FERC Commissioner Calls PJM Status Quo Untenable as Suppliers Say the Market Is Delivering
A regulator's gridlock warning collides with supplier data showing 55 GW of PJM generation ready to build, reframing whether the largest US power market is failing or adjusting.
PJM Interconnection's "status quo is really untenable," FERC Commissioner David LaCerte said Thursday (2026-07-09) at a WIRES meeting in Philadelphia, describing a stakeholder process that has "continued to just grind into gridlock."6
The timing matters. The remark landed two days after the trade group for competitive generators pushed back on exactly that framing. Todd Snitchler, who runs the Electric Power Supply Association and represents suppliers owning more than 225,000 MW across the US, argued in commentary dated 2026-07-07 that recent criticism mistakes a market under strain for a market that has failed.5,6
That gap between the two views is the trade. PJM serves 67 million people across 13 states and has become the reference point for whether wholesale power markets can absorb a demand shock they were not built to price.4 Spot power at the PJM Western Hub was $72.38 as of 2026-07-11's reading, reflecting tight capacity rather than any single day's swing.5
Snitchler's case rests on supply queuing up, not walking away. More than 55 GW of new generation has cleared the interconnection queue and is ready to build, he said, another 220 GW entered the latest review cycle, and when PJM asked developers whether they would contract directly with large loads, over 130 GW came forward.5
Those are large numbers against a demand backdrop that is genuinely new. Data centers now account for about half of incremental US electricity demand growth, according to the IEA's global energy assessment.1 Global data center electricity demand grew 17% in 2025, the agency reported, with AI-focused consumption up 50%, and Grid Strategies projects the US data center market to expand by at least 65 GW and as much as 90 GW by 2029.1,2
The skeptics have arithmetic of their own. Of the capacity that submitted interconnection requests between 2000 and 2019, only 13% had reached commercial operation by the end of 2024, while 77% had been withdrawn.1 A queue full of paper megawatts is not steel in the ground, and PJM's own history shows how much of it never gets built.1
That is why LaCerte's "gridlock" language carries weight. FERC directed regional operators outside Texas back in late 2021 to upgrade transmission and improve capacity, and several have since asked for more time.2 The US needs roughly 5,000 miles of new high-voltage transmission a year to keep pace, with permitting, supply chains and queues all running slower than the load they serve.1
The demand side offers another release valve PJM has been slow to open. Utilities in the footprint have spent nearly $6 billion installing about 12 million smart meters, yet the data those devices generate is largely not shared in ways that would let virtual power plants and demand response ease the strain.3 Independent analysts say those tools can materially relieve grid pressure, but only if the meter data flows.3
So both things can be true. The price signal is working well enough to pull 55 GW of ready generation and 130 GW of load-contracting interest into view, which is what a functioning capacity market is supposed to do.5 The interconnection and transmission machinery around that signal is moving too slowly to convert interest into delivered megawatts before the demand arrives.1
The near-term risk is that the political read hardens before the physical one does. If FERC concludes the stakeholder process cannot self-correct, pressure shifts toward imposed remedies, which competitive suppliers like EPSA's members would rather avoid because they tend to blunt the price signals they depend on.6,5
Watch two things through the rest of the summer. First, how much of that 55 GW ready-to-build reaches financial close rather than joining the 77% historical withdrawal rate.5,1 Second, whether PJM utilities start releasing smart meter data at scale, since that is the cheapest capacity in the system and the one least dependent on the queue.3