The underlying problem is one Parselion has tracked since the June show-cause orders. PJM's conventional capacity auctions procure against a forecast of system-wide demand. Large-load growth of the magnitude now queued does not fit that model: the demand is lumpy, concentrated, and attributable to identifiable counterparties rather than diffused across millions of customers.
A dedicated auction attempts to solve this by procuring new capacity specifically against data center load and, implicitly, allocating its cost to the loads that caused it. That is the right instinct. FERC's objection concerned the execution rather than the premise.
The Companion Decisions
FERC's October rulings drew a clear line. The Commission approved PJM's request to create a data center database, giving the operator visibility into what is actually queued rather than relying on speculative interconnection requests. It rejected a separate request to curtail power to data centers during grid stress.
Read together, the two decisions say that FERC will support better information and better market design, but will not authorise involuntary curtailment of a class of customer by administrative fiat. Flexibility, in the Commission's view, is something to be contracted for rather than imposed.
What the Delay Costs
Parselion's assessment: the five-month slip pushes resource adequacy clarity for the 2028 to 2030 window into the second quarter of 2027, which is uncomfortably late for projects now entering final investment decisions. Developers should not assume the March auction clears cleanly; a second FERC rejection is a live possibility, and the contingency worth planning against is that PJM enters 2027 still without a working mechanism for pricing large-load capacity.