Another White House Power Grab: PJM

Source: Image generated by Google Gemini (and its data centers).

Fresh on the heels of the White House takeover of Venezuela and its “uninvestable” oil sector, President Trump, Energy Secretary Chris Wright, and the rest of the National Energy Dominance Council have turned their sights on the largest wholesale electricity market in the United States – PJM.  Their concern is high prices, which continue to rise in significant part because of all the data center demand in the region and the deeply problematic structure of the PJM markets.  Along with various other initiatives on “affordability” that the President is suddenly pushing, this one at least recognizes implicitly that electricity is an essential service and that people are pissed about their rising utility bills.  The last election cycle made that clear and rising electricity prices will surely be a major issue in the coming midterms.  Needless to say, President Trump’s election pledge to cut energy bills in half by the end of his first year in office has turned out to be total bullshit.

First established in the 1920s as a tight power pool to coordinate and share generation capacity among the investor-owned utilities serving Pennsylvania and New Jersey (and later extended to Maryland), PJM was refashioned as a Regional Transmission Organization (RTO) in the wake of FERC’s efforts to deregulate and restructure the power sector in the late 1990s and early 2000s.  PJM is now the largest RTO in the country, operating across all or part of thirteen states.  It’s two main responsibilities are to manage the transmission grid for the region and run the wholesale power markets. In discharging these responsibilities, its number one priority is to ensure that there is enough electricity generating capacity to meet demand and ensure reliability.

But this is suddenly getting much harder because PJM also happens to contain the largest concentration of data centers in the world, mainly in northern Virginia’s so-called data center alley, with many more planned across multiple states. As a result, PJM has been ground zero for the controversy over data centers and their links to rising electricity prices. In its most recent January 2026 report on PJM’s auction results for future capacity, PJM’s independent market monitor concluded that “[t]he combination of existing data center demand and forecast data center demand has resulted in significant increases in costs for other PJM customers (emphasis added).”  More specifically, the independent market monitor found that the total increase in costs to other PJM customers from data center demand over the last three auctions was more than $23 billion. Even for a market serving 67 million people, $23 billion in added costs is a lot of money.

In fact, things are so bad that the market monitor even filed a complaint with FERC calling for a moratorium on new data center development pending sufficient new generation capacity. While PJM itself has convened various workshops and stakeholder processes to try to find a solution, various Governors fearing the political impacts of high prices have mounted relentless attacks, chief among them Democratic Presidential candidate Josh Shapiro of Pennsylvania, who has sued PJM because of high prices and even threatened to pull his state out of the market.  But Governor Shapiro clearly wants it both ways.  At the same time that he is hammering PJM on high prices, he has rushed to take credit for billions of dollars of new investment by Amazon and others in new data centers as well as a special arrangement between Amazon and Talen Energy to provide the entire 1,920 MW of electricity from the Susquehanna Nuclear Power Plant to Amazon’s co-located data centers up though 2042.  That’s basically 2 GW of existing carbon-free electric generating capacity that is being taken out of the PJM supply mix and dedicated to Amazon for the next 15 years. Seems like someone might need to explain to Governor Shapiro that this is actually bad for the PJM customers he claims to be protecting from high prices.  But, hey, who said incoherence and hypocrisy were only for Republicans!

As the name suggests, the PJM capacity market is a market for future generating capacity. Generators (both existing and prospective new plants) that clear the capacity market auction get paid to have a certain amount of generation capacity available in the future (typically three years from the date of the auction), which would then be available to bid into the wholesale electricity markets (and receive whatever additional compensation they can from the sale of energy in those markets).  It’s basically an additional revenue stream that seeks to create enhanced incentives for merchant generators to build more power plants to ensure that PJM has sufficient capacity to meet forecasted load growth.  Several RTOs around the country have adopted capacity markets in the hope that this would encourage new investment. But for a whole bunch of reasons, that investment has not materialized fast enough even as prices in the capacity markets continue to rise.

Turns out that while the electricity markets have been good at sweating existing assets, they have not been very effective at driving new investment. Economists and other market enthusiasts have long blamed the price caps adopted in the energy markets in order to protect consumers as the main reason, pointing to the inability of generators to drive prices as high as possible during periods of peak demand and extract the massive rents available as “the missing money problem.”  Seriously?! Translated:  if we simply let the merchant generators make massive windfall profits during periods of peak demand, which is precisely when electricity customers need their electricity the most (think Winter Storm Uri in Texas during February 2021), they will invest more.  True as far as it goes, but irrelevant when it comes to essentials.

