Investors in private credit funds have identified a specific vulnerability that did not register clearly at underwriting: significant loan exposure to PE-owned software companies at the precise moment AI is pressuring that category’s revenue assumptions. The response—redemption requests, secondary market discounts, fund gates—has run since the fourth quarter of 2025 and shows no sign of abating.
How PE Built the Exposure Through Insurance
The architecture is one that CEPR co-director Eileen Appelbaum mapped in April 2026. Starting around 2018 and 2019, large PE firms moved into life-insurance and annuity underwriting, gaining control of policyholder reserves. Those reserves—stable, long-duration liabilities—became the funding base for proprietary private credit funds that operated with limited disclosure and no mark-to-market requirement. The credit funds deployed at scale into PE-owned portfolio companies, particularly mid-market software businesses, during a period when those borrowers could support six-to-eight-times leverage on the strength of their subscription revenue.
The question that emerged in late 2025: does that revenue hold? Generative AI has created a realistic path for enterprise buyers to automate workflows that they previously addressed with purchased software. The revenue hit may not materialize uniformly or quickly—but underwriting that assumed 2022-level SaaS growth rates through 2028 is more exposed than it appeared 24 months ago.
The LP Calculus
Limited partners cannot quantify this risk from fund disclosures because the disclosures do not contain the necessary data. Software exposure appears as a percentage of the total portfolio. Within that percentage, the distribution between infrastructure software (relatively defensible), horizontal application software (higher AI-displacement risk), and vertical SaaS (variable, depending on lock-in) is not reported. An LP trying to size their AI-displacement exposure across a private credit allocation is operating without inputs.
When the risk is unquantifiable, some fraction of the LP base concludes that the appropriate response is to reduce exposure. Redemption requests at major funds climbed through the first quarter of 2026. Two perpetual private credit vehicles moved to cap quarterly withdrawals in March. A third followed in April. All three made the announcements without disclosing credit losses—the gates are preemptive, not reactive to confirmed borrower distress.
The Secondary Market Signal
Secondary buyers of private credit fund interests have priced discounts that exceed stated NAVs. This spread represents the market’s estimate of where marks will eventually settle if software borrower revenue deteriorates. It is not a prediction that losses will materialize—it is a price for uncertainty. The size of the discount has widened with each gate announcement, as each announcement adds information about the direction of fund manager sentiment.
The Differentiation Within the Asset Class
Not every private credit portfolio carries uniform AI-displacement risk. The funds that lent most aggressively to horizontal application software during the 2022–2024 vintage face the most direct exposure. These are tools where AI substitution is near-term and the buyer’s switching cost is low. Funds that built books around infrastructure software, mission-critical vertical SaaS, and asset-backed structures are in a materially different position—different enough that aggregate industry statistics obscure more than they reveal.
The manager community’s structural argument holds that private credit workouts are more controlled than public high-yield distress. Covenants are tighter. Negotiations are private. Marks can be managed without forced sales. All of this is true and none of it has been tested at scale in a software-credit stress environment. The next two quarters of NAV prints will begin to answer whether the structural advantages hold under real pressure, or whether they simply delay the price discovery that the secondary market is already performing.
Source: Private Credit Fund Redemptions Climb Sharply, Some Caps Now in Place