Three of the largest perpetual private credit vehicles have imposed quarterly redemption caps since March 2026—a sequence that has reordered the way limited partners think about exit liquidity in an asset class long marketed on its stability relative to public markets.
A Transparency Deficit Meets a New Risk Category
The sequence started with a diagnosis. Eileen Appelbaum of the Center for Economic and Policy Research published a structural analysis in April 2026 tracing how private equity firms, over the previous seven years, redirected life-insurance and annuity policyholder reserves into private credit funds. Those funds lent at scale to PE-owned portfolio companies, many of them mid-market application software businesses that took on high-leverage debt between 2022 and 2024.
The question that followed was predictable: if generative AI erodes the revenue base of enterprise software companies over the next 36 months, how many of those borrowers can service their debt loads? No fund letter currently answers that question with specificity. Software exposure appears as a category total. AI-displacement-risk breakdown does not appear at all.
The Gate Announcement as Its Own Event
Fund managers that moved to gate argued they were managing orderly liquidity, not signaling distress. The market disagreed, at least partially. Each cap announcement triggered secondary-market repricing of fund interests, which in turn created new incentive for LPs on the fence to file redemption requests before conditions tightened further.
None of the three funds that capped withdrawals have disclosed material credit losses. The repricing is forward-looking—investors are discounting the possibility that marks will eventually reflect borrower stress, not the confirmed reality of it. Whether those marks actually deteriorate depends on what happens to enterprise software revenues over the next two to three years.
Where the Portfolio Risk Concentrates
Not every private credit book carries the same exposure. The portfolios most at risk lent heavily to horizontal application software—project management tools, CRM platforms, productivity software—categories where AI substitution is most direct and near-term. The portfolios with more defensible positions focused on infrastructure software, vertical SaaS with regulatory lock-in, and asset-backed lending where the collateral is physical rather than recurring-revenue.
The manager community is drawing a distinction between this cycle and public high-yield stress events of prior years. Direct lending covenants are tighter. Workout negotiations happen privately rather than through distressed-debt markets. Both arguments are genuine. Neither is testable until a forced-sale environment materializes.
Signals to Watch Through Mid-2026
Two data points will clarify the picture by the end of the third quarter. First, NAV prints from the largest perpetual vehicles—if marks hold, the redemption cycle may stabilize. If they move, the secondary discount widens and more LPs exit. Second, LP letters: when fund managers start disclosing AI-displacement-risk metrics by portfolio segment, it signals that the LP base has asked loudly enough to force a response. That kind of disclosure does not happen pre-emptively.
Private credit accumulated roughly $1.7 trillion in assets under management through the post-2015 expansion. The asset class has not been tested by a broad mark-down cycle. The next six months represent the most informative stress test it has faced.
Source: Private Credit Fund Redemptions Climb Sharply, Some Caps Now in Place