DSCR is the single number that decides whether a deal pencils. Coupon, term, and proceeds all flow through it, and in 2026 the floor is creeping up across nearly every lender bucket. Treasuries that refuse to settle, insurance allocations that have grown selective, and a CMBS market still digesting 2024-2025 office losses have all pushed underwriters to ask for more cushion. Brokers who quote yesterday's minimums to sponsors are losing deals at the LOI stage. This piece walks the current floors lender by lender.
Quick refresher: what DSCR is and why lenders weight it now
You know the formula: net operating income divided by annual debt service. What changed is how lenders treat the output. Through the cheap-money decade, DSCR was a sanity check that almost every stabilized deal cleared. With the SOFR curve where it sits and cap rates that have only partially adjusted, DSCR is now the binding constraint on proceeds for a majority of loans. Most lenders size to the lower of LTV, DSCR, and debt yield, and DSCR wins the tug-of-war more often than not.
Two structural shifts matter for 2026. First, lenders are stress-testing DSCR at a constant rate above coupon rather than at coupon, which mechanically lifts the required in-place number. Second, refi waves from 2021 vintage paper have made every credit committee allergic to thin coverage, even on assets they like. The result is a market where a 1.20x quote that would have closed in 2022 now needs 1.30x or a rate buy-down.
Lender categories at a glance
Five buckets cover the institutional CRE debt market. Each has a different cost of capital, regulatory regime, and tolerance for stress, which is why their DSCR floors diverge.
- Agency (Fannie Mae, Freddie Mac, HUD): Multifamily-only, government-backed pricing, the deepest pool of perm capital. HUD 223(f) and 221(d)(4) sit alongside the GSEs as the cheapest non-recourse money for qualifying assets.
- CMBS / conduit: Fixed-rate, ten-year non-recourse paper pooled into bond trusts. B-piece buyers and rating agencies set the underwriting tone, and that tone is conservative right now.
- Life insurance companies: Match-funded balance-sheet lenders chasing institutional-quality assets in primary markets. Lowest leverage, lowest spreads, longest memories.
- Regional and community banks: Recourse and limited-recourse construction, bridge, and mini-perm. Deposit-relationship driven and currently capital-constrained after 2023.
- Debt funds and bridge lenders: Floating-rate transitional capital, often warehouse-financed. They live where the agencies and life cos won't, and they price accordingly.
DSCR minimums by category and asset class
The table below reflects in-place stabilized minimums for 2026 vintage originations. Numbers are qualitative ranges from broker checks across primary and secondary MSAs; tertiary markets and lease-up deals see tighter requirements.
| Lender type | Multifamily | Industrial | Retail (necessity) | Office (CBD) | Hospitality | Net lease |
|---|---|---|---|---|---|---|
| Agency (Fannie/Freddie) | 1.20-1.25x | n/a | n/a | n/a | n/a | n/a |
| HUD 223(f) | 1.176x (statutory) | n/a | n/a | n/a | n/a | n/a |
| CMBS / conduit | 1.30-1.35x | 1.30-1.40x | 1.40-1.50x | 1.50-1.65x | 1.55-1.75x | 1.25-1.35x |
| Life insurance co | 1.35-1.45x | 1.35-1.45x | 1.45-1.55x | 1.55-1.70x | not active | 1.30-1.40x |
| Regional bank (perm) | 1.25-1.30x | 1.30-1.35x | 1.35-1.45x | 1.40-1.55x | 1.45-1.55x | 1.25-1.30x |
| Debt fund (bridge) | 1.00-1.10x going in | 1.05-1.15x | 1.10-1.20x | 1.15-1.25x | 1.10-1.20x | 1.05-1.15x |
A few notes on reading the table. CMBS numbers assume a constant-payment underwrite at the actual coupon; some shops still publish "stressed" minimums based on a refi constant, which run 10-15 bps higher. Life co ranges tighten quickly as you move down the asset-quality stack, and several major lenders have effectively redlined office below trophy class. Debt fund minimums look low because they are quoted on going-in coverage with a stabilized pro forma showing 1.25x or better at exit; sponsors who can't show that path don't get the loan.
How rate environment moves the floor
The mechanical link between rates and DSCR floors is straightforward, but worth restating. When the ten-year sits at 4.40-4.70% and conduit spreads are 175-225 bps, all-in coupons on CMBS perm paper land in the 6.15-6.95% range. At those coupons, sizing to a 1.30x DSCR on an asset trading at a 6.0% cap means proceeds top out around 60-65% LTV, so the LTV constraint stops binding and DSCR takes over as the proceeds governor.
Floating-rate paper amplifies the effect. Bridge lenders quoting SOFR plus 350-500 bps require interest-rate caps, and they size DSCR to the strike rate of the cap, not the spot rate. A sponsor buying a cap with a 5.50% strike on a 4.30% SOFR pays for the privilege of a higher in-place coupon for sizing purposes. The math is unforgiving: every 25 bps of stressed coupon is roughly 3-4% of proceeds at constant DSCR.
