Powerstack: In PJM and ERCOT, it's no cuts, no buts, no coconuts

It’s peak summer – the World Cup is in full swing, and we’re getting a lot of out-of-office replies. But this week, we’ve got a cool one, with the lessons learned from the PJM fast-track queue, just as ERCOT’s batch zero for large load connections starts up this week.

This is Currence’s weekly newsletter on the moves and motives shaping the load growth era. Not a client yet?

FERC just rejected a waiver for design changes to the turbines of a proposed 1.3GW Ohio gas plant, leaving the $2bn project at a crossroads.

The Carroll County project -- developed by Advanced Power Services, owned by ArcLight Capital Partners -- joined PJM's fast-track Reliability Resource Initiative (RRI), a process intended for shovel-ready projects to meet near-term demand, in May 2025. In its application, Advanced Power detailed plans to use GE Vernova's HA.03 (430MW/unit) turbine and be online by May 2030.

But whoa there. Demand for that turbine model is hot, and it’s backordered, pushing the timeline out by at least two years. So Advanced Power asked FERC to let it switch to the HA.02 (384MW/unit), reduce output by 55MW to 1.245GW, and trim its capacity interconnection rights accordingly. But the RRI specifically bars mid-process changes to size or interconnection rights -- when you're in line, you stay in line as-is or get out. The result is that FERC agreed with PJM that granting the waiver would delay review of other RRI projects.

The stakes go beyond one project. The RRI was PJM's one-time fast-track program for large load requests, launched while closing the doors on its standard queue for three years to clear its backlog. It initially attracted 51 projects totaling about 11.8 GW of nameplate capacity. Of that initial set, 39 are uprates of existing plants -- the remaining 12 are new builds, and 6 of those are gas combined cycle. Chestnut Run is one of them. The program has already shrunk to 41 projects -- 10 gone before Chestnut Run's situation even became public -- and the next study phase of the process has a decision point next Tuesday (14 July) where more may opt out. Given that gas turbine lead times are running 5 to 7 years industrywide, Chestnut Run probably isn't the only RRI gas project running into the same problem. 

PJM has suggested Advanced Power could apply through the Expedited Interconnection Track, not part of the RRI, but that would mean effectively starting over on a project that was already supposed to be shovel-ready. The developer hasn't said publicly what it plans to do.

This seems like foreshadowing.

I saw this the other day and immediately had four thoughts:

Batch Zero is a one-time cohort ERCOT is using with the aim of clearing its large-load queue, which has ballooned to over 400GW. It’s for load, not generation like RRI, but it’s similar in that it has a hard readiness or commitment deadline – Friday! 10 July –  a single commitment window, and capacity that locks once you commit. And interestingly, there are different pathways applicants are able to take to get their preferred set-up. They can specify that they want more capacity but elect for at least part of that to be curtailable. Or they can elect to use on-site generation to add to the base-level capacity they draw from the grid – this is called Withdrawal-Limited Private Use Network, or WLPUN. Really rolls off the tongue.

But the thing to remember is that in both set-ups, there’s a base-level, or floor of power drawn from the grid with no curtailment requirement, the low power consumption, or LPC.

The first check is the deadline. If you miss the deadline, you’ll likely get bumped to Batch One, which can mean years added to your project timeline. And the second check is sticking to your plan. Unlike RRI which will kick you out of the queue, it’s looking like ERCOT will be more prone to not kick you out, but cap you at the LPC. Missing either check could be devastating to a project. Start date is most often the most important part of any project economic model – making money sooner than later is crucial. But that’s only slightly more important than utilization. If a project is capped at its LPC, that would mean it’s simply going to operate less and make less money. So in both cases, I would not be surprised if projects that miss either check drop out rather than wait.

So the turbine and equipment bottleneck will loom large here, just like RRI. The most exposed are the WLPUN crowd. That on-site generation has to be an approved, modeled resource by the deadline, and once you're approved, you can’t quietly back out. And in ERCOT, there’s no FERC waiver as a safety net. If you get hit by the turbine supply bottleneck – or bottlenecks across the supply chain – ERCOT could unilaterally decide to put you in LPC-land and crush your project economics.

