Artificial intelligence is forcing a painful reckoning on Wall Street, rattling its exposure to software lending. As AI threatens the core economics of the sector, private credit funds that aggressively financed software companies are now facing mounting redemption pressure and industry-wide markdowns on collateral valuations.
JPMorgan Chase’s decision to lower collateral values on software loans vulnerable to AI disruption was the canary in the coal mine. The move dragged a public market fear—the existential threat of AI—squarely into the private credit space. Now, fund managers must confront the uncomfortable reality that many traditional software business models may not survive automation at this scale.
This predicament was years in the making. After interest rate hikes in 2022 drove a flood of capital into direct lending, private credit funds chased software deals with abandon. Lenders compressed spreads and accepted borrower-friendly terms, building portfolios now dangerously exposed to obsolescence. Morgan Stanley’s analysis paints a grim picture: roughly 50% of the $235 billion software loan market carries B- ratings or lower, and nearly half of this debt needs refinancing by 2030 in what is likely to be a persistently high-rate environment.
The redemption requests tell the story. In the fourth quarter of 2025, investors pulled funds at nearly three times the previous quarter’s pace, reaching 4.5% of fund net asset value industry-wide. The pain was acute for tech-focused players like Blue Owl, whose fund absorbed redemption requests of 15.4%—the highest among non-traded vehicles tracked by Fitch Ratings. Morgan Stanley was forced to cap redemptions at 5% of outstanding shares, ultimately meeting only 45.8% of withdrawal requests for the quarter.
The core risk here is structural, not cyclical. Software’s traditional moats, such as switching costs and network effects, are being eroded by agentic AI. The threat is a massive consolidation, where enterprises could deploy a single AI agent to orchestrate workflows across their entire infrastructure rather than buying 15 specialized applications. This “substitution effect” mirrors the cloud transition but is unfolding at a far more brutal speed.
Not everyone, however, is predicting extinction. IBM’s Chen Xudong argues the industry confuses automation with elimination, noting that IT modernization involves far more than code translation—it requires hardware upgrades, organizational restructuring, and business process redesign. He points to IBM putting over 10,000 of its own developers on AI tools with measurable productivity gains, suggesting the transition creates new work instead of simply erasing old functions.
For investors, this environment demands surgical precision over broad exposure. Managers with restructuring expertise and rigorous underwriting standards will likely outperform. The companies that survive this shakeout will be those offering genuine competitive advantages in an AI-native world—particularly platforms managing governance and security across distributed AI agents, where traditional software orthodoxies no longer apply.
[References & Sources]
- primebuchholz.com
- investing.com
- businessinsider.com
참고문헌
