The Case of AI Disruption and CLO Dispersion

Software’s Impact

While software exposure is often used as a proxy in CLO analysis, AI-related disruption increasingly impacts adjacent sectors, such as business services and healthcare. We estimate that, on average, approximately 11% of U.S. CLO portfolios and 7% of European CLO portfolios are exposed to near‑term AI disruption, with meaningful levels of dispersion across managers. However, we do expect that certain borrowers within these sectors will demonstrate viable long‑term paths forward.

In the U.S., declines in software loan prices have driven CLO equity valuations near post COVID-era troughs, pressuring BB tranches within CLO structures. For example, average levels on BB market value overcollateralization (MVOC) tests recently fell by approximately 1.2 percentage points to 104%—implying only a four percent cushion before full repayment of BB tranches becomes impaired (Exhibit 1).¹

Notably, the share of BB tranches with thin MVOC cushions (i.e., below 3%) has doubled this year to roughly 23%. Meanwhile, approximately 13% of BB tranches exhibit cushions of less than zero—indicating potential principal impairment under further collateral deterioration and reflecting reduced market tolerance for incremental downside risk. Select BBB tranches are also beginning to show pressure in market‑value terms as overcollateralization cushions compress.

Exhibit 1

Lower Software Loan Prices have Pressured MVOC Cushions

Exhibit 1 highlights two CLO managers that experienced outsized price moves relative to the broader market during the recent increase in market volatility, though for different underlying reasons. Despite underweight exposure to the software sector relative to the market (“A”, highlighted in the shaded box above), the first CLO manager experienced higher price volatility driven by idiosyncratic exposure to specific, underperforming software issuers. In contrast, the second CLO manager (“B”, highlighted in the shaded box above) had more concentrated software exposure, resulting in greater sensitivity to sector-wide price declines as well as higher overall portfolio and sector-driven volatility.

Increased Differentiation

While AI‑driven disruption likely represents a genuine fundamental risk to certain corporate credits and may lead to issuer-specific downgrades, the market is increasingly differentiating risk across software business models and subsectors rather than treating the sector as a single risk bucket. As a result, dispersion is expanding across CLOs, with investors increasingly focused on issuer‑level exposure, underwriting quality of managers, and general portfolio construction—allowing them to distinguish between managers rather than focusing on cursory information on software portfolio weights.

For context, Exhibit 2 compares software exposure of 10 U.S. CLO managers (ranked by their weighted average price change since the end of 2025). Exhibit 2 not only shows broad dispersion in performance, but it also reveals significant variations in software allocations, rating exposure, and price changes to software holdings.  Given the dispersion evident across these 10 CLOs and mounting concerns about AI disruption, performance dispersion is likely to persist, most visibly across mezzanine and equity tranche pricing.

Exhibit 2

Software Sector Breakdown by Average Portfolio Price Change

Another key consideration is the maturity profile of the underlying loans, which influences refinancing risk and price sensitivity. Managers with a higher share of software exposure coming due over the near-term face elevated downgrade and default risk amidst increasingly stringent market selectivity. This maturity skew is creating meaningful differentiation across platforms. For example, approximately 50% of the software exposure owned by several of the CLO managers in Exhibit 3 will mature in fewer than three years. Conversely, other managers face far fewer software-related maturities over the same period. 

Exhibit 3

Software Exposure vs. Near Term Maturities 

European Exposure

Software exposure in European CLOs is far lower than in the U.S. However, the broad market selloff in software-related issuers has still weighed on European mezzanine tranches while underscoring the dispersion across European CLO managers.

Exhibit 4 focuses on differing dynamics across European managers. One CLO manager (“A”, highlighted in the shaded box below) experienced larger price moves within its software exposure, despite less concentrated software exposure. In contrast, two others (“B”, highlighted in the shaded box below) saw both higher portfolio‑level price movements and larger software‑driven moves relative to the broader market, consistent with more concentrated exposure driving overall portfolio performance.

Exhibit 4

Declines in Software Loan Prices are Increasing Manager Tiering

Exhibit 5 again highlights the top 10 European managers most exposed to recent software‑driven volatility.

Exhibit 5

Software Sector Breakdown by Rating Bucket

While several managers maintain more concentrated software positions relative to the broader European market—where average exposure is approximately 7%—software loans in Europe generally benefit from longer‑dated maturity profiles, providing additional runway before refinancing pressures emerge (Exhibit 6). That said, European CLO portfolios tend to be more concentrated than their U.S. counterparts. As a result, despite lower headline software exposure, individual issuer risk is higher, particularly in scenarios where idiosyncratic pressure leads to lower recoveries.

Exhibit 6

European Software Exposure vs. Near Term Maturities 

The repricing of software loans represents another key inflection point for the CLO market across both the U.S. and Europe. AI‑related disruption has highlighted meaningful differences in portfolio construction, risk management, and manager underwriting discipline. The impact has been concentrated in equity and mezzanine tranches to date, and we expect senior tranches to remain well insulated. Ultimately, this environment is likely to reward CLO investors that have the ability to underwrite underlying loan portfolios at the issuer level, while penalizing those relying primarily on headline levels of sector exposure.