Newsletter · · Ashutosh Agarwal

PIMCO Says a Default Cycle Has Begun - The Private Credit Boom (and Cracks) - Week of July 22, 2026

The Private Credit Boom (and Cracks) newsletter for the week of July 15–22, 2026. PIMCO's Lotfi Karoui says a default cycle in credit has begun even with the economy fine, the bear thesis migrates from banks to insurance balance sheets and Apollo's Athene, and the visible stress is in the plumbing (redemptions to 40% of quarterly capacity, 6–7 point mark dispersion, traded BDCs at discounts) rather than in defaults.

The Private Credit Boom (and Cracks)

Week of July 15–22, 2026: PIMCO Says a Default Cycle Has Begun

TL;DR

  • A credible, non-hysterical voice just made a regime call. PIMCO's Lotfi Karoui said flatly that "a default cycle in credit has begun," and, crucially, that it's happening while the economy is fine. That directly reverses the "credit is actually getting better" story we heard last week, and it lands on direct lending specifically. PIMCO Accrued Interest
  • The bear case moved houses, from banks to insurers. The scariest podcast of the week wasn't about bank loans; it was about life-insurance balance sheets stuffed with private credit, and it put Apollo's Athene squarely in the crosshairs. On the very same day range, Apollo's own wealth executive was on another show making the bull case for that exact model. Same balance sheet, opposite verdicts. Monetary Matters / Simple, but Not Easy
  • The real damage is showing up in the plumbing, not the fundamentals. Redemptions running to 40% of a quarter's capacity, marks on the same loan differing by 6–7 points between managers, and traded BDCs (public "private-credit funds") selling off to discounts, while the actual default rate is still ~1% and portfolio revenues are mostly growing. It's a liquidity story first, a credit story second. Loud week; the operators showed up in force.

What's new

1. "A default cycle has begun," and it doesn't need a recession to keep going. The single most important thing said all week came from PIMCO's Lotfi Karoui on PIMCO Accrued Interest (July 16). He didn't hedge: "we feel like a default cycle in credit has begun." What makes that striking is his second point: normally you need an economic downturn to spark defaults ("take a 40-year time series... you see four humps, basically... the four last recessions"). This time is decoupling: "the economy is still in great shape... but this time might feel a little different." His reasons are concrete and mechanical: most private-credit loans carry floating interest rates, so the Fed's "500 basis points of cumulative hikes... between 22 and 23" are still squeezing borrowers; software is a slow-burn problem; and, as the asset class ballooned, "you start loosening your underwriting standards... you start to write bigger checks." Why it matters: This is the credible middle of the road, not a permabear, saying the cycle has turned. It updates (and partly contradicts) last week's Slok/Eisman "credit is getting better" call, and it points the finger at direct lending specifically rather than public high-yield, which he calls "remarkably disciplined."

2. The bear thesis relocated to insurance, and it's an Apollo/Athene story now. On Monetary Matters (July 20), analyst Nick Nemeth argued the danger isn't primarily on bank balance sheets, it's on insurers. His framing: insurers are "a $10 trillion balance sheet... $1 trillion on those balance sheets is private credit," and unlike banks "there's no FDIC." He was blunt about Apollo's insurance arm, Athene: "$300 billion of assets... I've calculated the real capital as more like $4 to $6 billion," roughly "50% Level 3 assets" (Level 3 means hard-to-value, no market price, you trust the model). On Apollo the manager he was almost admiring ("good underwriters... amazing lawyers... they are sharks, and if you're in a deal with Apollo, look out, they will cut you out"), while warning they're "running risk at an F35 level." He also flagged MassMutual at "25% private credit" and Carlyle's Fortitude for a balance-sheet exemption. Why it matters: This is a direct challenge to the permanent-capital story that underpins APO's valuation, the idea that insurance money "can't be pulled." It moves the debate from "banks are fine" (which most now concede) to "the risk just migrated somewhere with less oversight."

