When GLP-1 weight-loss drugs started showing real efficacy, the market sold off snack-food companies as if the world were going to suddenly stop binging on cookies. The long-term implications may prove real. But big disruptions rarely happen overnight, and investors who price it that way tend to be wrong. The same mistake is now happening in private credit.
Earlier this year, a UBS research report warned of private credit’s vulnerability to software exposure.[1] The report estimated that roughly 25% to 35% of private credit portfolios faced elevated artificial intelligence (AI) disruption risk, with technology accounting for approximately 24% of holdings at business development companies, the publicly traded vehicles in which a significant portion of private credit is held. The press picked it up and the momentum crowd sold.
It is a clean story. It is also mostly wrong.
When a private credit fund makes a loan to a software company, it is not making a bet that the company will exist forever. It is making a bet that the company will exist long enough to pay back the loan. Those are very different bets, and right now the market is confusing them.
The contractual maturity on new loan averages are around five years. Given the frequency of refinancings and early payoffs, the actual hold period is typically closer to three to four years. That’s the clock the market is ignoring. The question is not whether AI will eventually disrupt mid-market software. It will. The question is whether it will do so completely and catastrophically before those loans mature. Private credit exposure would be a relatively minor concern in a disruption scenario with that speed and scale.
Technology does not kill companies overnight. It kills them slowly, then all at once, but the “slowly” part matters enormously if you are a lender rather than an equity holder. A software company losing customers to AI does not go dark on a Tuesday. Revenue erosion unfolds over years. That is painful for an equity investor. For a lender, it can still mean years of visible contractual cash flow sitting in front of the maturity date. The loan gets paid. The equity gets wiped. Those are two different outcomes that the market is currently treating as one.
Then there is the sponsor. Private equity firms do not abandon portfolio companies gracefully. They have fund economics, investor relationships, and hard-won reputations that make fighting for these assets a rational move. The private equity industry is sitting on roughly $1.7 trillion in dry powder globally. When a software company starts showing stress, a private equity fund may cut costs or inject capital if necessary. They may find a strategic buyer who wants customer relationships even if the product is weakening. The loan gets taken out in a sale process, not a bankruptcy. The equity cushion that everyone is ignoring is doing real work here.
Here is what the market is missing. Private credit managers have a solution available to them that is so obvious and so boring it apparently has not occurred to anyone selling the stocks. If a manager genuinely believed AI was going to devastate their software book, they could simply stop making new software loans and let the existing portfolio run off. Three to four years from now, the exposure is largely gone. No fire sale. No systemic crisis. Just time doing what time does to short-duration assets.
Credit is a duration business. Refinancing risk emerges at maturity, not immediately. Markets are pricing a terminal outcome into short-duration instruments. Even accepting elevated default assumptions, the loss severity implied by current share price moves would require recovery rates far below historical experience for senior secured lending, which has averaged 60 to 80 cents on the dollar across multiple credit cycles, including the financial crisis. To justify where some of these stocks are trading, you would need to believe that software companies will go bankrupt en masse, that sponsors will not defend them, and that lenders will recover pennies on senior secured paper. That is not a stress scenario. That is a simultaneous failure of every structural protection in the capital stack.
And when credits do go sideways, they do not go to zero. They go to the restructuring group. A restructuring professional does not see a dead company. They see a cash flow management problem. You cut costs aggressively, run out the revenue stream, and use every dollar of incoming cash to pay down the loan before the terminal decline arrives. The market is pricing outcomes that would require these companies to simply vanish, debts unpaid, into the AI-disrupted void. That is not restructuring. That is magic.
Sure, bad private credit managers exist. They always have. They do bad deals in good times and bad deals in bad times. When the cycle turns, they produce carnage right on schedule. Some of that carnage is coming. AI will accelerate the timeline for the weakest credits at the weakest managers, and there will be defaults that were entirely avoidable. Interest coverage ratios across private credit borrowers have fallen materially since 2021 as higher rates have worked through capital structures. But that is an argument for manager selection, not an argument for indiscriminate selling across the asset class.
The losses, when they come, will be more contained than the current selloff implies, concentrated in the funds that were already making mistakes before AI gave them a new way to make them.
The private credit lenders who made these software loans are not sitting in a dark room waiting for the apocalypse. They are sitting across the table from sponsors who have every incentive to fix these companies quietly, watching performance evolve, and managing against a maturity schedule that gives them considerably more runway than the market seems to understand.
AI is going to change a lot of things. The math on private credit loans probably is not one of them.
References
[1] https://finance.yahoo.com/news/private-credit-exposure-ai-disruption-195654032.html
© Copyright 2026. The views expressed herein are those of the author(s) and not necessarily the views of Ankura Consulting Group, LLC., its management, its subsidiaries, its affiliates, or its other professionals. Ankura is not a law firm and cannot provide legal advice.
