FERC Order 2023 and Cluster Studies: What Replaced the Serial Queue, and What It Costs to Stand in It
On 31 July 2026, the US Court of Appeals for the DC Circuit upheld FERC Order No. 2023 in its entirety. Three years after it was issued, the reform that replaced the first-come-first-served interconnection queue is settled law.
The queue it was designed to clear has not moved as much as anyone hoped. Interconnection queues held roughly 2,600 GW of proposed capacity at the end of 2023 and remained above 2,060 GW at the end of 2025, with more recent readings around 2,200 GW — including 111 GW of hybrid solar-plus-storage added in a single year.
The reform's real effect is not on the aggregate number. It is on what it costs a developer to stay in the queue. Order 2023 converted queue position from a date into a series of financial commitments, the last of which is twenty percent of a network upgrade estimate that does not exist when the decision to proceed is made.
This post covers what changed, what "ready" now means in dollars, and how to model the capital a project has at risk at each stage.
ℹ️ Note: Implementation varies by region and compliance filings are still progressing unevenly. The figures here are the Order's pro forma requirements; check the applicable tariff for what actually applies.
What Did Order 2023 Change?
It replaced a serial, first-come-first-served queue with a clustered, first-ready-first-served process, and attached money and deadlines to both sides. Projects are studied in groups rather than one at a time; customers must demonstrate site control and post escalating deposits; and transmission providers face financial penalties for late studies.
The single-sentence version is that the old queue rewarded being early and the new one rewards being committed. Everything below is the detail of what commitment costs.
From Serial to Cluster
Under the old process each request was studied individually, in order, with every study dependent on the assumptions of the one before it. A withdrawal near the front invalidated the studies behind it, triggering restudies, which triggered further withdrawals. The queue did not clear because it was structurally unable to.
Order 2023 replaced this with a single-phase cluster study on a defined clock:
| Stage | Duration |
|---|---|
| Customer Request Window | 45 calendar days |
| Customer Engagement Window | 60 calendar days |
| Cluster study | 150 days |
Requests arriving in the request window are grouped; the engagement window allows the group to be refined; and the transmission provider then has 150 days to study the cluster collectively. Costs and network upgrades are allocated across the cluster rather than assigned wholly to whichever project happened to trigger a constraint.
That last point is the substantive improvement. Under a serial queue, one project could be assigned the entire cost of an upgrade that half the queue behind it would benefit from, which made position in the queue a lottery with very large prizes and penalties. Cluster allocation spreads it.
What Does "First-Ready" Actually Require?
Site control, and the thresholds are specific.
A generator must demonstrate 90% site control when it submits its interconnection request, and 100% by the time it executes the Facilities Study agreement. Critically, no deposit may substitute for site control — the only exception being where federal, state, Tribal or local regulatory barriers make timely acquisition infeasible.
That is a genuine change in development sequencing. Under the old regime a developer could enter the queue on a thin land position and firm it up while waiting, because waiting took years and the queue position was the scarce asset. Now the land has to be substantially secured before the request is filed, which moves real money forward by years and removes the cheap option value that made speculative queue entries rational.
The elimination of that option value is the mechanism by which the reform is supposed to clear the queue. Whether it has is a question the aggregate figures answer ambiguously.
The Deposit Ladder
This is the part to model, because the commitment escalates and the final step is the largest and the least knowable.
Study deposit, at entry:
| Project size | Study deposit |
|---|---|
| >20 MW and <80 MW | $35,000 + $1,000/MW |
| ≥80 MW and <200 MW | $150,000 |
| ≥200 MW | $250,000 |
Initial commercial readiness deposit, to enter the cluster study: two times the study deposit.
LGIA deposit, at agreement: 20% of estimated Network Upgrade costs — replacing the previous formulation of nine times the study deposit — and credited against actual network upgrade expenditure.
Acceptable forms of security include cash, an irrevocable letter of credit and surety bonds, which matters more than it sounds: a letter of credit consumes facility capacity rather than cash, and a surety bond consumes neither, so the choice of instrument is a working capital decision rather than a formality.