3. Apollo, on a different show, made exactly the opposite case for the same model. The mirror image: Apollo's Lucy Xin on Simple, but Not Easy (July 21) laid out the bull case for the wealth-and-insurance flywheel. Apollo runs "about $1 trillion in assets," and its edge is "origination, flexibility, and alignment," the alignment coming precisely from Athene: "we buy our own book." Her TAM pitch: individuals are "more than 50% of global assets under management... $150 trillion" but hold "less than 5%" in alternatives versus ~20% for institutions and "40 to 50%" for family offices, implying "$10 trillion in money in motion over the next 10 years." She tied it to AI directly: "in one of our direct lending strategies, we provided financing to a company that rents out GPUs... works directly with xAI." Why it matters: You now have the operator's growth story and the bear's blow-up story pointed at the identical balance sheet, in the same week. For anyone with a view on APO, this is the debate to resolve.

4. The visible stress is in the plumbing, redemptions and marks, not (yet) in fundamentals. Two operators with genuine look-through said the fundamentals are holding while the liquidity machinery is grinding. J.P. Morgan Asset Management's Brad DeMong on Insights Now (July 16): semi-liquid funds are "designed to effectively only provide redemption requests of roughly 5%," yet "we've seen anywhere from high single digits to up to 40% redemptions in any given quarter," forcing BDCs "to sell assets into the market." (A BDC, business development company, is essentially a private-credit fund, some public, some not.) And the marks are all over the place: PIMCO's Karoui said that for the same loan held by multiple BDCs, the gap between the most optimistic and most pessimistic manager is "well over six, sometimes seven points." Why it matters: This is why publicly-traded BDCs are selling off even though loans are mostly performing: it's a forced-selling-and-mistrust problem, not a wave of defaults. It also tells you where to watch for the first real breakage: whoever has to sell good assets to meet redemptions.

5. Data-center lending, the growth engine, just hit a very physical wall. The AI-infrastructure origination story that Blackstone and Blue Owl have leaned on ran into local politics. On The Promote Podcast (July 15), the hosts detailed Blackstone walking away from a Prince William County, Virginia project "bigger than Hudson Yards... 22 million square feet of data centers, 34 buildings, and 1.7 gigawatts," a court voided the rezoning, Blackstone appealed to the Virginia Supreme Court, then decided "the litigation wasn't worth pursuing." A CoreWeave/Blue Owl data center in Lancaster, Pennsylvania is drawing similar pushback. Yet demand hasn't blinked: across the four biggest US markets, supply grew "2 gigawatts in the last 12 months... a 33% increase," but "there are fewer than 75 megawatts available," and Starwood just raised a "$10.2 [billion]" fund, "more than a third" earmarked for data centers, "2x what it was last time." Why it matters: The bottleneck for the highest-growth slice of private-credit origination is shifting from money and power to permitting and community consent, a slower, messier constraint that doesn't show up in a fundraising number. Blackstone, the hosts noted, "is now an AI infrastructure company above everything else."

The debate

The honest framing this week is that both sides got stronger, and the fault line is now clear.

Cycle-starting deterioration (the bears). Karoui's "a default cycle has begun" is the marquee quote because it comes without drama and without a recession. Whalen kept the bank-exposure charge alive on The Julia La Roche Show (July 18): "$1.5 trillion for the loans to non-depository financial institutions... $3 trillion worth of undrawn commitments... about $4 trillion" of bank exposure, "half of that's good. The other half... is problematic," describing "zombie portfolios" kept alive by "extend and pretend." Nemeth escalated it into an insurance-solvency thesis. And the market's own plumbing (40% redemption quarters, 6–7-point mark disagreements, traded BDCs at discounts) is exactly what early-stage stress looks like.

Idiosyncratic noise in a growing, well-underwritten asset class (the bulls). Eagle Point's Tom Leveen, who finances BDCs and holds CLO equity (giving him look-through into thousands of loans), was the data-rich rebuttal on The Credit Edge (July 16): default rate "around 1%... well below the long-term average," portfolio "revenue is growing in many cases, EBITDA is flat to growing," and the structural backstop that "across the 46 years since the BDC laws have been in place... there have been two defaults ever by BDCs, Allied and ACAS. And both of those were a 100-cent recovery to the creditor." Frank Danieli of MA Financial argued the problem isn't the asset class but the discipline: diversify across "38 different subsectors" and you avoid the monoline software trap.

