A bridge extension gets budgeted the way it gets described: a fee and a signature. Twenty-five or fifty basis points to the lender, a new maturity date, and you keep executing. The extension fee on a $3.5 million loan at 50 basis points is $17,500 — real money, but a rounding error against the asset.
That is not usually the cost of the extension. The cost is the replacement rate cap, and it is a different kind of number, because it is an option price. It moves against you exactly when you most need the extension, it is quoted rather than posted, and it does not show up in a single coverage ratio in your model.
The argument here is not that caps are expensive. Sometimes they are cheap. The argument is that the price is knowable twelve months before you need it, and not knowable sixty days out — and that the difference between those two moments is the difference between a decision and a bill.
The Mechanic Sponsors Under-Model
Floating-rate bridge debt comes with a hedge requirement. The lender requires it because a floating coupon on a levered asset is the fastest path to a collateral problem: if SOFR runs, debt service runs with it, coverage breaks, and the lender is holding a non-performing loan on an asset that operationally did nothing wrong. The cap caps that.
Here is the structural part. The lender sets the requirement — the strike, the notional to be covered, the term, the acceptable counterparties — and the sponsor buys the instrument. The party choosing the specification is not the party paying for it. On the original closing that gets absorbed into the capital stack alongside every other closing cost. At extension, twelve or twenty-four months later, it arrives on its own, at a price nobody modeled, against a curve nobody expected.
Which is why the number to know is not the extension fee. It is what the replacement cap will cost when the extension test comes due.
Why the Price Moved
A cap is not a fee. It is a strip of European call options on SOFR — one caplet per reset date, each one paying the difference if SOFR sets above the strike. The premium is what an options desk charges to write that strip, and it is a function of five things: where the forward curve says SOFR is going, implied volatility around that path, how far the strike sits from it, the term, and the credit of the counterparty on the other side.
Four of those five are market variables. None of them is stable, and the first two both moved this year in the direction that makes caps more expensive.
The forward curve is the one you can verify. On July 29 the Committee held the target range at 3.50%–3.75% — but the vote was 9–3, with Beth Hammack, Neel Kashkari, and Lorie Logan each preferring a quarter-point increase, the first three-way hawkish alignment since 2016. The long end took the point: the 10-year Treasury finished July at 4.75%, back to its highest level since January 2025. A market that spent 2025 pricing the next cut spent this summer pricing whether the next move is a hike.
That repricing is not a forecast about your loan. It is a direct input to your cap premium. Every caplet in the strip is worth more when the curve underneath it shifts up, because more of the strip finishes in the money. A cap premium quoted against a curve that priced cuts is not the premium you will be quoted against a curve that prices hikes, and a budget line carried forward from your 2024 closing is not an estimate — it is a number from a different market.
I am going to be careful about the second variable. Implied volatility is the other half of an option price and it moves premiums hard in both directions, but I do not have a sourced read on where rate vol sat at the end of July, and I am not going to assert one to make the paragraph land better. The mechanism is real; treat the vol channel as a reason your quote can move without the curve moving at all, not as a claim about this week.
(The population this matters most to — loans past maturity but current on interest — we sized three weeks ago in Past Maturity, Still Paying. Trepp's June print is the most recent as of this writing; the July report had not posted when this went up.)
Notional Is Linear. Term Is Not.
Two mechanics with real planning consequences, and sponsors reliably get the second one backwards.
Notional is roughly linear. Premium is quoted as a percentage of notional, and at a constant strike and term it scales close to proportionally. Double the loan, roughly double the premium. This is the intuitive one, and it is why a percentage-of-notional quote travels reasonably well across deal sizes — with the caveat below about where published ranges actually come from.
Term is not linear. A five-year cap costs more than twice a two-year cap. The reason is the caplet structure: a longer cap is not the same protection stretched thinner, it is more options, and the later ones are individually worth more than the earlier ones, because uncertainty about where SOFR sits compounds with time. Each additional month of protection is priced above the month before it.
Now put that against a twelve-month extension, because the planning implication inverts. A one-year cap is the cheapest cap on the board in absolute dollars. It is also the most expensive cap on the board in carry, because the entire premium is absorbed over a single year. There is no second or third year to spread it across. That is the whole of the next section.
The Worked Example
Same file as the last post — one operator, one fact base. Value-add multifamily, Northeast secondary market, business plan executed, stabilized. Everything below traces from two numbers: a $3,500,000 maturing bridge balance and $315,000 of stabilized annual NOI.
Where the loan sits today
The loan floats at SOFR plus 350. SOFR was 3.66% for July 31, 2026 per the New York Fed — the most recent published value as of this writing, and it resets daily, so treat it as a snapshot.
