PJM Moves Up Reliability Auction as Data Center Power Costs Loom Over the Grid

PJM accelerates a high-stakes power procurement

PJM Interconnection is moving up a planned backstop reliability auction from March to September as it tries to respond to fast-rising demand linked to data centers and other large new loads. The shift, announced by PJM’s board, is more than a scheduling change. It shows how urgently the grid operator sees the need to line up supply while a deeper dispute remains unresolved: who should pay for the capacity and infrastructure required by hyperscale expansion.

PJM’s service territory spans 13 Mid-Atlantic and Midwest states plus the District of Columbia, placing it at the center of one of the most closely watched electricity growth stories in the United States. Data center development has become a major driver of new load forecasts, especially in regions already grappling with reliability concerns and political sensitivity around rising power prices. PJM’s message to stakeholders was blunt. States should immediately develop frameworks that shield residential and other existing customers from costs tied to the new procurement if those costs are being driven by incoming data center demand.

The cost allocation problem is still unresolved

The most consequential issue is not whether PJM can run the auction. It is whether the costs that emerge from it can be assigned in a way regulators and market participants accept. In the board’s words, if states have not created frameworks by the time the backstop procurement runs, it may be unclear which customers should bear those costs.

That warning captures a growing tension in power markets. Data centers can bring jobs, tax revenue, and long-term infrastructure investment, but they also create demand shocks that can force systems to procure new generation, transmission upgrades, or backup reliability resources faster than planned. When those costs flow into broad rate bases, residential and legacy commercial customers may end up subsidizing growth they did not cause.

The Utility Dive report also notes that analysts questioned how a reliability auction’s costs could realistically be allocated only to hyperscalers. That skepticism is important. It suggests that even if policymakers want a neat “cost causer pays” solution, the design of wholesale markets may make precise isolation difficult in practice. The political appeal of protecting households is obvious. The implementation is much harder.

From bilateral contracting to a faster backstop

PJM first proposed a one-time, two-part procurement process in April. One part would allow bilateral contracting between large loads and suppliers between September and March. The other would serve as a backstop auction to acquire any remaining shortfall. Under the updated timeline, the auction will now happen first, in September, to address near-term reliability risks.

The original plan was aimed at adding roughly 14.9 gigawatts of new resources to serve data centers and other large loads expected to be online by the summer of 2029. Under the revised structure, the auction will cover shortfalls emerging from PJM’s next base capacity auction, which is scheduled to begin June 30 for the 2028-29 delivery year.

That sequencing matters because it prioritizes certainty over flexibility. Bilateral deals can in theory create tailored arrangements between large customers and suppliers. But they take time, depend on willing counterparties, and may not materialize at the scale needed. By accelerating the backstop process, PJM is signaling that it does not want reliability planning to wait on bilateral market formation.

Why this matters beyond PJM

The dispute inside PJM reflects a broader structural shift in electricity planning. For years, many grids operated under assumptions of relatively modest load growth. The AI and cloud computing boom has changed that. Large data centers can add concentrated demand rapidly, sometimes outpacing traditional interconnection and procurement timelines. The result is a collision between digital infrastructure ambitions and electric system economics.

What happens in PJM will be watched closely by utilities, regulators, developers, and large technology firms elsewhere. If the region finds a workable mechanism to procure capacity quickly while assigning costs in a politically durable way, it could become an early template. If it fails, the controversy may deepen around whether power market rules are equipped for the scale and speed of modern computing buildouts.

There is also a competitive angle. Regions that cannot provide clear, timely answers on interconnection, reliability support, and cost treatment may become harder places for large-load developers to plan investments. At the same time, regions that move too quickly without resolving cost allocation risk creating backlash from existing customers and state officials.

Reliability first, governance second

PJM’s revised schedule solves only one problem: time. It does not solve the governance challenge beneath it. The operator can accelerate procurement, but states and stakeholders still have to determine how the resulting obligations should be shared. Until they do, every additional step to support data center growth will carry a second question behind it: who benefits, and who pays.

That makes the September auction an important milestone, but not the final answer. The immediate goal is reliability. The longer-term test is whether the region can adapt market rules to a new era in which electricity demand growth is being shaped not by diffuse consumption patterns, but by a relatively small number of extremely large, strategically important loads.

This article is based on reporting by Utility Dive. Read the original article.