So, what I'm watching is the shape of the cohort. If we get any transparency on it, it will be such a tell on the market as to how much of the pipeline is phantom projects vs ones that are actually ready to go. Cobble together as you go isn’t really a strategy anymore, at least in Batch Zero.

A quick read on the numbers shaping the market. The capex, the contracts, the regs, all anchored in the so-what.

$64.49/MWh // Solar PPA prices at their highest levels since 2018, up 4.6% in Q1 2026 and 13% year-over-year. Wind hit $79.40/MWh, up nearly 8% quarter-over-quarter and 24% annually. The OBBBA's July 4 construction deadline just passed, with developers racing an estimated 170GW of projects into construction to lock in the 30% ITC before the window closed. Projects that haven’t started will likely have to make up for the shortfall with higher power purchase agreement prices if they continue. Storage still qualifies for credits, so battery deployment could cover the differences. 

1GW // Kodiak Gas Services’ first tranche of Baker Hughes turbines in new agreement. The multi-year rolling framework — covering NovaLT16 and Frame 5 gas turbines plus BRUSH generators — is structured to flex with data center project timelines, giving Kodiak supply chain certainty as turbine lead times stretch and giving Baker Hughes a large-volume anchor customer for years to come.

4.2GW // SGE bid to build 14 BWRX-300 SMRs across three UK sites, with Google Cloud among the backers alongside Samsung C&T, Laing O'Rourke, Aecon, and Fermi Development. It’s financed via Contract for Difference with National Wealth Fund backing. Hyperscaler nuclear appetite is moving beyond direct US PPAs into financing and development stakes in SMR fleets abroad.

380MW // Arizona Public Service converting two retired coal plants to natural gas by 2029, reusing existing site and interconnection infrastructure rather than permitting from scratch. With gas turbine slots sold out into the 2030s at GE Vernova and Siemens Energy, brownfield conversion is emerging as a fast path to new dispatchable capacity for utilities sitting on retired coal fleets, and a solution to the same turbine backlog that hit Chestnut Run this week.

Explore more Signals on Currence here.

The policies, rulings, and company moves worth watching.

NRC's most sweeping reactor licensing overhaul in a generation. The commission issued a 553-page proposed rule bundling 17 modernization measures across licensing, safety oversight, and siting, everything from the legal definition of "construction" to emergency preparedness, siting criteria, and license renewal terms, after lots of talk about shorter, cheaper licensing timelines. The comment period comes next, and NRC aims to finalize the rule by EOY or 2027. 

Eleven Western governors signed a bipartisan permitting agreement. It’s the first multistate coordination framework for transmission built entirely below the federal threshold, an explicit design choice.  The governors of Nevada, Arizona, Colorado, Idaho, Montana, New Mexico, North Dakota, Oregon, Utah, Washington, and Wyoming endorsed the Western Transmission Expansion Coalition roadmap and stood up a new Permitting Alignment and Coordination Task Force to move it forward.

The latest research, features, and data drops on the Currence platform.

The H1'2026 Investment Trends report is out today: the best first half since 2022, clean firm power IPOs breaking records, and startups looking at adaptation and resilience seeing record investment — all while low-carbon data centers made up 34% of total VC. Fervo and X-Energy's public market debuts are pulling a second wave of companies toward IPOs, and the data and Earth Observation sectors are seeing a surge as physical AI hungers for real-world training data. Download the full PDF and data pack here.

Data centers are becoming the premium buyers funding VPPs, with Voltus, Octopus Energy, and the Renew Home/Sunrun/Tesla partnership all structuring deals around hyperscaler money. Our first deep dive in a new series on flexibility breaks down how each model works across wholesale markets and utilities, what the 16.8GW headline number actually means, and what commercial milestones will separate real platforms from compelling announcements. Read it here.

Interested in diving deeper? Talk to our team and learn how the teams that finance and build the energy system use Currence to stay ahead in the energy and AI buildout, including Southern Company, Tokyo Gas, Jefferies, Galvanize, B Capital, and others.