The Private Credit Debate Isn't Going Away. Here's What Investors Should Know.

The article argues that private credit remains fundamentally sound, with Brookfield and Blackstone saying they see no systemic risk and highlighting strong historical returns, including Blackstone's 9.4% annual net return in non-investment-grade private credit since inception. The main caution is concentrated exposure to unsecured software lending, where AI disruption could pressure margins and increase defaults. Brookfield says it has no material software exposure, while Ares Capital says 85% of its software portfolio has low AI disruption risk.

Analysis

The market is being too coarse in treating private credit as one bucket. The real spread is between asset-backed senior lending and unsecured software lending: the former is effectively a pass-through on contracted cash flows and replacement-cost collateral, while the latter is underwriting to business models that AI can reprice faster than covenant packages can adjust. That means the next wave of losses, if it comes, is more likely to show up as idiosyncratic markdowns and fund-level dispersion than a systemwide credit event.

For BN and BX, the important second-order effect is not just lower default risk, but stronger capital allocation optionality. If public-market lenders get more cautious around software and lightly secured growth credit, the large platforms with real-asset origination pipelines should capture share and negotiate better terms, especially on AI infrastructure where financing is tied to tangible collateral and utility-like demand. That can support fee-related earnings even if headline private-credit AUM growth slows.

The underappreciated risk is that AI disruption will surface first in the leverage finance ecosystem, not in software equity multiples. Once lenders begin stress-testing renewal risk, borrowers with weaker sticky revenue will see tighter spreads, shorter maturities, and lower advance rates within the next 2-4 quarters, which can force dilutive equity raises or asset sales before outright defaults appear. That is a better setup for relative-value trades than outright sector shorts because the damage should be concentrated, not broad-based.

Consensus appears to be missing that the 'problem' is less about private credit as an asset class and more about underwriting duration versus technology duration. If AI compresses software margins faster than credit documents can reprice, the most exposed lenders will look fine until refinancing windows open, then underperform abruptly. That argues for staying constructive on large diversified platforms with collateral-heavy books while fading managers with outsized software exposure or opaque covenant-light structures.

Key Decisions for Investors

• Long BN vs. short a software-heavy private credit proxy over the next 3-6 months; target 10-15% relative outperformance if AI-driven dispersion widens and real-asset lending re-rates premium.
• Add BX on pullbacks for 6-12 months; risk/reward favors fee-earning platforms with broad origination and the ability to rotate away from software toward AI infrastructure financing.
• Reduce exposure to BDCs or private-credit managers with visible software concentration; use 1-2 quarter horizon for refinancing risk to show up in non-accruals and spread widening.
• For a cleaner pair, long ARCC against a basket of smaller, less transparent direct lenders; ARCC’s disclosed stress testing and portfolio mix should support relative resilience if defaults rise modestly.
• Avoid shorting the whole private-credit complex outright; the better expression is short names with unsecured growth-credit books, because broad systemic-risk fears look overstated and could unwind quickly.