Guaranteed Dates, Longstop Dates and the Right to Walk Away

There is a window in most project schedules where the contractor has stopped being liable and the sponsor has not yet earned the right to do anything about it. The delay LD cap has been exhausted, so further delay costs the contractor nothing. The longstop date has not arrived, so nobody can terminate. Every day in that window is carried entirely by equity, and it exists because two dates in two documents were negotiated separately.

This is the third post in the completion-and-delay series. The first set out the three completion milestones and why they drift. The second covered whether the LD regime is actually funded. This one is about the date ladder itself: what separates a guaranteed date from a longstop date, the four clocks a project runs on simultaneously, and the dead zone that opens when one of them runs out before another begins.

ℹ️ Note: This describes how these mechanisms work in practice. It is not legal or tax advice — drafting and applicable rules vary by deal and by jurisdiction.

What Is the Difference Between a Guaranteed Date and a Longstop Date?

A guaranteed date is the date on which money starts changing hands. A longstop date is the date on which the relationship can end. Missing a guaranteed date triggers damages and the project continues; missing a longstop date hands a counterparty the right to walk away from the contract entirely.

They are different instruments for different problems. The guaranteed date prices delay. The longstop date caps how long the other side has to tolerate it. A project can miss its guaranteed date by a wide margin and never approach its longstop, and the commercial consequences of those two events are not on the same scale — one is a cost line, the other is the end of the offtake.

The gap between them is the tolerance the counterparty has sold you. How wide that gap is, and whether it lines up with the gap in the neighbouring contract, is the whole subject of this post.

What Are the Four Clocks on a Power Project?

Four documents each impose their own schedule, with their own start, their own extension mechanics, and their own remedy. They are negotiated by different teams, often months apart, and none of them references the others.

Clock	Lives in	The date that matters	What missing it does	Who can extend it
Construction	EPC agreement	Guaranteed substantial completion	Delay LDs accrue until the cap	The parties, by change order
Offtake	PPA	Guaranteed COD, then longstop	Delay damages, then the offtaker may terminate	The parties, often with a force majeure extension
Grid	Interconnection agreement	Milestone dates; in-service and commercial operation dates	Suspension consequences; deemed termination on prolonged inactivity	The transmission provider, within tariff limits
Tax	Statute and IRS guidance	Placed-in-service deadline under the continuity safe harbour	Loss of safe harbour for the credit	Nobody

The last row is the one people discover late. Three of these clocks are contractual, which means a counterparty can be persuaded, paid, or renegotiated. The fourth is not a negotiation.

Why Must the EPC Date Sit Before the PPA Date?

Because the contractor's obligation is the only thing funding the sponsor's exposure, and it cannot fund a liability that starts before it does. If substantial completion is guaranteed after COD is guaranteed, the sponsor owes the offtaker on days the contractor owes nothing.

Taft Law's guidance on aligning the two agreements is explicit: "The EPC guaranteed substantial completion date and PPA guaranteed COD should either match or require delivery prior to the guaranteed COD." Their worked illustration is as plain as it gets — "if the PPA has a June 1st guaranteed COD, then the EPC guaranteed substantial completion date must occur either on June 1st, or some earlier date." An EPC milestone landing after the guaranteed COD "may lead to lost revenue and schedule liquidated damages."

Two mechanics follow from that.

Align the amounts, not just the dates. Taft's second point is that "the owner should align the liquidated damages amounts under both agreements so that PPA damages will flow through to the contractor." Matching dates with mismatched rates still leaves a shortfall — the subject of the previous post in this series.

Control the change order. Dates move by agreement, and the agreement that moves them is the change order. Taft recommends the EPC "narrowly define change order conditions," because a broadly drafted change order regime lets the contractor extend past the guaranteed COD without ever breaching. The date you negotiated is only as firm as the mechanism for moving it.

ℹ️ Note: Matching the dates exactly is the minimum, not the target. A buffer between EPC substantial completion and PPA guaranteed COD absorbs commissioning, testing and interconnection energisation — all of which sit after the contractor's obligation ends and before the offtaker's clock stops.

What Happens When the LD Cap Runs Out Before the Longstop?

You enter a period with no contractual protection and no exit. The contractor's liability is capped and exhausted, so additional delay is free to them. The longstop has not been reached, so the PPA cannot be terminated. The days in between are funded entirely by the sponsor.

