It’s been nearly a year since capacity market prices in the mid-Atlantic spiked, reaching the price cap and putting the region at the center of a national debate over whether energy markets are prepared for the AI era.
Grid operator PJM is juggling a veritable storm of change this summer, driven by demand that the RTO itself acknowledges requires a much faster response than its current frameworks allow for. PJM expects peak summer electricity demand to increase by 3.6% annually over the next decade, fueled in part by the massive growth in data center interconnection requests.
In response, PJM is reworking the generation interconnection queue, undergoing a stakeholder process to get large loads online more quickly, grappling with the question of whether data centers can be relied on for load flexibility, and scrambling to organize an “emergency backstop procurement”, designed to facilitate bilateral deals between new generation and power hungry data centers.
But simmering in the background, fueled by a comment from Trump-appointed FERC chair Laura Swett, is a more foundational question: Is the region just too big to be effective?
PJM’s status quo process for grid modernization is “unacceptable and has been for some time,” Swett said at the grid operator’s annual meeting last month. If the RTO’s “huge and diverse footprint” is preventing it from operating an efficient market amid “shifting fuel mixes and rapid technological change,” she added, “perhaps it has simply grown too big to function.”
The argument for breaking up PJM stems from the reality that, compared to other markets in the country, PJM is enormous, serving nearly a fifth of the American population. With 13 member states plus Washington, D.C., it’s very diverse both geographically and politically, and is home to some of the nation’s large data center hotspots.
And as the question of who should foot the bill for America’s AI dominance efforts has become increasingly politically charged, elected officials in PJM member states have begun to openly challenge PJM’s structure and size. That includes threatening to pull out of the RTO altogether.
PJM “drew the short straw” in many ways when it came to the AI boom, explained Elizabeth K. Whitney, an energy market expert and regulatory consultant. “They had an outlandish concentration of data centers before we really realized what the outcome of that was going to be.”
Without the massive weight of data centers bearing down on the market, its underlying “structural rust” wouldn’t have led to the type of problems PJM is currently experiencing, Whitney added. It isn’t data centers’ fault that PJM was slow to process generator interconnection requests, she clarified, but it did exacerbate things, causing the RTO to have to address many problems at once to meet the demand.
FERC is hosting a technical conference on PJM’s governance structure and stakeholder process in late July. While it’s not yet known what form a solution for PJMs problems would take — much less whether it would actually involve breaking up the RTO — it’s clear that the region’s capacity market will be a primary target of any redesign.
Reevaluating the capacity market
PJM added its capacity market in 2007 because its preexisting setup, a year-ahead model, wasn’t providing enough incentive to build and keep generation where it was needed. This was particularly true for gas peaker plants. The old model essentially bought capacity for the whole region at one price, one year out, which provided a weak signal for developers and high risk for investors.
The year-ahead model also bumped up against state retail price shields, which meant many customers didn’t see real-time or locational prices. That muted demand response and other demand-side flexibility.
So the region landed on the capacity market. The current design is a forward market that buys capacity three years out and sets different prices by location. It’s designed to tell developers when and where to build, reducing investment risk and lining up new generation with grid needs.
That design worked reasonably well in an era of flat, predictable demand and steady natural gas expansion. But accelerating load growth, transmission constraints, an increasingly diverse generation mix, and the retirement of existing generators caused prices to skyrocket, as the market failed to attract enough new power plants. It’s a perfect storm that has PJM to the present moment, in which its capacity market prices are the subject of nation-wide controversy.
Whether the capacity market itself has to change is something even the RTO board itself is examining. In a white paper published in May, the board acknowledged significant structural change is likely needed.
However, according to Julia Hoos, head of the eastern U.S. at Aurora Energy Research, the paper also highlighted the RTO’s reticence to be the director of a wholesale change. “I think the biggest statement that they made in that paper was that PJM is not going to make a judgement call,” Hoos said.
That said, the paper did outline three potential paths forward. Two of them keep PJM’s existing capacity construct at the center of the system, while the third would see PJM slowly back away from reliance on the capacity market altogether.
That would mean more of the revenue share of a generation asset would need to come from the energy and ancillary services markets. Other places, including ERCOT in the U.S. as well as in Australia, operate energy-only models, allowing wholesale electricity prices to spike freely when power is in short supply, showing developers when and where building a new plant or battery will be most profitable.
Moving towards a more ERCOT-style approach in PJM could be a boon for demand-side flexibility, as well as for renewable energy, Hoos added. Stronger, more volatile price signals in the energy market would better reward resources that can ramp quickly or shift consumption: both good things in a moment of load growth.
But actually making that work in PJM would still require ways to lock in stable revenues for projects that don’t have guaranteed returns but still need support from risk-averse developers, she said: “I worry that we’ve moved into a world where nobody is able to finance new projects without a long-term contract.”
