Feb 2, 2026 7 min read

Does The BDC Sector Have A Software Credit "Shock" Coming?

Bloomberg has once again published an article warning of big headaches ahead for Private Credit (including BDC) lenders. This time the culprit is said to be AI and the borrowers affected software companies. Will this be the next credit "shock"?

February 2, 2026

On February 1, 2026 Bloomberg columnist Paul J. Davies who covers banking and finance and used to be a reporter for the Wall Street Journal and the Financial Times, wrote an article warning that "Private Credit Is Headed For A Software Shock". He contends - principally on the strength of information provided by a UBS analyst - that "many enterprise software companies could soon be put out to pasture due to the artificial intelligence race" and that "some 25%-35% of private credit portfolios face heightened threats from the AI boom". That last statistic is drawn from the UBS analyst's "analysis of the portfolios of listed BDCs". The article also argues that Private Credit's extensive exposure to software borrowers dates back to the period just after Covid when "many buyouts and the loans behind them were done at what turned out to be peak valuations". That's when the syndicated market was offline and Private Credit stepped into the breach. "The market share taken by direct lenders in funding buyouts jumped to between 40% and 70%, depending on the quarter, between 2022 and 2023" - as per UBS.

INTRODUCTION

This is a very serious concern to raise, given that technology generally and software specifically are very important components of BDC portfolios. Here is what our research in the public record shows:

The percentage of BDC investment assets allocated to software companies varies significantly depending on classification methodology and reporting standards, ranging from 6% to 26% of total portfolio investments. When software is combined with related technology and business services sectors—which often contain software-enabled companies—the aggregate exposure reaches 25-35% of total BDC portfolios, representing one of the most concentrated sector exposures in the private credit market.

Furthermore Mr Davies - and his UBS source - seem to believe recovery rates on loans made to soon-to-fail companies will be disappointing:

For lenders, a particular problem is the potential for very low cash recoveries if a company goes bust. The assets are mostly intangible – skills, licenses and intellectual property – and not as easy to seize and sell as inventory, buildings or other physical equipment.

In this article, we'll look closer at the claims made in the article and offer our own view as to whether the roll-out of AI will have anything like the devastating effect on Private Credit financed software companies that Mr Davies' suggests.

The Origin

The seed of this contention is the fact that "since late October[2025} the software segment of the S&P 500 is down by more than a fifth, while the full index is roughly flat". Furthermore, even the bonds of certain software companies - 3 are mentioned - have "taken a beating". As often happens with Bloomberg articles, the source given is an earlier article from colleagues of Mr Davies on January 28, 2026. The headline says all you need to know: "Software Company Bonds Drop as Investors’ AI Worries Mount".

Signaling

However, Mr Davies is mostly concerned with Private Credit. He contends that the supposedly "lack of regular reporting from private equity-owned, or private credit-funded, firms means the trouble isn’t immediately obvious", but that there are "signs". The most potent "sign" is that a non-traded technology BDC fund of Blue Owl Capital - an asset manager much in the news recently - has faced much higher than usual redemptions of late. The implication is that the withdrawals - apparently led by Asian investors - were motivated by concerns about the credit quality of the BDC's portfolio.

Maybe Not

Much more likely, though, is that investors are reacting to the continuing drop in interest rates, which was largely responsible for the well known drop in public BDC prices in 2025. Or, the investors in the non-traded BDC might be concerned that Blue Owl would take the entity public by merging it with its existing public technology fund - ticker OTF. Blue Owl got into trouble in a similar way last year with its flagship public BDC - ticker OBDC - and its proposed merger with another of its non-traded BDCs. That merger was called off as the latter's shareholders revolted at being folded into the public BDC when it was trading at a discount to book. We very much doubt that the redemptions in the Blue Owl technology non-traded BDC has anything to do with investors awakening to catastrophic software credit risk.

Not Following

The article mentions other private BDCs redemption requests at Ares Capital and Blackstone in 2025, but admits lower interest rates or broader credit worries might be the cause. "But Blue Owl’s technology fund had the largest withdrawals, which is telling". We'd say that's a stretch of a conclusion.

