Over PJM's last four base capacity auctions, Monitoring Analytics, the grid operator's independent market monitor, attributes $29.4 billion of the $63.6 billion in total capacity charges to data centers. That is 46 percent, close to half of what the region paid to keep capacity on the system, assigned to a single customer class. In the most recent auction the monitor puts the data center share at $6.3 billion of $16.4 billion, or 38 percent. PJM has not adopted that split and the method behind it is not public, but it is the number the policy fight is now organized around.
Joseph Bowring, who runs the monitor, calls it "a paradigm shift" and says PJM spent too long treating data center growth as business as usual. At the end of July, PJM stopped.
On July 31, PJM filed a proposal at FERC for a one-time reliability backstop auction, running September 30 to October 21, to fill the 6.8 GW shortfall left by its last base auction for the delivery year that starts in mid-2028. The overall price cap is $555/MW-day, up from $325/MW-day in that base auction. Those costs land on load-serving entities inside each zone. Whether they reach data centers specifically depends on each state writing the cost-allocation rules that PJM's filing openly says it is counting on. Results are due by December 2.
That is one half of the plan. PJM also said it would file separately to curtail large loads that do not bring their own power supply when the system nears emergency conditions, aimed at data centers coming online after June 1, 2027 and backed by a mandatory registry of every large load at or above 50 MW. PJM estimates large load in its territory could grow by as much as 70 GW by 2038, which is the pressure driving all of it.
Read the two filings together and the signal to anyone building large load in PJM is plain. Bring firm supply of your own, or take a spot near the front of the curtailment line and pay into a backstop procurement at the higher cap.
Bowring's recommendation is that data centers and other large loads contract for their own generation, and that the ones who cannot should be procured separately under 15-year contracts instead of riding in the shared auction. PJM's curtailment proposal points the same way from the opposite side: the loads that bring their own power are the ones it does not plan to cut first. PJM is also facilitating bilateral deals between large loads and generators, and it will lower the backstop target for any capacity locked up that way before the auction runs.
For a developer, dedicated on-site generation takes the project out of the shared capacity auction, out of the state-by-state fight over who absorbs the cost, and off the priority list when PJM runs short. That is the case for going behind the meter, and right now it is being made by the grid operator and its own monitor rather than by suppliers.
Building your own generation answers the capacity and curtailment problem and opens a different one. The turbines are the straightforward part. Firm gas supply and firm transportation to the site are the long pole, and on many of these sites the local system cannot carry the load at all, which puts a new lateral and an interconnect on the critical path, both of which take longer to build than the plant. Then there is where the gas actually prices. Henry Hub is a reference. Your delivery point is a basis market that can turn against you hard in a cold snap, and a self-supplied site carrying a five-nines uptime obligation cannot treat that as a footnote.
Most large load developers underwrite the generation equipment closely and leave gas supply as a line item to solve later. That order is backwards. The supply path decides whether the plant runs when you need it, and it takes the most lead time to lock down.
If you are in planning or pre-construction on a data center or other large load and looking seriously at self-supply, the fuel side is worth structuring early, while there is still room to shape it. That is the part I work on: firm delivery, real basis exposure at the delivery point, and a supply stack that holds when the system is tight.
Digby Ferrara is a natural gas originator and Director of Energy Services at Aggressive Energy in Brooklyn, NY, focused on behind-the-meter on-site generation and grid-connected power. He structures fuel supply around how a project actually consumes it: firm delivery, firm transportation, basis hedges at execution, and multi-supplier stacks built to satisfy project finance lenders. Contact Digby to talk through your project's fuel side.