The 50-megawatt threshold in PJM's proposed Large Load Registry is the line that determines which data centers enter the grid's new curtailment hierarchy — and which remain outside it. The threshold was confirmed in PJM's July 27 board-directed FERC filing, and its operational consequences are immediate for any facility planning to site in the 13-state footprint. [1]

Facilities at or above 50 MW of peak demand at a single delivery point enter the regime. They face curtailment during supply shortages, cost allocation for reliability backstop procurement, and — starting with the 2029/2030 delivery year — removal from future capacity auction demand calculations if they have not secured qualifying supply. Facilities below 50 MW remain outside the framework entirely. [1]

The threshold's placement is not arbitrary. PJM's own analysis attributed $21.3 billion of the last three auctions' $47.2 billion total to forecast data center load. The 50 MW line sorts that exposure into two tiers: large enough to affect grid reliability, and small enough not to. Monitoring Analytics, PJM's independent market monitor, has tracked the growth of data center load as the single largest driver of capacity price increases — from $28.92/MW-day in 2024/25 to a FERC-capped $333.44/MW-day in 2027/28. [1]

The DOE's emergency orders in June and July — authorizing PJM to curtail facilities above 50 MW and shift them to backup generation — established the threshold as a federal operational line, not just a regulatory proposal. The July heat wave saw PJM lean on approximately 6,000 MW of demand response, including data center backup generation, to hold peak demand below the grid's all-time summer record. [2]

For the paper, the 50 MW threshold is the instrument that converts data center growth from an abstract grid-reliability concern into a concrete cost-allocation mechanism. The August 10 stakeholder workshop and the September-October auction window are the next data points. The question is not whether the threshold will hold — it will — but whether state regulators will use their authority to allocate the costs it creates, or whether the default peak-load methodology will distribute them across all ratepayers.