Notice the two sides aren't actually contradicting each other on facts. Bulls are talking about today's fundamentals (defaults low, EBITDA growing); bears are talking about the forward path (floating-rate squeeze, software wall in 2028–29, marks that haven't caught up, insurance leverage). The swing factor is time and refinancing, and whether the marks converge down or the fundamentals hold up first.

Pull-quote of the week, from PIMCO's Karoui, because it captures the whole tension in one line:

"Typically it does take an economic downturn to actually fuel a rise in financial distress in corporate credit. But this time around... that default cycle is kind of starting to decouple a little bit because the economy is still in great shape."

Stocks in play

Apollo (APO), the week's main event.

  • Bull: $1T AUM, a genuine origination edge, and the Athene "buy our own book" alignment; a retail-alts TAM Xin frames as "$10 trillion in money in motion over the next 10 years," with AI/GPU lending (e.g., financing a GPU-rental company tied to xAI) as a fresh origination lane. Even the bear conceded Apollo are "good underwriters" who "make marginal moves" out of hot sectors.
  • Bear: Nemeth's core attack is on Athene's capital adequacy ("real capital... more like $4 to $6 billion" against ~$300B of assets, ~50% Level 3) and on the premise that insurance money is truly "permanent" (he notes surrenders: "7% in year 1, 5% in year 2," with "1/3 of Athene's completely outside the window of any penalty whatsoever").
  • Next catalyst: Apollo earnings (FRE growth, Athene spread and Level 3 disclosure, retail/wealth net flows). Xin noted she'd listened to "Apollo's recent earnings call" and Rowan's point that hyperscalers are becoming the biggest debt issuers, watch for that framing to recur.

Blackstone (BX)

  • Bull: The scaled AI-infrastructure play, the "$10 billion take private of QTS" as "Blackstone's lens into the AI boom," a fundraising magnet for sovereign and retail capital; data-center demand still wildly outstripping supply (33% supply growth absorbed, under 75MW available across the top four markets).
  • Bear: Execution risk is now political. Walking away from the Prince William County mega-project (after a court loss and community backlash) shows the constraint has moved to permitting. Nemeth's aside, "Blackstone seems to love the hottest Momo trade of the year in private credit," is the concentration worry in a sentence.
  • Next catalyst: More permitting/zoning decisions on flagship data-center sites; data-center segment fundraising and deployment pace.

Blue Owl (OWL)

  • Bull: A core beneficiary of the same data-center-credit tailwind.
  • Bear: Named directly in the backlash, the "CoreWeave and Blue Owl Data Center" in Lancaster, PA "has seen a lot of pushback." Same permitting risk as BX, on a smaller base.
  • Next catalyst: Progress (or stalling) on named data-center projects; any read on origination pace into the AI build-out.

Annaly (NLY) / AGNC, tangential but named.

  • Whalen disclosed NLY as his "biggest position" and that he owns AGNC, reiterating the mortgage-REIT playbook: "the actual level of interest rates... don't matter... the spread is what matters," and the higher rate complex "beyond 2 years is actually good for them." Not private credit per se, but a rates-spread read from a vocal bear.
  • Next catalyst: The rate path, Whalen and others see cuts off the table and inflation sticky.

Barings BDC (BBDC), a real-money retail signal.

  • RIA's Lance Roberts said outright they "bought BBDC a month, two months ago" near "98% of market decline," on the view that traded BDCs are being unfairly "lumped into the triple C" as "retail is freaking out and they're selling BDCs." A concrete example of the discount-to-NAV-as-opportunity trade.

