There is a possibility of private loan crash
Despite investors’ fears, private credit is far from recession as not all risks are created equal.
The cracks in the private debt market appear to be widening.
Private debt is an important alternative to syndicated bank loans as a source of corporate capital, provided primarily by private equity (PE) firms. The market is heavily involved in financing data center capacity, which is growing alongside demand for artificial intelligence. Investors fear that a surge in artificial intelligence capital spending poses a threat to the software industry and could create a market bubble that leaves private credit funds overexposed.
Yet there are reasons to believe that the potential damage to the private debt market is manageable and contained.
This article is published in the May 2026 issue global finance magazine. .
As First Brands Financing shows, banks as well as PE firms are involved in private debt, either through financing from investment funds sponsored by Ares Capital, Antares, Apollo, Blackstone, Blue Owl and their ilk, or through their own funds. With pension funds, insurance companies and individuals increasingly investing in private debt, law firm Quinn Emanuel warned in a March client memo that the trend could create systemic risk, even though private debt is still a relatively small part of the overall debt market.
The memo’s authors warned, “The result is a transmission chain that runs from technology companies, through private credit originators, to the regulated banks that lend to them, to the insurers and pension funds that invest with them, and potentially to the retirement accounts of ordinary Americans.”
Only a minority of small corporate borrowers are in distress, and companies with EBITDA of $25 million or less experienced a significantly higher default rate than larger companies – 15.8% in 2025. Default rates are higher in healthcare and consumer companies. Fitch also notes that actual losses for first-lien lenders have been limited, with full or high-percentage recoveries in most cases.
In particular, private loan default rates have historically run higher than those for widely syndicated loans, a trend that some observers attribute to more tailored, and sometimes distressed, lending terms. According to Fitch, the surge in January was largely driven by “distressed” exchanges and payment-in-kind (PIK) interest.
AI concerns
There are growing concerns about PE funds being exposed to software. Investors worry that AI software will disrupt the industry, causing the sector to default on portfolios of private-credit loans. But most such funds are diversified, and even those that are not may not be as sensitive to disruption by AI as investors fear. This is because large language models based on AI require application program interfaces to operate, so software may still be needed to facilitate use of the technology.
“Implementing AI still requires significant effort to make it work in a particular environment,” Alan Lin, senior director of technology, North America corporates at Fitch Ratings, told the audience at a recent webinar hosted by the firm.
Of course, much depends on the type of application involved. As Fitch notes, companies making software that is either deeply embedded in enterprise technology systems, leveraging proprietary data, or operating in more regulated industries such as health care and financial services could benefit from AI developments. In contrast, those producing software for applications that are not so embedded, such as digital content creation or certain types of analytics and visualization tools, are more vulnerable to AI disruption.
Even if the AI bubble bursts, that risk is unlikely to go away, Lyle Margolis, senior director in Fitch’s corporate group, where he manages its personal credit business, said in an interview. global finance. “AI is here to stay and is going to be disruptive in certain areas of the software market,” he says.
Yet the risks may be exaggerated. Whether measured by leverage, interest coverage, or EBITDA, “the trends in the software sector have actually been somewhat positive,” he said. Refinancing risks for the sector are relatively benign. And the data-center build-out provides one of several “significant tailwinds” for private credit in the software sector, said Dafina Dunmore, senior director of Fitch’s North American non-bank financial institutions division.
Another mitigating factor: redemption risk, which could see massive outflows of capital. However, this is largely limited to business development companies (BDCs), a more liquid, retail-oriented variety of private-credit investment vehicle. For example, Blue Owl recently blocked redemptions in one of its BDCs and terminated some others. And the $33 billion Cliffwater Corporate Lending Fund, the largest U.S. private-credit gap, received redemption requests at 14%.
Although defaults are increasing for these portfolios, redemption risk is not an issue for most credit funds, as investors remain locked in until maturity. In addition, stress focuses on direct lending: corporate loans that fund working capital and growth.
hidden risks
Of course, given the opaqueness of private credit, many such risks may remain hidden. Blue Owl’s exposure to software loans is one of the highest in the industry, about twice as extensive as its public filings indicate, according to a recent analysis by the Wall Street Journal. The paper also found other PE firms whose credit funds exhibit more than publicly disclosed software exposures, including Blackstone, Ares and Apollo.
Investor concerns could compound Blue Owl’s redemption woes because its data center financing deals involve accounting practices that obscure the risk involved. The main source of concern is probably Blue Owl’s $27.3 billion financing for Meta’s Hyperion data center in Louisiana.
Still, S&P rates the bonds backing the deal, called beignet, as an obligation of Meta, indicating it carries the risk of default. In fact, investors are liking the fact that cash-rich Meta is standing behind the beignets. This bond was recently spread over a bond financing the CoreWeave data center, which is not backed by hyperscalers.
Still, some wonder whether the risks in these issues have been adequately priced.
Quinn Emanuel warned that irregularities in Meta’s accounting treatment could lead to litigation between the parties over who would suffer losses if AI fails to meet expectations and Meta chooses not to renew the lease. Blue Owl similarly finances an Oracle data center, but that bond is trading at a discount to the meta, partly because Oracle does not support it and partly because the end tenant is the less financially stable OpenAI.
“When we rate data centers, to some extent we look at the credit quality of the end tenant,” says Victor Leung, vice president of project finance at ratings firm DBRS Morningstar.
This type of complexity is why Quinn Emanuel warned in his March 13 memo that, “AI data center buildout—estimated to require $5.2 trillion in infrastructure investments by the end of the decade—has given rise to complex financing structures that are creating significant litigation risks.”
Mark Koziel, CEO of the International Association of Certified Professional Accountants and president-CEO of the American Institute of CPAs, says he will raise the issue of existing accounting rules for such financing arrangements at an upcoming meeting with the Financial Accounting Standards Board. Also last month, the US Treasury Department said it would meet with industry and investor representatives to discuss the potential risk of private debt to the financial system.
So far, warnings of a private credit slowdown seem exaggerated.
Credit funds focused on asset-backed finance (ABF), which is based on the value of the borrower’s assets and is the fastest-growing sector in the market, are relatively resistant to stress, thanks to their self-liquidation feature. Unlike direct loans, the principal on asset-backed financing is paid over the lifetime of the loan. As a result, ABF funds do not face the same refinancing risks as direct lenders.
Fitch’s Margolis says sponsors of direct loan funds “do not get the benefit of cash flow directed to repaying the loan.”
Apart from First Brands’ receivables deal with Jefferies, the ABF segment has not yet been fully tested. But a test may soon begin: Beignet is also asset-backed. Or kind of.
The loan principal remains outstanding at each renewal point, so it is not fully self-amortizing. As a result, says DBRS Morningstar’s Leung, “you face the risk that your facility will lose a source of revenue.” Therefore, Meta is guaranteed to compensate investors for any losses incurred if it fails to renew the lease and the residual value of the facility falls below a certain threshold.
This scenario is not far-fetched, Quinn Emanuel warns, noting that converting AI data centers to general-purpose cloud computing or other uses is expensive: “If there is demand for AI computing contracts, these facilities could act as stranded assets with limited alternative uses and low liquidation value.”
