Private credit's default rate fell last quarter and hit an all-time record last quarter. Both statements are true. The gap between them is the entire trade — and today one of the biggest BDCs quietly wrote a $140 million check against GPUs.
Private credit has not avoided a default cycle. It has reclassified one.
Proskauer's Private Credit Default Index printed 2.51% for Q2 2026, down from 2.73% in Q1, across 716 loans representing $195.6 billion of original principal. Two days later Fitch printed a US private credit default rate of 6.0% on a trailing-twelve-month basis — an all-time record, up from 5.7%.
That is not a data error. It is a definitional one, and the definition is where the money is. Of Fitch's 32 private credit default events in Q2 2026 — 20 new defaulters across roughly 1,300 tracked US borrowers — more than half were maturity extensions rather than missed payments or interest deferrals. Proskauer counts payment and covenant defaults. Fitch counts every accommodation a creditor was forced to grant.
So: borrowers are paying. Lenders are paying for them to keep paying, in the currency of time. And the price of that time is now measurable.
| Series | Latest print | What it counts |
|---|---|---|
| Proskauer Private Credit Default Index | 2.51% (Q2 2026, down from 2.73%) | Payment and financial-covenant defaults only |
| KBRA Middle Market Default Monitor (count) | 3.3% (TTM to 6/30/26) — record | Adds imminent defaults and interventions that prevented a payment default |
| KBRA Middle Market Default Monitor (debt) | 2.4% (TTM to 6/30/26) — series peak | As above, dollar-weighted; 92 borrowers, >$30bn of debt |
| Moody's private credit estimate | 1.6%–4.7% (FY2025) | Publishes the range, as a function of whether distressed exchanges count |
| Fitch US private credit default rate | 6.0% (Q2 2026) — record | Full rating-agency definition: distressed exchanges + maturity extensions |
A three-point spread created by a definitional choice. Moody's states the mechanism outright: distressed restructurings accounted for roughly 65% of all 2025 defaults.
If you only remember one figure from this piece, make it this one, from Lincoln International's 31 March 2026 private market data:
Loans that carried no PIK at close but carry PIK today are now 5.9% of all direct-lending loans, up from 2.5% at the end of 2021. That cohort's average loan-to-value has gone from 39.4% at inception to 76.1%.
Read that mechanically. A first-lien loan underwritten at 39% of enterprise value — the equity cushion private credit marketing decks are built on — now sits at 76% of a smaller enterprise value, and the borrower is paying interest with more debt. The cushion did not erode by 5 points. It erased roughly two-thirds of itself. Lincoln's own term for the metric is a shadow default rate.
Consequence one: recovery math inverts. At 39% LTV, a first-lien lender is money-good in almost any liquidation. At 76% LTV in a business whose EBITDA is why it needed PIK relief, the first lien is the fulcrum security. Octus puts average realized recoveries in private credit restructurings near 50 cents, well below the 70% figure most managers cite; Fitch, measuring its own monitored portfolio, still sees 70–90% on consensual first-time restructurings. Both are current. Which one applies is decided by whether the credit has already been restructured once.
Consequence two: PIK screens miss half the problem. PitchBook LCD's 21 July 2026 review of more than 170 BDCs found that of roughly 5,000 companies held at 31 March 2026, 538 — 10.6% — showed signs of credit pressure, a 15% increase in one quarter, with first-lien and unitranche investments under pressure up 44% to $35.4 billion. Half of those 538 companies had used no PIK in the preceding twelve months. The most popular early-warning indicator in the sector has a 50% false-negative rate.
Amend-and-extend volume was $106 billion in H1 2026 against roughly $84 billion in H1 2025, up 26%, with $27 billion in June alone. The near-term wall genuinely cleared: leveraged loan volume maturing through end-2027 narrowed to $32 billion from $62 billion at the end of 2025, while loans maturing in 2029 and beyond grew by $129 billion over the same six months.
Here is the part the headline volume hides. Of 2026 amendments, 30% went to issuers rated BB− or higher, up from 11% in 2025 — while the B− share collapsed to 27% from 44%.
Record extension volume plus a collapsing B− share means one thing: the extension window is open, and the credits that need it are being cut out of it. Private-equity-backed borrowers drove 74% of institutional maturity-extension amendments in Q2 2026 while accounting for just 44% of new-money activity. Sponsors are defending, not funding.