That flexibility deserves more attention than it usually gets, because for a portfolio developer it is the difference between a manageable programme and an unfundable one. A developer running twelve projects through clusters simultaneously is carrying twelve deposit ladders at once, and the aggregate at the LGIA stage can exceed the equity available for construction. Posting cash against all of them is rarely possible; posting letters of credit consumes a facility that is also needed for PPA credit support, as the Series D post on that subject described; surety bonds, where the market will write them, leave both intact at the cost of a premium.
The instrument choice is therefore a portfolio-level capital planning question, not a project-level administrative one, and it belongs in the development company's funding plan rather than in any individual project model.
The shape of that ladder is the point. A 200 MW project enters with $250,000, commits $500,000 to be studied, and then faces a deposit of twenty percent of a network upgrade number produced by a study it does not control and cannot predict. If upgrades come back at $40m the deposit is $8m. If they come back at $120m it is $24m.
The decision to enter the queue is therefore a decision to accept an unquantified obligation at a later date. That is not a criticism of the design — it is the design, and it is how the process filters for genuine commitment — but it is a fact a development model should represent explicitly rather than as a single line called "interconnection costs."
Withdrawal Penalties
The other half of the filter, and the mechanism that makes a cluster study stable.
Customers face increasing penalties based on study costs for withdrawals before a cluster restudy, and on network upgrade cost estimate increases for later withdrawals. The money does not go to the transmission provider: it is applied first to the cluster's study expenses, then toward network upgrades for that cluster, with anything left over refunded to the remaining participants.
That allocation is well designed. A withdrawal imposes costs on the projects that remain — restudies, reallocated upgrades — and routing the penalty to exactly those projects internalises the externality that made the serial queue unworkable.
It also means the penalty scales with how disruptive the withdrawal is, and when. Leaving early, before the cluster has done expensive work, is comparatively cheap. Leaving after network upgrade estimates have risen — which is precisely when a developer most wants to leave — is expensive, because the estimate increase is the basis of the penalty.
The next post takes this mechanism apart properly, because it is where most of the real money is.
Do the Penalties on Transmission Providers Bite?
Less than the symmetry suggests, and it is worth being precise about why.
Order 2023 eliminated the "reasonable efforts" standard — the provision under which a transmission provider could miss a study deadline without consequence provided it had tried — and replaced it with firm deadlines and financial penalties:
| Late study | Penalty |
|---|---|
| Cluster study | $1,000 per business day |
| Cluster restudy / affected system study | $2,000 per business day |
| Facilities study | $2,500 per business day |
Each penalty is capped at 100% of collected study deposits.
Set those against the scale of what is being studied. A cluster containing several gigawatts of projects represents billions of dollars of development capital, and delay in a market like PJM costs developers far more per day than the penalty transfers. A thousand dollars a business day is a compliance incentive, not compensation.
The deadlines themselves are the substantive reform; the penalties make them enforceable rather than aspirational. But a developer modelling delay risk should not treat the penalty as a hedge. It is not sized to be one.
There is a second-order concern that has been raised and is worth noting honestly: a penalty regime creates an incentive to litigate the reasons for delay, and litigation is itself a source of delay. Whether that materialises is not yet clear.
What Happened to the Projects Already in the Queue?
They were transitioned, and the ordering principle was the same one the new process runs on.
The transition was designed to give transmission providers and customers time to adjust, prioritising and processing existing requests "based on how far they have advanced through the interconnection process and their level of commercial readiness."
That is a defensible rule and it produced a real reallocation. A project that had been waiting four years on a thin land position and no offtake did not keep its place ahead of one filed later with site control, a PPA and financing in progress. Seniority stopped being the currency.
For developers the practical consequence was a period of genuine uncertainty, because a queue position that had been treated as an asset — and in some cases sold as one — was revalued according to criteria that did not exist when it was acquired. Projects that were substantively ready gained; projects holding position speculatively lost, which is precisely what the reform intended.
The residual point worth carrying into any acquisition diligence is that a queue position is no longer a transferable asset with a date attached. Its value is a function of the readiness attributes of the project behind it, and those attributes travel with the project rather than with the position. Buying a project for its queue number, without the site control and commercial readiness that now determine how that number is treated, buys considerably less than it used to.
