The recovery challenge in software credit
The collateral is confidence, the market has stopped clearing, and waiting compounds against the claim. The evidence is public.
- The collateralSoftware value is recurring revenue carried on confidence. It does not decline; it evaporates.
- The market33,575 unsold sponsor-backed companies worldwide; in the US alone, eight-plus years of inventory.
- The stakesA failed process is a failed recovery, and the mark that follows is permanent.
Why software is the hard case
Software is roughly a quarter of direct-lending books, and 20% or more of the loans in many funds. Nonaccruals across the large business development companies sit at five-year highs; watchlists are the longest since 2022–23 (aggregate figures; see the sourcing notes). And the 2021 peak-multiple vintage is the cohort sponsors least want to crystallize.
The deeper challenge is what the collateral is. Software enterprise value is recurring revenue carried on confidence: customers renewing, engineers staying. There are no hard assets underneath, no residual value to liquidate toward. When the confidence goes, value does not decline; it evaporates. Which is why intervention timed to formal distress, built for companies with assets, arrives too late for companies that are mostly belief, and why the recovery question has to be answered earlier than the covenant answers it.
The queue your collateral must sell into
At June 30, 2026, PitchBook counted 33,575 unsold sponsor-backed companies worldwide, up from 32,451 at the end of 2025 and from 15,923 a decade ago, with 13,509 in the United States alone: more than eight years of inventory against the recent US clearing rate of roughly 1,500 sales a year. And every channel narrowed at once: seventy sponsor-backed US IPOs since 2022 against 424 in the five years before, sponsor-to-sponsor demand largely dried up, strategic buyers selective and slow. No one expects the queue to clear in order, or at par. The question inside every credit book is the same: which names get out, in what condition, and at what mark.
The cost of delay
The visible artifact of the wait is the amendment. Across private credit, maturities are extended, covenants reset, cash interest converted to PIK, each amendment buying time in exchange for a larger claim on a company that is not growing into it. Waiting feels free because its price arrives later. It is not free. Customers renew more carefully each quarter a stressed name stays stressed; engineers leave; the buyer universe thins. That is the cost of delay, and it compounds.
The trap inside it: the fastest route to market (launching now, as-is) is often the slowest path to cash. Time-to-market and time-to-cash are different clocks, and only one of them pays.
One credible process
A credit-stressed software company gets one credible trip to market, and its owners get one exit from it. Processes are visible: bankers talk, buyers compare notes, employees read the room. Fail once and every later attempt starts from the broken-process discount: the price the market charges a company that has been shopped. That is the one-exit problem, and it inverts the usual logic of timing. The expensive process is not the late one. It is the unprepared one. For a lender it reads the same way in different units: a failed process is a failed recovery, and the mark that follows it is permanent.
The smooth sale is the fast one, and it is underwritten before anyone goes to market, not hoped for after.