A Tale of Two Eastern Grids

In this week's episode, we explore a tale of two very different eastern power grids: PJM and ISO New England.

For years, PJM's interconnection queue was hopelessly snarled, with the grid operator getting overwhelmed by small renewable projects and speculative developers. Project reviews were taking five to seven years under a "first-come, first-served" model, leading to massive delays and withdrawn projects.

Now, PJM has overhauled its process to a "first-ready, first-served" approach, requiring developers to provide meaningful upfront financial commitments and proof of site control. The results of their latest application window are staggering: PJM recently announced 220 gigawatts (220,000 MW) of proposed capacity across 811 new projects.

Watch to learn more about:

The new PJM fuel mix: Why gas-fired generation (106 GW) and battery storage (66 GW) are leading the queue, alongside a surprising 27 nuclear projects.

Google's AI grid intervention: How PJM is deploying Tapestry’s HyperQ AI software to expedite the review of these massive data sets.

The "Phantom Load" problem: Why speculative queue behavior is inflating capacity numbers for new data centers by an estimated 3 to 10 times.

The supply chain reality check: Why much of this approved supply won't come online soon, as gas turbines are sold out through 2030 and projects like Commonwealth Fusion don't even have a working reactor yet.

The New England contrast: Why ISO New England is facing the exact opposite scenario, having recently downgraded its load growth forecast to just 9% through 2035 due to a lack of data centers and slowing EV and heat pump sales.

Finally, we discuss the unpredictable wildcards in grid forecasting. If the ongoing conflict in the Strait of Hormuz triggers an enormous petroleum price shock, New Englanders paying over $500 for 100 gallons of heating oil will race for heat pumps, instantly altering these long-term projections.

In a world where everything is constantly in flux, the only thing grid planners can truly count on is the accelerating pace of change. Watch now to understand the regionally differentiated reality of our power grid!

