Sponsorship: Private Debt News reaches institutional investors, credit professionals, and LP decision-makers. Early sponsor rates available. Contact us here or reply directly to this email.
At the end of last year, Blue Owl’s OBDC carried Loparex’s first-lien loan at par and its second-lien near 88 cents. Nine months later the first-lien sits at 22 cents. The second-lien is at 5.
Nothing about the underlying business changed that fast. Loparex, a Pamplona-backed maker of adhesive liners, had been staggering under its debt for years. S&P called the capital structure “unsustainable” back in 2024 after a distressed exchange it deemed tantamount to default. The company skipped a June interest payment and is under forbearance through September. Moody’s now sees Chapter 11 as a live possibility. What changed was a rescue M&A deal that fell apart, and with it the story that justified the marks.
OBDC’s overall non-accruals remain low at 0.8% of fair value. That’s not the point. The point is the trajectory: 88 to 63 to 5 on the junior paper in three quarterly marks. Loparex had been out sounding private credit lenders for $1.5 billion of refinancing as recently as January.
Half a world away, private credit’s first big non-software blowup arrived in Sydney. Bathla Group declared insolvency after amassing A$3.3 billion from more than 40 private credit funds, some promising 15% returns, some backed by personal guarantees from a founder who started as a taxi driver. Australia’s A$200 billion private credit market has as much as 60% of its lending in real estate, against 15-20% in North America. Prices in Sydney have fallen five straight months. Several local funds gated to contain the panic.
The redemption grind continued in the US. BCRED capped again at 5% after 10% sought out. Cliffwater capped at 5% after 16%. Cox Capital’s tender for five BDCs at a 26% average discount deployed just $5 million of $90 million available, and John Cox says he’s coming back anyway. KKR’s non-US fund saw requests fall to 2.5%. Fitch’s default rate ticked to 6.1%.
And Fed Chair Kevin Warsh suggested he’d consider raising rates if inflation stays elevated. Floating-rate borrowers already stretched at current levels don’t need the reminder.
Key Market Themes
1. Loparex Shows How Fast Marks Move Once the Story Breaks
OBDC put Loparex on non-accrual after the second quarter. The second-lien loan, marked around 88 cents at year-end and 63 at March 31, now sits at about 5 cents. A first-lien loan marked near par at year-end is at 22. Moody’s deemed the company in default and flagged potential Chapter 11.
OBDC President Logan Nicholson said on the August 6 call that Loparex “had been pursuing a transformative M&A transaction, which would have recapitalized the business with fresh equity.” When that collapsed, so did the marks. Loparex had sought $1.5 billion from private lenders at the start of 2026 to refinance first- and second-lien debt coming due early next year. Managers argue stable marks benefit investors and that recovery could exceed current values if a buyer surfaces. Skeptics point to Zips Car Wash, valued near par months before bankruptcy.
What the marks were really pricing
The Loparex marks weren’t wrong because Blue Owl was hiding something. They were wrong because they priced a rescue that hadn’t happened yet. That’s the structural flaw in mark-to-model: a loan carried at par on the strength of a pending transaction is really an option on that transaction, and options expire. S&P called this capital structure unsustainable two years ago. The 88-cent second-lien mark last December was betting the market wouldn’t notice. Every BDC holding a stressed credit “in process” on a refinancing or M&A rescue just got shown what happens when the process fails.
2. Bathla Takes 40 Funds Into the Sydney Housing Bust
Bathla Group declared insolvency last week owing A$3.3 billion to more than 40 private credit funds. The developer took land loans, construction loans, and residual stock loans, many promising around 15% returns, some carrying personal guarantees from founder Bhart Bhushan. PAG extended more than A$300 million and hasn’t gated, believing its debt is well secured. Centuria Bass halted redemptions on two vehicles while insisting the impact isn’t material.
The wider problem is concentration. Real estate accounts for as much as 60% of Australia’s A$200 billion private credit market. Post-crisis rules pushed banks out of construction lending and private funds filled the gap. Sydney house prices have dropped almost A$126,000 this year. Qualitas’ Andrew Schwartz says the cost of development debt had already moved 30-50 bps higher before Bathla fell and is rising further. Mirvac’s Campbell Hanan is fielding calls from private-credit-backed syndicates frustrated by rising holding costs. Colonial First State’s Jonathan Armitage said his firm avoided the sector precisely because of “concentration risk within Australian private credit around real estate.”