And extension does not repeat for free. Lincoln estimates 30–40% of direct lending deals maturing in the next two years have already extended once, with leverage up roughly half a turn from inception across all vintages and closer to a full turn for surviving 2019–2020 vintages. Of 148 liability management transactions tracked by Covenant Review, 37 — one in four — ended in Chapter 11 anyway.
What replaces extension is ownership. Direct lenders foreclosed on $24.2 billion of principal in 2025 and a further $15.2 billion in Q1 2026 alone — against $13.6 billion across the preceding three years combined. Nearly 75% of those change-of-control transactions relate to 2021 and 2022 vintage deals. A single quarter at $15.2 billion annualizes to roughly two and a half times all of 2025.
Per Robert A. Stanger & Co. data published 10 August 2026:
One early counter-signal, and it deserves honest weight: three NAV BDCs reporting so far for Q3 show repurchase requests of 4.6% of NAV versus 7.9% for the same funds in Q2, with two meeting 100%. Three funds is not a trend. It is the single most important data point to watch in early November.
This is the transmission mechanism almost nobody prices. A perpetual-capital vehicle managing a quarterly gate has structurally less appetite to fund a delayed draw, a covenant holiday or a rescue tranche — regardless of the credit merits. Redemption pressure at the fund level converts directly into less forbearance at the borrower level, in exactly the year the borrowers need more of it.
On 20 August 2026, Wingspire Equipment Finance — the equipment finance arm of Wingspire Capital, a portfolio company of Blue Owl Capital Corporation (OBDC) — closed a $140 million equipment financing with an unnamed private-equity-backed GPU cloud provider, funding high-density GPU servers for a platform serving AI labs, enterprises and public-sector customers.
Take the deal on its own terms: it is senior, asset-backed, short-tenor equipment paper at attractive spreads, and the demand is real. Now stack it on the sector context above. Private credit's existing problem is a 2021–22 software vintage whose collateral is recurring revenue that AI is actively repricing. Its new growth engine is lending against the depreciating physical asset that is doing the repricing. Wingspire's own VP framed the pitch as needing "capital partners that understand both the equipment and the pace of the market." The pace of the market is precisely the risk: GPU residual value is a function of the next accelerator generation's ship date, and no BDC discloses a residual-value assumption.
This is not a reason to short OBDC today — the exposure is small against a manager with $319 billion of AUM as of 30 June 2026. It is a reason to demand disclosure, and a reason to treat "AI infrastructure lending" in BDC portfolios as an unmarked risk bucket rather than a diversification benefit.
Closing prices, Thursday 20 August 2026:
| Ticker | Close 8/20 | Day | Latest NAV/sh | P/NAV | Non-accruals | Yield |
|---|---|---|---|---|---|---|
| ARCC (Ares) | $19.78 | −0.10% | $19.94 (12/31/25) | ~0.99x | Q2: 5 borrowers moved to non-accrual, led by AmeriVet | 9.65% |
| OBDC (Blue Owl) | $11.28 | −0.44% | $14.26 (6/30/26) | 0.79x | 2.8% cost / 0.8% fair value (from 2.0%/1.0%) | 10.66% |
| FSK (FS KKR) | $11.91 | −1.24% | $18.30 (6/30/26, −2.8% QoQ) | 0.65x | 3.8% | 12.86% |
| BXSL (Blackstone) | $24.67 | +0.45% | $26.92 (12/31/25) | ~0.92x | ~3.6% and rising; coverage slipping below 100% | — |
| TCPC (BlackRock) | $4.03 | −1.47% | $7.07 (12/31/25) | 0.57x | NAV −50.8% from 2021 peak, 10 straight declining quarters | — |
| MAIN (Main Street) | $58.40 | +0.40% | $33.33 (12/31/25) | 1.75x | ~1% at fair value; NAV +5% in 2025 | — |
Sector context from the Raymond James BDC Weekly Insight dated 13 August 2026: average BDC price/NAV of roughly 0.82x and an average dividend yield near 12.3%. NAV figures dated 12/31/25 are the latest in the cross-sectional dataset used; June-quarter NAVs for OBDC and FSK are from company Q2 2026 disclosures.