Market practice puts delay LD caps at roughly 10–15% of contract price, and once the cap is reached the contractor generally has no further financial exposure for delay absent gross negligence or wilful misconduct. Lenders understand this clearly: when LD caps are exhausted, the contractual protection is gone and the lender is relying on sponsor support, additional equity, or a workout.

The dead zone is arithmetic, not drafting. Divide the sub-cap by the daily rate to get covered days, add that to the guaranteed substantial completion date, and compare the result to the PPA longstop. If cap exhaustion lands first, the difference is the uncovered window.

It is worth being precise about what a longstop date actually gives you when you reach it. Terminating a PPA because the project was late is not a remedy that makes anyone whole — it removes the contracted revenue that the debt was sized against. As Bracewell notes, development milestones and delay damages exist to "give the offtaker a measure of certainty about commercial operation dates while providing the developer flexibility, backed by liquidated damages if deadlines are missed and termination rights if a later deadline is not met." The termination right is the offtaker's protection, not yours.

What Does Reaching the Longstop Actually Give You?

A right, not a remedy. Terminating removes an obligation; it does not replace the revenue the obligation carried. For a financed project the exit is rarely even available unilaterally, because the lenders sit between the sponsor and the door.

Work through who holds what. If the offtaker terminates, it draws whatever completion security it holds and goes to the market for replacement power. The sponsor is left owning a plant with no contracted offtake and debt that was sized against that offtake — completion security is calibrated to the buyer's switching cost, not to the project's capital structure, so the draw does not come close to filling the hole.

If the sponsor holds a termination right, exercising it is a different problem. Skadden's analysis of lenders' relationships with project counterparties sets out the structural position: material breach of a project agreement is "universally an event of default under a credit agreement," with those covenants applying to agreements the lenders designated as material. Walking away from a material project agreement is therefore an act the credit agreement forbids, whatever the PPA permits.

The same structure runs in the other direction. Direct agreements create a relationship between lenders and the offtaker, EPC contractor and operator, and they typically include a right for lenders to step in and cure a project company default before the offtaker can terminate — Skadden describes this as a near-universal requirement in project financings. The offtaker's termination right is real, but it is queued behind a lender cure period that exists precisely to stop the revenue contract disappearing.

What the longstop genuinely provides, then, is leverage. It converts an open-ended delay into a dated negotiation with a defined downside for both sides, which is usually resolved by an extension, a price concession, or a schedule reset rather than by anyone actually terminating. Model it as the date your negotiating position changes, not the date you leave.

What Does the Interconnection Clock Do That the Others Don't?

It can terminate itself without anyone deciding to terminate it. The EPC and the PPA end when a party exercises a right; the interconnection agreement can lapse by operation of its own terms while everyone is busy negotiating the other two.

The pro forma Large Generator Interconnection Agreement records milestone dates in its appendices alongside the in-service date and commercial operation date, and it gives the interconnection customer a right to suspend work. The consequence of using that right is the part worth reading. Under the suspension provision carried in the standard form — Article 5.16 in the CAISO version — if the customer suspends work and has neither requested recommencement nor itself recommenced within three years of the suspension starting, the agreement is deemed terminated.

No notice is required from the counterparty. No default needs establishing. A project that pauses for financing, permitting or supply chain reasons, and pauses for long enough, can lose its grid position without anyone sending a letter.

That matters for the date ladder because the natural response to a construction delay — slow down, preserve cash, resolve the dispute — is exactly the behaviour the interconnection clock penalises. The three contractual clocks reward patience. This one does not.

Why Is the Tax Clock the Only One You Cannot Renegotiate?

Because there is no counterparty to negotiate with. Every other date on the ladder exists because someone agreed to it and can agree to move it. The placed-in-service deadline is set by guidance, and a project that misses it does not get an extension by paying for one.

Under the continuity safe harbour, a facility is treated as satisfying the continuity requirement if it is placed in service by the end of the fourth calendar year following the year construction began. Baker Tilly's reading of Notice 2025-42 states the rule directly — the facility must be "placed in service by the end of the fourth calendar year following the calendar year in which construction began," so construction beginning in 2025 requires placement in service by the end of 2029. The same analysis notes the notice applies to "facilities the construction of which did not begin, under prior guidance, before Sept. 2, 2025," and that the 5% safe harbour was eliminated for wind and for larger solar facilities. Update: that element of Notice 2025-42 was vacated with universal effect by the US District Court for the District of Columbia on 6 June 2026, restoring the 5% safe harbour, with an appeal expected — see the beginning of construction post for the current position. The four-year continuity safe harbour discussed here was not disturbed.

