Software Lending and the Recurring Revenue Premium

Outstanding loans by private credit funds to software-as-a-service firms increased from almost $8 billion in 2015 to over $500 billion, or 19% of total direct loans in the United States, by the end of 2025, according to a March 2026 analysis by the Bank for International Settlements.¹ That trajectory places software at the center of the private credit market’s growth story, and it has elevated the question of how — and how well — lenders are underwriting subscription-based businesses into one of the more consequential debates in middle-market finance. By the end of 2025, a third of private credit funds had extended loans to the SaaS sector, according to the same BIS analysis, and business development companies — the most transparent segment of private credit by virtue of their quarterly SEC filings — extended over 15% of their loans to SaaS firms that year.¹

The structural logic for this concentration is not difficult to explain. SaaS businesses built around enterprise software tend to produce what private credit investors prize most: visible, contractual, recurring revenue streams that renew automatically, require minimal physical collateral to maintain, and allow lenders to underwrite against a predictable cash flow base rather than against a liquidation value. A business-to-business software platform with multi-year subscription contracts and high switching costs presents a fundamentally different credit profile than a manufacturer with cyclical revenues and inventory pledged as collateral. The credit analysis is different in kind, not just in degree.

Why Subscription Revenue Earns a Structural Premium

The premium that subscription-based software commands in debt markets is grounded in a specific characteristic of the business model: net revenue retention. Net revenue retention measures how much revenue a software company keeps and grows from its existing customer base over a twelve-month period, capturing renewals, upsells, price increases, and cross-sells in a single metric. SaaS Capital’s 2025 survey of more than 1,000 private B2B SaaS companies found a clear and exponential correlation between net revenue retention and growth rate: companies reporting NRR of at least 110% posted median growth rates 83% higher than the population median, while those with NRR below 100% fell below the median.² The same survey found that increasing NRR from the 90%–100% range to the 100%–110% range improved median growth by 5 percentage points.²

For credit underwriting, these relationships matter because they translate directly into the durability of debt service capacity. A company whose existing customer base is expanding faster than it churns is, in credit terms, building enterprise value even when it is not booking new logos. That compounding dynamic gives lenders both an income floor and a recovery cushion that traditional businesses with linear, project-based revenue streams simply cannot offer. AllianceBernstein’s private credit team noted in a May 2025 analysis that customers with 90% or higher retention effectively represent assets generating revenue for a decade or more — but that these assets are invisible on standard financial statements because accounting rules do not permit their capitalization.³ The result is a persistent mismatch between reported earnings and underlying business quality that makes traditional credit tools insufficient for assessing the true risk profile of a fast-growing software borrower.

The ARR Lending Structure and Its Mechanics

The financing structure that emerged to bridge this gap is the annual recurring revenue loan, which sizes debt as a multiple of a company’s contracted ARR rather than as a multiple of current EBITDA. As AllianceBernstein’s analysts explained, a software company with $100 million in recurring revenues seeking a loan of twice that amount would borrow $200 million — a ratio that would be difficult to justify on an EBITDA basis for a growth-stage business reinvesting most of its cash flow into customer acquisition.³ The rationale is that lenders can determine the appropriate loan size by modeling what EBITDA the business would generate if run for profitability rather than growth, then expressing that figure as a debt-to-ARR multiple.

The covenant architecture for these loans reflects the special demands of the model. Rather than relying solely on traditional leverage and coverage tests, effective ARR loan covenants test recurring revenue trajectory and include provisions — sometimes called “covenant flips” — that require borrowers to demonstrate profitability by a specified date.³ AllianceBernstein’s framework, for example, structures five-year ARR loans with a requirement that the borrower meet predetermined earnings thresholds within two years of close.³ This covenant design acknowledges that the borrower is deliberately sub-optimizing current profitability in favor of growth, while also ensuring that the lender retains the ability to redirect the business toward cash generation if conditions deteriorate. Covenants requiring minimum liquidity to support payroll and core operations alongside revenue-based tests are also important monitors, since even financially sound software businesses can become illiquid if they fund unprofitable growth with debt.

