Investors requested $13 billion back from private credit funds in the first quarter of 2026. They received less than half. One fund saw 41% of its investors ask for the exit. The managers running these funds say the underlying loans are performing well. Both things are true, and that is exactly the problem.
A quick primer
Private credit is lending done outside of banks. Instead of a company borrowing from JPMorgan, it borrows from a fund run by Blackstone, Apollo, or Blue Owl. These funds pool money from investors, lend it to mid-size companies, and collect interest. The pitch is simple: earn 10-11% yields in a world where Treasuries pay 3-4%.
The industry barely existed fifteen years ago. After the 2008 financial crisis, new regulations restricted how much banks could lend. Private credit filled the gap. It grew from $300 billion a decade ago to $1.8 trillion today, a six-fold expansion that makes it one of the fastest-growing corners of finance.
The redemption wave
Starting in late 2025, investors in private credit funds began asking for their money back at rates the industry had never seen.
Blue Owl Capital's tech-focused fund received redemption requests from investors holding 40.7% of its shares in Q1 2026. Its flagship fund, OCIC, with $36 billion in assets, saw 21.9% of shares submitted for redemption. Both funds capped withdrawals at 5% per their fund documents. Investors in the tech fund got back roughly one-eighth of what they asked for.
Blackstone's BCRED, the largest fund in the space at $82 billion, received 7.9% in redemption requests. Rather than enforce its 5% cap, Blackstone raised it to 7% and had its executives personally contribute cash to cover the remaining gap. It was the third time they had exceeded the cap.
BlackRock, Apollo, Aries, Morgan Stanley, and KKR all faced similar pressure. Most enforced the 5% cap. Blue Owl's parent company stock fell 42% year-to-date.
As Silas Brown, a Bloomberg Television analyst covering private credit, put it: the industry is in "a high-stakes blinking contest. No one wants to be the first to restrict investor withdrawals, and yet everyone agrees with the premise of them."
The product design problem
The core issue is not whether the loans are good. It is how the product was sold.
Private credit funds hold illiquid loans: five-to-seven-year commitments to companies that cannot be easily traded. But in the last few years, the industry packaged these illiquid assets into "semi-liquid" vehicles designed for retail investors and wealth management platforms. These vehicles promised quarterly redemptions, typically capped at 5%.
Goldman Sachs refuses to use the term "semi-liquid." In a shareholder letter, they noted they call their funds "evergreen" because "semi-liquid paints a somewhat misleading portrayal of what these funds are."
The structure worked as long as inflows exceeded outflows. New money coming in funded departing investors. But when sentiment turned, the math broke. The funds hold long-dated loans they cannot sell quickly, and investors want cash the funds do not have on hand.
Armen Panossian, Co-CEO of Oaktree Capital Management, described the dynamic plainly: "Anytime that in a short period you do get meaningful inflows, and cash drag creates a big problem for returns, then you do see excessive risk-taking in any asset class."
What spooked investors
Two forces converged. The first is artificial intelligence.
Roughly 25% of private credit loans went to software companies that private equity acquired between 2018 and 2022. Those acquisitions were made before ChatGPT launched in November 2022. Nobody underwriting those deals anticipated that AI could make legacy SaaS products partially or fully obsolete.
But the reported 25% may understate the real exposure. According to a Wall Street Journal investigation cited by Steve Eisman, the investor who famously shorted subprime mortgages before the 2008 financial crisis, Blackstone's actual software exposure is closer to 33%, Blue Owl's is roughly 21% (versus the reported 11.6%), and Aries is at 30% rather than 23.8%. The discrepancy comes from companies classified under healthcare, business services, or government services that are, in practice, software companies.
Panossian confirmed this pattern: "Some software companies are tagged as business services or government services and are not labeled as software, masking the true AI risk concentration."
The second force is the Iran war. Oil prices surged above $100 per barrel after the Strait of Hormuz was disrupted. Fertilizer prices jumped 43%. The prospect of sustained high energy costs raised recession fears, which would hit leveraged borrowers hardest.
What the defenders say
The managers running these funds make three arguments.
The loans are performing. Jon Gray, President of Blackstone, pointed to BCRED's underlying portfolio: "EBITDA growth was 10%. Debt service coverage went up 25%. Those are not the signs of something that's nearly as problematic as you read about." Michael Zawatsky, Blackstone's Global CIO for Credit, added that portfolio company interest coverage is up 25% and that Blackstone's software loans carry only a 37% loan-to-value ratio, meaning 60% of a company's value would need to be destroyed before the loan is impaired.
This is nothing like 2008. Gray noted that banks were leveraged 30-40x before the financial crisis. Private credit funds operate at less than 1x leverage. Subprime mortgage defaults exceeded 20%. Private credit defaults today are in the low single digits. Zawatsky called the comparison "completely misguided."
The selling is emotional, not fundamental. Gray characterized the redemption wave as a "disjointed environment between what's happening on the ground with underlying portfolios and what's happening in the news cycle."
But each of these arguments has a counterpoint sitting in the same pile of transcripts.
The loans may be performing today, but Eisman points out that the funds themselves set the marks: "Marks on its portfolio are priced by the private credit funds themselves." JPMorgan has already started marking down software company loans held by private credit, while many funds still carry the same loans at par. The divergence between how a bank values a loan and how the fund holding it values the same loan is a gap that will eventually close.