The new White House initiative urging PJM to adopt a one-off “emergency auction” will do little to fix these longstanding problems with the PJM market (and electricity markets generally as discussed here and here).  The initiative itself is very short on details—a one-page “Statement of Principles Regarding PJM” signed by Doug Burgum and Chris Wright (the chair and vice chair respectively of the White House National Energy Dominance Council) and the Governors of the thirteen PJM states. The statement calls on PJM to “expeditiously” file with FERC proposed revisions to its tariff (the rules that determine how its markets function) that will provide “15-year price certainty” for new generation capacity by holding what is called a “Reliability Backstop Auction” no later than September 2026.  In effect, PJM is being urged to run a special backstop auction (what some have called an “emergency” auction”) that would provide 15-year revenue guarantees to new generation.  The costs of the new capacity that is procured through these special auctions will then be allocated to the various load serving entities (LSEs) that are mandated to pay for new capacity under PJM rules according to how much data center demand they are serving. Finally, the statement commits the various Governors to ensure that their state public utility commissions adopt new retail rate structures that ensure that the costs incurred in the auctions to provide service to the new data centers will be properly allocated to the data centers rather than residential customers.

While some of the reporting on the initiative has suggested that the data centers and/or their hyper-scaler patrons will be bidding directly on new generating capacity in the backstop auction, the mechanics of how that would happen (if accurate) are not spelled out.  What seems more likely is that the LSEs that sell power to the data centers will be able to bid on new 15-year power supply contracts through the auction. But, again, the details are sparse.

As for PJM itself, the grid operator’s governing board, which was not invited to the White House event last week, released its own guidance on the data center problem shortly after the White House announcement: a 13-page “Board Decisional Letter on Critical Issues Fast Path – Large Load Additions.”   The letter contains a bit more detail on how this might work.  Specifically, it calls for a new voluntary “Bring Your Own New Generation (BYONG)” approach that will be paired with expedited interconnection as the way to deal with rising data center demand. Leaving aside the rather hilarious frat house phonetics of the BYONG acronym, the basic message is clear: data centers need to cut deals with new generation and bring that new generation to the table as it were.  In return they will get long-term supply commitments at fixed prices and expedited interconnection.  Sounds like a great deal for the data centers, even though the White House and the various Governors are congratulating themselves for making the data centers pay for the costs they impose on the PJM system (which, just for the record, has always been a bedrock principle of just and reasonable rates).  So while this is all being hailed as a great victory for the public because data centers will now have to secure and pay for new generation capacity if they want to connect to the PJM grid, which clearly is better than taking out existing resources such as the Susquehanna plant, in reality the data centers will get priority access and price certainty for the generation they so desperately need.

Buried in the letter is also some very revealing language on the deeper structural problems with the PJM capacity market that are now coming to a head because of all the new data center demand. It is worth quoting at length:

PJM has concluded that the investment environment for new generation has evolved materially in recent years, and in ways that differ significantly from conditions that prevailed at the inception of the Reliability Pricing Model (RPM) capacity market. As it is designed, the three-year forward capacity market was intended to produce a transparent price signal reflecting the value of capacity in a given delivery year and, among other purposes, to inform bilateral contracting between willing buyers and sellers.

The resulting single-year commitment three years forward, however, provides a comparatively short-term signal relative to the long-lived nature of the generation resources that underpin resource adequacy. On its own, this structure may not provide the stable revenue streams needed to justify new investment in today’s volatile and uncertain investment environment, particularly when elevated costs are combined with external constraints or intervention.

In addition, a combination of recent factors including heightened price volatility, external headwinds and frequent rule changes has increased investment risk and reduced certainty around long-term procurement decisions. Recent examples of existing resources entering into longer-term bilateral arrangements provide evidence that such contracts are increasingly relied upon to support investment decisions within the independent power producer community.

To date, the current framework of a one-year capacity commitment procured three years forward has not produced a sufficient level of bilateral contracting or price certainty to support the scale of new resource investment required to maintain resource adequacy. (PJM Board Letter at pp. 6-7).

