I owe Governor Josh Shapiro an apology.
When his administration sued to cap PJM capacity prices at $325/MW-day in early 2025, I was one of the many voices calling it bad policy. Price caps are the worst of both worlds: they allow a high price to be collected from consumers but provide a low price as a market signal. If the market cannot do its job, I argued, it should be unwound, not capped.
But after reading PJM’s capacity market reform proposal this weekend, it’s clear Shapiro’s intervention was the stress test that finally forced PJM to admit the fundamental model is broken. PJM calls it the “Credibility Trap” — high prices that are “economically necessary to trigger investment…become the trigger for governmental intervention.” Well, yeah.
PJM isn’t the first grid operator to confront the fact that pure economics doesn’t play nicely with political reality. Just a few years ago, ISO-New England faced a similar reckoning when policymakers grew frustrated that its capacity market was working against state policies designed to support lower-emissions generation, and forced a transition to a prompt/seasonal market for capacity. But PJM’s transformation could be much broader—and as the global epicenter of data center concentration, its approach matters for the entire tech and electric sectors.
Across many pages admiring the problem, PJM acknowledges that the issues facing the capacity market are structural (in fact, the word “structural” appears 55 times in the document) and that “the era of purely merchant generation development is largely over.” This is huge. But there are still some fundamental realities that are glaringly absent from PJM’s view of the path forward.
Taking a step back for a moment, PJM’s proposal is really three separate (but potentially interrelated) proposals for reform:
Path A: Keep the Capacity Market, but Let Us Fix It
In this scenario, utilities would be required to procure long-term contracts for their anticipated needs, and the capacity market would become truly residual. Spot prices could climb, but the vast majority of load would be hedged from feeling the financial impact of that signal. This approach doesn’t really reduce costs, PJM acknowledges, it spreads them over a longer time horizon. And it doesn’t escape the cycle of endless rule changes and lawsuits that characterized the capacity market before data centers exposed its fragility—something PJM doesn’t explicitly acknowledge.Path B: Ration Reliability Based on…Some Guiding Policy Principle, TBD
PJM sees this path as allowing more blackouts to occur, but establishing a more orderly framework for how to shed load when it happens. Kind of bold to put that out there as an actual option. But they point out this is pretty much the status quo, so why not establish some principles for who gets dropped first? This is also the basic idea behind “connect and manage” regimes in places like Texas, where large loads accept occasional curtailment as a condition of faster interconnection, and that have already been under discussion in PJM. The door is left open to a discussion about whether some states might want a cheaper, less reliable product, or if other types of customer classes might pay more for greater reliability.Path C: Unwind the Capacity Market and Let the Energy Market Price Scarcity
A longer term approach is to transition away from having a separate market for capacity, returning to more fulsome energy and ancillary services markets that allow scarcely-used generators to recover their costs on a few high-cost days per year. More than likely, this path must also incorporate the other two reforms: a hedging requirement (everyone bring your own power!), accompanied by a shedding principle, and an understanding that spot prices may spike dramatically on occasion. Otherwise, a simple return to uncapped energy market prices is subject to the same political intervention that led to the current caps and the creation of the capacity market as a cost recovery tool for resource adequacy.
Maybe if bilateral hedging and load-shedding prioritization had been a requirement from the start we wouldn’t be in this situation, but we can’t solve structural problems with another tweak. But as far as PJM has come to acknowledge its own existential crisis, they are still ignoring the most flexible asset on the grid: the consumer.
For decades, grid operators have treated residential and commercial load as a single, immutable, and “inelastic” demand. PJM’s visibility ends where the distribution system begins, so they view the consumer as a static constraint rather than a partner to be utilized. The proposal sees the potential for large loads to curtail, but ignores the fluid potential of the “prosumer.” A capacity market reform proposal that takes structural change seriously should treat full Order 2222 implementation as foundational, for starters. Operationally complex, to be sure—but so is redesigned scarcity and rationing, which PJM is already willing to entertain.
When PJM refers to demand that is “insulated” from price volatility, they are describing a suppression of agency. People choosing to invest in rooftop solar, home batteries, and EVs deserve visibility into the price signals that dictate their upfront costs and tolerance parameters. Whether through aggregators, utilities, or consumer-owned cooperatives, we need a market design that stops treating the distribution edge as a “black box” and demand response as an emergency escape hatch.
We cannot build an era of energy abundance on a structural foundation of scarcity and rationing. PJM’s proposal is a triage plan for a grid under duress, when there are real options for an energy future built on resilience, sustainability, and equity. If we are going to rewrite the rules, we should be aiming for a positive, high-agency vision of the grid. That means embracing a fluid system where every home battery and rooftop array is a visible, valued participant.
Fundamentally, PJM has understood that in order for markets to exist, they need durable rules that are trusted by stakeholders. That includes the elderly couples on fixed incomes, apartment dwellers, and homeowners who are underrepresented in policy proceedings but top of mind for elected officials. These people are demanding affordability, not economic purity.
So let’s not replace administrative constructs with a more organized version of scarcity. Let’s build a marketplace that actually enables the abundance we’re all supposed to be working toward: transparent enough for policymakers to trust and flexible enough for consumers to participate in. The current opacity of the market leaves residents with confounding bills and policymakers with few policy tools to protect them, so it should not be a surprise when they turn to blunt instruments like price caps to halt the machine.
Governor Shapiro, I’m sorry. And thank you.
There seem to be mixed signals about what kind of role the White House intends to play in reviewing AI models before they’re released. (Politico Pro)
Anthropic has contracted to use the compute from SpaceX’s Colossus megacampus as demand for Claude swells. I’m…disappointed, tbh. (Business Insider)
Speaking of disappointments, Microsoft is reportedly considering walking back its clean energy pledge in the face of runaway demand. (TechCrunch)
EPA has proposed letting data centers and power plants start construction before receiving air permits, which could add a new wrinkle to the “permit certainty” debate…but that’s for another year. (Politico Pro)
Make data centers beautiful again! It’s a real thing…well, the memes are real. (Fast Company)
Fermi Energy’s second largest shareholder is resisting departed CEO Toby Neugebauer’s bid to return and reshape the company. But true to the drama surrounding the company, that shareholder is Rick Perry’s son. Who has the movie rights to this? (Politico Pro)
Breakthrough Energy Ventures is betting quantum computing can transform the way AI uses energy. (Latitude Media)
Catch up on how Texas is adjusting to the growth in data center load. (Politico Pro)
Read this status report on grid enhancing technologies and whether they are poised to scale up. (Latitude Media)
Maine Senate hopeful Graham Platner released an energy platform that includes a gas tax holiday—but not a moratorium on data centers. (Politico Pro)
That’s all for this week! I hope you’re finding this newsletter informative and entertaining, and if you are I hope you’ll share it with others and send me some comments. Thanks for reading!
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.