Summary
- PJM plans filings for a reliability backstop procurement process and new operational treatment of large loads.
- The large-load framework would include a registry for 50MW-plus sites and potential emergency curtailment for some new demand.
- The proposals show how data centre load is moving from passive connection queue to active grid-management issue.
PJM Interconnection is preparing to change how large electricity users, including data centres, are integrated into the largest US power market, with proposals covering a large-load registry, emergency curtailment, and additional resource procurement.
The PJM board has directed filings with the Federal Energy Regulatory Commission for two related measures: a reliability backstop procurement process and operational rules for large loads. PJM serves 67 million people across 13 states and the District of Columbia.
The grid operator is responding to fast-growing demand, resource adequacy pressure, and the need to understand where large loads are connecting. Reuters reported that the proposal would create a registry for data centres and other large users to track location and electricity use.
Large loads are becoming controllable objects
The reliability backstop procurement would create a route to bring forward additional capacity where PJM sees a near-term resource adequacy problem. Eligible resources could include generation, demand response, uprates, and other capacity able to support reliability.
The more direct data centre element is the large-load framework. PJM has been working on rules for major new loads, including a registry for sites of 50MW or more. Better data on where those loads will connect, when they will energise, and whether they are paired with supply resources would give the operator a firmer view of demand than a simple queue position.
Some new large loads could also face emergency curtailment if they do not bring sufficient supporting supply or cannot otherwise be integrated without threatening reliability. That approach would require coordination with state authorities because retail load, customer obligations, cost allocation, and emergency powers sit across different parts of the regulatory system.
The change is a marked shift from treating data centres as large but conventional customers. Under tight capacity conditions, the grid begins to separate firm demand, speculative demand, self-supplied demand, and demand that can be reduced in emergencies. Connection applications become only one part of the evidence required.
Europe is moving in the same direction
UK and European markets are dealing with the same physical problem through different institutional structures. Data centre developers are seeking grid connections at a scale that can exceed available network capacity, while system operators and governments try to separate credible projects from speculative queue positions.
The UK’s connection reform process already points towards a readiness-based system where projects must show land rights, planning progress, and strategic value before retaining priority. Ireland, the Netherlands, Germany, and Spain are each dealing with versions of the same issue: how to host digital infrastructure without allowing large loads to overwhelm grid planning.
PJM’s approach gives the debate a more operational edge. A registry improves visibility. Backstop procurement addresses supply. Curtailment introduces a harder question about whether some data centre loads should be interruptible unless they are backed by deliverable generation or other reliability measures.
Developers will need to treat power procurement and grid strategy as part of the earliest site filter. Land without deliverable electricity is weak collateral. A grid application without evidence of readiness may no longer hold its place. A customer contract without a firm power path can become difficult to finance.
Large digital loads are entering the same planning language used for industrial electrification, generation, and transmission investment. The data centre sector may still argue that its services are economically critical, but the electricity system will increasingly ask what flexibility, supply, or infrastructure contribution comes with the load.