Put simply, times have changed.  The original PJM capacity market design, which never worked as intended but was supposed to provide additional revenues on top of what generators received in the day-ahead and real-time energy markets in order to stimulate more investment (a response to the missing money problem), is no longer sufficient because three years is not enough time given the long-lived nature of power plants, “heightened price volatility,” and the uncertainty created by all of the political interventions and rule changes.  All of which means that the much vaunted “transparent price signals” that the markets are supposed to provide to encourage new investment are not at all transparent and could be “adjusted” at any time. Generators are thus looking for long-term contracts because they can’t trust the markets.  And so the solution is (you guessed it) yet another market construct (this time a new emergency auction for 15-year contracts) and a pledge by PJM to do a “holistic review of the investment incentives in PJM’s markets” in order to return to “market fundamentals” as the White House statement of principles put it.  But as others have remarked in other contexts, trying to fix markets with more markets “is just another way that neoliberals have of never having to say they’re sorry.”

The problem, in other words, is with the markets themselves and it’s a problem that can’t really be fixed by creating new market-like constructs.  Here it is worth recalling that these “markets” are highly contrived and regulated constructs that are supposed to look like markets even if they involve more regulation than the traditional cost-of-service ratemaking they were supposed to replace.  Thus, the capacity markets use an “administratively determined demand curve” together with price collars that constrain the generators’ offers in the auction, which are supposed to somehow keep this all close to what is known as “net CONE” (where CONE is “cost of new energy”).  That is, PJM uses its own administrative load forecast to determine the demand curve for new capacity, which it then allocates to the various Load Serving Entities across PJM who are in turn responsible for paying for a fixed amount of capacity depending on how much load they serve.  PJM then uses a series of bidding rules and a price collar specifying minimum and maximum offer prices that are supposed to ensure that the offers to sell new capacity are close to the estimated all-in revenue requirements for new generating capacity (just as an aside, the price collar mechanism was adopted in part to placate Governor Shapiro).

If this sounds like a bad version of cost-of-service ratemaking, you are correct! Imagine how much easier it would be if we just empowered a regulator to evaluate the costs of new generation (call it net CONE if you must) and then used that to develop a revenue requirement and set electricity rates accordingly. Turns out we have a phrase for that: public utility regulation.

But there’s more.  As with virtually all electricity markets all over the world, the capacity markets also use a single-clearing price design, which means that all of the generating resources that submit offers to sell future capacity that are below the clearing price (which is set by the last increment of generating capacity ranked by price needed to meet the administratively determined demand) will receive the clearing price associated with the marginal offer rather than the actual offer they submitted. The single clearing price design has sometimes led to very high prices in the day-ahead and real-time electricity markets when generation is highly constrained or when natural gas prices spike (given that natural gas generators are almost always on the margin setting the clearing price). In the capacity markets, when the “administratively determined demand curve” is pushed out significantly because of a substantial increase in forecasted demand, the clearing price will be set by higher (perhaps much higher) priced offers, which will then create significant rents for the lower cost “inframarginal” generators.  Turns out we have a word for that too: windfall.

To be clear, there are sound theoretical and practical reasons for adopting a single-price auction design (as I have written about here and here), and there is a fascinating history here about how this particular market device migrated from the lofty reams of auction theory and mechanism design to actually existing electricity markets.  The design itself encourages bidders to make offers at their short-run marginal cost (honest bidding you could call it), which avoids at least in part a situation where bidders are trying to guess and game the clearing price. Likewise, the inframarginal rents captured by those generators whose costs are below the clearing price can provide revenue needed to cover their fixed costs.

But there are lots of problems with trying to impose this on electricity, not least the political problem of trying to explain to the public why low-cost generators should receive the same price that the most expensive generators receive.  This problem became acute in Europe during the Ukraine invasion when natural gas prices reached unprecedented levels and, due to the single price auction design, drove electricity prices to all-time highs.  There are also technical reasons that are specific to electricity, where the grid operates as one big machine that has to be balanced perfectly in real time all the time, which means that when generation is constrained (during periods of peak demand) generators can sometimes exert significant market power by withholding generation and driving the clearing price higher.  Finally, there are longer term problems here in a world where most of the generation has very low or even zero marginal costs (as is the case for nuclear power, wind, solar, an storage). None of these problems are addressed in any way by the new emergency power auction that the White House is pushing.