Read-throughs

  • BDCs (ARCC, BXSL, OBDC, generically discussed, not named this week): The whole traded-BDC complex is caught in a sentiment washout. Roberts: junk bonds "trading fine" but BDCs "at pretty big discounts to their net asset value," because "B minus" BDCs are getting "lumped into the triple C." Leveen's structural comfort (the 150% asset-coverage test and "two defaults ever... both 100-cent recovery") is the counter. The tell to watch: whether NAV marks drift down toward the market's discount, or the discount closes. Biggest gap remains that no ARCC/BXSL/OBDC executive spoke on the record; the specific non-accrual/PIK/NAV numbers still live in filings.
  • Insurance balance-sheet partners: This is now the fault line. Athene (Apollo) is the lightning rod, with MassMutual ("25% private credit") and Carlyle's Fortitude also named. The debate is capital adequacy and whether surrenders can ever spike: Nemeth thinks penalties ("7% year 1") won't stop a panic; the industry says the capital is permanent. Any insurer disclosure on Level 3 share and private-credit allocation is now price-sensitive.
  • Regional and money-center banks: Whalen's "$4 trillion" exposure and "extend and pretend" charge keeps the bank-counterparty risk alive, but the more interesting read is competition: Zegen said Madison's back-leverage arm "competes with JP Morgan, Goldman Sachs, and banks" for loan-on-loan financing. Non-banks are taking the warehouse-lending relationship, not just the loan. Separately, banks are using "regulatory capital relief" / significant-risk-transfer trades (Leveen is a first-loss buyer) to keep loans while shedding capital charges, a quiet channel where banks and private credit are partners, not rivals.
  • Syndicated-loan / CLO market: Spread compression, not defaults, did the damage: Leveen cited CLO equity at "negative 15% total return last year" as loans repriced tighter ("$375 over... repriced to $325"). The first European CLO default "since an overhaul of the market more than a decade ago" (a Bain-managed deal) is the outlier to watch. And "well over 100 liability management exercises" in 18 months ("lender-on-lender violence") means the headline 1% default rate understates real creditor pain.
  • Data-center / ABF borrowers: The origination TAM is enormous and getting more structured: Morgan Stanley's Anish Shah on Thoughts on the Market (July 16) sized it as "more than 30 gigawatts... around $2 trillion of aggregate cash capex" over two years, with "$30 billion priced across 15 deals since fall 2025" in public high-yield data-center bonds, and AI funding potentially "top[ping] 15% of the total issuance across all credit products." He flagged the risks bulls gloss over: "circularity, vendor financing, and technology obsolescence." Pair that with the permitting backlash above and you get the full picture: demand is real, structuring is innovating fast, but the constraints (power, permits, obsolescence) are multiplying.

What changed vs last week

Last week (July 8–15) the story was: banks retreating (HSBC/JPM/GS/Barclays pulling leverage), a strong bull rebuttal from Apollo's Torsten Slok and Steve Eisman ("credit is actually getting better," defaults falling for 12 months), and operators finally appearing (Sixth Street's Julian Salisbury, Blackstone's Farhad Karim) to call the bear commentary "irresponsible."

This week, three things moved:

  1. The bull's central claim got contradicted. Last week's headline bull point was that defaults, distressed exchanges and LMEs were all falling, "credit is getting better." This week PIMCO's Karoui said "a default cycle in credit has begun," and Nemeth claimed "defaults above 2008 levels." Leveen's ~1% default rate still supports the bulls on level, but the direction is now contested by a credible source. That's a genuine change, not noise.
  2. The bear case migrated from banks to insurers. Last week the sourced bear action was banks de-risking (Snider relaying HSBC/JPM/GS/Barclays). This week the sharper, and more novel, bear thesis is insurance-balance-sheet solvency, aimed at Athene/Apollo. It's a different, arguably more systemic, place to worry, and it directly targets APO rather than the banks.
  3. The redemption and marks story got quantified. Last week we flagged redemption-gating at Apollo/Ares qualitatively. This week we got numbers from multiple operators: DeMong's "up to 40%" redemption quarters, Karoui's "six, sometimes seven point" manager-to-manager mark dispersion, Leveen's ">5% redemptions" non-traded BDC headline, and Roberts' retail-selling-BDCs read. The plumbing stress is now measured, not just asserted.

New this week that wasn't in last week's issue: the data-center permitting backlash (Blackstone's Prince William County exit; Blue Owl/CoreWeave in Lancaster) as a concrete brake on the AI-origination growth engine, versus last week's Karim framing of data-center deals as simply "massive."

Still-missing (now three-plus weeks running): no named BDC (ARCC/BXSL/OBDC) executive with specific non-accrual/PIK/NAV data; no manager-provided new-loan-spread-vs-BSL comparison (Zegen's CRE spread grid and DeMong's secondary BSL marks are the closest we got); and this week, silence on KKR, Ares, Sixth Street, HPS, Golub and Oaktree by name. Those numbers live in earnings and filings, and the next round of BDC/manager prints is where they'll surface.