PJM Drops Oklo From Its Queue as Capacity Shortfall Hits Three Years Running
With a fast-track gas plant also blocked by regulators, PJM's supply pipeline is shedding projects across technology categories while spot traders focus on mild weather.
PJM has dropped Oklo's advanced nuclear project from its interconnection study cycle, removing what was a relatively small tranche of future dispatchable capacity but marking another departure from a queue that the grid operator has spent two years trying to make workable.6
Traders focused on PJM real-time prices are reading mild temperatures as bearish. The consensus runs 58% negative on near-term power, and with weather providing no particular stress, that lean is defensible. PJM Western Hub spot power closed at $73.72/MWh on 2026-08-31. But the supply architecture underpinning reliability over the next few years is eroding in ways the real-time market does not immediately reflect.5
PJM's capacity auction fell 6.8 GW short of its reliability target for a third consecutive year, according to results reported on 2026-07-15 — the largest U.S. power grid serving 67 million customers coming up empty again despite documented urgency around data center load growth. Three straight misses is not a trend to explain away. The grid operator responded by moving its backstop reliability auction to September rather than waiting until 2027, citing surging demand from data centers and insufficient future supply commitments, as E&E News reported on 2026-05-21.5,1
PJM attempted to accelerate supply through its Reliability Resource Initiative, a fast-track interconnection process that initially drew 51 projects totaling about 11.8 GW, all supposedly shovel-ready. That process is fraying. Advanced Power Services' Chestnut Run plant, a roughly $2 billion gas-fired facility owned by ArcLight Capital Partners, sought a FERC waiver on 2026-06-29 to reduce its output by about 550 MW through a switch to the GE Vernova HA.02 turbine model, a change that would have allowed it to hold close to its original commercial date. PJM opposed the waiver. FERC denied it on 2026-07-02.3,4
A developer absorbing a turbine-model switch driven by supply chain realities, offering to bring more than a gigawatt of gas capacity online near schedule, hit a regulatory wall in the one process designed for exactly this kind of urgency. The RRI was built to be fast. It has proven inflexible where speed most matters.3
Oklo's removal adds a different technology to the same pattern. Advanced nuclear represents the dispatchable zero-carbon capacity that utilities and data center operators have been including in longer-term reliability models. A removal from the interconnection study cycle does not simply pause a project. It typically restarts the process, adding years rather than months to any potential in-service date.6
The broader queue is not visibly healthier. PJM's first reformed cycle attracted 811 projects totaling 220 GW of proposed capacity, a figure that sounds like abundance until you read what Canary Media found in reporting from 2026-08-24: many developers had suspended all permitting, siting, and equipment procurement during the prior queue's prolonged uncertainty, and those practical delays do not disappear because a new framework has launched. Equipment orders and permits run on their own timelines.2,6
The spot market's current bearish posture reflects real conditions. But the September backstop auction is where longer-term pressure either shows up in price signals or doesn't. Project count and clearing price in that auction will be more informative than the temperature forecast. The Oklo departure, the Chestnut Run denial, and a development pipeline stalled by years of queue paralysis collectively describe a grid dropping future supply across technology categories at the same moment its operator is scheduling unplanned auctions to shore up reliability.1,56