Falling outside the safe harbour does not automatically forfeit the credit — continuity can still be demonstrated on facts — but it converts a bright-line protection into an evidentiary argument, at exactly the moment the project is already late. For a tax equity investor pricing off credit certainty, that shift is expensive well before anyone loses anything.

ℹ️ Note: Treat the placed-in-service date as a hard constraint in the schedule and work backwards from it, rather than treating it as an output of the construction programme. It is the only date on the ladder with no negotiating counterparty.

A second tax date now sits alongside it. Under the One Big Beautiful Bill Act, wind and solar facilities must begin construction on or before 4 July 2026 to preserve the four-year continuity safe harbour described above. A project beginning construction after that date must instead be placed in service by 31 December 2027 — which allows roughly eighteen months from start to operation and is well short of a normal construction programme for anything beyond a few megawatts.

That converts the tax clock from one date into two, and the earlier of them runs at the front of the project rather than the end. For a project in development now, the binding constraint is frequently not whether it can be built in four years but whether it can establish beginning of construction before a fixed calendar date — which is a procurement and contracting question, not a construction one.

What Is the Float Actually Worth?

The question the ladder makes answerable and that a daily rate cannot.

Schedule contingency — float — is usually valued at a rate per day: so many days of buffer, so much per day of delay cost. That is wrong here, and the ladder is the reason.

The cost of delay does not accrue smoothly. It steps at each date. Before the guaranteed substantial completion date, delay costs nothing contractually. After it, liquidated damages begin. When the cap exhausts, recovery stops and the sponsor absorbs the full daily exposure. At the longstop, a termination right arises — leverage rather than money. And at the placed-in-service deadline there is no damages regime at all, because there is no counterparty; the consequence is the credit.

So sixty days of float is not worth sixty times a daily figure. It is worth whatever it prevents, and what it prevents depends entirely on which segment of the ladder those sixty days would have been spent in.

Two practical consequences.

Float must be measured against each date separately. A project with ninety days of slack to the longstop and thirty to the placed-in-service deadline is exposed on the second, whatever the first says — and it is the second that has no remedy.

The binding date is rarely the one the programme is managed to. Construction teams manage to substantial completion because that is the contractual milestone with damages attached. The tax date is usually the one that actually determines whether the project works, and it is frequently absent from the schedule the project team reviews each week.

How Do You Model the Date Ladder in Excel?

Lay the four clocks on one timeline, derive the dead zone, and price it. The whole exercise is a dozen rows, and it answers a question no single contract can.

The inputs

All of these are assumptions taken from the executed documents — label them as such:

Guaranteed substantial completion (EPC)     2027-06-01
Guaranteed COD (PPA)                        2027-06-01
PPA longstop date                           2028-06-01
Delay LD daily rate                         $45,000
Delay LD sub-cap                            $9,000,000
Daily carrying cost (uncovered)             $65,000
Construction began                          2025-03-15

The ladder

Covered_Days      = Sub_Cap / Daily_Rate
                  = 9,000,000 / 45,000            = 200 days

Cap_Exhausted_On  = Guaranteed_SC + Covered_Days
                  = 2027-06-01 + 200              = 2027-12-18

Longstop           = 2028-06-01

Dead_Zone_Days     = Longstop − Cap_Exhausted_On  = 166 days

What the dead zone costs

Dead_Zone_Exposure = Dead_Zone_Days × Daily_Carrying_Cost
                   = 166 × 65,000                 = $10,790,000

Larger than the entire LD cap. The protection you negotiated covers 200 days; the contract tolerates 366. The uncovered tail is bigger than the covered head, and nothing in either document says so.

The tax constraint

PIS_Deadline  = 31 December of (Year(Construction_Began) + 4)
              = 31 December 2029

Slack_vs_Longstop = PIS_Deadline − Longstop = 578 days

Comfortable here. Compress the schedule — construction starting a year earlier, or a longstop pushed out in a renegotiation — and this is the row that turns negative first, silently, because nobody re-runs it when a date moves.

