Your Private Credit Fund May Be Worth Less Than It Says

The Money You Cannot Easily Price

Private credit, meaning lending done by investment funds rather than banks, has grown to between $1.5 trillion and $2 trillion in assets, according to the Financial Stability Board’s May 6 report on the sector. For years this was the preserve of pensions and insurers. Now it is sold to the rest of us, through business development companies that trade like ordinary shares and through “interval funds” that take your money regularly but let you withdraw it only at set windows. The pitch is a bank-beating yield without the daily lurch of the stock market. The trouble sits in how these funds work out what they own.

A bank can mark a traded bond against whatever the market will pay that morning. Private credit cannot, because its loans almost never change hands. So the funds estimate. They mark to model, in the trade’s own phrase, relying on their own read of a borrower’s health rather than a price agreed between a buyer and a seller. The FSB does not dress this up. Valuations, it writes, are “often conducted less frequently and may involve significant discretion, which can amplify unceAI software lending private creditrtainty during times of stress.” Put plainly, the body that watches over the global financial system, chaired by the governor of the Bank of England, is saying the prices these funds report rest partly on judgment, and that the judgment wobbles precisely when it is tested.

How AI Walked Into The Loan Book

A good chunk of private credit went into software. The Bank for International Settlements, the central bankers’ bank in Basel, set out the numbers in its March Quarterly Review. Lending from private credit funds to software-as-a-service firms, the subscription businesses behind much of the modern office, climbed from under $8 billion in 2015 to more than $500 billion by the end of 2025, about 19% of all direct lending. A third of these funds have lent to the sector.

The bet looked shrewd. Subscription software throws off steady, predictable cash, exactly what a lender likes to see. Then the picture changed. Investors began to worry that AI tools could hollow out those very businesses, letting customers build in-house what they used to rent by the month. If a mid-sized company can prompt a model to produce the tool it once licensed, the subscription lapses, and the borrower’s revenue, the thing the loan was lent against, starts to look less solid than the spreadsheet assumed. Between October 2025 and February 2026, software shares fell by almost 30%, the BIS found. Shares in the listed funds that lent to them dropped about 10%. And the gap between the loans’ stated worth and what the market would pay for the funds themselves widened.

The scale of the exposure is easy to underplay. The BIS notes that BDCs, the listed slice of this world, account for roughly a fifth of all direct lending in the United States, and that they extended more than 15% of their loans to software firms in 2025. These are not fringe vehicles holding an odd software loan or two. They are mainstream income funds, sitting in retirement accounts and brokerage portfolios, with a meaningful share of their lending pointed at the one sector the market has decided AI may rewrite. When the BIS looked closer, it found that the BDCs most exposed to software had lagged their peers by about 5 percentage points since October 2025. The funds carrying the most of this risk, in other words, were already the ones doing worst.

The Market Is Marking Its Own Doubt

The fund’s reported value, its net asset value, leans heavily on the book value of those hard-to-sell loans. If the borrowers are worth less than the model claims, the figure flatters whoever holds it. The market has its suspicions. As of late April, the average listed business development company traded at roughly 85 cents on the dollar of its stated NAV, a discount of about 15%, with the median nearer 80 cents, according to figures compiled by the data firm With Intelligence. Those are vendor calculations rather than an official tally, but the direction is hard to miss: investors are quietly betting the stated numbers are too high. They have been wrong before, and a frightened market often is. Still, a discount that wide, holding for months rather than days, is not nothing.

BlackRock TCP shows the abstract turning concrete. The firm’s own filing records the 19% drop, concentrated in a handful of holdings, with about 91% of the damage tied to loans underwritten in 2021 or earlier, businesses that thrived on pandemic-era demand and then faded. None of that is alleged or whispered; it sits in the company’s published results. The class action, which the firm has not yet had the chance to answer in court, argues investors were not warned soon enough. Whatever a judge eventually decides, the episode is a tidy lesson in the mechanics the FSB worries about: marks that hold firm on paper, then move all at once when reality forces the issue. A 19% write-down is not the sound of a portfolio gently drifting. It is the sound of the gap between the model and the world growing too large to ignore.

Reasons Not To Panic, And One Not To Relax

It would be easy, and wrong, to read this as a crash in waiting. Defaults remain low. Borrowers are mostly paying. The FSB is careful to say private credit “remains untested to a prolonged economic downturn and so warrants close attention,” rather than forecasting trouble. PIMCO, the bond house, argued in March that the strain belongs to one corner, direct lending to mid-market firms, rather than the whole asset class, and that plenty of private credit still earns its keep; it called the BDC discounts “a direct lending problem, in our view, not an indictment of private credit as a whole.” That is a fair point, and worth keeping in view. The awkward part is that the corner under strain is the very one sold hardest to ordinary investors chasing yield, often through the interval funds and listed vehicles now trading below their own stated worth.

There is a second worry buried in the FSB report, quieter but arguably sharper. The watchdog wants firmer oversight, then concedes it cannot get the numbers to provide it. “Data challenges currently hinder effective monitoring,” the report states; there is a “lack of harmonised private credit definitions,” and authorities work with only “limited granular fund- and loan-level data.” The regulator raising the alarm is, in the same breath, admitting it is partly in the dark. So much for the comforting assumption that someone official is keeping a careful tally of the loans inside your fund. The FSB has proposed a common set of metrics to close the gap. Proposing a fix and having one are not the same thing, as anyone who has waited on a builder’s promise can tell you.

What A Sensible Holder Does Now

The useful move is to treat the headline yield as the start of the homework, not the end of it. Ask how the fund values its loans, and how often, because quarterly marks checked by an outside valuation firm are sturdier than a manager’s in-house guess. Ask how much of the portfolio sits in software, technology and other corners exposed to AI, since the funds disclose their largest holdings and a glance tells you whether you are lending to businesses the technology might undercut. And if the fund is listed, look at its price against NAV, because a steep, stubborn discount is the market telling you it doubts the marks, and that is worth more than any brochure.

There is an old line in the City about marking your own homework, and that, stripped of jargon, is what a private credit fund does every quarter. A price you cannot test against a willing buyer is a belief dressed as a fact. For the better part of a decade, the belief held, and the loans were worth what the spreadsheet said. AI has poked a hole in that comfort, not by sinking the loans outright, but by making the businesses behind them suddenly hard to value with a straight face. The funds may yet prove right about their numbers. Millions of savers are now wagering, many without quite knowing it, that they are. The case against BlackRock TCP is the first clear sign that some of those savers have begun to ask, in the one place that forces an answer, whether the number on the statement was ever the number at all.