- All-in rate: 3.66% + 3.50% = 7.16%
- Interest-only carry: $3,500,000 × 7.16% = $250,600/year ($20,883/month)
- Coverage on current debt service: $315,000 ÷ $250,600 = 1.26×
- Debt yield on the maturing balance: $315,000 ÷ $3,500,000 = 9.00%
- Annual free cash after debt service: $315,000 − $250,600 = $64,400
That last number is the one the rest of this hangs on. Sixty-four thousand four hundred dollars is what the property produces in a year, after paying its debt, before anything else. Not distressed — 1.26× coverage is comfortable and no servicer is calling. But it is the budget the extension has to come out of.
One disclosure before the table. What follows prices the consequence of a premium. It does not assert what a cap costs today, and you should not read it as a quote. We did not have live dealer pricing at this notional when this was written; cap pricing is quote-driven, specific to your strike and term, and stale within days. Secondary commentary through early 2026 put CRE cap premiums broadly in a 1%–4%-of-notional band across strikes and terms — but most published pricing is derived from $25M–$50M notional, roughly an order of magnitude above a $3.5M bridge, and scaling it down is an assumption, not a measurement. The table is built as a sensitivity for exactly that reason: it works regardless of where your quote lands.
A twelve-month extension, replacement cap, premium absorbed in one year
| Premium (% of notional) | Cash cost | Effective all-in coupon | Drag | As multiple of annual free cash |
|---|---|---|---|---|
| 0.50% | $17,500 | 7.66% | +50bp | 0.27× |
| 0.75% | $26,250 | 7.91% | +75bp | 0.41× |
| 1.00% | $35,000 | 8.16% | +100bp | 0.54× |
| 1.50% | $52,500 | 8.66% | +150bp | 0.82× |
| 1.84% | $64,400 | 9.00% | +184bp | 1.00× |
| 2.00% | $70,000 | 9.16% | +200bp | 1.09× |
| 2.50% | $87,500 | 9.66% | +250bp | 1.36× |
The finding, and it needs no quote to be true: on this loan, any twelve-month cap priced above 1.84% of notional costs more than the property earns in a year after debt service. That is the break-even, and it is arithmetic, not a forecast.
There is a reason the break-even row lands on a round 9.00%. For an interest-only loan, the debt yield is the all-in rate at which free cash goes to zero. Whether that zero arrives through the coupon or through a one-year cap premium is a matter of which line of the budget it hits — the property does not care. Your debt yield is your one-year cap budget, expressed as a rate. On this file that budget is 184 basis points, and every basis point of premium above it is funded by the sponsor, not by the asset.
And the version with the number in it, for anyone pricing the tail: on a $3,500,000 bridge loan producing $315,000 of NOI, a replacement cap priced at 2.50% of notional costs $87,500 for twelve months of protection — roughly 1.36 times the property's entire annual free cash flow after debt service.
The same dollars on a three-year cap
| Premium | Cash cost | Amortized drag | Effective all-in |
|---|---|---|---|
| 2.00% | $70,000 | 67bp/yr | 7.83% |
| 2.50% | $87,500 | 83bp/yr | 7.99% |
| 3.00% | $105,000 | 100bp/yr | 8.16% |
Identical cash out the door. Very different effective cost, purely because of the window it is absorbed over. A sponsor weighing a twelve-month extension against a longer recapitalization should be comparing these numbers against each other — not the headline coupons, which are the same in both cases.
The nuance that explains why this surprises people. The cap premium is upfront cash. It is not debt service. It does not depress DSCR, it does not touch debt yield, and it will not trip a covenant. It comes out of equity and reserves. Run this deal through a model that solves for coverage and leverage — which is every model — and the entire cost is invisible. The sponsor sees 1.26× before the extension and 1.26× after, and writes an $87,500 check that appears nowhere in the analysis that produced the decision.
What Is Actually Negotiable
Sponsors treat the hedge requirement as fixed. Some of it is. Much of it is not, and the pieces that move are the ones that price.