What Order 2023 Did Not Fix
Three things, and each of them can still stop a project that has done everything the Order requires.
It allocates the cost of transmission; it does not build any. Cluster studies produce better cost allocation across a group of projects, which is a real improvement over one project bearing an upgrade the whole queue benefits from. It does nothing about the underlying scarcity. If the network needs a new line, the study says so and assigns the cost; nothing in the process causes the line to exist faster.
Affected system studies run on their own clock. Where a project's output affects a neighbouring system, that system runs its own study, on its own timetable, outside the host region's 150-day clock. The Order attaches a penalty to late affected system studies — $2,000 per business day — which acknowledges the problem without collapsing the two processes into one. For a project near a seam, this remains a genuine and poorly bounded source of delay.
The service election still decides whether you have capacity rights. As the accreditation post set out, a project taking energy-only interconnection service may hold no Capacity Interconnection Rights and therefore no capacity revenue. That choice is often made to reduce network upgrade obligations or to move faster — a rational trade, and one that deletes a revenue line. Order 2023 governs the process; it does not change that the election is a revenue decision disguised as a technical one.
To which should be added the obvious: state and local permitting, land use approvals and environmental review sit entirely outside this framework and are, in many jurisdictions, now the binding constraint.
Has It Worked?
Partially, and the aggregate figures are genuinely ambiguous.
The queue fell from roughly 2,600 GW at the end of 2023 to above 2,060 GW at the end of 2025 — a real reduction, and the first sustained one. But more recent readings have it back around 2,200 GW, with 111 GW of hybrid solar-plus-storage added in the past year.
Two readings of that are both defensible. The reform is clearing speculative entries and the queue is being replaced by better projects. Or the reform is clearing the queue more slowly than new demand is filling it, which is the same problem with better paperwork.
Implementation status supports caution. Compliance filings have progressed unevenly across regions, with at least one large utility requesting an extended compliance deadline, and PJM only expecting to begin processing new requests under its reformed system during 2026 — targeting an additional 63,000 MW of review through the year.
The honest summary for a developer is that the rules are now settled, upheld on appeal, and only beginning to operate at scale. The queue data from 2027 onward will answer the question; the data available now does not.
How Do You Model Queue Capital in Excel?
As a schedule of capital at risk by stage, with the LGIA deposit modelled against a range rather than a point.
The inputs
Assumptions, labelled as such:
Project size 200 MW
Study deposit (≥200 MW) $250,000
Initial commercial readiness deposit 2 × $500,000
Network upgrade estimate — LOW $40,000,000
Network upgrade estimate — CENTRAL $75,000,000
Network upgrade estimate — HIGH $120,000,000
LGIA deposit 20% of estimate
Capital at risk by stage
Stage 1 Request submitted cumulative = $250,000
Stage 2 Cluster study entry cumulative = $750,000
Stage 3 LGIA — low case cumulative = $8,750,000
Stage 3 LGIA — central case cumulative = $15,750,000
Stage 3 LGIA — high case cumulative = $24,750,000
Plus site control, which under the 90% rule is committed before stage 1 and is not a deposit at all — it is land cost, and it is usually the largest number on this list before the LGIA.
The step that matters
Commitment before cluster study results = $750,000
Commitment after, central case = $15,750,000
Multiple = 21×
A developer commits three quarters of a million dollars to learn a number that then requires fifteen million. The decision structure is therefore an option: the first $750,000 buys information, and the LGIA deposit is the exercise price.
Valuing the decision as an option
P(upgrades acceptable, project proceeds) assumption = 45%
Cost if abandoned at stage 2 $750,000 + withdrawal penalty
Cost if proceeding LGIA deposit, credited against actual spend
Expected stage-2 spend per project that reaches LGIA
= $750,000 ÷ 45% = $1,666,667
That last line is the number a development budget should carry. A portfolio developer does not spend $750,000 per interconnection agreement; it spends $750,000 divided by its hit rate, and on a 45% assumption that is $1.67m of queue cost per project that actually reaches an agreement.