As for the winners in all of this, two stand out: tech companies, who will get long term contracts for electricity at fixed prices together with priority access under the new initiative, and natural gas.  Indeed, the various proponents of the emergency auction have made no secret of their desire to push for baseload power, which in their world means natural gas, at least in the short to medium term. This is why the stock price of GE’s Vernova, the leading supplier of natural gas turbines, jumped significantly in the wake of the White House announcement. The problem, here, as has been widely remarked upon for most of the last year, is that new orders for gas turbines are facing at least a three-year backlog. If you want to build a new gas plant, in other words, get in line: GE Vernova is now taking orders for 2029 delivery.  Once you get your turbines, of course, you still have to build your plant, which requires all the siting and permitting that everyone complains about not to mention the rest of the materials and labor force you need (both of which are constrained).  And then you have to connect to the grid.  We have all heard plenty about the long wait times in the interconnection queues, with some suggesting that this is all a conspiracy against renewables.  In reality, it reflects the challenges of connecting new power sources to a complex interconnected system that must be perfectly balanced in real time all the time. But rest assured that as the interconnection delays continue to hamper new generation, the Trump White House will almost certainly declare some sort of emergency and move natural gas plants to the front of the queue.

The climate costs of continuing to build natural gas plants are obvious.  So too the economic costs given that solar, even without the IRA tax credits, is considerably cheaper than natural gas and, when paired with increasingly cheap storage, can provide the same firm power that natural gas plants can.  The longer natural gas stays on the system and continues to set the clearing price in these markets, moreover, the longer customers in PJM and other markets will be left paying the gas price for all the cheap renewables that over the longer term can actually reduce their bills.

Stepping back, then, what the White House and the PJM Governors are actually doing with their proposed emergency auction is ensuring that data centers will get their own special auction with fifteen-year deals at fixed prices with priority grid access, which will in turn ensure that gas stays on the system longer than it would have, keeping prices higher than they would otherwise be.  The whole effort could also work to crowd out efforts to build new generation for the rest of PJM. Every megawatt-hour of electricity that is procured through Trump’s emergency PJM auction will go to the data centers, not to the rest of the PJM customer base.  And every construction worker and gas turbine and interconnection spot that goes to a natural gas plant dedicated to a data center is a construction worker, gas turbine, and interconnection spot that is not available to the rest of the PJM customer base.

All of which means that for PJM as a whole, there will be no net gain in new generation as a result of the emergency action. The market is being segmented to ensure that the data centers get their power. If anything the whole effort will likely make it harder to build new generation serving the rest of PJM.

At the end of the day, while this is all being framed as an effort to ensure that electricity hungry data centers pay their own way by enabling BYONG, thereby making good on President Trump’s promise that the American people will not be on the hook for rising power prices as a result of massive data center buildout, in reality this looks like another great deal for the data centers and the AI overlords standing behind it all.

This is also happening of course at the same time that Trump is trying to kill offshore wind (and renewables generally), including projects that would supply significant electricity into (wait for it) the PJM grid!  Luckily, for now, the courts have pushed back.  But to stop 2.6 GW of new capacity from coming online to supply PJM, which is the planned capacity for Dominion Energy’s Coastal Virginian Offshore Wind project that is 70% complete, while pushing for a new emergency auction to deliver $15 billion in new capacity to supply PJM is just sheer stupidity and shows how little the President actually cares about lowering prices.

One thing we have learned about prices is that when it comes to necessities they are deeply felt. As much as the President might try to bullshit his way out of this, Americans are not going to buy it when he tells them that prices are going down when their utility bills (along with food prices, health insurance, and so many other essentials) are going up.  And this from the man who likely won the election because of how out of touch the Biden White House was on prices.

Prices are more than just signals.  They are also relationships and sometimes those relationships can be coercive, especially when it comes to essentials such as electricity.  As the price of electricity rises to the top of the political agenda for the first time in at least a generation, it is time to recognize that the forty-year experiment with neoliberal electricity has failed.  Going forward, we would do well to remember that struggles over the techniques and practices of price making have long animated some of the most important distributional struggles in our society.  Careful attention to the ways of price making and the systems of provisioning that they support, in other words, helps to surface vitally important questions about politics, fairness, and the public interest that have lain dormant for too long.