Strike. The single biggest lever, and the one worth solving before you accept a number. Lenders generally set the strike so the loan clears a minimum coverage test at the capped rate — which means you can back into it. On this file, at a 350 spread over a $3.5M interest-only balance:
| Required DSCR at capped rate | Max annual debt service | Implied all-in ceiling | Implied strike | vs. spot SOFR 3.66% |
|---|---|---|---|---|
| 1.00× | $315,000 | 9.00% | 5.50% | +184bp |
| 1.10× | $286,364 | 8.18% | 4.68% | +102bp |
| 1.15× | $273,913 | 7.83% | 4.33% | +67bp |
| 1.20× | $262,500 | 7.50% | 4.00% | +34bp |
| 1.25× | $252,000 | 7.20% | 3.70% | +4bp |
Read that table before you panic about the one above it. Every plausible strike on this file sits at or above where SOFR is today — the tightest test, 1.25× coverage, still lands four basis points out of the money. This cap is not intrinsically in the money; there is no embedded cost you are simply obligated to pay. Its entire premium is time value, and time value is negotiable in a way intrinsic value is not. A lender who will move from a 1.20× to a 1.10× test moves your strike 68 basis points further out and takes real money off the premium. The loans getting hurt worst in this cycle are the ones where a low legacy strike sits below spot — those caps carry intrinsic value and price like it. Know which one you are before you budget.
Notional coverage. Whether the cap must cover the full balance or only the funded portion. On a value-add deal with unfunded capex holdback, that distinction is worth real dollars, and it is worth asking about explicitly rather than assuming the answer is the full commitment.
Term. Match the cap to the extension, not to the outside date. Buying a three-year cap to support a twelve-month extension is buying two years of optionality on a loan you intend to retire — and per the section above, those are the two most expensive years in the strip.
A swap instead. Some lenders will take one, and a swap has no upfront premium, which solves the cash problem cleanly. It also has an asymmetry a cap does not: if rates fall, you are paying above market with a breakage cost to get out, and that cost is largest in exactly the scenario where you would most want to refinance. A cap is an option; a swap is an obligation. On a loan you plan to retire inside a year, that asymmetry usually argues for the cap.
Basis. The cap index must match the loan index. This is a checkbox, not a strategy, but a mismatch means you are hedged against something adjacent to your actual exposure, and it is the kind of thing that gets discovered at the worst possible moment.
Which lender you are having this conversation with matters too, and bank appetite in particular now depends on where the institution sits on CRE concentration. Our NJ Bank CRE Lending Tracker computes it quarterly for every New Jersey–chartered bank.
The Actual Argument: Price It Early
Everything above is a setup for one operational point.
A cap quote twelve months before your extension test is an input. It sits next to the takeout math and tells you which path is cheaper. If the cap prices at 60 basis points, the extension is trivially the right answer and you stop thinking about it. If it prices at 250, you have eleven months to refinance early, negotiate the strike down, restructure, or market the asset into a bid you control.
The same quote sixty days out is a bill. Nothing about the number changed. Everything about your optionality did.
This is why "price the cap early" is not generic advice about being organized. Cap premiums are the one major cost in a bridge extension that is both large and fully knowable in advance — you can get an indication today for a cap you will buy next June. Sponsors get an indication on the takeout coupon a year ahead as a matter of routine and would find it strange not to. Almost nobody does it for the hedge, and the hedge is the line item that actually moves.
And note that the argument does not depend on direction. If the curve reprices lower and caps get cheaper, the sponsor who priced early captures that too — they just learn their extension is cheap and proceed with confidence. What they never do is discover the number at the moment they have lost the ability to act on it.
The Counter-Case
Five things cut against the above, and the second one is the most important paragraph here.
A cap that pays off is worth every basis point. Nothing here is an argument against hedging. If SOFR runs 200 basis points, the cap is the reason the loan survives and the sponsor keeps the asset. The premium is the price of optionality in a market where three FOMC voters just dissented in favor of a hike. Read the tables as a cost to be sized and negotiated, not as waste to be avoided.
An expensive cap does not automatically favor refinancing. This is the comparison that actually decides the question, and it is worth doing precisely. We sized the takeout on this same file last week: at a representative 6.25% takeout and 1.25× coverage, supportable proceeds came to $3,410,657 against the $3,500,000 payoff — an $89,343 cash-in requirement to refinance at par. Set that beside a cap at 2.50% of notional, $87,500. Those two numbers are $1,843 apart. They do not buy remotely the same thing. The $89,343 retires the bridge into a fixed-rate loan with a known constant and ends the exercise. The $87,500 buys twelve months of protection, after which you still owe $3,500,000 and still need a takeout. When the cap premium approaches the equity a par refinance would require, the refinance is usually the better trade — and a sponsor who never runs both numbers side by side never sees that.
Some lenders will genuinely negotiate. Strike, notional coverage, and term are all live, and the strike table above shows how much a coverage-test concession is worth. Sponsors who treat the hedge specification as non-negotiable leave money on the table.
