Bridge Loans: Floating-Rate Risk, Rate Caps, and the 2026 Exit Assumption

Key Takeaways

The 2021-Vintage Bridge Loan Time Bomb

The most-stressed slice of the CRE maturity wall is the 2020–2022 vintage of floating-rate bridge debt: loans originated at SOFR + 250–350 bps when SOFR was near zero (0.01–0.11% in the first half of 2021, per New York Fed SOFR data), producing headline coupons of roughly 2.5–3.6%. By the time many of those loans' rate caps expired, SOFR had moved to a 5.30–5.40% range (late July 2023 through mid-September 2024), which on the same spreads means headline coupons of roughly 7.8–8.9% on properties underwritten at coupons under 4%.

The 2021-vintage bridge time bomb is one of the central CRE finance stories of the current cycle. This article is the practitioner reference for what bridge debt is, how it prices in 2026, why rate caps cost more than sponsors expect, what the four exit options look like for a maturing loan, and how to underwrite a deal that actually clears in the current environment.

The audience is institutional value-add and opportunistic CRE sponsors, debt brokers, capital markets analysts, LP investors evaluating sponsor underwriting, and asset managers facing maturity on existing bridge debt. The article works through the structural anatomy, 2026 pricing reality, the rate cap cost that gets underestimated, the four exits available to a maturing loan, and the worked $25M multifamily bridge that the AQ-141 Opportunistic Pro Forma model runs end-to-end.

Anatomy of a CRE Bridge Loan

A bridge loan is short-term debt against a transitional CRE asset: one that isn't yet at stabilized cash flow but is expected to reach stabilization through renovation, lease-up, or repositioning. The bridge finances the acquisition plus the value-add capex; the take-out at stabilization (typically agency, life company, or CMBS) refinances both into long-term fixed-rate debt at the property's stabilized economics.

The structural features that matter:

• Term. 24–36 months initial term, with one or two 12-month extension options contingent on performance milestones (DSCR thresholds, occupancy targets, debt yield benchmarks). A common plan is to stabilize within the initial term and hold the extension as insurance; the worked example below assumes exactly that.

• Rate structure. Floating, typically indexed to one-month CME Term SOFR, a forward-looking rate derived from SOFR futures. TheNew York Fed publishes overnight SOFR, which this article uses as the index proxy; one-month Term SOFR differs slightly. Spreads vary by sponsor tier and property type (covered below). SOFR floors are common in post-2022 loans and vary by lender; pre-2022 deals often had no floor and benefited from the rate compression in 2020–2021.

• Sizing. 60–80% loan-to-cost at acquisition, depending on tier (see the table below); the construction component (capex) is funded in draws against documented progress. Debt yield at stabilization is the lender's exit underwriting metric; this article uses an illustrative 8.0% minimum, and thresholds vary by lender.

• Origination fees. 50–150 bps of loan amount upfront for most tiers and up to 250 bps for heavy-lift deals, plus exit fees of 0–50 bps at refinance or sale. The exit fee is sometimes waived if the bridge lender provides the take-out (seen with debt fund originators that keep loans through stabilization).

• Recourse. Non-recourse with bad-boy carve-outs is common for institutional bridge loans, though terms vary by lender. Some higher-leverage or higher-risk deals carry partial recourse during the construction period that burns off at stabilization.

• Rate cap requirement. Lenders commonly require borrowers to purchase a rate cap (or rate swap, though caps are more common) for the initial loan term, with a renewal cap as a condition of any extension. Cap pricing covered below.

Pricing in 2026: SOFR Spreads by Sponsor Tier

Bridge loan pricing varies materially by sponsor track record, property type, leverage, and lender source. The ranges below are illustrative modeling assumptions, not a published survey; actual quotes vary by lender and deal.

With overnight SOFR at 3.90% on October 6, 2026 (per theNY Fed), headline coupons (SOFR + spread) run roughly 6.65–10.90% across the spread bands above, or 6.75–11.00% where a 4.00% SOFR floor applies. For comparison, the same spreads in the first half of 2021, with SOFR at 0.01–0.11%, produced headline coupons of roughly 2.8–7.1%. The repricing has shrunk the universe of bridge-debt-eligible deals materially: a deal that penciled at a coupon near 3.5% now has to carry a 7.50% headline in the mid-tier example below.

CRE CLO collateral from the 2021–2022 originations is among the most exposed, because those loans were sized to coupons and cap rates that no longer hold.

