Why is the construction facility the wrong long-term loan?
Because it was priced for a risk that no longer exists. A construction facility carries completion risk, contractor risk, and a project with no operating history, and its margin, its fees, its reserves and its covenant package all reflect that. On the day the asset achieves commercial operation, the largest single risk in the credit is retired, and the loan sitting on top of it is still charging for it.
The tenor is the second mismatch. Construction facilities are short by design, typically running through building plus a defined operating period, while the asset behind them has a concession or a useful life measured in decades. That gap is deliberate: lenders willing to take construction risk are not always the lenders who want to hold twenty-year paper, and the structure assumes a handover.
So the refinancing is not opportunism. It is the second half of a plan made at financial close, and a sponsor who treats it as an afterthought discovers the assumption was load-bearing. The exit from the construction facility was priced into the original deal whether or not anyone said so out loud.
What is a mini-perm, and what is the difference between hard and soft?
A mini-perm is a facility deliberately shorter than the asset, written on the expectation that it will be refinanced once operations stabilize. The distinction between the two forms is what happens if the refinancing does not arrive, and it is the single most consequential term in the document.
Under a hard mini-perm the loan matures and must be repaid: failure to refinance before maturity is an event of default. Under a soft mini-perm the loan does not mature on that date; instead the economics turn against the sponsor, with the margin stepping up and a cash sweep applying all surplus cash to principal, which extinguishes distributions until the debt is refinanced or repaid. Norton Rose Fulbright set the distinction out in these terms in November 2012, describing hard structures as requiring sponsors to refinance prior to maturity and soft structures as ones in which a failure to refinance is not an event of default; the definitions are stable market vocabulary, though that particular commentary is now old enough that it should not be read as a statement of current pricing.
The practical reading for a sponsor is straightforward. A hard mini-perm transfers refinancing risk to the sponsor absolutely and prices accordingly. A soft mini-perm keeps the lenders in the asset and punishes the equity instead, which is cheaper at close and more expensive if the market shuts. Neither is wrong. Choosing one without noticing which one is.
When does the refinancing window actually open?
When the operating record is long enough to be evidence. Lenders refinancing an operating asset want performance data against the forecast that sized the original loan: availability, output, cost, and the behaviour of the revenue contract through at least one full cycle. Until that data exists, the new lender is underwriting the same forecast as the old one and will not pay much for the privilege.
That points at a window rather than a date. Too early and the asset has no track record to sell, so the refinancing prices against projections and the saving is thin. Too late and the sponsor is refinancing under a deadline, which is the one condition under which a lender does not have to compete. The work in between, assembling the operating evidence, resolving construction snags and completion claims, and confirming that the revenue counterparty is performing, is what converts a maturity into a negotiation.
The market's own shape says the same thing. Green Street's IJGlobal league tables put project finance refinancing at $191.45bn in the 2024 financial year, 23.96% of $799.16bn of total project finance value, and up 54.3% on the $124.07bn recorded for 2023 (Green Street, IJGlobal league tables FY24, published 27 January 2025). On the broader infrastructure finance measure the same tables put refinancing at $473.98bn for 2024, up 51.45% on $312.96bn in 2023. Refinancing is not a corner of this market. In 2024 it was close to a quarter of it.
| FY2023 | FY2024 | Q1 2026 | |
|---|---|---|---|
| Total project finance value | Not stated in the FY24 release | $799.16bn | Absolute figure gated |
| Project finance refinancing | $124.07bn | $191.45bn | Down about 31% year on year |
| Refinancing as a share of project finance | Not stated | 23.96% | Not published |
| Infrastructure finance refinancing | $312.96bn | $473.98bn | Not published |
| Overall activity | Not stated | Refinancing up 54.3% on 2023 | Value down about 5%, count down about 20% |
“The saving on a post-completion refinancing is earned in the eighteen months before it, in operating data nobody thought to collect properly. Lenders pay for evidence, and evidence has to have been recorded.”
What is an operating track record actually worth?
Enough to change the lender universe, which matters more than the margin. An asset with completion risk is bid by construction lenders and a small group of banks. The same asset, complete and performing, is bid by institutional infrastructure debt funds, insurers and pension capital looking for long-dated matched liabilities, and by development finance participants where the asset qualifies. Competition, not persuasion, is what moves terms.
Three things typically improve together. Tenor extends, because lenders willing to hold operating risk will hold it for far longer than lenders holding construction risk. The covenant package loosens, because the coverage ratio can be tested against actuals rather than a stressed forecast. And the reserve requirements shrink, because the reserves were sized for uncertainty that has been resolved. Each of those releases cash to the sponsor independently of the headline rate.
What a sponsor should not assume is that the market will be there. Green Street's Q1 2026 league table report describes global infrastructure finance activity down roughly 5% by value against the first quarter of 2025 with transaction count falling around 20%, and singles out a drop in refinancing of about 31% as the main weight on the quarter (Green Street, Q1 2026 infrastructure and project finance league table report, published 10 April 2026; that report publishes percentage moves publicly and keeps its absolute volumes behind a registration gate). A window that was wide in 2024 was materially narrower at the start of 2026, which is an argument for running the process early rather than at maturity.
How should a sponsor run the process?
As a financing, not as a renewal. The incumbent lenders have an obvious advantage, which is that they already know the asset, and an obvious interest, which is keeping it at the price they have. The only way to test whether their proposal is competitive is to put the asset in front of lenders who have to win it, and the incumbents price differently when they know that is happening.
The preparation is unglamorous and it is where the money is. Operating data reconciled to the original model; the completion certificate and any outstanding contractor claims closed out; the revenue contract confirmed and its remaining term stated plainly; insurance, environmental and permitting positions current; and a refreshed model built on actuals rather than the close-date forecast. That package is what allows a new lender group to move at the speed of a competitive process instead of a first-time diligence.
Then the two structural questions. Whether to term the debt out fully or to accept another intermediate maturity, which is the same hard-or-soft judgement made once more with better information. And whether refinancing is the right transaction at all, because an operating asset with a resolved risk profile is also the most saleable it will ever be, and the comparison between refinancing, recapitalizing and selling is one calculation, not three separate conversations.
As of September 2026
Sources: project finance and infrastructure finance refinancing volumes and shares, and the FY2023 comparatives, are from Green Street, IJGlobal league tables FY24, published 27 January 2025. The first-quarter 2026 activity, count and refinancing percentage moves are from Green Street, Q1 2026 infrastructure and project finance league table report, published 10 April 2026. The hard and soft mini-perm definitions are from Norton Rose Fulbright, Project Bonds and Mini-Perms: A New Era in the Middle East, November 2012, cited for the structural distinction only and flagged in the copy as dated. No full-year 2025 volume is quoted because the publicly circulating figures could not be reconciled against the primary report. No pricing, margin or saving figure is given, because none is published as a general series. Nothing here describes a transaction or engagement.
This position sits within our project finance practice.