Solar Interconnection Backlog Crisis 2025: Grid Constraints Delaying Projects Nationwide

Last Updated: 2026 • Sources: LBNL "Queued Up" reports, FERC Orders 2023/2023-A, ISO/RTO public filings and status reports

The US solar industry's biggest bottleneck in 2025 wasn't modules, inverters, labor, or capital. It was a queue. More than 2,600 gigawatts of proposed generation and storage sat in interconnection queues waiting for grid studies — more than double the capacity of the entire existing US power plant fleet. Projects that used to clear interconnection in under two years were routinely waiting four, five, or six. Every year of queue time is carrying cost, option payments, and PPA price inflation — and for a growing share of projects, it's the difference between getting built and getting withdrawn.

We live downstream of this backlog. When a developer's interconnection agreement finally lands, the equipment order that follows is a sprint — and the projects that survive the queue are the ones whose sponsors planned procurement around the wait instead of pretending it away. This guide maps the crisis as it actually stood in 2025–2026: the numbers, FERC Order 2023's reform architecture, how each ISO/RTO was actually performing, what cost allocation does to project math, and the strategies separating projects that energize from projects that expire.

Lawrence Berkeley National Laboratory's "Queued Up" research series is the definitive scoreboard, and its end-of-2023 dataset captured the system at maximum strain: 11,597 active projects representing 2.6 TW of generation and storage across the seven organized markets plus 44 non-RTO utilities. For every project that had ever made it through the queues to commercial operation since 2000 (~4,155 projects), nearly three had withdrawn along the way (~17,873). That withdrawal count isn't a footnote — it's the system's immune response to a process that had become unplannable.

Digest the last row. The Inflation Reduction Act did exactly what it was designed to do — it made a terawatt of projects financially viable. But the queue infrastructure those projects had to pass through was designed in an era of dozens of requests a year, not thousands. Policy pulled demand forward into a process with no capacity to absorb it. That's the crisis in one sentence, and everything else in this article is detail.

The interconnection process was engineered for a twentieth-century grid: a few dozen large, dispatchable power plants per year, each studied individually against a stable load forecast. Three forces broke that design simultaneously. First, the cost collapse — solar, wind, and storage became the cheapest new capacity in most of the country, and development went from a utility-scale trickle to a flood of thousands of projects. Second, the policy accelerant — state RPS targets, corporate procurement, and then the IRA's 2022 credits made marginal projects financeable, and every one of them filed for interconnection. Third, the process itself — serial first-come-first-served studies meant every restudy cascaded: when project #47 in line withdrew, projects #48 through #200 often needed re-analysis, and queue positions became speculative assets held by developers with no intention of building.

The result was a queue that measured itself in terawatts while the construction industry downstream of it measures itself in tens of gigawatts a year. Everyone who touched the system — developers, utilities, RTOs, equipment suppliers, financiers — knew by 2021 that the queue had stopped being a list of future power plants and had become a parking lot. Order 2023 is, at bottom, the formal admission of that fact: readiness deposits exist to evict the parked cars, cluster studies exist to clear the rest in batches instead of one restudy at a time.

On July 28, 2023, FERC issued Order No. 2023 — the most comprehensive interconnection reform in two decades. Effective November 6, 2023, with compliance filings due April 3, 2024 (and clarifications following in Order 2023-A, March 2024), the rule rebuilt the queue around three pillars:

The cluster process runs on gates and money: a defined cluster request window, a cure period for fixing deficient applications, and — the teeth of the reform — commercial readiness deposits in cash, letters of credit, or surety bonds, escalating as a project advances. Speculative projects that used to camp in the queue for free now pay to stay. That single change is why queue composition started cleaning up: when holding a queue position costs real money, only real projects hold positions.

We've heard the grumbling from smaller developers about deposit requirements, and it's not wrong — readiness capital is a barrier that favors the well-funded. But the alternative was a queue where a serious project sat behind forty paper projects, and everyone in that line paid the carrying costs. Order 2023 chose the devil that at least moves.

Reform on paper is not reform in the field. By 2025–2026, each grid operator's Order 2023 transition had produced genuinely different outcomes — and for developers choosing where to site the next project, the regional divergence is now a site-selection variable as real as irradiance.

