backslide—and shows that PJM can’t even do the bare minimum and hold data center companies accountable for derailing reliability and affordability in the PJM region.

How Does PJM’s Updated Proposal Fall Short?

There are notable omissions from PJM’s updated proposal, including:

Failing to Directly Address High Energy Bills: By making grid flexibility and exclusion from the demand curve voluntary, PJM is allowing data centers to pass on their capacity costs to other customers in the coming years. This bill for ratepayers adds up to more than $100 billion, according to some estimates, if not addressed by PJM.
Limited Scope: PJM’s proposal only addresses the capacity market, and not all the other ways data centers are passing on costs to other ratepayers. That includes transmission and other grid upgrades needed to serve these large loads.
Actions to Prevent New Dirty Fossil Gas Infrastructure: PJM still needs to enact guardrails for any interconnection fast tracks to ensure dirty and expensive gas projects aren’t jumping in front of clean energy projects that have already been waiting years in the queue. PJM has repeatedly failed to process a backlog of new, cost-effective clean energy projects waiting to connect to the grid. In this new proposal, PJM is proposing an expedited interconnection track for sponsored generation—showing it could move quickly, it just hasn’t paid the same level of attention to greenlighting much-needed clean energy projects in the existing queue.

What Should PJM Do Next?

This updated proposal is a bandaid on a much bigger issue: PJM’s chronic mismanagement and its inability to clear its clogged interconnection queue to approve cost-effective clean energy solutions. PJM will largely finalize its proposal (PDF) this month and will submit it to FERC this December for approval. Right now, PJM should:

Require new large loads to “Bring Your Own Capacity” at a bare minimum. Households deserve affordable energy—and shouldn’t be forced to shoulder the cost of data center development. Data centers should prioritize deployment of renewables and battery storage (the fastest way to add energy to the grid), as well as pay for their own infrastructure costs and network upgrades. As part of this, PJM should also address the costs of transmission and other infrastructure upgrades.
Require flexibility from data centers in all years and exclude new large loads from the capacity auction. Only then will consumers be more protected from higher capacity auction prices caused by data centers. Voluntary participation is not enough.
PJM should not allow data center generation projects to cut in front of projects already in the queue. Instead, it should focus on speeding interconnection for those energy projects already in the queue and allow data centers and others to take advantage of “surplus interconnection service,” or extra grid capacity at existing and retiring power plants, as well as on-site battery storage.
Prioritize energy affordability, which this proposal currently fails to do. PJM should also implement guardrails to ensure that new fossil gas infrastructure—which is slow, costly, and harmful to public health—isn’t prioritized over affordable, quick-to-deploy clean energy.

Outside of the CIFP, PJM must fix its interconnection queue and fully comply with FERC Order 2023. While addressing data center demand and requiring flexibility is essential in keeping costs down, it does not take away from the need to act on the supply side by fixing the broken interconnection queue and approving energy projects more quickly.

It’s simple: Consumers should come first, not data centers. Consumers should not be subsidizing Big Tech’s speculative data center boom. PJM must strengthen its CIFP proposal to protect consumers and protect the grid.

Author: Mattea Mrkusic. Contributors: Julia Kortrey and Charles Harper.