Everything About AI
NextFin News - Tony Yoseloff, chief investment officer of Davidson Kempner Capital Management, warned on Wednesday that private credit is facing bigger problems than public markets, as higher interest rates, slowing business growth and a wall of debt coming due in 2027 and 2028 converge on the roughly $2.3 trillion asset class. Speaking at the 2026 Global Financial Leaders' Investment Summit in Hong Kong, Yoseloff joined a growing group of hedge-fund managers treating private credit not as the steady-income substitute it was marketed as, but as the corner of the credit complex most exposed to a rate cycle that never really ended.
The warning carries weight because of what private credit has become. Assets under management reached about $2.31 trillion in 2026 and are projected to expand toward $3.36 trillion by 2031, according to a market research report published in October. A decade ago the strategy was a niche for institutional specialists; today it is a core holding for insurers, pension funds and wealth platforms that bought it for bond-like income with equity-like protection and mark-to-model valuations that promised to spare investors the daily volatility of public bonds. That promise is now being tested by the same three forces Yoseloff named - rates, growth and maturities - at the same time.
The evidence of stress is already visible in individual credits. First Brands Group, an auto-parts maker funded heavily by private credit, filed for bankruptcy in September 2025 with approximately $10 billion in debt. In the same month, subprime auto lender Tricolor Holdings filed to liquidate after a bank alleged fraud. This year, BlackRock's TCP Capital wrote down the value of a private loan to Infinite Commerce Holdings, an aggregator of online Amazon sellers, to zero from full value. Each case is idiosyncratic, but together they trace the same fault line: borrowers that looked sound when money was free are breaking when it is not.
The Maturity Wall Meets a Rate Cycle That Never Reset
The pressure point is timing. A wave of debt maturing from 2028 is forcing borrowers to refinance at rates far above those available when the loans were originated during the cheap-money years of 2020 and 2021. Fund managers are racing to negotiate extensions on much of an estimated $300 billion wall of debt coming due for private-equity-backed companies, and a large share of that exposure sits in software borrowers whose cash flows now face a second threat: displacement by artificial intelligence.
Moody's Ratings framed the same problem in April, warning that business development companies with outsized exposure to software and tech loans face rising refinancing and credit risks as the maturity wave builds, and describing leveraged software borrowers as "a particular challenge" amid inflation concerns and AI disruption. The transmission mechanism is unforgiving. The Federal Reserve lifted its policy rate by 525 basis points from March 2022 to a range of 5.25% to 5.50%, and private credit loans float with that benchmark. A software company that borrowed when rates were near zero now pays a much larger share of its revenue in interest - just as AI competitors erode the subscription revenue that justified the loan in the first place. With an asset-light balance sheet and few hard assets to liquidate, the lender's recovery rate in a default is structurally weak.
This is where the cyclical and the structural separate, and confusing the two is the most common mistake investors make here. The rate cycle is cyclical: rates rise, then eventually fall, and borrowers that survive the refinancing window can deleverage. But three features of private credit make this cycle structurally more dangerous than the last one. First, a meaningful share of the book was underwritten in a near-zero-rate world with looser covenants, so the credit quality of the vintage is weaker than public high-yield debt issued in the same period. Second, valuations are mark-to-model rather than mark-to-market, which delays loss recognition and encourages borrowers and lenders alike to extend and pretend. Third, the investor base has changed: insurers, pension funds and retail wealth channels now hold private credit as a low-volatility bond proxy, and they did not underwrite a drawn-out high-rate regime.
Default Rates Are Rising - But the Reported Number Understates the Stress
The official statistics already show deterioration, and the arguments over methodology are themselves a warning sign. Fitch Ratings put the U.S. private credit default rate at a record 6.3% in August 2026. Moody's reported a 4% default rate among 607 private credit borrowers in the 12 months to June - but that figure rose to 5.8% when payment-in-kind amendments and maturity extensions were counted, because the rating agency treats those tools as distressed exchanges equivalent to defaults. Proskauer's Private Credit Default Index, tracking 697 loans totaling $189.2 billion, recorded a 2.73% default rate in the first quarter of 2026, up from 1.84% in the third quarter of 2025. For context, the Morningstar LSTA U.S. Leveraged Loan Index recorded a payment default rate by amount of 1.23% in 2025, up from 0.91% in 2024.
The gap between 4% and 5.8% is the story in one number. Borrowers are using PIK - paying interest with additional debt rather than cash - and maturity extensions to avoid being counted as defaulted. That keeps the headline default rate artificially low while the underlying cash burden compounds. For a floating-rate lender, PIK interest is income on paper even as the borrower's ability to pay cash deteriorates. The practice is a classic marker of quiet distress: the loan stays "current" while the collateral value erodes, and the fund's reported yield stays intact while the borrower's debt load grows.
Yoseloff's own forecast sits in the middle of this range. In an interview published in October, he said he expects the true default rate for public and private loans to settle at around 5% to 6% over the long term, with direct corporate lending producing returns in the mid-single digits. That is a sharp departure from the environment that built the industry. The cycle he says today's market most resembles is the dot-com bust: from 2000 to 2002, high-yield bond defaults soared from 5% to 12.8% in the wake of the technology bubble and the telecom collapse.