The All-In Coupon Math Nobody Shows

The headline rate on a bridge loan is SOFR plus spread, plus floor if applicable. The actual all-in cost is higher because of amortized fees, rate cap premium, and exit fees. The full calculation for the illustrative mid-tier bridge used in the worked example below ($18M loan, 24-month term, take-out at month 24, no extension exercised):

• SOFR (October 6, 2026): 3.90%
• Floor adjustment (assumed 4.00% SOFR floor): +0.10% (SOFR is below the floor, so the floor binds)
• Spread (mid-tier multifamily sponsor): 3.50%
• Headline coupon: 7.50%
• Origination fee ($135K, 75 bps) amortized over the 24-month hold: +0.375%
• Rate cap premium ($90K, 50 bps for a 2-year 5.5%-strike cap) amortized over 24 months: +0.25%
• Exit fee ($45K, 25 bps) amortized over 24 months: +0.125%
• All-in effective cost: 8.25%

The 7.50%-headline bridge loan is really an 8.25% effective cost loan after fees, a 75 bps gap. Swap in the 2-year 4.5%-strike cap at 75 bps ($135K) and the cap line becomes +0.375%, for an all-in of about 8.38% and a gap of about 88 bps. If the 12-month extension is exercised, the 25 bps extension fee ($45K) and a renewal cap for months 25–36 come on top; neither is in the figures above. Sponsors who model only the headline coupon understate their cost of bridge debt by at least this much.

Rate Caps: The Hidden Cost

A rate cap is a derivative contract that pays the borrower the difference between SOFR and a strike rate whenever SOFR exceeds the strike. For a 2-year cap at a 4.5% strike on $18M of notional, each monthly period pays: max(0, SOFR − 4.5%) × $18M × days in period ÷ 360 (Actual/360). If SOFR were 5.5% for a 30-day period, the cap would pay $15,000. The cap protects against catastrophic rate moves; it doesn't protect against the SOFR + spread structure becoming uneconomic at rate increases below the strike.

Illustrative cap premiums used in this article (assumptions for modeling; actual pricing moves with the SOFR forward curve and volatility):

• 2-year cap at 4.5% strike on $18M notional: ~75 bps premium (~$135K upfront)
• 2-year cap at 5.5% strike on $18M notional: ~50 bps premium (~$90K upfront)
• 3-year cap at 4.5% strike on $18M notional: ~125 bps premium (~$225K upfront)
• 3-year cap at 5.5% strike on $18M notional: ~85 bps premium (~$153K upfront)

The institutional issue with rate caps in 2026 is the renewal. Sponsors who bought caps at the trough of cap pricing in 2020–2021 paid far less for them than a renewal costs today. Assuming a trough premium of 10–20 bps, a renewal at the 50–125 bps illustrative grid above is about 2.5–12.5x the original premium in bps (strikes and tenors are not matched, so treat this as a rough range). Many 2021-vintage bridge loans hit a cap renewal moment in 2024–2025 and had to either purchase a substantially more expensive cap or refinance into less-rate-exposed permanent debt. The sponsors that couldn't refinance and couldn't afford the renewal cap are the ones showing up in the special-servicer pipeline.

The Four-Exit Waterfall

A maturing bridge loan has four exit options, in decreasing order of borrower preference. The weights shown for each path are illustrative base-case assumptions for a probability-weighted sensitivity table, not market statistics; they sum to 100% so they can drive a weighted IRR.

• 1. Refinance into permanent debt. Property has reached stabilization, NOI supports permanent debt at the current rate environment, agency or CMBS take-out clears. The sponsor's planned exit. Illustrative weight: 50%.

• 2. Property sale. Stabilized or near-stabilized property is sold; the bridge is paid off from sale proceeds. Common when the take-out math doesn't clear at sponsor-acceptable LTV but a buyer exists at the marked NAV. Illustrative weight: 20%.

• 3. Extension (with conditions). Sponsor exercises a 12-month extension; lender consents, typically conditioned on a rate cap renewal, possibly a paydown to a target debt yield, occasionally sponsor equity replenishment. Often used as a path-to-perm bridge. Illustrative weight: 20%.

• 4. Modification / workout / foreclosure. Sponsor cannot afford the cap renewal, cannot clear refinance, cannot sell at a level that pays off the loan. Path runs through forbearance, loan-to-own structures, deed-in-lieu, or judicial foreclosure. Illustrative weight: 10%.

The 2026 Refinance Gap

The "refinance gap" is the dollar amount by which a maturing loan exceeds the loan a current lender would underwrite on the same property. The math is mechanical: at acquisition, the property was valued at a 5.0% cap rate and the bridge was 75% of that value (75% LTV, which equals 75% LTC when cost equals value at purchase). At maturity, the same property at a 5.75% cap rate (cap rate expansion) and 65% LTV (tighter underwriting) supports a meaningfully smaller permanent loan. The difference is the gap the sponsor has to fill with equity or accept as loss.

An illustrative example. A $30M acquisition valued at a 5.0% cap rate on $1.5M of NOI. The bridge funded 75% = $22.5M. At maturity in 2026, NOI is $1.5M as planned; a 5.75% exit cap implies a value of $26.1M (down from $30M). Take-out lenders sizing to 65% LTV on the $26.1M value write a $17.0M permanent loan. The refinance gap is $22.5M − $17.0M = $5.5M of additional equity required at refinance, about 25% of the original loan balance.

