By Telis Demos

With fears erupting that software companies will be disrupted by artificial intelligence, investors in private asset managers are shooting first and asking questions later.

Private investment firms have piled into the software industry in recent years. At the end of last year, almost 9% of private-equity backed companies were in the software space, according to PitchBook tracking. The exposure is even more significant on the loan side: Within the private-credit universe tracked by ratings firm KBRA, the firm classifies about 17% of borrowers as software companies, representing about 22% of the $1 trillion-plus debt exposure in that universe overall.

Private investors prize the recurring subscription revenue of companies that sell software over the cloud, in part because it lends itself naturally to leverage. Debt investors need more long-term clarity than equity investors, because they want to be sure they will be paid back in future years. Loan funds are hungry for that paper to match their fundraising surge of recent years. And more leverage in turn enhances the return for equity investors.

Nowadays, however, the core assumption of predictable revenue has come under fire. If these companies are going to be displaced by cheaper, better, AI-powered tools, then in theory customers might have enough incentive to go through the headache of canceling subscriptions, or at least demand a fee cut. So earlier this week, Anthropic's announcement of an AI tool for legal services was like a match tossed into a dry forest.

Compounding that is the inherent limited transparency for companies owned or funded by the private markets, which is currently lending itself to panic among investors.

This week, earnings reports by private-asset companies, including managers Blue Owl Capital and KKR, and business-development companies Ares Capital, Golub Capital BDC and Oaktree Specialty Lending have been opportunities to explain why their software portfolios are insulated -- or are even set to benefit -- from AI disruption.

Executives made many points: That they have been monitoring AI risk for some time. That good software companies can be embedded so closely with customers' business processes that they can't be easily ripped out. That some software companies use proprietary data that AI can't replicate. That software companies that can operationalize AI tools will end up even more valuable for their customers. That equity values are still up relative to when the loans were underwritten several years ago.

Broadly speaking, the software companies in private-loan portfolios also enter this period of uncertainty in solid shape. The median annualized growth in earnings before interest, taxes, depreciation and amortization among software borrowers in the private-credit universe tracked by KBRA has been 36% over roughly the past two years.

But shareholders in listed lending vehicles aren't yet reassured. The VanEck BDC Income ETF, which tracks business-development companies, on Thursday erased a Wednesday bounceback and is down more than 5% on the week. The largest alternative-asset managers also resumed their sharp declines. Blackstone, KKR and Apollo Global Management were all down more than 5% on Thursday.

It will be difficult for investors to get much further concrete, numerical assurance in the near term. For one thing, the public credit markets aren't helping as a benchmark. The volume of leveraged loans to software companies considered distressed doubled in January, according to tracking of that market by PitchBook LCD. And the recent examples of private loans held near cost that were marked down dramatically shortly thereafter might have investors in a doubtful frame of mind.

To be sure, there is a big difference between owning the equity of a company and being a lender to it. A company that has had its growth prospects cleaved off could see its equity valuation plummet, but it could also still be collecting steady checks from existing customers, making it a safer credit. In theory an enterprise software company with subscription revenue fits exactly into that model.

But the current mood of investors seems to be to presume that software credits could be impaired until their performance can demonstrate otherwise.

What is more, some private managers have in recent years sought to broaden their lending to software companies that haven't turned a profit yet. So-called recurring revenue loans are based on future revenue projection rather than earnings. At some point these aim to convert to loans based on profit.

"Having more competition, their product being more vulnerable, could affect that trajectory of when the company becomes profitable," says Alen Lin, Fitch's senior director covering North American technology credit research.

Of course, the managers and funds whose portfolios prove out through the cycle will likely end up with bigger shares of inflows and assets in the future. But some investors right now don't seem inclined to stick around and find out who that will be.

Write to Telis Demos at Telis.Demos@wsj.com

(END) Dow Jones Newswires

February 06, 2026 05:30 ET (10:30 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.