Hard assets, same problem
Bathla is the software crisis with different collateral. A single sector absorbed most of a market’s private credit, rates rose, valuations fell, and a borrower that grew fastest in the boom broke first. The 15% coupons were the warning. Nobody pays that when cheaper money is available, and Balmain’s CEO admitted his firm scaled back lending because Bathla looked “too stretched.” The uncomfortable part for the industry is that this happened with hard assets. The land is real, the buildings half-built. Collateral didn’t prevent the loss. It just changes who spends the next two years arguing over it.
3. The Q3 Gates Hold, and the Backlog Doesn’t Shrink
BCRED received an estimated $4.3 billion of repurchase requests in the third quarter, about 10% of shares, and capped at 5%. Some of that demand came from investors resubmitting after getting half of $4.5 billion in the second quarter, with $2.3 billion left outstanding. Cliffwater’s $31 billion fund capped at 5% after roughly 16% sought out, returning about a third. The firm noted that investors who’ve requested cash since the first quarter have received 78% of their capital.
KKR’s non-US evergreen fund, KIT, saw requests drop to 2.5% from over 5% in the second quarter and reported $170 million of year-to-date net subscriptions. Its letter noted that muted BDC flows have “reduced marginal buying power, creating a more favorable supply-demand backdrop for scaled lenders with available capital.” Blue Owl Technology Finance raised $150 million at 7.6%, its highest coupon since 2023, bringing debt raised since June to $800 million.
Stalemate, not recovery
The redemption queue has stopped growing and stopped shrinking. Repeat requesters are slowly getting paid out while new exits replace them, so the $15 billion backlog holds roughly steady. Cliffwater’s 78% figure is the useful frame: a patient investor who asked in March has most of their money back by now, which is exactly the argument for not selling to Cox at 26% off. OTF paying 7.6% for unsecured notes is the cost of keeping that machine running. Funds are borrowing at junk-adjacent coupons to pay out investors at par.
4. Cox Comes Up Empty and Doubles Down
Cox Capital’s tender for shares in five BDCs run by Blue Owl, Ares, Apollo, and HPS deployed roughly $5 million of the $90 million on offer, at discounts averaging 26%. That follows the February Saba-Cox bid for OBDC II that captured under 1% of shares. John Cox is undeterred: “I don’t think the discounts were too steep. People are getting familiar with the idea of valuing securities at a price that’s not NAV.” He’s raising a $150 million fund for more of the same and plans another tender as soon as possible.
The institutional secondaries market is telling a different story. Volume doubled to $20 billion in 2025 and Carlyle sees $80 billion by 2030. HarbourVest’s Greg Ciesielski said software and redemption worries improved pricing on LP-led deals by about 300 bps. BlackRock’s TCPC sold $523 million to Pantheon. Ares is working on one of the largest credit continuation vehicles ever. Bridgepoint and Jefferies Credit Partners have each explored $1 billion-plus deals.
Two markets, two clearing prices
Retail investors keep refusing the 26% haircut while institutions quietly take smaller ones through structured secondaries. Both are rational. A retail holder who waits two quarters gets most of their capital at par. An institution with a $500 million position can’t wait, and a 5-10% discount through a continuation vehicle beats a fire sale. Cox’s persistence suggests he’s betting that a rate hike or a few more Loparex-style markdowns turn the math. Two failed tenders say the panic he’s pricing for hasn’t arrived. His new $150 million fund says he thinks it will.
5. Banks Hedge the AI Buildout as Communities Fight It
Deutsche Bank is working on its first project-finance SRT, tied to about €2 billion of loans including data center projects. SocGen, ING, and BBVA have done or considered similar deals; BNP, RBC, and TD have explored AI-specific structures. Crescent Capital expects SRT sales to hit a record $45 billion this year. Deutsche’s CFO said SRTs already provide 75-80 bps of CET1 relief.
The physical side is getting harder. Seven in ten Americans say they don’t want to live near a data center. Activists stopped at least 75 sites in the first quarter. Texas Governor Greg Abbott ordered an audit of local impacts less than a year after calling the state AI’s epicenter. Blackstone’s QTS and Brookfield-backed Compass walked away from the $100 billion Virginia Digital Gateway after a zoning-notice error voided a hearing. A Scientific Climate Ratings study of 1,000 facilities found location drives physical risk more than operator, with wind exposure implying a 5.5% discounted cost of inaction by 2050 and Japan’s sites averaging 2.78% annual value impact against 0.6% in the US.
