Private credit leaders push back on AI panic: exposure is limited, cushions are real, and the math still matters
Software stocks have been hit hard as investors price in the risk that AI will gut legacy vendors. That fear spilled into alternative managers with big credit businesses. Executives at Blue Owl and Ares say the market is overreacting.
Blue Owl co-CEO Marc Lipschultz told investors that Fortune 500 buyers are not "ripping out" their software to ask a chatbot. He pointed to industry context from Nvidia's Jensen Huang and added, "I can do it in math," arguing that the loss assumptions being priced in don't match portfolio data.
Blue Owl crossed $300 billion in AUM, yet its stock fell about 4% on the day and 27% over the past month. The firm says software loans represent roughly 8% of assets, with average loan-to-value around 30%. Since those deals were underwritten, equity values for the borrowers are up about 23%, and management says they see "largely green flags" on credit quality.
Lipschultz framed the market's implied hit as extreme: you would "have to destroy 70% of the value of every one of these software companies" for the current marks to be justified. In that scenario analysis, Blue Owl sees the impact as lower returns for a year or two, not portfolio impairment.
Ares Management echoed the message. CEO Michael Arougheti said software exposure is less than 9% of its private credit AUM and reminded investors that "not all software exposure is the same." Ares crossed $600 billion in AUM in the quarter, with just over $400 billion in credit, and said its earnings growth outlook is unchanged by AI risk in the current book. Shares still fell 9% on the day.
Pressure isn't isolated to those two. Other alternative managers with software exposure, including Apollo, Blackstone, and KKR, were pulled into the same sell-off. That matters because enterprise software has been a favorite hunting ground for private equity and private credit for a decade given high margins and sticky subscription revenue.
What the positioning actually looks like
- Debt vs. equity: Managers emphasized that their exposure is largely debt, not common stock. With ~30% LTV, equity cushions absorb a lot before lenders take pain.
- Sector concentration: At Blue Owl, software loans are 8% of assets; at Ares, software is under 9% of private credit AUM. For BDCs more broadly, about 18% of loans are tied to software, according to PitchBook.
- Flow dynamics: Blue Owl allowed more withdrawals in its tech-focused BDC and saw 15% of net assets redeemed. It also halted a planned fund merger in November due to volatility-reminders that liquidity optics can amplify price action even when fundamentals hold.
The bear case still exists-here's the credible version
UBS strategists flagged that in an "aggressive disruption" path-think recession-like default rates combined with industry transition-private credit could see higher defaults than other credit segments. That is a real risk if AI compresses pricing power or slows new sales at mid-market software vendors over multiple quarters.
The difference is speed and depth. Most mid-market software credits are underwritten to recurring revenue, high gross margins, and resilient retention. AI would need to drive a sharp step-down in ARR growth and renewal rates, not just create product noise, to push many deals into distress at current leverage levels.
What finance teams should track next
- Revenue durability: Net dollar retention, logo churn, and enterprise renewal timing in 2H and FY27 cohorts.
- Pricing pressure: Evidence of discounting or lengthening sales cycles where AI alternatives are credible substitutes.
- Underwriting cushions: Entry LTV, interest coverage, and covenant headroom by vintage; expect tighter terms in new issue.
- Sponsor behavior: Will PE owners add capital or cut aggressively if growth slows? Watch cash burn, R&D reallocation to AI features, and hiring freezes.
- Refinancing calendar: 2026-2028 maturities, forward rates, and access to club deals; note any migration from term loans to unitranche.
- Marks and flows: BDC NAV marks versus public comp moves; retail redemption trends and gating policies that can feed volatility.
- Second-order effects: Valuation resets on software exits, implications for GP fundraising, and knock-on to origination volumes.
How to position without overreacting
- Re-underwrite AI risk by subvertical: Dev tools, SMB apps, and content-heavy categories carry more substitution risk than mission-critical back-office systems.
- Stress tests: Model 300-500 bps ARR growth hits, 5-10 pts lower net retention, and 100-150 bps higher spreads. Recompute interest coverage and default probabilities; focus on recovery given LTV.
- Opportunistic credit: Spread widening in quality senior secured software names with real ARR and low churn can be attractive if covenants are clean.
- Equity discipline: Wait for a base-building phase and proof on renewals before leaning into public comps or PE add-ons exposed to AI substitution.
- LP messaging: Share the math-exposure percentages, LTV, and scenario outcomes-to keep redemptions tied to facts, not headlines.
Two useful references if you want primary context: Nvidia's recent keynote on enterprise AI direction is available on its site (NVIDIA GTC Keynote), and UBS maintains an overview of its credit and disruption research here (UBS Research Insights).
Bottom line: pricing has moved faster than fundamentals. If the loan books are truly sitting at ~30% LTV with stable ARR engines, the downside looks like lower returns for a few years, not broad impairment. Keep your eye on renewals, coverage ratios, and the 2026-2028 refi wall-that's where this story will be decided.
If you're pressure-testing your team's AI thesis for software exposure or building an internal toolkit, this curated list of AI tools for finance is a practical starting point: AI Tools for Finance.
Your membership also unlocks:
