- The share of private software debt trading at less than 80% of its original value peaked at 6.1% in September 2025 — a five-year high — according to MSCI analysis of $73 billion in institutional private-credit holdings; this stress predates the SaaS-pocalypse (the 24% Q1 2026 selloff in software stocks driven by AI-displacement fears) and was driven by software companies that took on heavy debt against pandemic-era revenue peaks that didn’t sustain when employers cut back on SaaS subscriptions.
- Software now accounts for 17% of institutional private-credit portfolios — second only to industrials and up from 12% in 2019 — and about 25% of major business-development company (BDC) portfolios that sell private credit to individual investors; this concentration means AI-driven SaaS disruption is not a marginal risk but a core portfolio exposure for anyone invested in private credit, at a moment when MSCI estimates it takes two quarters for private markets to fully price in events that public markets reprice instantaneously.
- MSCI’s Patrick Warren says the record markdowns through September 2025 don’t yet reflect AI-agent adoption fears: “Even when we get the Q1 numbers, we’re not likely to fully see the reaction to Claude Cowork — the data lags there are keeping us in suspense.” KKR’s BDC FS KKR Capital has already taken a $560 million write-down (10% of NAV) from defaults including software maker Medallia; Blackstone Secured Lending Fund marked a healthcare loan down to 80% of face value.
- Healthcare debt marked down more than 20% also hit a five-year high of 7.4% last year, and consumer-discretionary loans suffered their worst markdowns since early in the pandemic — meaning the software stress is part of a broader private-credit reckoning after three years of elevated interest rates, with the AI disruption now arriving on top of an already-fragile foundation; paradoxically, MSCI found no significant outperformance from more experienced fund managers, though larger funds contained less stressed debt than smaller ones.
What Happened?
New MSCI data covering $73 billion in institutional private-credit holdings reveals that private software loans were already deteriorating well before AI-displacement concerns triggered the SaaS selloff in early 2026. The share of software debt marked down more than 20% peaked at 6.1% in September 2025 — a five-year high — driven by companies that had borrowed heavily against pandemic-era revenues, then saw those revenues plateau or decline as employers scaled back software spending. Many of these loans were structured by private-equity firms that took software companies private at high valuations and leverage levels, assuming recurring revenues would compound indefinitely. When employers cut back, the debt load remained. Software has grown to 17% of institutional private-credit portfolios and approximately 25% of BDC portfolios — making it the second-largest sector exposure after industrials — which means the AI-disruption risk now sitting on top of pre-existing leverage stress is a major unresolved vulnerability across the $2 trillion private credit market.
Why It Matters?
Private credit has been sold to pension funds, insurance companies, sovereign wealth funds, and increasingly individual retail investors as a high-yield, low-volatility alternative to public bonds. But the MSCI data exposes a structural lag: private-credit marks take approximately two quarters to reflect events that public markets price in immediately. The SaaS-pocalypse — a 24% decline in software stocks in Q1 2026 driven by fears that AI agents will replace traditional SaaS tools — has not yet appeared in private-credit valuations. When it does, it will hit portfolios that are already stressed from three years of high rates and pandemic-era leverage. For individual investors in BDCs, this timing gap means the worst markdowns may still be ahead, not behind. And if private-credit funds need to restrict withdrawals to preserve capital (as several already have), those investors may find themselves locked in at precisely the wrong moment.
What’s Next?
Watch Q1 2026 private-credit marks — expected in the coming weeks — for the first quantified look at how much AI-disruption fears have filtered into portfolio valuations. The key risk is non-linear: a wave of simultaneous markdowns could trigger an acceleration of the BDC redemption cycle already underway, forcing fund managers to sell assets to meet withdrawals at discounted prices, which triggers further markdowns — a dynamic that in public markets is called a run but in private markets unfolds over quarters rather than days. The IMF and several central banks have flagged illiquid private-credit funds as a systemic concern precisely because of this feedback loop. Whether private credit’s AI-software exposure becomes a managed write-down cycle or something more acute depends heavily on how quickly and deeply AI agents actually displace enterprise software spending — a question that remains open.
Source: The Wall Street Journal