The Scale of Concentration and the Emerging Risks

The BIS analysis published in March 2026 placed the growth of private credit SaaS lending in the context of a market that has moved from niche to systemic in less than a decade.¹ The report noted that software companies’ stocks fell by almost 30% between October 2025 and February 2026 as markets priced in concerns about AI-driven disruption to traditional SaaS business models. Over the same period, BDC stock prices fell by roughly 10% on average, and BDCs with higher SaaS exposure underperformed those with lower exposure by approximately 5 percentage points.¹ These developments reflect investor concern that the credit quality assumptions embedded in existing software loan portfolios — particularly assumptions about customer retention durability and competitive moat — may require revision in an environment where generative AI tools can replicate certain software capabilities at materially lower cost.

Proskauer’s Private Credit Default Index, which tracks senior-secured and unitranche loans across nearly 700 loans representing $189.2 billion in original principal, reported an overall private credit default rate of 2.73% for the first quarter of 2026, up from 2.46% in the fourth quarter of 2025.⁴ Proskauer’s Private Credit Group noted that while overall default levels remained below those in the broadly syndicated loan market, and that software and technology sector default rates had “remained relatively stable” despite investor concerns, the first quarter result continued a “modest upward trend in defaults that began in mid-2025.”⁴ The distinction is meaningful: software is not producing defaults at elevated rates by historical standards, but the market is pricing in the possibility that it will.

What the Operator Data Shows About Retention

The KeyBanc Capital Markets and Sapphire Ventures 16th Annual Private Company SaaS Survey, released in November 2025 and covering more than 100 private SaaS companies, offered a ground-level view of the underlying operator metrics that support — or complicate — credit underwriting assumptions.⁵ The survey found that year-over-year ARR growth was expected to accelerate from 15% in 2024 to 20% in 2025, the first acceleration in three years. Gross revenue retention, which had declined to 86% in 2023, was expected to approach the 90% threshold in the near term, while net revenue retention remained above 100% through the survey period and was expected to show modest further improvement.⁵ These trends, if they hold, support the thesis that the SaaS credit risk profile remains manageable: a business retaining roughly 90% of its revenue base on a gross basis and expanding above 100% on a net basis is not exhibiting the rapid deterioration that would translate into imminent default pressure.

The EBITDA margin trajectory reported in the survey is also relevant to the lending case. Profitability metrics had improved consistently since 2022, and the survey found that EBITDA margins were expected to cross into positive territory for the private SaaS cohort in 2026 — a development that would accelerate the natural migration of ARR-based loans toward more traditional EBITDA-based covenant structures as borrowers mature.⁵ That migration is itself a healthy sign for the sector: a software business that has grown its ARR base while also achieving positive EBITDA is, by definition, a business that lenders can now underwrite with both a subscription-quality revenue base and a cash flow coverage ratio, compressing credit risk from two directions simultaneously.

Underwriting Discipline in a Concentrated Market

For lenders active in software, the concentration data from the BIS and the market volatility of late 2025 argue for more rigorous sector-level underwriting rather than less. The case for software as a credit asset has not collapsed; the structural arguments about recurring revenue quality, switching costs, and contractual cash flows remain valid. What has changed is the environment in which those arguments must be tested. AI-driven competitive pressure is a genuine uncertainty, and underwriters who treated NRR as a permanent structural feature of a given software product rather than an outcome that must be earned and defended each renewal cycle are now revisiting that assumption.

Effective underwriting of ARR-based loans requires understanding the profitability of each service line the borrower provides, not just the aggregate revenue trajectory. AllianceBernstein’s framework distinguishes between the variable gross profit of an existing customer, the cost to acquire a new one, and the retention rate — a three-variable system that determines whether the economics of growth are additive to enterprise value or dilutive of it.³ A software business with strong aggregate NRR but deteriorating unit economics in its fastest-growing product line may be a materially worse credit risk than its headline retention number suggests. The same logic applies in reverse: a business with temporarily suppressed NRR driven by a specific cohort issue rather than structural competitive erosion may warrant more favorable treatment than surface-level benchmarks imply.

