Adam Torkildson
Fund gates were supposed to provide breathing room. In practice, each new cap on quarterly withdrawals at a major private credit vehicle has produced a fresh wave of exit requests from limited partners who read the gate as an admission of vulnerability. The mechanism is working in reverse.
How the Exposure Built Up
The chain runs through the insurance industry. Over the seven years ending in 2025, large PE firms acquired life-insurance and annuity businesses and channeled policyholder reserves into their own private credit funds—a structure that CEPR co-director Eileen Appelbaum documented in April 2026. Those credit funds deployed capital into mid-market software businesses, primarily between 2022 and 2024, at high leverage multiples during a period when enterprise software revenue growth appeared durable.
The question now is whether those borrowers remain creditworthy as AI reshapes the enterprise software market. Code-generation tools are compressing development costs. Workflow-automation platforms are reducing seat counts for business applications. The assumption baked into 2022–2024 underwriting—that SaaS revenue would compound at historical rates—is being tested in real time. No fund letter currently quantifies the exposure by AI-displacement category. “Software” is a line item. The sub-category breakdown does not exist in disclosed form.
Three Funds, Three Gates
Two of the largest perpetual private credit vehicles disclosed quarterly outflow caps in March 2026. A third moved in early April. All three framed the decision as orderly liquidity management. All three did so without disclosing material credit losses—meaning the market is pricing risk, not confirmed losses. Secondary buyers of fund interests have moved the discount to reflect what the primary fund has not yet marked.
The feedback loop that follows gate announcements is well understood in other asset classes. The gate is news. News increases the number of LPs who had been watching and decide to act. The next quarter’s redemption total comes in higher. The discount in the secondary market widens. The new cohort of potential exiters now faces a worse secondary price, creating pressure to file before conditions deteriorate further. Each gate announcement, in this environment, has functioned as a catalyst rather than a circuit breaker.
The Portfolio Split That Matters
The risk is not distributed evenly across the asset class. Portfolios concentrated in horizontal application software—productivity, collaboration, CRM, project management—face the most direct AI-displacement pressure. These categories have the shortest path from LLM capability to revenue disruption. Portfolios anchored in infrastructure software, vertical SaaS with deep workflow integration, or physical-asset collateral face a different and less immediate risk profile.
The Marks Have Not Been Tested
Private credit managers point to structural advantages over public high-yield: direct loan covenants are tighter, workouts can be negotiated confidentially, and forced-sale dynamics do not apply in the same way. Those arguments have merit. They are also untested in a broad software-borrower stress environment. The last meaningful test of public high-yield—2015 in energy, 2008 broadly—preceded the current PE-insurance-private credit structure by a full cycle.
NAV prints from the largest perpetual vehicles over the second and third quarters of 2026 will provide the first hard data point. If AI-displacement-risk metrics begin appearing in LP letters before then, it will indicate that the LP community has pushed hard enough to force a disclosure upgrade—a significant shift from the current posture, where the aggregate “software” line item tells limited partners almost nothing about what they are actually exposed to.
Source: Private Credit Fund Redemptions Climb Sharply, Some Caps Now in Place
