Funds are locking the doors at 5.2%. That's the current private credit default rate, and managers are already imposing caps and gates on redemptions. Think about what that means. The stress hasn't arrived yet, and the liquidity tools designed for emergencies are already deployed. If defaults double to 10%, which ING's own research flags as plausible under sustained AI disruption, those gates won't be managing outflows. They'll be trapping capital inside a repricing that nobody modeled.
Goldman Sachs says private credit poses limited systemic risk. JPMorgan agrees. I've heard this song before. In 2006, the consensus on subprime was that losses would be "well contained" because the mortgage market was diversified and individual exposures were small. The problem was never the size of any single position. It was the correlation. When the underlying assumption broke, everything moved together.
The Correlation Nobody Priced
Some private credit managers hold 25 to 30% of their portfolios in software and technology. That's not diversification with a sector tilt. That's a concentrated bet on a single thesis: that recurring SaaS revenue is recession-proof and AI-resistant. February 2026 broke that thesis. Software stocks cratered. The equity story changed overnight, and now the credit story is catching up on a delay.
The reassurance is that these are senior secured loans, so recovery rates protect lenders even in default. Fine. Senior secured recovery rates averaged roughly 70 cents on the dollar historically. On a $100 million loan, that's a $30 million loss. Multiply that across hundreds of software positions in a market that has grown past $1.7 trillion, and the arithmetic stops being comforting.
Publicly traded private credit vehicles have dropped 25 to 40% year-to-date. The market is telling you something. Borrower EBITDA growth looks strong on paper, but EBITDA in private credit is almost always "adjusted" EBITDA, a number the borrower and sponsor negotiate with the lender. I'd like to see the add-backs. I'd like to see how many of those double-digit growth figures survive a quarter where enterprise software budgets get slashed by companies deploying AI agents instead of buying seats.
The Insurance Channel Is the One That Worries Me
US insurers allocate 10 to 25% of their assets to private credit. A mid-sized life insurer with $40 billion in assets could have $10 billion parked in these funds. If redemption gates prevent that insurer from accessing liquidity when policyholders file claims or annuity holders withdraw, the stress doesn't stay inside private credit. It migrates into the insurance system, which touches every retirement account and pension in the country.
The BIS is already investigating ratings on some of these instruments. The ECB is starting checks on banks' private credit holdings. The Bank of England ran system-wide stress tests this month. Regulators don't launch coordinated probes into markets they consider healthy. They do it when the data they're seeing privately doesn't match the story being told publicly.
A 400 basis point funding cost increase for high-beta borrowers means a company paying 8% on its debt is now paying 12%. On $50 million in borrowings, that's an extra $2 million a year in interest. For a profitable, growing company, that's manageable. For the marginal borrowers that got funded during the 2021-2024 boom when private credit was desperate to deploy capital, it's a death sentence on an installment plan.
I'll grant that the structural features of private credit, locked-up capital and low fund-level leverage, make a 2008-style overnight cascade unlikely. But "unlikely to cascade overnight" is not the same as "safe." The 1998 LTCM crisis didn't cascade overnight either. It ground forward for months while everyone insisted the positions were fundamentally sound, until the counterparty web tightened and the Fed had to organize a bailout in a conference room.
The Harvard Kennedy School recommends expanding the regulatory perimeter to cover significant private credit funds. That's the minimum. The SEC should require standardized default and recovery reporting, ban adjusted EBITDA as the sole covenant metric, and mandate quarterly mark-to-market disclosures for any fund with insurance company capital. Opacity is not a feature of private markets. It's the mechanism by which risk gets hidden until it can't be.
At 5.2% defaults, the gates are already up. The question isn't whether private credit is a time bomb. The question is how much of the fuse has already burned.
