TL;DR
The retail redemption wave is cresting — Q3 withdrawal requests fell across every major semi-liquid fund — but the deeper story this cycle is convergence getting physical: Apollo closed the "largest private financing ever" for Broadcom's AI platform, Blackstone is about to price the first euro hybrid CLO blending private credit with public loans, and Apollo's own John Zito conceded "all the marks are wrong." Meanwhile the NAIC put a $1.2 trillion number on insurer private credit and is building a framework to police the private ratings underneath it.
The retail liquidity crunch is decaying, not resolving
The redemption wave that defined 2026 for semi-liquid private credit is showing its first broad-based signs of cresting, with every major non-traded BDC reporting sequentially lower or steady Q3 withdrawal requests. Apollo Debt Solutions saw requests fall from 16.8% to 14.7% of fund value, and aggregate requests across the 13 non-traded BDCs reporting came in at $9.6 billion, or 10.1% of fund value, down from 11.2% in Q2 evergreen-private-credit-redemptions-liquidity-gating-2026.
"Following third-quarter repurchase payments, investors who have sought liquidity during 2026 will have received an estimated 75% of their requested capital." — evergreen-private-credit-redemptions-liquidity-gating-2026
But the easing is relative. Apollo has gated withdrawals for a third consecutive quarter, Cliffwater's $31 billion fund capped roughly 16% of requests at 5% for a third straight time, and BCRED's requests are running at about double its cap — the 5% quarterly limit mechanically keeps request ratios elevated as investors re-tender unfulfilled asks evergreen-private-credit-redemptions-liquidity-gating-2026. Barclays' Peter Troisi captures the mood: "It's too early to call a bottom on the redemptions, but it's clear that the trend is that the headline redemption numbers are declining" (FT). The listed channel tells the same story from the other side — Blue Owl down 36% year-to-date, Carlyle down 33%, against the S&P 500 up 12% evergreen-private-credit-redemptions-liquidity-gating-2026
.
What to watch: Ares and Blue Owl Q3 data, due in coming weeks — they're the last big checkpoints before anyone can call the crest a bottom.
AI financing graduated from MOU to closed deal
The compute-financing buildout crossed from concept to closed transactions this cycle, and the Nvidia consortium is no longer the only show. Apollo-managed funds are leading an initial $35 billion capital solution for Broadcom's AI XPV Platform, alongside Blackstone and global banks, designed to enable over 20GW of compute capacity through 2028 and to fund Anthropic's expansion of more than 1GW starting mid-2026 ai-infrastructure-data-center-private-credit-financing-2026.
"AI compute is rapidly emerging as one of the most compelling new asset classes in finance, characterized by contracted cash flows, mission-critical utility and a supply-demand dynamic that continues to intensify." — ai-infrastructure-data-center-private-credit-financing-2026
The tranche architecture is the convergence thesis in one deal: Wells Fargo as global coordinator, Goldman Sachs, Bank of America and Morgan Stanley as placement agents — private credit as lead principal, bulge brackets as arrangers ai-infrastructure-data-center-private-credit-financing-2026. The $500 billion Nvidia consortium, meanwhile, remains at the MOU stage with no funded vehicles disclosed — the closed Broadcom deal is the template it would industrialize, and the risks (rapidly depreciating chip collateral, counterparty concentration in a few AI labs, circular vendor financing) are just as concentrated.
What to watch: Whether the Octus-reported Blackstone junior tranche — reportedly over $9 billion to fund Anthropic's Google TPU purchases — prices, since it's the cleanest read on how the capital stack values AI risk.
The valuation problem is now conceded from the inside
The price-discovery debate moved from critics to the industry's own leadership this cycle. John Zito, Apollo's co-president of asset management, told the Wall Street Journal that "if you don't mark your book, you actually lose trust with the clients" — and then went further: "I literally think all the marks are wrong." Apollo later clarified the comment referred to software valuations jpmorgan-markdown-software-private-credit-loans-2026.