The 2008 comparison may be structurally wrong in terms of leverage ratios, but The Economist's Wall Street editor offered a subtler point: "In my view, it's the assets people think are safe that are the problem, not the assets people think are risky." Private credit was sold as a safe, yield-generating alternative. If it turns out to be riskier than marketed, the reaction will be disproportionate precisely because expectations were so high.
And the claim that selling is purely emotional ignores that Goldman Sachs and JPMorgan have assembled baskets of listed companies to allow hedge fund clients to bet against private credit. When the same banks telling retail investors to stay calm are building tools for institutions to short the sector, the incentive mismatch speaks louder than the reassurance.
What the skeptics say
Eisman put it bluntly on his April 2 podcast: "An entire generation of private credit executives have mistaken a lack of a credit cycle for genius."
His argument: there has not been a significant wave of loan losses in the United States for 17 years. Every private credit executive running a fund today built their track record in an environment where nothing went wrong. They have never managed through actual defaults at scale. That is not genius. It is favorable conditions.
Leonard Tannenbaum, who founded Fifth Street Finance and sold it to Oaktree for roughly $100 million, was even more direct on Bloomberg: "It's bursting. The industry will be here. It's a good industry. It's just not going to be quite this big."
Tannenbaum flagged a detail most sources gloss over: stated leverage is roughly 5x EBITDA, but an S&P report found that real leverage, after accounting for add-backs in how EBITDA is calculated, is closer to 7x. "Covenants are like brakes on a car," he said. "Lately they've been driving at 65 with no brakes."
The Economist took the most measured skeptical position. Their editors argued this will not be a single Lehman-style explosion. It will be "a slow burning" problem that "may crop up for quite some time." The risk is not a sudden crash but a gradual erosion of confidence, where insurance companies and pension funds that hold private credit discover their assets are worth less than reported.
Three things nobody is connecting
The most interesting observations come not from any single source but from reading all of them together.
The generosity trap. When Blackstone honored 100% of redemption requests by having executives write personal checks, they set a precedent. Investors at every other fund now expect the same. But as Bloomberg's Silas Brown noted, "the level of generosity that you're seeing in funds in terms of investor withdrawals will have to subside at some point." The first fund to enforce the 5% cap strictly will face a reputational penalty, even though it is doing exactly what the fund documents allow. Blackstone's short-term fix created a long-term industry problem. Blue Owl, which enforced the cap, may have been more honest about how these products actually work.
The refinancing cliff nobody discusses. Eisman flagged that 11% of software-related private credit loans need refinancing in 2027 and another 20% in 2028. That is nearly a third of the software book rolling over within two years. These loans were originated at lower rates, against higher company valuations, before AI disruption was priced in. Who refinances them? At what terms? If the new rate is materially higher and the company's value is materially lower, the math does not work. Every source discusses current defaults (low single digits) and current EBITDA growth (double digits). Almost nobody discusses what happens when these loans mature. The current performance metrics describe a portfolio that has not yet been tested by its own maturity schedule.
The informed insiders saw this coming. In October 2025, five months before the redemption crisis, Stephen Dulake, Co-Head of Global Fundamental Research at JPMorgan, said on a podcast: "I do worry a little bit about growing retail participation in a deeply illiquid asset class." He flagged the 2020-2021 vintage as a watch item and noted rising PIK ratios. Goldman's own research analyst said in March 2026 that "defaults have been zero. They have literally nowhere to go but up." These are not bears. These are the institutions that profit from private credit acknowledging, in measured language, that the risks are real. When the people selling the product warn about the product, it is worth listening carefully.
What this means if you lend or borrow
Private credit turbulence does not stay contained in private credit.
If funds face sustained outflows, they slow their lending. Companies that relied on private credit for financing, particularly mid-market businesses that cannot easily access public bond markets, may find it harder and more expensive to borrow. Spreads on new private credit loans have already widened 25 to 50 basis points, according to Churchill Asset Management.
For traditional lenders and credit unions, this could mean two things. First, an influx of borrowers who previously went to private credit now showing up at their doors. Second, a reminder that lending discipline matters. The private credit boom was built on the assumption that defaults would stay near zero forever. That assumption is being tested.
For borrowers, particularly small and mid-size businesses, the practical impact is higher borrowing costs and slower approval timelines. The era of plentiful, cheap, no-covenant credit is ending. Lenders who evaluate borrowers on their full financial picture rather than just a credit score will be the ones still writing loans when the cycle turns.
Key takeaways
- Private credit investors requested $13 billion back in Q1 2026 and received less than half. The product design, where illiquid loans are packaged into vehicles promising quarterly liquidity, is the structural fault line.
- Software exposure is likely higher than reported. AI disruption risk was not priced into loans made before November 2022. A third of the software book matures in 2027-2028 with no clear refinancing path.
- The insiders are hedging. Goldman and JPMorgan built short baskets for hedge fund clients while their own podcasts tell retail investors to stay calm. That gap between what institutions do and what they say is the most reliable signal in this story.
- This is probably not 2008. But it does not need to be. A slow, grinding repricing of an $1.8 trillion asset class, where marks catch up to reality over quarters rather than days, can still cause real damage to portfolios, pensions, and the companies that depend on this capital.