Mandatory covenant reporting in this sector should be calibrated to the specific risk signals that subscription businesses generate. Quarterly disclosures should include ARR, gross revenue retention, net revenue retention, customer acquisition cost, and top-customer revenue concentration. A covenant framework designed for a traditional manufacturing borrower, applied without modification to a SaaS platform, will systematically mis-read the warning signals that matter most, because the leading indicators of credit stress in subscription businesses — accelerating churn, rising CAC payback periods, declining net expansion — manifest in operating metrics months before they appear in income statement or balance sheet data.

Conclusion

The growth of private credit lending to software companies reflects a genuine structural credit advantage in subscription-based business models. That advantage is real, measurable, and supported by loss experience accumulated over more than a decade. But the BIS’s finding that SaaS loans now represent 19% of U.S. direct lending¹ — a figure that has grown roughly 60-fold since 2015 — also means that the sector’s next credit cycle will be a private credit event, not merely a software sector story. The BIS data also show that the market is already registering stress: software equity prices down sharply, BDC valuations under pressure, and a regulatory community beginning to ask pointed questions about concentration. Lenders who treat subscription economics as a framework for rigorous, ongoing underwriting rather than as a rationale for compressed diligence are well-positioned for that cycle. Those who treated the sector’s favorable historical loss experience as a permanent feature rather than a condition that requires maintenance through disciplined covenant design, active monitoring, and genuine borrower-level transparency face a more difficult road ahead.

—

Footnotes

1. Bank for International Settlements, Sebastian Doerr, Egemen Eren, Ingomar Krohn, and Karamfil Todorov “Outstanding loans to SaaS firms increased from almost $8 billion in 2015 to over $500 billion, or 19% of total direct loans, by end-2025”; “a third of private credit funds have extended loans to the SaaS sector”; “BDCs … extended over 15% of their loans to SaaS firms in 2025”; “Software companies’ stocks collapsed by almost 30% between October 2025 and February 2026”; “BDCs with high exposure to software firms have performed around 5 percentage points worse than those with low exposure.” (BIS Quarterly Review, March 2026, Box B1, “Private credit’s software lending meets AI disruption.”)
2. SaaS Capital, Nick Perry “Companies with the highest NRR report median growth that is 83% higher than the population median”; “Increasing Net Revenue Retention (NRR) from the 90% to 100% range to the 100% to 110% range improves growth rate by 5 percentage points.” (“What is a Good Retention Rate for a Private SaaS Company in 2025?”, September 18, 2025, drawing on SaaS Capital’s 14th annual survey of more than 1,000 private B2B SaaS companies.)
3. AllianceBernstein, Jay Ramakrishnan and Shishir Agrawal “a software company with recurring revenues of $100 million seeking a loan of twice that amount would borrow $200 million”; “retention rates of 90% or higher are common”; “a lender can determine the appropriate loan size based on the ratio of debt to earnings before interest, taxes, debt and amortization (EBITDA) of the business if it was run for profitability”; “a five-year loan might be made on a recurring-revenue basis but with a covenant that requires the borrower to meet or beat a predetermined earnings threshold after two years.” (“Software Recurring Revenue Lending: Flexibility and Skill Required,” AllianceBernstein, May 9, 2025.)
4. Proskauer Rose LLP, Stephen A. Boyko “The Index revealed a default rate of 2.73% for the period of January 1, 2026 – March 31, 2026. The rate represents an increase from 2.46% in Q4 2025”; “This quarter’s Index encompasses 697 loans representing $189.2 billion in original principal amount”; “Despite concerns about the performance of loans in the software and technology sector, default rates in that industry have remained relatively stable.” (Proskauer Private Credit Default Index, Q1 2026, released April 27, 2026.)

5. KeyBanc Capital Markets and Sapphire Ventures “YoY ARR growth is expected to accelerate from 15% in 2024 to 20% in 2025 for the first time in the last three years”; “Gross Retention is expected to approach the 90% threshold in the near term after declining to 86% in 2023, while net retention has continued to remain above 100%”; “EBITDA margins have continued to improve since 2022 and are expected to breach the threshold of profitability in 2026.”