Uncertainty around AI's impact on software companies is creating two distinct CLO markets. Older deals with high software exposure are heavily discounted and risky, while newly issued, software-light CLOs offer superior risk-adjusted returns, even if they aren't trading at a discount.

Unlike the public equity markets, software exposure in credit markets is concentrated in private, not public, companies. An estimated 80% of these issuers are private, and 50% are rated B- or lower, creating a unique and more challenging risk profile due to lower credit quality and less transparency.

Despite fears of AI disruption, private credit software loans have significant downside protection. With loan-to-value ratios around 30-40%, there is a substantial equity cushion. A company's value must erode by nearly 70% before the lender's principal is at risk, highlighting the structural safety of debt versus equity.

Private equity giants like Blackstone, Apollo, and KKR are marking the same distressed private loan at widely different values (82, 70, and 91 cents on the dollar). This lack of a unified mark-to-market standard obscures true risk levels, echoing the opaque conditions that preceded the 2008 subprime crisis.

In private markets, there's a perverse incentive for both private equity owners and private credit lenders to avoid marking down asset values. This "mark to make-believe" system keeps valuations artificially high, hiding underlying financial stress and delaying the recognition of losses.

The absence of daily pricing in private credit removes an essential discipline. Mark-to-market in public markets acts as an honest, early warning system that forces managers to scrutinize underperforming assets, a mechanism private lenders lack.

A consistent 2-5% of Europe's public high-yield market restructures annually. The conspicuous absence of a parallel event in private markets, which often finance similar companies, suggests that opacity and mark-to-model valuations may be concealing significant, unacknowledged credit risk in private portfolios.

Jeff Gundlach argues private credit's attractive Sharpe ratio is misleading. Assets aren't priced daily, hiding risk. When an asset is finally marked, it can go from a valuation of 100 to zero in weeks, exposing the “low volatility” as a dangerous fallacy.

A significant portion of private credit portfolios consists of loans to software companies, which were underwritten based on predictable, recurring revenue. AI is now fundamentally disrupting these business models, threatening to devalue the very collateral that underpins billions of dollars in these 'safe' loans.