Indicative pricing is not a quote, and this post shipped without live ones. Said plainly because it should be: the premium levels in the tables are a sensitivity, the 1%–4% band is secondary and dated, and published cap pricing is mostly derived from notionals ten to fifteen times the size of this loan. The break-even — 1.84% of notional on this file — is the only number here that requires no market input at all, which is why it is the one to carry.
The curve can reprice again, quickly. Markets are currently leaning toward a hike; a soft July employment print on Friday, August 7 would cheapen caps in a hurry, and the energy-price path could do it from the other direction. That is not a hole in the argument. It is the argument. Nothing about the timing case depends on which way the curve moves — only on whether you learn your number while you can still do something about it.
FAQ
Who pays for the interest rate cap on a bridge loan? The sponsor. The lender sets the requirement — it is protecting its collateral and its minimum coverage test — and the borrower buys the instrument and delivers it at closing or at extension. The sponsor bears a price it did not set, on terms it usually did not negotiate, which is the structural reason this line item gets under-modeled.
Why did rate cap costs increase in 2026? A cap is a strip of options on SOFR, so its premium is a function of the forward curve, implied volatility, how far the strike sits from where SOFR is expected to be, and the term. Premiums rise when the curve shifts toward higher rates. Through 2026 that is the direction it moved: the FOMC held at 3.50%–3.75% on July 29 with three voters dissenting in favor of a quarter-point increase, and the 10-year Treasury finished July at 4.75%, its highest since January 2025. A premium quoted against a curve that priced cuts is not the premium you will be quoted against a curve that prices hikes.
How much does a rate cap cost relative to loan size? Premium scales roughly linearly with notional — double the loan and you roughly double the premium at the same strike and term — but it grows non-linearly with term, because a cap is a strip of caplets and each successive caplet carries more time value than the one before it. A five-year cap costs more than twice a two-year cap. That asymmetry is why a short extension cap looks cheap in dollars and expensive in carry: the whole premium is absorbed over a single year.
Does a rate cap premium affect DSCR? No. The premium is upfront cash, not debt service, so it does not appear in a coverage ratio or a debt yield calculation. It comes out of equity and reserves. A sponsor who models only DSCR and debt yield will not see this cost anywhere in the model — which is precisely why it surprises people. On a $3,500,000 loan producing $315,000 of NOI and $64,400 of free cash after debt service, a twelve-month cap priced above 1.84% of notional costs more than the property earns in a year, and coverage never moves.
Send Us the Loan Terms and the Extension Test
If you have a floating-rate bridge loan with an extension option inside the next eighteen months, the useful exercise takes an afternoon: pull the hedge requirement out of your loan documents, back into the strike your lender's coverage test implies, and get an indication before the number is load-bearing. Send us the balance, the spread, the NOI, and the extension conditions, and we will size the strike, tell you where the break-even premium sits against your free cash, and run the extension against the takeout so you are comparing the two decisions rather than defaulting into one. If the answer is that your cap is cheap and you should extend, that is a good afternoon too.
We placed a $5.55M Newark bridge for a sponsor taking a stabilized asset toward permanent financing, and the hedge conversation on that file ran exactly this way. Pricing the cap is arithmetic once you have the quote. Getting the quote — and knowing which of 100-plus bank, agency, and private lenders will move on a strike — is the part worth a phone call.
Dominick Prevete — Founder, Blue Sky Capital Advisors. 31 years in real estate finance; 100+ bank and private-lender relationships across the bridge, agency, and private-credit spectrum. 4 Sutton Ct, Hamburg, NJ 07419 · (908) 220-6404. NMLS information available upon request. Lending in all 50 states.
SOFR is as published by the Federal Reserve Bank of New York for July 31, 2026 and resets daily. The rate decision, 9–3 vote, and dissents are from the FOMC statement of July 29, 2026. The 10-year Treasury close is from Advisor Perspectives' Treasury Yields Snapshot for July 31, 2026. Delinquency figures referenced are from Trepp's June 2026 CMBS delinquency report; the July report had not posted as of August 4, 2026. Cap premium levels shown are a sensitivity, not quotes — no live dealer pricing was obtained at this notional, and the 1%–4%-of-notional band cited is secondary and dated to early 2026. Takeout proceeds and the $89,343 cash-in figure are carried forward from our July 30, 2026 analysis. Every deal prices to its own quote on its own day. This is market commentary, not a commitment to lend, an offer to transact in derivatives, or investment advice.
Loans are for business purposes only and are not subject to TILA, RESPA, or HOEPA. Not for primary residences. Equal Housing Opportunity. All loans subject to underwriting approval. Rates and terms shown for illustration; actual rates depend on deal specifics. We do not lend to borrowers with credit below 600 or on owner-occupied properties.