FSK is the textbook case from the filings. PIK income went from 18.6% of net investment income in FY2021 to 33.1% in FY2025; dividend coverage went 107.8% → 100.0% → 83.6%, and the 30% dividend cut arrived exactly when the math said it would. Non-accruals of 3.8% at 30 June 2026 sit roughly five times the ~0.7% sector average measured at year-end 2025. Q2 2026 adjusted NII of $0.43 against a $0.44 dividend is 98% coverage — no margin, again. Analysts raised targets into that print (JPMorgan to $11 Neutral on 8/19, RBC to $13 Sector Perform on 8/12); the targets moved, the credit metrics did not.
OBDC is the opposite configuration. Management took the pain first: a 16% base dividend reset restored coverage to roughly 116%, non-accruals at fair value are 0.8% — down from 1.0% — NII was $0.36/share against a $0.33 declared dividend, portfolio yield 9.9%, leverage 1.11x, and the company repurchased ~$35 million of stock accretively in the quarter. Insiders bought in the open market at $11.21–$11.31 in May and June 2026, which is within 1% of today's $11.28 close. The stock is at 0.79x a NAV that just fell only 1%.
The market is charging a 21% NAV discount for OBDC and a 35% discount for FSK. Given a 4.75x difference in non-accrual rate and the PIK divergence, that spread is too narrow.
Today the Treasury's intervention failed in public. After Wednesday's announcement that buybacks would at least double from $2bn to $4bn per operation, the 10-year rose more than 5bp to 4.704% and the 30-year rose more than 5bp to 5.248% — both above where they sat before the announcement. The 2-year barely moved at 4.185%. JPMorgan's Maia Crook: the interventions "belie the underlying structural challenges and do nothing to address them." Evercore's Krishna Guha called it "a weak form of Operation Twist" that "could backfire if it is seen as signaling concern about the ability to fund longer-term at acceptable cost."
Three downstream consequences for private credit specifically:
Trigger conditions: Q3 non-traded BDC repurchase requests print between 6% and 10% of NAV (Stanger, early November); Proskauer Q3 index stays in the 2.3–2.9% band; Fitch's TTM rate stays at or above 5.7%. Public BDC discounts stay wide (sector ~0.80–0.85x NAV), one to three more mid-tier dividend cuts land — BXSL is the most telegraphed, with coverage slipping below 100% and non-accruals near 3.6% — and dispersion between best-in-class and worst-in-class widens further. No systemic event. Alt manager multiples grind lower on decelerating fee-related earnings growth rather than credit losses.
Trigger conditions: Fitch's own July guidance proves right — it stated the H1 decline in leveraged loan and high-yield default rates was driven mainly by base effects rather than credit improvement, and that default activity should accelerate in H2 2026. Watch for: Proskauer Q3 above 3.0%; the share of the Morningstar LSTA index trading below 80 cents pushing past its March 2026 high of 7.23%; a fourth consecutive quarter of non-traded BDC net outflows; a Q3 direct-lending foreclosure print above $12 billion. In this path the 2028 wall stops being a 2028 problem because the extension market has already priced out B− credits — you get 2027 restructurings on 2028 maturities. FSK-type names go to 0.55x NAV; HYG breaks $78; alt managers de-rate another 15–20%.
Trigger conditions: the early Q3 signal generalizes — sector-wide repurchase requests fall below 6% of NAV and sponsors meet 100%; Q3 fundraising recovers above $4 billion; ARCC/OBDC report flat-to-lower non-accruals; the 10-year retraces below 4.45%. Then the 20%+ discounts on quality BDCs are simply wrong, and OBDC at 0.79x NAV with 116% coverage is one of the better risk-rewards in income. This is the scenario that pays best if you own the right names, which is why the trade below is a pair and not a short.
Structure: equal dollar, long Ares Capital / short FS KKR Capital.
Entry zone: ARCC $19.30–$20.10 (last $19.78) / FSK $11.70–$12.50 (last $11.91). Current ratio 1.66.
Target: ratio 1.90 (~14% gross spread gain), e.g. ARCC $20.50 / FSK $10.80. FSK's 52-week low of $9.72 (18 March 2026) is the stretch objective on the short leg.
Invalidation (one condition, measurable): FSK's Q3 2026 report shows non-accruals below 3.0% at fair value and NII coverage above 105%. Either alone is noise; together the thesis is wrong and the pair comes off same day. Secondary kill-switch: ARCC non-accruals above 2.5% at fair value.
Timeframe: 1–3 months (Q3 prints land in the first week of November).