The AI Mask - And Why It Could Slip
The most counter-intuitive part of Yoseloff's argument is that the same force driving this market cycle has also been hiding its damage.
We could have gone into a big recession in 2022, and we didn't. I think in large part that's because of AI. There's been so much AI capex spending in this country over the last few years that's bailed out the rate hikes. But you can imagine an environment where spending slows and rates continue to be high, and that's very analogous to what happened in the 2000s.
That is a second-order point most investors are not pricing. The conventional wisdom is that private credit is resilient because its loans float with rates: higher benchmark rates mean higher coupons, which cushion lenders. That holds only as long as the borrower's revenue keeps growing. If AI capex slows while rates stay high, the cushion disappears on both sides at once - the borrower's interest burden stays elevated while the growth that justified the leverage fades. The 2000-2002 comparison is the template: a capex boom masked over-leverage, then both reversed together.
The exposure is concentrated exactly where the risk is highest. Software and technology borrowers make up a large share of BDC and direct-lending portfolios, and they are the companies most vulnerable to AI-driven disruption of recurring revenue. When a software company's subscription cash flow proves less durable than the underwriting model assumed, the loan is effectively unsecured - there is little to liquidate. The same logic applies beyond software: any borrower whose earnings were modeled on a permanent low-rate world faces the same arithmetic shock when the rate regime changes and does not quickly revert.
The Counter-Thesis: Industry Managers Say Stress Is Contained
The strongest case against the alarmist read is that the industry has been here before and the losses are manageable. Fund managers at some of the largest private credit firms have argued publicly that floating-rate structures protect lenders, that underwriting has tightened since 2022, and that the default rates now being recorded remain below the double-digit peaks of previous credit cycles. The argument has real force: a 6.3% default rate with a 50% to 70% recovery rate on asset-backed loans produces mid-single-digit net losses, which the higher coupons in the book can absorb. The sector was priced for a higher-rate environment, and many lenders have already pulled back from the riskiest vintages.
There is also a selection argument that deserves to be stated plainly. The managers warning loudest are distressed-debt specialists whose business model depends on dislocation. Davidson Kempner built a roughly $40 billion franchise by finding opportunities in crisis and keeping its distance from the direct-corporate-lending boom. Yoseloff has been loading up on asset-backed loans - "loan pools, it can be real estate, it can be ships, it can be aircraft" - precisely because "there's tremendous competition in what's called direct corporate lending. There's much less competition in asset-backed lending." The bearish diagnosis and the bullish positioning are the same trade: avoid the crowded, opaque book; buy the collateralized one.
The counter-thesis breaks, however, on two observable facts. First, the PIK and extension activity that Moody's flagged means the reported default rate understates realized stress - the losses are being deferred, not avoided. Second, the investor base has changed in a way that turns a credit problem into a liquidity problem. Redemption requests have already risen across the sector. In March, Blackstone allowed investors to pull a record 7.9% of shares - roughly $3.8 billion - from its flagship BCRED fund, meeting the requests by expanding a tender offer to 7% of the fund's shares and having the firm and employees absorb the remaining 0.9%. Earlier in the year, at least one large manager paused quarterly redemptions after it could not sell assets fast enough. A widening of redemption gates would force a repricing that mark-to-model accounting has so far prevented.
What to Watch - And the Signal That Would Prove This Wrong
The forward path splits by time horizon, and the same event points in different directions across them. In the short term, liquidity is the risk: any widening of redemption gates or tender-offer queues would be the first visible crack. In the medium term, the 2027-2028 refinancing window is the test - how many borrowers can roll debt without a PIK amendment, a covenant waiver or a maturity extension. In the long term, the question is structural: whether private credit's opacity and mark-to-model pricing survive as regulation tightens. The Federal Reserve launched its first formal monitoring survey of the private credit sector in 2026, reflecting growing supervisory attention toward non-bank lending.
The base case is a contained but visible deterioration: default rates grind toward the 5% to 6% range Yoseloff forecasts, concentrated in software and subprime consumer exposures, with losses absorbed by coupons but returns compressing to the mid-single digits. The upside case is that AI capex stays strong, growth outpaces interest costs, and the maturity wall is refinanced without a spike in defaults - in which case today's warnings look like a distressed investor talking his book. The downside case is that capex slows while rates stay high, PIK usage expands further, and the reported 6.3% default rate climbs toward the 12.8% peak of the early 2000s.
The single falsifying signal is specific: if Fitch's private credit default rate stays below 7% through the end of 2027 while PIK amendments and maturity extensions decline quarter over quarter, the structural-stress thesis is wrong and this is a cyclical wobble that coupons absorbed. If instead the default rate breaks above 7% while PIK use keeps rising, the losses are being masked and the repricing has barely begun.
The real risk in private credit is not that defaults rise - they always do after a credit boom. It is that the industry's accounting lets it pretend they have not, right up until the maturity wall makes pretending impossible.
Explore more exclusive insights at nextfin.ai.