Transcript

Why Interconnection Queues Broke

PJM’s New Queue Rules

What’s In The New Applications

Using AI To Speed Studies

Why Approved Projects Still Wait

New England’s Demand Forecast Shift

Energy Shocks And Planning Uncertainty

Wrap Up

SPEAKER_00 0:02

I've got your video for this, the first week of May 2026. Well, in late April, Mid-Atlantic Rid Operator PJAM announced that it finally overhauled its interconnection process with a redesigned approach aimed at improving the certainty, speed, and discipline of project review. For those of you who don't follow this at all, lucky you. The process of interconnecting new supplier projects involves getting in line, the queue, with a proposed project to be studied at multiple levels. These studies include the impact of a new project on existing power flows, whether or not new transmission infrastructure will be required, who pays for that, and the potential impact on neighboring systems. Queues have become hopelessly snarled in recent years, and not just in PGM. But for a while, PGM's review process took as long as five to seven years to get to a signed interconnection agreement. And at one point, the grid operator simply stopped taking new applicants. Delays were so long that almost three-quarters of the projects in the queues were withdrawn, including many that had achieved interconnection approval status, but had become uneconomical due to the lengthy time frames. A large part of the issue arose from the huge number of projects that caught planners by surprise. A decade or more ago, you had a much smaller number of larger projects, but as smaller solar, wind, and battery storage projects entered the picture in large numbers, PGM got overwhelmed by the sheer volume of cats and dogs and rats and frogs. There was also a lot of speculation in the mix. Developers might have a finite amount of capital to deploy in the U.S. while dealing with these clogged queues. So they might put different projects in various queues across the country, and if one got approved, they'd pull the others. They often didn't have to put up much, if any, capital, so they enjoyed a free option that ended up costing everybody. By the way, the dynamic on the demand side with data centers and large load interconnection cues for hookup to the grid is very similar. As a consequence, you have large amounts of phantom load inflating capacity numbers for new data center loads by as much as an estimated 3 to 10x. Some utility and grid operators are now requiring hefty deposits if you want to stay in the game, but more rigor and higher fees are likely necessary to deter rampant queue speculation. Back to the supply side. For those who've been watching PGM and who doesn't love observing interconnection queues, you may remember the old approach involved a first come first served model. And it also didn't require much by way of proof of financial viability. So the quicker you got in line, the better, even if you weren't really ready to develop a project. Now, with its new queue, PGM has a first ready first served approach, according priority to projects that are more advanced and have a better chance of actually being commissioned and flowing power. These projects must also show proof of financial viability prior to entering the queue, and that proof includes meaningful upfront financial commitments and a demonstration that the developer actually controls the project site through lease or ownership. The goal is to dramatically cut the number of speculative projects lined up, increase predictability, and hasten the timelines for interconnection. The deadline for new applications closed on April 27th, and shortly thereafter, PGM announced it had 220,000 megawatts, so 220 gigawatts of nameplate capacity in new applications representing 811 projects. Gas fire generation led the mix with 106 gigawatts of capacity and 157 projects, followed by energy storage with 66 gigawatts and 349 projects. Perhaps surprisingly, 27 nuclear projects were in the mix with an associated 18 gigawatts of capacity, followed by 15 gigawatts of solar and 9 gigawatts of solar storage hybrids, together totaling 187 projects. 65 wind projects also made the list with about 4.7 gigawatts of capacity. And then there were a few other types of supply as well. PGM must now commence validating the applications to determine whether the necessary technical and financial information has been provided. Last year, it announced that it had teamed up with Google's Tapestry to aid in the interconnection review process. And in its recent announcement, PGM said it will deploy Tapestry's Hyper Q AI-enabled software to improve efficiencies by reviewing huge reams of data associated with the interconnection process and expediting the associated studies. To be clear, not all the projects in the queue, even if approved, will come online anytime soon, and that's for multiple reasons. To take one example, Commonwealth Fusion is in that queue for a project in Virginia to come online sometime in the early 2030s, but it doesn't even have a working demo fusion reactor yet. And the gas turbine projects, well, G Vernova's largely sold out through 2030, per its recent earnings report, and Siemens and Mitsubishi enjoy similar positions. So if developers don't have turbines, they'll have to get in line. BGM also notes that state permitting processes slow things down. The grid operator's goal is not to be the critical bottleneck to getting much needed supply online, citing an expected increase in demand of 30 gigawatts between 2024 and 2030, driven largely by data center loads. Thus, they intend to review projects in this first new interconnection cycle in what is designed to be a one-to-two-year process, depending on the impact of an individual project. Meanwhile, looking further to the northeast, we see an entirely different case showing just how regionally differentiated this power puzzle has become. ISO New England is also eyeing changing demand, but in this case, it's forecasting a reduction from previous load projections. Few, if any, large data centers will locate in New England, and the ISO's large load forecast foresees only a couple hundred megawatts at present. Instead, projected load growth is anticipated to come from EV sales and heat pump deployments. The grid operator has cut its forecast, though, a couple times in response to the changing facts on the ground. For example, there are no longer any federal incentives for EVs, so sales have fallen. In 2024, ISO New England forecasted about 17% load growth over the coming decade. Then, last year it ratcheted that back to about 11%, and with its just released capacity, energy, load and transmission, Kelt report, it's dialed the number back yet again to 9% through 2035. In a recent blog, ISO New England commented that the region's net annual energy use has actually trended downward over the past two decades, a result of more efficient heating and cooling systems, appliances, and lighting, as well as the growth of on-site rooftop solar power. That dynamic shifting as a result of state policy goals for carbon emissions reductions that will drive electrification technologies. However, that transition is occurring more slowly than anticipated just a few years ago. This stuff is hard to get right in a world in which everything is constantly in flux. Let's take the Iran conflict as an example. If the Straits of Hormuz stay blocked for another two months, and the concerns of some forecasters about an enormous economic petroleum future shock, even more than we're seeing today, come true, then EVCLs could boom again, and they're already inching up. And anybody on oil heat in New England will be racing for the economic refuge of heat pumps. A close relative of mine just got a bill for$526 for only a hundred gallons of oil. In this world, it seems we and grid planners can count on little else besides the accelerating pace of change. Well, thanks for watching, and we'll see you again soon.