Two stories stand out. ISO-NE demonstrated that a region with modest queue volume and aggressive compliance can actually finish the job — proof the Order 2023 architecture works when the inputs are manageable. And PJM, carrying the largest queue in the country, showed that scale plus automation plus enforcement can cut a backlog by more than half in two years. The lesson transfers: the regions still drowning are mostly the ones whose load growth (looking at you, data centers) is refilling the queue as fast as reform drains it.

The least understood part of interconnection is the part that kills the most projects: network upgrade cost allocation. When a cluster study determines that your project requires transmission upgrades — a reconductored line, a new substation, reactive compensation — the cluster assigns those costs to the projects causing them. The number lands late in the process, and it can move a project's economics by more than every module price swing of the last five years combined.

That third row is the queue's mortality mechanism. A developer spends two to four years and seven figures on studies, land, and deposits, then receives an upgrade assignment that adds 40% to capex — and withdraws. The withdrawal frees the capacity for the next cluster, which is exactly what Order 2023 intends: capital and grid capacity recycle toward projects that pencil. Painful for the individual sponsor; healthy for the system. If you're evaluating queue positions today, the single best due-diligence dollar you can spend is on early power-flow screening of candidate points of interconnection before you commit site control money.

Watching projects move through — and die in — these queues for years, a clear pattern separates the energizers from the withdrawals:

• They screen before they queue. Serious developers run independent power-flow and short-circuit analysis on candidate POIs before filing, targeting substations with headroom instead of famous transmission corridors everyone else is also crowding.
• They bring storage or go hybrid. Solar-plus-storage projects that can shape output around constraints clear studies more cleanly than standalone solar in saturated zones — and the IRA's standalone storage credit made the storage leg financeable on its own.
• They treat deposits as portfolio decisions. Readiness deposits at multiple POIs is expensive; readiness deposits at multiple screened POIs is a strategy.
• They plan procurement around the queue, not against it. Long-lead electrical gear — transformers, switchgear, breakers — gets spec'd and reserved early against the expected energization window, with safe-harbor tax strategy aligned to the same dates. When the interconnection agreement lands, the schedule sprint is won or lost on equipment lead times.
• They communicate honestly with offtakers. PPAs with queue-delay provisions survive; PPAs that pretend 5-year waits don't exist end up in renegotiation.

Every year a project sits in the queue, its procurement plan ages. Module prices fall (good), electrical gear lead times stretch (bad), and the tax-credit clock keeps ticking (urgent). The projects that handle this best treat the queue window as procurement runway:

The safe-harbor dimension deserves emphasis: under the commence-construction rules, incurring at least 5% of project cost in the right tax year can lock credit treatment — which turns early equipment purchases from a schedule risk into a tax strategy. Our safe-harbor guide covers the mechanics. The dangerous version is buying hardware for a queue position that later dies; the smart version is safe-harboring with transferable, resale-friendly equipment classes rather than project-specific custom gear.

We've been on both ends of this. We've rushed a transformer order for a developer whose IA landed nine months early (rare, happy day), and we've warehoused switchgear for a project whose cluster study slid twice. The difference in outcomes was never luck — it was whether the contract had delivery-flexibility clauses written for a queue world.

Honest answer: the backlog as a process failure is improving; the backlog as a capacity constraint is not. Order 2023's machinery — cluster studies, deposits, penalties — is demonstrably draining speculative volume, as ISO-NE and PJM prove. But two forces keep refilling the pipe. First, load growth: data centers, electrification, and industrial reshoring are pulling new generation demand forward faster than transmission gets built, and interconnection queues ultimately surface the transmission system's limits, not just its paperwork. Second, the transmission build-out itself: queues clear faster when the network has headroom, and transmission permitting remains the slower crisis behind this one.

The 2025–2026 retrospectives capture the split well — see our 2025 H2 industry retrospective for the market context, the 2025 consolidation analysis for what queue stress did to developer balance sheets, and the 2026 workforce analysis for the labor side of actually building what clears the queue. For tax-timing strategy on projects approaching energization, our safe-harbor and commence-construction guide is the operative reference.

Our read from the equipment desk: plan for a world where interconnection is the pacing item through the rest of the decade. Reform will keep shaving the administrative half of the delay, but the physical half — transmission headroom — only moves as fast as steel gets strung. The winners in this environment are the developers who site where the grid has room, the contractors who can mobilize the moment an IA signs, and the suppliers who keep long-lead gear honest. The queue crisis isn't ending; it's becoming a discipline.