Pebbles III.
$112 Billion Gated This Week. Bloomberg Calls It Prudent.
On March 26, I wrote that the next gating test would arrive when Q1 10-Q filings dropped between April and June. This week, three of the largest private credit vehicles in the world capped investor redemptions. On June 4, Blackstone Private Credit Fund — the largest non-traded BDC in the world at $79 billion in assets — capped at 5% after investors requested 10%. Three days earlier, Cliffwater capped at 5% after investors requested 17%. And on June 3, Switzerland’s Partners Group capped withdrawals on its $8.6 billion fund — shares closed down roughly 17% on the announcement. Combined: $120.6 billion gated across three funds in two regions, in a single week.
Bloomberg Intelligence — in a credit primer published the morning of the BCRED gate by senior credit analyst David Havens — called it “prudent fiscal management” in the June 4 Credit React publication.
That phrase is the entire piece. The institutional consensus voice — the analyst whose job is to read the same SEC filings I read, the same earnings calls I parsed, the same risk metrics I scored — looked at a $79 billion fund refusing to return investor capital and called it prudent.
I want to walk through what Bloomberg Intelligence accidentally documented in defending it.
---
The Reassurance-Collapse Sequence — Bloomberg Edition.
In Pebbles I (March 26 2026), I identified a pattern I called the Reassurance-Collapse Sequence: peak reassurance language from BDC management appears on the final earnings call before the gate drops. Per the Pebbles I parse of those calls (full transcripts cited in that piece), Blue Owl Capital co-CEO Marc Lipschultz described the fund as having ample liquidity on the February 5 2026 Q4 2025 earnings call, 14 days before Blue Owl OTIC halted redemptions on February 19. Apollo Global Management CEO Marc Rowan described the business as firing on all cylinders on the February 9 2026 Q4 2025 earnings call, 43 days before Apollo Debt Solutions capped redemptions at 45 cents on the dollar on March 24. The pattern was about insiders managing perception.
What I did not expect was to see the same pattern play out at the institutional analyst level. The Bloomberg Intelligence Credit Primer on BCRED — authored by senior credit analyst David Havens with secondary analyst Nicole Castelblanco — published three takes on Blackstone Private Credit Fund in roughly 75 days:
*Bloomberg Intelligence Credit Primer publications on BCRED, March-June 2026 (David Havens, lead analyst; titles and framing language quoted directly from the BI primer):*
| Date | BI Primer Title | BI Primer Frame |
|---|---|---|
| March 20 | As Private-Credit Gloom Settles in, BCRED’s Spreads Elevate | Bond spreads may have overshot; spreads may be topping out. |
| March 23 | All-Weather BCRED Can Manage Through Private Credit’s Tempest | Many of its strengths intact; nonaccrual rate half that of peers. |
| June 4 | BCRED Lowers Gates as Redemption Requests Hit 10%: Credit React | Prudent fiscal management. |
That is the arc of an institutional defender losing the argument. The March 20 BI primer defended the bond price (spreads may have overshot). The March 23 BI primer defended the portfolio (strengths intact). The June 4 BI primer defends the gate itself.
When the institutional consensus voice stops defending the price and starts defending the gate, the bear thesis has won the argument. The only thing unresolved is the timeline.
I built TonalityIQ to detect this drift in CEOs. Bloomberg Intelligence just published it about itself.
BI is not the only consensus voice doing this. The Wall Street Journal published a piece in the same window titled Why Blackstone and BlackRock Can Ride Out the Private-Credit Storm — the second-largest financial publication framing the gating as a survivability narrative (ride out the storm) rather than a portfolio-quality question. Two institutional defenders, same week, same defensive frame: the funds are large enough to survive. That is a solvency argument, not a portfolio-marks argument, and the bear thesis I am laying out concedes the solvency point. BCRED is not going to default tomorrow. The bear thesis is about the marks, the redemption-acceleration mechanic, and the principal-agent dynamic between the fund and its adviser. The consensus defenders argue against the strongest version of the wrong question.
---
Five Bear Receipts Bloomberg Intelligence Published in Their Own Defense.
The most damaging thing about the BI primer is not what Havens says. It is what the data inside the primer says when you apply his own analytical framework against it. The five receipts that follow are all sourced from the June 4 Bloomberg Intelligence Credit Primer publication on BCRED, with direct quotes attributed to Havens unless otherwise marked.
