Exploring the New Math Puzzle of a PJM Intervention

Commentary by Arushi Sharma Frank

Published January 16, 2026

Earlier today, the White House National Energy Dominance Council and partnering agencies released a "Statement of Principles Regarding PJM." These principles come out of "an agreement with governors across the Mid-Atlantic region to urge PJM to make electricity more affordable for residential customers and strengthen grid reliability by building more than $15 billion of reliable baseload power generation."

The Statement of Principles outlines a radical departure from market norms as follows:

• A new reliability backstop auction that will guarantee 15-year price certainty, a stark contrast to the market's traditional short-term signals.
• To protect ratepayers, the plan extends the existing price cap on existing generation for two more auctions, effectively freezing scarcity revenues for the legacy fleet.
• Costs for the new plants will be allocated 100 percent to new data centers-"whether they show up and use the power or not"-unless they can prove they have self-procured capacity or have agreed to be curtailable.
• The Trump administration explicitly directs PJM to bypass its own stakeholder process ("no further CIFP processes are necessary") and mandates that interconnection studies be completed in "150 days or less."

A proposal to auction 15-year generation contracts will do more than just fund new power plants. It will fundamentally alter who bears the risk of building out the American grid and keeping current power plants (thermal and non-thermal) open, creating a new, paradoxical reality for the technology sector: Participants are potentially "short" on power today, but "long" on liability tomorrow.

This shift moves regional transmission organizations (RTOs) into uncharted territory. RTO markets were designed to facilitate bilateral contracting and unitary pricing, not to host separate government-mandated procurements for a specific asset class. Whether that is the intent of the proposal, and what PJM files, is yet to be seen.

The policy implication is clear though: By carving out a dedicated market for certain loads while leaving the rest of the grid in the legacy auction, this proposal breaks the unitary market structure. On first read, the reality appears to be a hybrid: financially segregated but physically integrated. The data center might pay the mortgage (the 15-year contract), but the electrons flow into the shared grid. This distinction will be vital to understand the merits of the rules to come from PJM; it determines whether this intervention actually lowers prices for consumers-via a "denominator effect," discussed below.

Scenarios for PJM Money Problems Versus Physical Problems

As this issue unfurls in new or modified PJM rules, a few distinct futures could emerge.

In one scenario, the definition of "eligibility" could dictate: By explicitly listing self-procured capacity and curtailment as the only off-ramps, these principles create an immediate, existential financial incentive for large loads to engineer their way out of the auction. Building your own plants could be a path out of the cost allocation; an exemption for "curtailable" load puts pressure on PJM to enact pathways to validate and count on that behavior for economics and reliability.

Critically, a 15-year contract could solve the money problem, but it does not fix the physical problem. While the principles mandate that interconnection studies be completed in "150 days or less," industry insiders note that a memo cannot manufacture a transformer or staff a power study team overnight. This is a path for creating a financial instrument to pay for infrastructure that may still take years to physically connect.

Further, a structural risk looms in either scenario. Whether this commissioned generation is "siphoned" for large loads or allocated into the shared PJM loads pool, the influx of mandated supply-without a guaranteed match in net demand-creates a signal of artificial oversupply. Paradoxically, this could depress prices enough to motivate the retirement of the very baseload power the policy intends to preserve.

Denominator Effects for Consumer Electricity Rates

If the intent of the policy comes into being-the potential to protect residential retail ratepayers from steep increases due to data center demand-the proposed auction could operate like a capital injection: It forces data centers to cover the capital cost of these new plants upfront. When that generation eventually turns on, it bids into the PJM market with its construction costs already covered. It only needs to recover its fuel and operations. The result would be to inject massive amounts of supply that can bid at marginal cost. Assuming all other factors remain equal, this suppresses the wholesale price of energy for everyone.

Rates are assessed based on the costs of building, maintaining, and operating the grid. If some of those costs are already being absorbed by data center developers, then there is more power for everyone without increased costs, and rates can fall. This isn't magic-it's simply the math of the denominator effect (more energy hours to pay for a relatively fixed or new expense-insulated delivery infrastructure cost = lower rates for all impacted classes who can benefit).