The check that matters

Build one cell:

Protection_Gap = Longstop − Cap_Exhausted_On

Positive means an uncovered window. Negative means the cap outlasts the tolerance, which is the position you want and rarely the position you have. Run it again every time any date moves, because change orders move dates one contract at a time.

To run the full version — the ladder against sculpted debt service, force majeure extensions applied asymmetrically across the PPA and EPC, and the placed-in-service constraint as a binding date rather than a note — prompt Dezzmond with your executed dates and it will build the timeline and flag which clock binds first.

What Do Lenders Actually Check?

Lenders read the date ladder to find the first clock that breaks, because that is the one that determines whether the base case survives.

• Does EPC substantial completion sit on or before guaranteed COD? Anything else means the sponsor owes damages the contractor does not.
• Where does the LD cap exhaust relative to the longstop? The dead zone in days, and its cost at full carrying rate.
• Do the force majeure definitions match across the EPC and the PPA? Asymmetric relief means the contractor's clock stops while the sponsor's keeps running.
• How wide is the change order regime? A broad one lets the guaranteed date drift without a breach.
• What is the interconnection status, and has work ever been suspended? The three-year deemed-termination provision runs quietly.
• How much slack sits against the placed-in-service deadline? Measured after the longstop, not after the target COD.
• What does termination actually deliver? Drawing completion security does not replace a contracted revenue stream.

Frequently Asked Questions

What is a longstop date in a PPA?

The outside date by which commercial operation must be achieved. Missing the guaranteed COD triggers delay damages and the contract continues; missing the longstop generally gives the offtaker a right to terminate, and to draw whatever completion security it holds.

Should the EPC longstop be earlier than the PPA longstop?

An earlier EPC longstop gives the project company room to replace a failing contractor before its own offtake is at risk. If both dates fall together, the sponsor loses the contractor and the PPA in the same week, with no window to fix the first problem.

What happens if delay LDs are exhausted but the project is not yet terminable?

Further delay costs the contractor nothing and the offtake cannot yet be ended. Those days are carried by equity at full carrying cost — debt service, fixed operating costs, and any PPA delay damages still accruing.

Can force majeure extend a longstop date?

Frequently yes, often with a stated maximum. The risk is asymmetry: if the EPC and the PPA define force majeure differently or cap the extension differently, relief granted to the contractor may not be matched by relief the sponsor receives from the offtaker.

What is the 4 July 2026 deadline?

Under the OBBBA, wind and solar must begin construction on or before that date to keep the four-year continuity safe harbour. Starting later requires being placed in service by 31 December 2027 — about eighteen months, which is short for anything beyond a few megawatts.

Does missing the placed-in-service deadline forfeit the tax credit?

Not automatically — it means losing the safe harbour, so continuity must instead be demonstrated on the facts. In practice the loss of certainty is itself costly, because tax equity prices off predictability rather than argument.

Closing: Four Clocks, One Balance Sheet

The dates that end projects are rarely the ones that get negotiated hardest. Guaranteed dates attract attention because they are where damages start. Longstop dates attract less, because reaching one feels remote at signature. The interconnection clock attracts almost none, and the placed-in-service deadline is treated as a tax matter rather than a schedule constraint.

But the exposure sits in the relationships between those dates, not in any one of them. Cap exhaustion versus longstop. EPC substantial completion versus guaranteed COD. Longstop versus placed-in-service. Each of those comparisons spans two documents that were drafted by different people to solve different problems, and no single contract review catches any of them.

The first post in this series showed the completion dates drift apart. The second sized whether the delay regime funds the drift. This one maps where the protection stops and the tolerance keeps running. Next in the series: performance guarantees — capacity against energy, and where the LD caps leave exposure when the plant finishes on time and simply underproduces.

Sources: Taft Law — Aligning the EPC Agreement and Power Purchase Agreement · Bracewell — Offtake Agreements for Power Projects · FERC — Standard Large Generator Interconnection Agreement · CAISO — Appendix V, Standard LGIA · Baker Tilly — Safe Harbor Under Notice 2025-42 · SgurrEnergy — Preparing a Bankable EPC Contract · Ryan O'Connell — Project Finance Contracts & Documentation · Skadden — Lenders' Relationships with Project Counterparties