Five receipts:
One: PIK trajectory 5.3% → 7.2% over four quarters.
BCRED’s payment-in-kind income rate moved from 5.3% in Q1 2025 to 7.2% in Q4 2025 — figures sourced from the BI primer. That is a 35% relative increase in four quarters. The rate of change matters in Havens’ own analytical framework. In the March 23 2026 Bloomberg Intelligence publication on BCRED, Havens writes: “A high, or rising, rate may indicate gathering portfolio stress as leveraged borrowers resort to PIK to alleviate cash flow strain.”
His framework. His warning. His data. He just refuses to apply the framework to his own thesis.
In Pebbles II I documented that 1,020 holdings across 18 BDCs were currently on PIK, with historical conversion rates of 15-25% to default over 24 months. BCRED specifically being on the steepening end of that trajectory is the structural confirmation that the PIK pipeline is doing what the historical data says it does.
Two: 25% software exposure, 7 percentage points above peer mean.
BCRED has the highest software-segment exposure of any large BDC at 25% of holdings, per the BI primer. In the same March 23 2026 Bloomberg Intelligence publication on BCRED, Havens writes: “This may be problematic, at least in terms of perception, as the software sector may be at risk of disruption from the adoption of potentially less costly AI-based alternatives. This is a real risk for lenders, as some of their portfolio companies could sustain existential threats or declining valuations.”
For a credit book, what Havens calls a risk of disruption is not a perception problem. It is a fundamental cash-flow problem. Borrowers losing pricing power because their software is being replaced by AI-based alternatives means EBITDA compression means covenant breaches means PIK amendments means non-accruals. That is the mechanical pipeline I described in Pebbles II. BCRED’s largest single-sector concentration sits in the AI-disruption crosshairs, on a credit book at 0.74x leverage that targets 0.9-1.25x. The fund either grows the book back into target leverage at higher rates against a deteriorating sector or it accepts permanent shrinkage.
Three: Bond spreads 150bp wide of triple-B finance peers.
BCRED’s bonds are rated Baa2/BBB-. Its adviser Blackstone Inc. is rated A+/A+. And yet BCRED’s bonds trade 150bp wide of triple-B finance comparables (BI primer, March 20 2026). Havens frames this in the March 20 BI primer as market skepticism that may eventually position the bonds for spread compression.
The bond market has been voting on these credits since 2H 2025. It has not changed its vote. Havens is asking readers to believe that the bond market is wrong about the marks; the bond market is saying the marks are wrong. One of them updates eventually. The bonds are real-money positions held by institutional credit investors who can sell tomorrow if they want to. They are not holding because Havens told them to. They are holding because they cannot exit cleanly. The wide spread is the cost of that constraint.
Four: Blackstone Inc. extracted $1.18 billion in fees from BCRED in 2025 — 13% of BX fee income, 8% of total BX revenue.
This is the line in the BI primer that does the most work for the bear thesis, and it is the line Havens does not develop.
BCRED paid Blackstone Inc. $1.18 billion in management and incentive fees in 2025 (BI primer, March 20 2026). Blackstone Inc. has no explicit support agreement with BCRED. Havens notes in the BI primer that implied support is strong, citing shared branding and strategic importance in BCRED’s private wealth growth strategy. That is not contractual support. That is reputational alignment as long as the math works.
When a fund pays the adviser 8% of the adviser’s total revenue, the adviser has every incentive to keep the fund’s NAV high, every incentive to mark assets generously, every incentive to gate redemptions to preserve the AUM that pays the fee. There is no contractual mechanism preventing this. There is only reputational risk, and reputational risk is a soft constraint.
This is the principal-agent confession Havens publishes without recognizing. The fund’s adviser is the entity that benefits most from delayed loss recognition. The fund’s adviser is also the entity producing the marks. The fund’s adviser is also the entity that decides whether to bridge redemptions with executive capital.
In Q1 2026, Blackstone Inc. and key Blackstone employees put $400 million of their own money into BCRED to meet 7.9% redemption requests. In Q2 2026, they did not do that, and the fund gated at 5%. The change in loyalty signal is the diagnostic. When the parent and its executives stop bridging the fund’s liquidity gap with their own capital, they are telling you what their internal probability is on the fund’s recovery.
Five: Average mark 96.1¢, bottom 5% marked 68.3¢.