Left Wanting

The other evidence given that we face a software company credit apocalypse (my term, not Bloomberg's) is drawn from the UBS analyst's contention that "bankruptcy filings from technology and business-services firms are increasing". The analyst believes "private credit default rates to rise by about 2 percentage points this year to around 6%". We'd like more details about all of this. We're not especially aware of any special increase in bankruptcies at "technology and business service firms". In our corner of Private Credit - we've only only heard about one bankruptcy filing of a BDC-financed company in the first month of 2026. That was STG Logistics, a logistics company whose failure had nothing to do with technology or AI. Also, how is Private Credit defined and who says the default rate was 4.0% previously?

You Know What They Say About Statistics

These are numbers just tossed off without any sourcing or context. Do the numbers include restructurings? Do they include smaller companies? What proportion are in software - the subject of the article after all? Our experience is that there are no common standards or agreed upon methodologies amongst the many groups tracking such things, so we'd take any attempt at quantification with a bunch of salt.

Our Turn Now

Now, let's discuss what the BDC Credit Reporter knows from the time spent on the $550bn BDC industry and its 5,000- 8,000 borrowers.

Analysts Are Worried/BDC Managers Say They're Not: Multiple times during the public BDCs earnings season, analysts addressed this subject. In all cases, the BDCs involved insisted that there was no cause for concern. Here's one example of the back and forth:

John Douglas Hecht Jefferies LLC, Research Division – MD & Equity Analyst

Another good quarter. Congrats. First one, I mean, I guess these are sort of both kind of industry level kind of questions. Number one is, there's just increasing concern on legacy software companies because if they weren't developed with AI in mind, then they could get disrupted very quickly. I'm wondering your perspectives on that. And if you could talk about your portfolio of that type and how it is kind of positioned for the AI revolution?

Scott Bluestein Hercules Capital, Inc. – CEO, Chief Investment Officer, President & Director

Sure. Thanks for the question, John. One of the more interesting aspects of our business and depending on how you look at this, it's either a positive or a negative is the fact that our duration on the portfolio side is really short. Over the last 21 years, our duration on the loan book has been somewhere between 15 and 24 months. Right now, the duration is right around 18 months. So for us, when we talk about legacy companies, these are not companies that are generally very old in terms of borrowers for Hercules. And when we think about the question that you just asked, we think that's a significant positive because our portfolio is turning every, generally speaking, 1.5 years. We do not have a lot of legacy companies that really haven't been able to benefit from the AI revolution that we've seen over the last year or 2.

We built in AI analysis into our underwriting, the deals that we have booked over the last 12 to 24 months, how these companies are using AI, how these companies can be impacted positively or negatively from AI has been a part of our underwriting thesis. So I think it's certainly something that we're watching closely, and our credit teams are doing a great job at monitoring that. But we feel pretty good about how our portfolio is positioned with respect to what we're seeing across the AI landscape right now.

Share Of Troubled Companies: Overall, we've identified just over 140 Important Underperforming companies in our database. Of those 15 are tagged as being in Software. That number increases to 28 when we include Business Services and Technology. These numbers are proportionate to these sectors share of BDCs overall lending. Of the 31 Important Underperformers in the database where BDC exposure is over $100mn, only 5 are in Software.

Biggest Single Important Underperformer Is In Software, But..: The single largest troubled company is Peraton, with $654mn in aggregate advanced by several BDCs. However, the company's troubles - as this report shows - have nothing to do with AI.

CONCLUSION

We have been disappointed in the past by how Bloomberg - and other venerable - financial publications can yell "Fire" with little evidence except for clippings from their own prior articles. In this case - as in many others - a very serious concern is being expressed by patching together disparate matters and with little supporting evidence. The subject will draw in readers. We certainly were. The matter is on the mind of many in the financial community as reflected in the analysts repeatedly returning to the subject on conference calls. But to suggest - as the article does - that lenders - and their investors - might be in the early innings of a software company credit catastrophe brought on by AI is not supported by the evidence. We are not dismissing the subject out of hand because the impact of AI is only just beginning to be felt in the broader economy. As IVQ 2025 BDC results begin to come out this week - and for weeks to come - we'll be keeping an eagle eye for any AI-related impacts and note them in our database and articles. Maybe Mr Davies is prescient but, more likely from the information at hand, the fears expressed are overblown. The same could be said of the "cockroaches" that Jamie Dimon suggested might follow the First Brands fiasco or the "cracks" in Private Credit that have failed to materialize six months later into the sort of credit crack-up implied.