ℹ️ Note: The LGIA deposit is credited against actual network upgrade expenditure, so it is not a sunk cost if the project proceeds. It is working capital, with a timing and a funding cost. Model it as a drawdown against the construction facility, not as a development expense, or the development budget will be overstated and the funding requirement understated.
To build the stage-by-stage at-risk schedule, the upgrade range and the hit-rate-adjusted development budget, prompt Dezzmond with your project size and cluster timing.
What Do Developers Actually Check?
- Is 90% site control genuinely in hand before the request window closes? No deposit substitutes for it.
- What form is the security in — cash, letter of credit or surety bond? It is a working capital decision.
- What range is assumed for network upgrades, and therefore for the 20% LGIA deposit?
- What is the withdrawal penalty at each stage, and when does it step up?
- Has the relevant region completed its Order 2023 compliance filing, and is it actually processing under it?
- What is the portfolio hit rate, and is the development budget divided by it?
- Is the LGIA deposit modelled as working capital or wrongly as a development expense?
Frequently Asked Questions
What is first-ready, first-served?
An interconnection process that prioritises projects by their demonstrated readiness — site control and financial commitment — rather than by the date the request was filed. It replaced the serial first-come-first-served queue under FERC Order 2023.
How long does a cluster study take?
The Order provides for a single-phase 150-day cluster study, preceded by a 45-calendar-day Customer Request Window and a 60-calendar-day Customer Engagement Window.
How much site control is required?
90% at the time the interconnection request is submitted and 100% by execution of the Facilities Study agreement. No deposit may be posted in lieu, except where regulatory barriers make timely acquisition infeasible.
What are the deposits?
A study deposit of $35,000 plus $1,000/MW for projects between 20 and 80 MW, $150,000 from 80 to 200 MW, and $250,000 at 200 MW and above; an initial commercial readiness deposit of twice the study deposit; and an LGIA deposit of 20% of estimated network upgrade costs, credited against actual spend.
Is a queue position still worth buying?
Only together with the readiness behind it. Since the transition prioritised existing requests by advancement and commercial readiness rather than by filing date, a position without site control and commercial readiness no longer carries the seniority value it once did.
Does Order 2023 speed up transmission construction?
No. It improves how the cost of network upgrades is allocated across a cluster and puts deadlines on studies. It does not cause new transmission to be built, and affected system studies in neighbouring regions still run on separate timetables.
What happens if the transmission provider is late?
It pays $1,000 per business day for a late cluster study, $2,000 for a restudy or affected system study, and $2,500 for a facilities study, each capped at 100% of collected study deposits. The "reasonable efforts" standard that previously excused delay was eliminated.
Closing: A Queue That Now Costs Money to Stand In
The old interconnection queue failed for a reason that is obvious in retrospect. It was free to join, position was valuable, and nothing forced a project to be real. Developers filed requests on options over land they had not bought, for projects they had not financed, because the cost of being wrong was a study deposit and the cost of being late was everything.
Order 2023 priced the option. Ninety percent site control before you file. Twice the study deposit to be studied. Twenty percent of network upgrades to sign. Withdrawal penalties that rise as your departure becomes more disruptive, paid to the projects you disrupt.
That is a coherent design, and the DC Circuit has now upheld it in full. What it has not yet done is clear the queue — 2,200 GW and rising with hybrid additions is not a solved problem, and the regions are only beginning to process at scale under the new rules.
For a developer the practical consequence is narrower and immediate. Queue capital is no longer a line item; it is a staged commitment with an unknown final step, and it should be modelled the way an option is modelled — with a hit rate, a range on the exercise price, and an honest account of what the first $750,000 is actually buying.
The next post takes the largest and least predictable part of that structure on its own terms: network upgrade cost allocation, cost certainty, and what the deposit at risk really funds.
Sources: FERC — Explainer on the Interconnection Final Rule · Troutman Pepper Locke — Summary of FERC Order No. 2023 on Generator Interconnection Reform · White & Case — DC Circuit Upholds Interconnection Reforms of FERC Order No. 2023 · Van Ness Feldman — FERC Moves Forward with First-Ready, First-Served Generator Interconnection Process