This will not, however, fully insulate ratepayers. Increasing rates are not usually driven by the cost of adding generation. It's the wires. If a utility builds a $50 million substation for a data center and spreads that cost to the legacy rate base, residential bills can rise despite low-cost generation. This is where state regulators must hold the line, with tariffs that require large data centers to pay demand charges covering specific infrastructure costs (including substations). Regulators can look to models like the Rappahannock Electric Cooperative's Large Power-Dedicated Facilities tariff, which bills based on the installed capacity of the delivery infrastructure, and incentivize loads to build private firming solutions that bring flexibility to the utility. The incentive has an effect of derisking overbuild of the utility infrastructure The policy announcement today from the White House seems similar in intent: Pay for the power plant capital costs, whether the energy hours are used or not.

Considerations for Large Loads Asset Class Fracturing

Even if the new math of PJM protects the residential consumer, the pain within the technology sector is expected to be uneven, as will the pain to be distributed between legacy generation assets (thermal and renewable) and new construction under contract. This is largely true of existing proposals in PJM's stakeholder processes: at least one such proposal posits market-oriented reforms consisting of a reliability backstop mechanism with multiyear price certainty (from Amazon, Calpine, Constellation, Google, Microsoft, Talen).

Early reactions to the announcements portend that this approach-a contract-specific auction-effectively has the economic impact of fracturing the electricity ratepayer asset class into at least three tiers while at least solving some of the problems of bringing enough capacity into the market to support gigawatts of load. Impacts on these classes, at least from the principles released today, also depend on the work that states will do to effectuate load-serving entity cost allocation.

Residential customers could remain shielded, provided state regulators enforce the strict tariffs described above and other theoretical wins become reality, including distributed energy and grid utilization programs that support lowering the average cost of utility operations. They get the increased reliability of a robust grid without the increased costs. Tech giants like Microsoft, Google, and Amazon are the "long capital" class. They are the most likely candidates who could possess the balance sheets to absorb a 15-year infrastructure liability, and if the rules of such a program or other PJM programs going forward allow it, the engineering depth to execute complex private firm delivery strategies needed to opt out (they can hedge the risk or pass it downstream).

The likely casualties of this new regime may include specialized infrastructure providers, co-location firms, and conventional industrial loads that lack the balance sheet of a hyperscaler. The administration's Statement of Principles explicitly demands that new loads provide "credit/collateral support" to be counted, a hurdle that favors the giants. Often locked into fixed-price contracts with tenants, these "middle-class" providers lack the credit depth to absorb a 15-year liability and the physical footprint to execute the "self-procure or curtail" opt-outs envisioned by the new rules. For them, the choice between a massive balance sheet liability and a regulatory blockade represents an existential margin squeeze.

The other likely casualty to insure against, is price signals for existing PJM generators, and their contracted status with loads that cannot benefit from the new policy. The principles document extends the existing price collar for two more auctions, effectively capping scarcity revenues for the legacy fleet just as reliability value is rising. Perhaps most concerning to existing resources is a future where existing generators face a "dumping" risk. If the 7.5 GW of new mandated capacity is built but not fully utilized by data centers, that excess power flows onto the grid at zero marginal cost. This artificial oversupply creates a price-suppressive signal that could ironically force the retirement of the very baseload units the policy claims to value. As is usually the case in PJM, when a major market intervention suppresses price signals for existing thermal assets, it drags existing renewables down with it. By capping capacity revenues and potentially flooding the energy market with subsidized "must-run" baseload, the policy erodes the investment case for all legacy generation, not just the coal and gas plants it aims to protect.

Ultimately, this intervention privatizes the cost of reliability, a move that stands in sharp contrast to the unitary market design PJM has administered to date. While the principles document calls for a future "Return to Market Fundamentals" by 2027, the precedent from the proposed approach is one in which reliability procurement is no longer for and from a shared pool of buyers and sellers; PJM's work over the next several months will be to develop a segmented product that is effectively procured by multi-state mandate and billed to specific asset classes. Market stakeholders invested in future-proofing retail and wholesale power deliverability solutions should find themselves even more invested than before in innovating solutions for their large and small customers.

Arushi Sharma Frank is a senior associate (non-resident) with the Energy Security and Climate Change Program at the Center for Strategic and International Studies in Washington, D.C.

Programs & Projects