Risk markets have remained resilient in the face of heightened geopolitical tensions this year, although one notable area of weakness has been in the technology sector.
Uncertainty around artificial intelligence (AI) is nothing new, and thus far it has mainly been focused on the pace and scale of capital expenditure needed from the hyperscalers, how this would pressure balance sheets, and what the execution risk of those investments would be given a largely unknown return on investment.
In recent weeks, however, and following the release of Anthropic's legal plug-in for its Claude Cowork AI tool, the market is re-underwriting the AI threat to a large number of software-as-a-service (SaaS) businesses, and the sell-off has seen equity multiples in this sub-sector compress significantly (see Exhibit 1). Credit, while in aggregate remains resilient, has certainly not been immune, with many software bonds and loans down a number of points this year.
Stepping back, one can see why the market has been happy to underwrite software credits in recent years. Solid non-cyclical organic revenue growth, driven by high retention rates and steady increases in average revenue per user (ARPU), have coupled with high Earnings Before Interest, Taxes, Depreciation, and Amortisation (EBITDA) margins and robust free cashflow. This fundamental profile has allowed issuers, particularly those that are sponsor-led, to layer leverage onto structures, especially for businesses that typically trade at elevated equity valuations. Net debt-to-EBITDA of 5-6x is not uncommon, with the market underwriting these credits on a loan-to-value (LTV) of less than 50% given average Enterprise Value (EV)/EBITDA multiples in the low-to-mid teens (as an example, net debt-to-EBITDA of 5x on a business valued at 13x EBITDA equates to an LTV of 38%).
However, the recent moves call into question whether the level of growth going forward is sustainable for many of these businesses, and there will likely be more scrutiny on LTVs after the sharp equity sell-off.
From an operating perspective we would be particularly nervous around sub-scale, single-feature SaaS companies. Firms with narrow moats and low switching costs will be most impacted by AI disruption, as customers look to both consolidate platforms or even build in-house. Generic or repeatable process driven SaaS companies already see the threat of Claude et al. on their shoulder, and the complexity of the models developed by the frontier AI labs (Anthropic, Open AI etc.) are evolving quickly.
Having said that, many software companies could benefit from those developments. Platforms and databases that are deeply integrated in workflow, offer proprietary data (data that is not easily scraped online), niche sector expertise, and comply with local regulation, could be well positioned to take advantage of AI advancements. Integrating AI tools into existing platforms will be critical, but companies that do this well would expect to increase the stickiness of their customer relationships and help protect ARPU.
So, we don't think we are at the start of an AI "cliff" within sub-investment grade credit. The risks are real, but so is the opportunity, both from a corporate perspective (if they can leverage AI through the business) and an investment perspective.
If we consider earnings, the most likely impact we see, at least in the credit space, is less on business model obsolescence and more on the ARPU side. Unless a company can demonstrate its unique selling point and prove its moat, either via proprietary data or sector expertise, then even platforms that are imbedded in workflow might struggle to offer costs that are many multiples higher than the frontier lab equivalent. Positively, costs are in any case relatively low as a percentage of operating expense for most of the corporate customers they serve. We believe that AI will also benefit the cost bases of the very SaaS companies they threaten on the revenue side, particularly in development and coding. However, we expect investors to stress their "base case” assumptions in the upcoming earnings round if they haven't done so already, and we expect the market to become more discerning between the potential winners and losers.
Across sub-investment grade markets, sector exposure is most concentrated in the private credit space (US Business Development Companies have an average exposure of 26%, according to Morgan Stanley data) and least concentrated in high yield bonds (only 3% of the European and US HY market is in the software sector), with the broadly syndicated loan market in between (US 10%, Europe 7.5%). Outside of private markets, therefore, we think exposures are manageable, and it is worth saying that software covers a wide remit, from consulting to outsourcing, productivity, risk analysis, operating system, multimedia, design, and so on. Some sub-sectors will of course be more exposed than others.
Ultimately, the rapid evolution of AI will clearly throw up both risks and opportunities for credits in the bond and loan markets. European credit is less exposed to the sector, though not immune, and we view the risks as manageable. Undoubtedly, volatility in the space is likely to continue as the technology continues to leap forward.
Important Information
The views expressed represent the opinions of TwentyFour as at 13 February 2026, they may change, and may also not be shared by other members of the Vontobel Group. The analysis is based on publicly available information as of the date above and is for informational purposes only and should not be construed as investment advice or a personal recommendation. References to holdings and/or other companies for illustrative purposes only and should not be considered a recommendation to buy, hold, or sell any security discussed herein.
Any projections, forecasts or estimates contained herein are based on a variety of estimates and assumptions. Market expectations and forward-looking statements are opinion, they are not guaranteed and are subject to change. There can be no assurance that estimates or assumptions regarding future financial performance of countries, markets and/or investments will prove accurate, and actual results may differ materially. The inclusion of projections or forecasts should not be regarded as an indication that TwentyFour or the Vontobel Group considers the projections or forecasts to be reliable predictors of future events, and they should not be relied upon as such. We reserve the right to make changes and corrections to the information and opinions expressed herein at any time, without notice.
Past performance is not a guarantee of future results. Investing involves risk, including possible loss of principal. Value and income received are not guaranteed and one may get back less than originally invested. TwentyFour, Vontobel Group, its affiliates and the individuals associated therewith may (in various capacities) have positions or deal in securities (or related derivatives) identical or similar to those described herein.
