In brief
Artificial intelligence is introducing a structural reassessment of software credit risk across private credit portfolios.
Software has historically been an attractive sector for investors due to its sticky recurring revenues, high margins, strong cash generation and low capital intensity. As a result, software has become one of the largest sector concentrations in private credit portfolios.
However, concerns about the potential for AI to disrupt the software industry have begun to play out across financial markets. Public equity software valuations have de-rated substantially from peak multiples in 2021 and spreads on software loans in the broadly syndicated loan (BSL) and direct lending markets are also adjusting to the new environment.
Managers who can distinguish between software companies that are genuinely vulnerable and those simply impacted by indiscriminate de-rating will be best positioned as the market reprices. The ability to make this distinction requires a robust analytical framework based on a wide range of inputs and expertise.
Over the first quarter of 2026, J.P. Morgan Asset Management conducted an extensive series of meetings with direct lending managers, private equity sponsors, third-party research providers, independent allocators and independent technology analysts.
Our research found several key considerations for estimating the risk from software debt in a portfolio:
- AI disruption transmits to credit impairment via four channels: revenue erosion, margin compression, valuation compression, and refinancing seizure
- Which software exposure is more important than how much
- Traditional fundamental metrics are inadequate for analyzing software credit risks
- Pricing model context is critical to interpreting margins
- Transmission of AI disruption to financial metrics can lead to the “software trap”
- Sponsor quality is increasingly important
- Assess disruption timelines relative to loan maturities
The result of this rigorous, multi-stage analytical process is a scoring framework designed to assess AI disruption risk at the individual loan level. This framework is a living tool to help our investors make decisions grounded in evidence, rather than sentiment, and may be updated as the situation evolves.