I have sat in enough investment committee meetings to know that the word illiquid is treated as a modifier long before anyone treats it as a warning. Private credit funds are sold on a specific promise. Investors give up daily liquidity, and the manager compensates them for it through an illiquidity premium, typically 200 to 400 basis points above what a comparable public credit instrument would pay. 

That premium was never a gift. It was a wager that the manager could match the fund's liability structure to the true duration of its assets. That wager was tested on a real scale in the first half of 2026, and I have reviewed enough family office allocation memos to know how easily this distinction gets lost between the marketing deck and the subscription document.

Blackstone's flagship vehicle, BCRED, received redemption requests equal to 7.9 per cent of shares in the first quarter, roughly $3.8 billion, against a standard quarterly repurchase cap of 5 per cent. Blackstone raised the cap to 7 per cent and covered the remaining shortfall with capital from the firm and its own executives, honoring every request in full. The market read that as strength. Then the second quarter arrived. Redemption requests climbed again, to roughly 10 per cent of shares, and Blackstone held the standard 5 per cent cap rather than raising it a second time. Investors who wanted out did not get out. That alone is not a crisis. It is a fund functioning exactly as its prospectus describes.

BCRED was not an outlier. It was the opening data point in a pattern that widened across the industry through the second quarter. Blue Owl's two flagship funds drew combined redemption requests of $4.7 billion, one equal to 38 per cent of shares outstanding. Apollo's Debt Solutions BDC capped a 16.8 per cent request at its standard 5 per cent. Cliffwater saw demand rise to 17 per cent and lowered its own cap from 7 per cent back to 5 per cent, tightening the door precisely as more investors tried to walk through it.

The trigger had a specific cause. Morgan Stanley's credit team projected direct lending default rates could climb to 8 per cent, more than triple the historical average, concentrated in the roughly 26 per cent of exposure sitting in software borrowers threatened by agentic AI. Apollo's book carries more than 12 per cent in software alone. Man Group's mid-year credit outlook put the industry default rate at 5.8 per cent, with bad payment-in-kind interest, borrowers paying interest with more debt instead of cash, rising to 6.4 per cent of borrowers from 2.5 per cent at the end of 2021. 

Private loans are not continuously priced the way a bond is, so a reported net asset value can lag deteriorating conditions by a quarter or more, and an investor who suspects a markdown is coming has every rational reason to redeem before it arrives. Once one investor believes others are heading for the exit, requesting a redemption stops being a bet on the credit and becomes a bet on everyone else's behavior.

None of this means a 5 per cent cap is, by itself, evidence of distress. It is a structural feature built into nearly every non-traded fund, so a manager is never forced to dump illiquid loans into a falling market. The distinction that matters is between a cap absorbing ordinary turnover and one absorbing two, three, or nearly eight times its design capacity while default forecasts climb in the same direction. Diversified funds with real financing capacity are built to withstand that pressure. Funds concentrated in software borrowers are the ones a family office should question harder.

I spent years on the lending side of private credit, pricing this precise duration mismatch before it acquired a retail-friendly name. The mechanism has not changed. What has changed is who now holds the paper. A pension fund with a 50-year horizon absorbs a gate without much notice. 

A family office funding a grandchild's tuition cannot always say the same, and neither can individual investors who bought into semi-liquid credit funds trusting that the word semi was doing more work than it actually does.

The macroeconomic backdrop makes the timing worse. The Bank for International Settlements' 2026 Annual Economic Report named an artificial intelligence capital expenditure bust and opaque circular financing among hyperscalers as top threats to global financial stability. S&P Global Ratings separately projects that maturities of B-minus and lower-rated debt, the cohort private credit dominates, will jump from $56.6 billion in 2026 to $215 billion by 2028. That wall guarantees the mismatch keeps being tested as maturities climb.

UBS's 2026 Global Family Office Report shows family offices continuing to raise allocations to infrastructure and power themes downstream of the same AI capital expenditure wave. Diversification across a single macro bet wearing different fund wrappers is concentration with a longer prospectus. The claim that private credit is uncorrelated to public markets holds up beautifully in a rising cycle. It has not been tested through a genuine credit event, and I would rather my clients discover the answer from a case study than from their own account statement.

This is where liquidity bucketing stops being a slide in an investment policy statement and starts being a fire drill. The standard family office framework divides capital by purpose and horizon: a liquidity bucket for near-term operating cash, a lifestyle bucket for ongoing spending, and longer-horizon buckets for the next generation and beyond.

Private credit typically lives inside the lifestyle bucket, funding annual spending that assumes capital returns on schedule. Stress test that assumption honestly. If an allocation cannot return capital for two additional quarters because several managers are gating at once, does the family still meet its spending obligations without selling a liquid asset at an inconvenient moment? For most family offices I have advised, the honest answer before this year was that nobody had run that scenario on paper.

None of this argues for abandoning private credit. The asset class still offers a legitimate premium for genuine duration risk, and managers who underwrote conservatively and stayed diversified away from software will be rewarded when the cycle turns, as cycles always do. 

It argues for treating the illiquidity premium as exactly what it has always been: compensation for a specific, quantifiable risk, and for sizing the liquidity bucket to survive a gate rather than assuming one will never arrive. Two consecutive quarters, multiple managers, and a default forecast moving from 2 per cent toward 8 per cent is not a rounding error. It is a pattern, and the families who stress tested their liquidity buckets against it before 2026 are having a noticeably calmer year than the ones discovering the arithmetic for the first time.
 

About the author
Jay Rogers is a financial professional with more than 30 years of experience in private equity, private credit, hedge funds, and wealth management. He has a BS from Northeastern University and has completed postgraduate studies at UCLA, UPENN, and Harvard. He writes about issues in finance, constitutional law, national security, human nature, and public policy.