When rates rally, the floors don't fall as fast as you'd hope. Credit committees moved to higher minimums on the way up and have been slow to migrate back down, both because they remember the 2022-2023 mark-to-market pain and because B-piece buyers in CMBS are still demanding cushion. Plan on a six-to-nine-month lag between a sustained rate move and a meaningful relaxation in published DSCR minimums.
Underwriting tips: stabilizing vs. stabilized assumptions
The most common mistake we see in broker packages is conflating in-place and stabilized DSCR. Lenders care about both, and they apply different floors to each.
For stabilized deals, the in-place DSCR drives proceeds. Use trailing-three-month NOI annualized, not T-12, and be ready to defend why the trailing period is representative. Strip out one-time items but expect the lender's inspector to add back any expense you understated. Real estate taxes deserve special attention: most lenders re-underwrite to the post-sale assessed value, which can compress NOI by 5-15% in jurisdictions with frequent reassessment.
For lease-up or value-add deals, the going-in DSCR may be below 1.0x and the lender will size to a stabilized number under a stress test. Agency stabilized underwrites typically apply a vacancy floor of 5% (multifamily) regardless of in-place occupancy, and they discount achieved rents to a market study they commission themselves. CMBS conduits will haircut new leases that haven't burned off free rent, and they'll re-amortize TI/LC reserves based on weighted-average remaining term. Build the pro forma the lender will build, not the one the sponsor wants to see.
bipsio lets you set DSCR as a configurable gate per mandate, so stabilized and value-add criteria can live side by side without one set of assumptions polluting the other. Hugo flags borderline deals before they go to LOI, which saves the conversation where a sponsor finds out at term sheet that the lender's underwrite is 18 bps tighter than the broker's.
Asset-class quirks
Multifamily: Agency dominates, and the 1.25x floor at Fannie is sticky for market-rate deals. Workforce and affordable allocations get DSCR relief, sometimes 1.15x, but only with rent restrictions and LURA recordation. Lease-up deals at the GSEs use a 1.00x going-in test combined with a stabilized 1.25x test; you need to clear both.
Industrial: Single-tenant credit deals price like net lease and cover similarly. Multi-tenant shallow-bay still gets life co attention at 1.35-1.40x, but lenders are watching tenant rollover schedules carefully after the post-2023 rent rollback in some Sun Belt submarkets. Last-mile assets in supply-constrained markets clear minimums easily; secondary logistics parks face tougher questions.
Retail: Necessity and grocery-anchored deals are the bright spot. Conduits will stretch to 1.40x for a strong-credit grocer with 10+ years remaining, and they'll go higher with shadow-anchored shop space. Power centers and unanchored strip retail face floors closer to 1.50x and shorter amortization. Mall paper is essentially uneconomic outside special situations.
Office: The hardest asset class to finance in 2026. Trophy CBD with credit tenancy and long WALT can clear life co at 1.55-1.65x with 50-55% LTV. Class B suburban office faces 1.65-1.75x minimums when it can find a lender at all, and most originators want recourse or a meaningful sponsor guarantee. Many regional banks have paused new office originations entirely.
Hospitality: DSCR floors look high because the cash flow is volatile. Conduits underwrite to a stressed RevPAR (typically a 5-10% haircut to trailing) and apply a 4% FF&E reserve before computing NOI. The 1.55-1.75x range reflects the lender's view that a 200-bp drop in occupancy can erase a third of NOI. Limited-service flagged hotels in primary markets sit at the favorable end.
Net lease: Single-tenant credit deals get the lowest DSCR floors because the cash flow is contractual. Investment-grade tenants with 15+ year leases price almost like corporate paper, and 1.25-1.30x DSCR is achievable. Sub-investment-grade or shorter-term deals push toward 1.35-1.40x, and lenders will require dark-value testing on special-purpose buildings.
When to expect waivers or carve-outs
DSCR minimums are not immovable. Lenders will accept lower in-place coverage in three common situations. First, hard cash collateral or a debt service reserve held by the lender, typically 9-12 months of payments, can buy 10-20 bps of DSCR relief because the lender treats the reserve as synthetic NOI for the trailing period. Second, a sponsor-funded interest reserve during lease-up, sized to carry the asset to stabilization, lets the lender underwrite to the stabilized number rather than going-in. Third, recourse or a partial repayment guarantee from a creditworthy sponsor can move a regional bank's floor down 15-25 bps.
What rarely works in 2026: arguing that the trailing period was anomalous, projecting rent growth above the lender's market study, or trying to add back ownership-level expenses the lender considers operational. Credit committees have heard every version of these pitches and the answer is no.
The waiver conversation is also where mandate-level configuration earns its keep. When you can show a capital source that your gate is set at the lender's published floor and the deal clears it on the lender's own assumptions, you remove the suspicion that the broker package was tuned to make the deal look better than it is. See /pricing for how bipsio handles this, or sign up to see Hugo flag a borderline deal during a live scan.
DSCR minimums in 2026 are not catastrophic, but they are firm, and they reward brokers who quote them honestly. Build the lender's underwrite into your package from day one and the term sheet will look like the LOI.
