The Software Maturity Wall

Recurring Revenue Was Sold as a Bond Substitute

The pitch was elegant. Software companies bill annually. They bill the same customers year after year. Net Revenue Retention above 115% — common in vertical SaaS through 2022 — meant the existing customer base alone grew the borrower's revenue by 15% a year, before adding a single new logo. From a credit underwriter's seat, that looked less like equity risk and more like an inflation-protected bond with a software wrapper.

So private credit funds wrote loans against it. A lot of loans.

One-third of every private credit fund in the United States now carries SaaS exposure. Business Development Companies (BDCs) — the publicly-listed wrappers that recycle private credit to retail and registered-investor money — extended more than 15% of their 2025 loan originations into the sector. And those numbers understate the concentration risk, because they treat all "SaaS" as one bucket. A vertical workflow tool for orthopedic surgeons sits in the same line item as a horizontal CRM that an AI agent can replicate over a long weekend.

Why 60× Earnings Was a Defensible Bid

Through 2021-2022, a sponsor-backed SaaS LBO had a particular shape. The buyer paid 12-18× revenue (or 50-70× earnings on the rare profitable target). The capital structure leaned heavily on private credit unitranche — typically 5-7× EBITDA in senior debt at SOFR+550 to SOFR+700. The model that justified the price required, on a five-year hold:

• 25-30% annual revenue growth (sustained, not declining)
• Gross margin expansion of 200-400 bps via AI-assisted cost reductions and offshored CX
• NRR maintained or expanded above 115%
• An exit multiple at or near the entry multiple — i.e., the asset class doesn't re-rate

None of those four assumptions hold individually if any one of them breaks. The model is multiplicatively levered to growth.

The lender's collateral — software-equity claims junior to the senior loan — was modeled at the entry multiple. The mark-to-market of that collateral has compressed by roughly 60-70%. The loan still pays interest, but the loan-to-collateral-value ratio has migrated from "comfortable cushion" to "underwater on a refinancing day," depending on the borrower.

The Customer Base Was Always Sitting Inside Another PE Portfolio

The single most important detail in private credit's SaaS story is one almost nobody writes down: the customers buying horizontal SaaS were themselves PE-backed companies. The same buyout funds that financed Vista Equity's and Thoma Bravo's roll-ups of CRM, marketing automation, sales-engagement, and document-workflow tools also owned the manufacturers, distributors, and services businesses that were the addressable market for those tools.

That was the cross-portfolio synergy story. It's now the cross-portfolio cannibalization story.

This is not theoretical. AI-referenced deals were 72% of all SaaS M&A in 2025, up roughly twelve-fold from 2018. AI-native challengers — many funded by the same sponsors who own the legacy SaaS targets — are explicitly pitching themselves as replacements for specific incumbents. And two-thirds of private-equity buyers, when diligencing claimed AI capability on the BUY side, say the AI claims are not passing diligence. The signal that's actually trustworthy is on the SELL side: the public software comp set down 30% in four months.

What the NRR Collapse Looks Like at the Borrower Level

A horizontal SaaS company with $200M ARR underwritten at 120% NRR was modeled to be a $250M ARR company in twelve months, before any new sales. At 95% NRR — a number that's becoming common in marketing automation, sales-engagement tooling, and entry-tier customer-support platforms — the same company is a $190M ARR company in twelve months. Combined with a sponsor cost structure built for the higher growth profile, a 25-percentage-point swing in NRR can move EBITDA by 40-60% in a single fiscal year.

That swing doesn't show up as a default. It shows up as a covenant breach.

When the Loan Has to Roll Into a Worse Market

This is where the slow-burning credit deterioration converts to a forced repricing event.

$12.7 billion of unsecured BDC debt comes due in 2026 — a 73% increase over 2025. That debt was originally raised at investment-grade-adjacent pricing on the assumption that the underlying loan book was high-quality and diversified. The same BDCs are now refinancing into a market where:

• Banks have reclaimed share of the corporate-lending market on cheaper funding and easier regulatory treatment
• The underlying SaaS loan collateral marks at one-third its origination valuation
• "Shadow" defaults — non-disclosed distress — are estimated at ~6% against a headline 2% rate
• Morgan Stanley's credit strategy team projects direct lending defaults could hit 8%, approaching the COVID-era peak

Who Pays First, Who Pays Last

The largest private-credit sponsors are not waiting for the maturity wall. They've already begun reshuffling the loss-bearing tranches.

The Order of Pain

In a private-credit unwind, the loss waterfall is well-defined on paper and entirely renegotiated in practice. On paper:

1. Sponsor equity — first to absorb loss
2. Junior tranches / second-lien / mezzanine
3. Senior unitranche (the bulk of private credit)
4. BDC unsecured noteholders
5. BDC common equity holders (retail, RIA channel)

In practice, the GP often has reputational and franchise reasons to absorb pain that contractually sits with someone else — see KKR/BlackRock/Apollo's current "fix struggling funds" activity. That decision has a limit. When the GP can no longer absorb without impairing its own balance sheet or fundraising prospects, the contractually-junior tranche takes the next hit.

The BDC retail investor — the registered RIA client allocated to a "private credit" sleeve in their model portfolio — is structurally last in line and structurally has the least visibility into where the underlying SaaS exposure sits. They are the bag-holder. They don't know it yet.

You Cannot Tell a Defensive Vertical SaaS from a Replaceable Horizontal One Until It's Too Late

The most damning structural feature of the current private credit ecosystem: the disclosure framework treats all SaaS loans as one asset class. An LP looking at a quarterly NAV statement can see "Software/SaaS: 19% of portfolio." They cannot easily see whether that 19% is:

• Defensive vertical: a workflow tool for radiology practices where the switching cost is regulatory and the replacement timeline is years (low AI-displacement risk)
• Horizontal commodity: a CRM, marketing-automation, or sales-engagement tool where the switching cost is one quarter and an AI agent can replicate the workflow (high displacement risk)

Both will have appeared to perform identically in 2023. By the time their performance diverges in 2027 quarterlies, the refinancing window has closed.

What to Watch in the Next Four Quarters