BCRED’s overall portfolio is marked at 96.1 cents on the dollar as of April 30. The bottom 5% of the private debt portfolio is marked at 68.3 cents.
Either the bottom 5% is wrongly marked too low, or the rest of the book is wrongly marked too high.
The bond market is voting for the latter at 150bp wide of triple-B finance peers. The historical recovery data is voting for the latter at 33 cents on the dollar for defaulted private credit. The PIK trajectory is voting for the latter at 5.3% → 7.2% in four quarters. The redemption requests are voting for the latter at 10% of NAV.
The average mark is the smoothed lie. The tail mark is the truth.
---
The Loyalty Signal Withdrew.
I want to expand on the executive-capital point because it is the single most important new datapoint in the BI primer and it does not appear in Pebbles I or Pebbles II.
In Q1 2026, BCRED faced redemption requests of 7.9% of NAV against a stated cap of 5%. Apollo’s ADS fund, in the same window, faced 11.2% and filled 45%. Ares Strategic Income Fund faced 11.6% and filled 43%. BlackRock HPS faced 9.3% and filled 54%. Every other gated platform pro-rated and returned less than what was requested.
BCRED met all of it. 100% fill rate. The Q1 2026 BCRED letter, which I have now re-read alongside the BI primer, attributes this to executive participation: Blackstone Inc. itself committed $250 million and key Blackstone employees committed $150 million to bridge the redemption excess. $400 million of insider capital was deployed to keep the fund’s brand promise of meeting investor requests intact.
That is a loyalty signal of the highest order. It is also a non-repeatable maneuver. You cannot extract $400 million from your own balance sheet and employees every quarter to defend the gate. The Q1 fill was the marketing exception, not the operating rule.
In Q2 2026, with 10% requested and only 5% allowed, the executives did not put their capital in again. The gate closed. Brad Marshall and Blackstone Credit leadership made the rational decision to preserve their own capital and accept the bad-PR hit of gating. The decision tells you what they think the next 12 months look like.
When the people closest to the asset stop putting their own money in, you should not be putting yours in either. That is the simplest sentence in this entire analysis.
---
The Contagion Mechanic Activated.
In Pebbles I and Pebbles II I described two contagion mechanics. The first was cross-fund overlap: the same companies appearing in multiple BDC portfolios such that one fund’s markdown forced auditor questions at every other fund holding the same name. The second was redemption acceleration: investors who could not fully exit a gated fund this quarter would resubmit larger requests next quarter, plus their gating experience would generate redemption requests at sibling funds where investors had not yet tried to exit.
Both mechanics activated this week.
Three gates in seven days. BCRED, Cliffwater, and Partners Group, in that order — combined AUM $120.6 billion. Combined redemption requests: 10%, 17%, and elevated respectively. Combined fill rate: 5% each, well below requested in every case.
The Cliffwater Corporate Lending Fund (CCLFX) is a $33 billion fund that did not appear on most retail BDC watchlists six months ago. The fact that it received 17% redemption requests in a single quarter — more than three times its stated 5% policy — confirms that the redemption pressure is not a Blackstone-specific or Apollo-specific phenomenon. It is the channel.
Partners Group, a Swiss-listed private markets firm, capped withdrawals on its $8.6 billion fund on June 3. Shares closed down roughly 17% on the announcement (per Bloomberg / Crypto Briefing / Bitget reporting; equity-market verdict, not a quote) — that’s the market verdict in real time. When an institutional manager gates and the stock drops 17%, that is not market confusion. That is investors marking down the entire platform’s franchise value because the gate signals what the fund manager believes about the next 12 months of liquidity. The inclusion of a European-listed name in this week’s gating cluster means the contagion is no longer just a US non-traded BDC channel story. It is the global private credit complex.
Bloomberg’s June 4 reporting (Olivia Fishlow) captures the forward dynamic clearly: “Across the $1.8 trillion private credit market, redemption requests are expected to increase this quarter as investors redouble efforts to claw back money after being restricted.”
Restricted investors generate larger future requests. That is not a hypothesis. It is mechanically what happens when retail and family-office investors learn their fund will not honor full redemption requests. They submit larger requests next quarter to try to get a higher pro-rated share. The pro-rating creates the incentive that compounds the pressure.
This is how an isolated event becomes a channel event becomes a sector event.
---