Why two independent signals agree: (a) firm-level filings — FSK PIK at 33.1% of NII vs a sector that averages ~8%, non-accruals at 3.8% vs ARCC's materially lower book, coverage at 98% with no margin; (b) sector-level structure — the amend-and-extend market's B− share collapsed from 44% to 27%, which is precisely the cohort FSK's problem loans sit in, and FSK's own 2021–22 vintage concentration matches the 75% of foreclosures now coming from those vintages.
Quantified risk: negative carry. You receive ARCC's 9.65% and pay FSK's 12.86% — roughly −3.2% annualized, about −80bp per quarter held. A 14% target against 80bp of quarterly carry is a 17:1 gross ratio, which is why this is a pair and not a naked short. The other real risk is a squeeze in the deepest-discount BDCs on any dovish Jackson Hole surprise (26–28 August); size accordingly and do not add on the first 5% adverse move.
Entry zone: $10.90–$11.45 (last $11.28). Insiders bought at $11.21–$11.31 in May and June 2026.
Target: $12.75 (0.89x the 6/30/26 NAV of $14.26) — roughly +13% plus a ~10.7% distribution.
Invalidation: Q3 2026 non-accruals above 1.5% at fair value (vs 0.8% today), or a weekly close below $10.40.
Timeframe: 1–3 months tactical, 6–12 months for full discount closure.
Why speculative and not high conviction: the setup is clean — dividend already reset, coverage at 116%, non-accruals falling at fair value, accretive buybacks, insider purchases at the current price, 21% NAV discount versus a sector average near 18%. But two things are unconfirmed. First, the Q2 fair-value non-accrual improvement (1.0% → 0.8%) came alongside a rise at cost (2.0% → 2.8%), which means the improvement is a marks story as much as a credit story. Second, the Wingspire GPU book is a new, undisclosed-residual-value exposure that did not exist in the underwriting history. One signal, unconfirmed.
No position yet. Levels set; the trigger is a data release, not a price.
Short trigger: Stanger's Q3 non-traded BDC data (early November) shows a third consecutive quarter of net outflows worse than −$2 billion. On that print, short OWL below $10.80 (last $11.40), target $9.00, invalidation weekly close above $12.60. Blue Owl is the purest expression: its fee-related earnings algorithm is the most dependent on perpetual retail capital, and OCIC alone fielded repurchase requests for 18.8% of shares outstanding in one quarter and honoured ~27%.
Long trigger: the same print shows net flows positive and repurchase requests below 6% of NAV. Then buy OWL above $13.00, target $16.00, invalidation $11.20.
Timeframe: 1–3 months to trigger.
High yield is not pricing any of this: HYG closed at $79.56, down 0.19%, on a day the alt managers fell 2.6–3.1%. Trigger: a daily close below $78.50 would confirm that public credit has begun to agree with the lenders' equity, and is the signal to press trade 1 and take the short leg of trade 3 early. Invalidation of the whole bear framework: HYG above $80.75 with the 10-year back under 4.45%.
Our 5 August piece, "Private Credit Is Now Lending Against Both Sides of the AI Trade," set a specific re-rating trigger: OBDC Q2 non-accruals at or below 1.2%. The Q2 print delivered 0.8% at fair value. That leg hit. It is the reason OBDC appears on the long side today rather than the short side — and the reason the position is sized as speculative rather than high conviction, because the same filing showed non-accruals rising at cost.
The consensus argument is that private credit's low payment-default rate proves the bears were wrong. It proves something narrower: that lenders with discretion over the definition of "default" and control over the timing of the workout have used both. That is a real advantage — it is genuinely why private credit has outperformed syndicated loans through this cycle — and it is not free. It has been paid for with equity cushion (39.4% → 76.1% LTV on the PIK-converted cohort), with cash yield converted to accrual, and with a maturity schedule pushed into two years whose base rate nobody can forecast.
KBRA's Q2 compendium contains the number that decides how this ends: median interest coverage held at 1.6x, but the share of borrowers with improving coverage plateaued after more than two years of gains, and median EBITDA growth fell to 24% from 27% — the largest quarter-on-quarter decline in the series. Earnings growth has been doing the deleveraging work since 2023. That engine just started slowing. Everything in this piece is downstream of whether it keeps slowing.
Not investment advice. Positions may be held in securities discussed. All price data as of the 20 August 2026 US close.