In this example LTV, not debt yield, is the binding constraint: $1.5M of NOI on a $17.0M loan is an 8.8% debt yield, and an 8.0% debt-yield test alone would permit $18.75M. At the original $22.5M bridge balance, the debt yield would be 6.7%. The gap scales with the size of the cap rate move and the leverage step-down; assets whose NOI also fell short of plan see larger gaps.

Worked Example: $25M Multifamily Bridge

A 200-unit value-add multifamily acquisition. $25M total cost: $20M purchase price, $4M renovation budget, $1M closing costs and reserves. Bridge loan: $18M (72% LTC) at SOFR + 350 bps with a 4.00% SOFR floor in October 2026; $7M sponsor equity. 24-month initial term with one 12-month extension. Rate cap at 5.5% strike for the 2-year term. Base case: stabilization and take-out at month 24, extension not exercised. The cap lapses at month 24, so an extension would require a renewal cap whose cost is excluded below.

All-in cost calculation:

• Coupon: 4.00% SOFR floor (SOFR 3.90%) + 3.50% spread = 7.50% headline
• Origination fee: $135K (0.75% of $18M), amortized 24 months → +0.375%
• Rate cap premium: $90K (0.50% of $18M), amortized 24 months → +0.25%
• Exit fee: $45K (0.25% of $18M), amortized 24 months → +0.125%
• All-in effective cost: 8.25%

Annual interest on the $18M loan at the 7.50% headline = $1.35M, assuming SOFR stays at or below the floor. Interest is funded by an interest reserve (assumed $1.5M, funded by the sponsor at closing on top of the $25M budget) plus operating cash flow from existing tenants during renovation. The 24-month interest accrual at 7.50% is $2.70M against the $1.5M reserve, so operating cash flow or additional sponsor equity needs to cover $1.2M. This assumes the full $18M is funded at close; drawing the capex portion over time would reduce the accrual.

Exit math at stabilization (month 24): the renovated property's stabilized NOI of $2.0M (an assumption, equal to an 8.0% yield on the $25M cost) / 5.75% exit cap = $34.8M stabilized value. Take-out at 65% LTV = $22.6M permanent loan (an 8.8% debt yield, above the illustrative 8.0% minimum), which repays the $18M bridge with $4.6M of gross refinance proceeds. After the $45K exit fee, that is about $4.56M before refinance closing costs; net of the $1.5M sponsor-funded reserve, and of the $1.2M interest shortfall if that was also funded with sponsor equity rather than operating cash, the net is about $1.9M. The deal works on the math but only because (a) the renovation produces the underwritten NOI growth and (b) exit cap rates don't widen further. Both are real risks; the AQ-141 Multifamily Opportunistic Pro Forma sensitivities are sized to exactly this scenario.

Six Mistakes Bridge Borrowers Make

• Modeling SOFR as static. SOFR is the floating rate the loan resets to monthly. Sponsors who modeled SOFR at 0.05% in early 2021 for the full 24-month bridge discovered material increases in interest accrual when SOFR moved. Always model SOFR as a forecast curve, not a point estimate.

• Quoting the headline coupon as the all-in cost. In this article's illustrative mid-tier example, the headline coupon is 75–88 bps below the all-in effective cost after origination, cap premium, and exit fees, and more if the extension is exercised. The all-in is what an honest IC memo should quote.

• Treating the rate cap as one-time cost. The cap is commonly required for the initial term and each extension. On this article's illustrative inputs, a renewal cap costs about 2.5–12.5x what a 2020–2021 trough cap cost. Budget for the renewal cost; don't assume the initial cap protects through extension.

• Ignoring the exit fee. Exit fees of 0–50 bps add to the all-in. A 25-bps exit fee on the $18M worked-example loan is $45K of additional cost at refinance. Small relative to the loan but real in the IRR calculation.

• Assuming the bridge lender will roll into take-out. Some debt funds offer bridge-to-perm programs; many do not. Sponsors who assume the bridge lender will provide the takeout without confirming the structure get caught at maturity needing a third-party refinance lender they haven't underwritten.

• Failing to underwrite the exit waterfall. The four exits (refinance, sale, extension, workout) have different IRR and equity implications. Underwriting only Path 1 (clean refinance) ignores the other 50% of outcomes in the illustrative weighting above (Paths 2–4). Build the sensitivity tables for at least the top three exit scenarios.

Do It in Apers

Related Articles

• Construction Loans: Draws and Interest Reserves: the construction debt that often precedes a bridge.
• CMBS Conduit vs SASB, Defeasance and YM: the takeout option after stabilization.
• IRR Calculator and Formula for Real Estate: the return math the bridge feeds into.
• IRR Sensitivity Analysis and Stress Testing: the stress tables for the bridge exit scenarios.

FAQ