Who ends up holding the dirt
Banks want the AI lending fees and none of the AI tail risk, which is what an SRT delivers: keep the loan, sell the first-loss to someone chasing double digits. The buyers of that risk should read the other two stories in this section. A data center that can’t get zoned is a loan against dirt, and a data center built in a floodplain is a loan against a very specific insurance claim. The Virginia project died on a clerical error after five years. The capital structures financing this buildout assumed the physical world would cooperate. It’s starting not to.
6. TCPC’s CEO Exits as BlackRock Liquidates the Legacy Book
Phil Tseng resigned as CEO of BlackRock TCP Capital Corp effective August 31 and leaves the firm October 1. Jason Mehring takes over as CEO with Dan Worrell as president. TCPC marked down NAV 19% in January and another 5% in May, drew a Manhattan US Attorney valuation probe, and saw its shares fall 26% this year against a 9% drop in the BDC index.
BlackRock is effectively winding the strategy down. It sold $523 million of loans to a Pantheon-backed vehicle last month and hired KBW to find buyers for the remaining $671 million. HPS executives, who arrived via the $12 billion acquisition, have been running the cleanup and cut BlackRock’s legacy US private debt team to about a dozen people.
The first fund put down
TCPC is the first fund of this cycle to be more or less put down. Not gated, not restructured, just sold off piece by piece with the CEO shown the door. It was small, roughly $1.5 billion, which is why BlackRock could afford to do it cleanly. The lesson for bigger books is less comfortable. The problems here were 2018-2021 vintage middle-market loans, e-commerce aggregators, and a valuation process prosecutors found worth examining. None of that is unique to TCPC. The difference is that BlackRock had HPS to hand the mess to and $15 trillion to make the losses irrelevant.
Deals of Note
Deutsche Bank SRT - Working on first project-finance risk transfer, tied to ~€2B including data center loans
Blue Owl Technology Finance - Raised $150M of 7.6% notes due 2032, third financing since June; $800M of debt raised since Q2
TCPC - Sold $523M of loans to Pantheon-backed vehicle; KBW marketing remaining $671M
Bathla Group - Sydney developer insolvent with A$3.3B owed to 40+ private credit funds; PAG holds $300M+
BBVA - Working on SRT tied to 100 billion lira of Turkish SME loans, a rare emerging-market use
Cox Capital - Deployed ~$5M of $90M tender across five BDCs; raising $150M fund for more discounted purchases
The Reality Check
Loparex and Bathla are the same story on different continents. A borrower that couldn’t service its debt got carried at flattering marks because a rescue was always just around the corner. Then the rescue didn’t come. Blue Owl’s 88-to-5 collapse on the second-lien and Bathla’s 15% coupons both scream the same thing: the lenders knew. They priced the risk, then marked as if they hadn’t.
The Australian blowup should end the argument that hard collateral makes private credit safe. Sixty percent of a A$200 billion market went into real estate because the assets were tangible and the returns matched equities. The land is still there. It just isn’t worth what 40 funds lent against it. Collateral changes the recovery fight. It doesn’t prevent the loss.
Cox failing twice while institutional secondaries boom tells you where the real repricing happens. Retail investors are waiting out the gates and getting paid, slowly, at par. Institutions are quietly taking single-digit discounts through continuation vehicles rather than waiting. The 26% haircut Cox keeps offering isn’t the market clearing price. It’s a bet that Warsh raises rates and the clearing price finds him.
TCPC is what a clean exit looks like: sell the book, replace the CEO, let a bigger balance sheet absorb the loss. Most managers don’t have a BlackRock behind them. They have a redemption queue holding steady at $15 billion, floating-rate borrowers facing a possible hike, and marks that, as Loparex just demonstrated, are only as durable as the next deal that has to close.
Great round up as always. Essential indeed. Thanks y'all!
Why are you defending bad marks? Did OWL perform a liquidation analysis to determine recovery on the loan? And if so how did that recovery vaporize in a matter of months? The problem w probate marks is they are not tested by the market. They are basically made-up numbers with no reference point.