Barriers to reform
Now, PJM is a long way from doing away with its capacity market. The RTO hasn’t endorsed one reform pathway over another (and indeed, an eventual plan could blend elements from each).
But retreating from the capacity market as it currently exists would be a significant, and likely very rocky path forward, even if done in the “deliberate, phased” approach PJM describes, said Ric O’Connell, executive director of GridLab.
That’s in part because it would likely result in a few years of high energy prices no matter what. Shifting wholesale market rules is essentially taking off the protective training wheels of capacity revenue, he added: “The energy only market is like, oh, there’s no protection, there’s no guard rail…It feels kind of like swimming naked.”
The reality is that most energy retailers are only exposed to a small percentage of wholesale energy prices, because they hedge most of their load with long-term bilateral contracts, O’Connell said. But when sustained grid stress pushes prices to scarcity levels, mismatches between hedges and actual conditions can leave retailers and their customers exposed to large residual risks, triggering political backlash and emergency interventions.
That dynamic is part of how PJM ended up in its current situation, he added, with state governors moving to cap market outcomes and shield consumers from auction price shocks.
Joe Bowring, the independent market monitor for PJM, has a different take: He told Latitude that the white paper incorrectly assumes the underlying issue driving PJM’s problems is market design — but the real problem is the huge wave of new data center load.
The capacity market, he explained, is a FERC-created administrative construct to make sure enough generation gets built and stays online. Peaker plants just don’t make enough money from selling energy at their marginal cost, and the capacity market exists specifically to solve that structural problem.
Bowring argues that shifting the so-called “missing money” from the capacity market and into the energy market, like the white paper suggests in one of its pathways, wouldn’t change the underlying economics. It just changes where customers see the cost. That means that prices for ratepayers could remain high without fixing the core issue, which is new load being added without enough new generation to serve it.
The only way to keep data center load from driving up bills for everyone else is to require them to bring their own new generation, Bowring said. Hyperscalers should be signing bilateral deals with developers to build dedicated new plants, so the costs of serving them aren’t socialized across all PJM customers.
Bowring has pushed this “bring your own new generation” approach for the region while also being sharply critical of what he calls the “regulatory fiction” of flexible data centers and of leveraging demand response to address a lack of new generation. Neither is a substitute for the firm, always-on capacity data centers need, he argued.
Whose job is it anyway?
The white paper, released in the midst of the conversation around whether PJM is too big to function, could also be seen as a signal that the RTO is abdicating responsibility for its own future. It essentially punts the heaviest questions about policy trade-offs to federal regulators, state lawmakers, and even hyperscalers themselves.
This could ultimately be a positive shift, Whitney said, one that expands participation in what has historically been a closed-door process dominated by transmission owners and merchant generators. States, despite taking on the political baggage of electricity rates and reliability, have largely been cut out of conversations about how market rules are written, she added. In stepping back, PJM is — perhaps inadvertently — creating space for a broader set of interests to shape the grid’s future.
And these days, those interests are going to step in no matter what. “If policymakers don’t like the outcome [of this process,] Whitney said, “no amount of perfect analytical and economic analysis and structure is going to overcome the fact that they are going to drive a stake into the heart of it.”
In other words, states in PJM — and elsewhere, to a lesser extent — are simultaneously grappling with their own authority over resource adequacy and generation choices. For a long time states have accepted market rules as fixed and binding, even when outcomes like gas-heavy capacity mixes, challenging entry for renewables, and high prices conflicted with policy goals.
That’s starting to change, Whitney added. ISO New England is a good example of what happens when states “spontaneously [remember]that the markets are supposed to work for them,” she said, pointing to that market’s move toward a prompt seasonal auction. That move was thanks to states reasserting their policy authority, and demanding that market rules accommodate their clean energy programs, she explained.
Ultimately, though, these choices may be in FERC’s hands, according to Bowring.
PJM shouldn’t be making fundamental policy for the region, he said, because they’re an operational entity that runs the market. “FERC is ultimately responsible, but I also think PJM should take a position on the policy and not just say ‘We can’t figure out what to do,’” Bowring added. “They should have a view and they should advance it, but PJM should not be making policy decisions.”
In late July, FERC will convene its technical conference to discuss PJM’s governance and stakeholder process, with a particular focus on “identifying and evaluating actionable reforms to improve PJM’s ability to address system needs in a timely and efficient manner.”
Bowring, for his part, is far from convinced that the outcome should be dismantling PJM, despite Chair Swett’s comments. “I think bigger markets are actually better; there’s more diversity across the footprint…it’s easier to manage as a diversity of resources, [and it] has a robust transmission system,” he explained. “[PJM]hasn’t been managed well for years and needs to get itself back on track, but it’s not ‘too big.’”