The outside evidence is piling up around the admission. S&P Global found EBITDA add-backs now account for up to 30% of EBITDA in deals versus roughly 10% a decade ago, with a direct correlation between add-back size and failure to service debt; JPMorgan has clamped down on lending against private credit collateral, imposing its own marks on fund NAVs; and MSCI data showed 10% of private credit loans marked down by at least 50% jpmorgan-markdown-software-private-credit-loans-2026. The market is answering the fiduciary question before regulators do — JPMorgan AM is launching a daily-NAV interval fund and Apollo has committed to daily mark-to-market pricing, betting that transparency becomes the asset class's next competitive moat rather than its liability.
What to watch: Whether daily-NAV structures migrate from differentiators to table stakes across the semi-liquid channel — the funds still gating at 5% have the most stale marks to defend.
Regulators map the insurance plumbing — and the number is $1.2 trillion
The insurance channel is now the most active regulatory front, with Senator Warren's September 10 letter to the NAIC drawing a 22-page response that defended state oversight while quietly detailing a tightening apparatus. The NAIC's own figures put US insurer private credit holdings at roughly $1.2 trillion at end-2025 — about 13% of cash and invested assets — with $544 billion in privately rated bonds, BDCs and funds flagged as warranting particular attention pe-owned-insurers-private-credit-short-sellers-2026.
"Rather than relying on a static regulatory framework, regulators have regularly updated capital requirements, reporting standards, supervisory tools, and analytical capabilities to address emerging risks." — pe-owned-insurers-private-credit-short-sellers-2026
The most consequential measure is a framework under development to assess whether private letter rating providers' methodologies remain appropriate for regulatory purposes — the most direct move yet against the PLR arbitrage that underpins the Athene/Global Atlantic model. Moody's puts private and illiquid holdings at $807 billion with the ten largest holders accounting for 44%, so the concentration question is live pe-owned-insurers-private-credit-short-sellers-2026.
What to watch: Whether the NAIC's PLR methodology framework lands with capital-charge changes attached — that combination would hit the economics of the insurer channel directly.
The hybrid CLO gets its first price test
The structural bridge between private and public credit markets is about to get a market-clearing price. Blackstone aims to price its first private credit–BSL hybrid CLO in October as an all-euro deal, with at least half a dozen similar issuances believed in the pipeline from Golub, Apollo and Blackstone private-credit-terms-spread-compression-covenant-lite-2026.
"BSL loans are priced daily in liquid secondary markets, providing a continuous external price signal that anchors the valuation of the private credit portion of the portfolio." — private-credit-terms-spread-compression-covenant-lite-2026
That quote is the tell: managers are importing public-market price discovery into private portfolios voluntarily, before regulators force the issue — a contractual answer to the same valuation critique Zito conceded. A clean euro print accelerates the pipeline; a wide, poorly distributed one would suggest public investors see the valuation risk too.
What to watch: The October pricing — spread versus conventional BSL CLOs is the market's clearing price for private credit risk in a rated, tradeable wrapper. (Carried over from last cycle: this is the same Blackstone hybrid thread.)
What surprised us
- The default-rate debate got a second, sharper axis. Beyond the soft-default composition problem from last cycle, Houlihan Lokey now puts the default rate below 1% on a size-weighted basis versus Fitch's record 6.3% — a gap of more than six percentage points where the number you quote says more about your positioning than the credit private-credit-default-rates-software-systemic-risk-2026
.
- The queue is self-perpetuating by design. Most Q3 redemption requests were investors resubmitting unfulfilled prior-quarter asks — the 5% cap doesn't just slow outflows, it manufactures the appearance of ongoing demand for liquidity evergreen-private-credit-redemptions-liquidity-gating-2026
.
- Apollo's ADS fund is outperforming while gating — Class I net return of 8.2% since inception, beating leveraged loans by 177bps and high yield by 377bps. The retail crisis and the product's relative performance are not the same story evergreen-private-credit-redemptions-liquidity-gating-2026
- Private credit may be eyeing your wallet next. A new thread this cycle: WSJ reports private credit firms are exploring partnerships with JPMorgan to underwrite rejected applications on co-branded credit cards — the first serious signal of expansion from corporate lending into consumer receivables via bank partnerships.
Open threads worth a vote
- Private credit's move into consumer credit: JPMorgan co-branded card partnerships — help steer whether to chase confirmation of the structure and counterparties behind the WSJ report, or wait for the convergence to announce itself.