Pebbles.

What four private credit CEOs said before seven platforms gated in 33 days

Mar 26, 2026

On February 5, 2026, Blue Owl Capital CEO Marc Lipschultz sat on his quarterly earnings call and told investors the fund had “lots and lots of liquidity.” He said the team had “met all investor requests for tenders as we have every quarter since inception.” He said the firm had “a very favorable view for 2026.” Fourteen days later, Blue Owl halted redemptions.

On February 9, Apollo Global Management CEO Marc Rowan told his quarterly call that business was “firing on all cylinders.” He reported record combined fee-related earnings and spread-related earnings of $5.9 billion. He used the phrase “principal’s mindset” eight times in a single call and told investors “we’re eating our own cooking.” His co-president, Jim Zelter, was asked directly by Goldman Sachs analyst Alex Blostein about rising redemptions and slowing gross sales in the non-traded BDC market. Zelter’s response was dismissive: “That’s about pebbles. What Marc is talking about is boulders. Don’t miss the bigger picture.” Forty-three days later, Apollo capped redemptions at 5% of net asset value and returned 45 cents on every dollar investors asked for back. Pebbles.

On February 26, FS KKR Capital CEO Dan Pietrzak told his earnings call that the “vast majority” of the portfolio “continues to perform in line with expectations.” He used the phrase “legacy investments” repeatedly to explain away deteriorating credit metrics, distancing KKR from the problems in the portfolio as though someone else had made the bad loans. Twenty-five days later, Moody’s downgraded FSK to junk.

And back in November 2025, Ares Management CEO Michael Arougheti assured investors on the Q3 earnings call that the credit events they were seeing “appear to be idiosyncratic and isolated.” He said there was “no credit cycle.” About 120 days later, Ares Strategic Income Fund turned down more than half the redemption requests its investors submitted, filling just 43% of what was asked for.

Seven platforms gated or capped redemptions within a 33-day window: Blue Owl, Apollo, Ares, BlackRock, Blackstone, Cliffwater, and Morgan Stanley. The CEOs who told you everything was fine are now returning less than half your money. And the one who called it “pebbles” appeared at Bloomberg Invest one day after gating and said, with no apparent irony, “It was foreseeable. It was predictable.”

---

What I Did

I built an independent stress analysis of the $3.5 trillion private credit market — the kind of cross-fund, network-level risk assessment that does not exist on any sell-side desk or in any consultant’s model. The sell-side covers funds individually. Rating agencies score creditworthiness. Nobody models the interconnections between funds, the contagion pathways created by overlapping portfolios, or the cascade mechanics that turn isolated defaults into systemic events. That is what this analysis does.

I cross-referenced SEC filings for 29 business development companies covering over $200 billion in assets, mapping portfolio overlaps, PIK exposure, leverage ratios, and NAV discounts to build composite risk scores for each fund. I identified 1,564 companies sitting in two or more fund portfolios and 378 contagion edges connecting fund pairs through shared borrowers. I built a Monte Carlo cascade simulation with 10,000 iterations, calibrated with default rate data from Fitch, Moody’s, and Morgan Stanley, and recovery rate data covering 25 years of middle-market lending. I then read 38 earnings call transcripts spanning 2021 through 2026 across these platforms to understand how management characterized the risk while the financial data was deteriorating beneath them.

What the financial data shows is unambiguous: defaults are rising, recovery rates have collapsed, 636 companies cannot pay cash interest on their loans, and the system is far more interconnected than it appears from the outside. What the transcripts show is that management at every platform that gated was telling investors the opposite of what the data indicated, in some cases just days before the gates came down. The financial analysis is the thesis. The transcript evidence is the confirmation that the people running these platforms either did not understand the risks or chose not to communicate them.

---

The Gating Scoreboard

Here is what happened across seven platforms in roughly one month:

|---|---|---|---|---|---|

| BlackRock | HPS | $26B | 9.3% | 54% | ~Mar 20 |

| Apollo | ADS | $25B | 11.2% | 45% | Mar 24 |

| Ares | ASIF | $11B | 11.6% | 43% | Mar 24 |

| Cliffwater | CCLFX | $33B | 14% | Capped | Mar |

| Morgan Stanley | NHPIF | $8B | 10.9% | Capped at 5% | Mar |

The combined AUM across these seven funds is approximately $159 billion. These are not illiquid micro-cap vehicles. These are the largest asset managers on the planet — Blackstone, Apollo, BlackRock, Ares, Morgan Stanley, Blue Owl — running funds they marketed as offering quarterly liquidity. Cliffwater’s $33 billion fund saw the highest redemption rate of any platform at 14%. The investors who put money into these funds were told they could redeem up to 5% of their shares per quarter. In March, they asked for 9-15%. They received less than half.

It is worth pausing on what this means in practice. If you invested $1 million in one of these funds and submitted a redemption request for the full amount, you received between $430,000 and $540,000, depending on the platform. The rest remains locked in the fund. You will have to resubmit your request next quarter and hope the fill rate improves. In the meantime, the underlying loans that your capital is invested in continue to age, and the borrowers on the other end of those loans are the same companies that drove the stress in the first place.

Apollo noted in its shareholder letter that it expects to apply the same 5% cap next quarter. It also disclosed that the $730 million in gross outflows from the redemptions was almost exactly offset by $724 million in new inflows — meaning net flows were approximately zero. New money is still coming in the front door while existing investors are being told they cannot leave through the back.

---

What Management Was Saying While the Numbers Deteriorated

The financial data tells you what happened. The transcripts tell you something arguably more important: what investors were being told while it happened. I analyzed the earnings calls of all four platforms that gated, and what I found was not ambiguity or nuance — it was a systematic pattern of reassurance that intensified as the underlying stress worsened. At EventHorizonIQ, I have spent the past two years building a framework called TonalityIQ that scores management language across 20 calibrated dimensions, tracking conviction shifts, deflection patterns, and structural risk signals across earnings calls. Applied to these four platforms, the framework identified three patterns that repeated identically across every one of them.

The first is what I call the Reassurance-Collapse Sequence. In every case, peak reassurance language appeared on the final earnings call before the gate dropped. The confidence does not gradually fade as stress builds. It does the opposite — it accelerates into the break, as though the need to reassure investors intensifies in direct proportion to the severity of the underlying problem.

|---|---|---|---|---|

| FSK | Feb 26 | “vast majority performing” | 25 | Moody’s junk downgrade |

The calls closest to the gate contain the strongest reassurance language, not the weakest. This is counterintuitive if you assume management is being candid. It makes perfect sense if you assume management is managing perception.

The second pattern is Deflection. When pressed on specifics by analysts, each management team redirected blame to something smaller than itself. FSK blamed “legacy investments” — assets predating KKR’s involvement — while KKR-originated non-accruals jumped from 3.4% to 5.1% in a single quarter. The problems were not legacy. They were current, and they were KKR’s own underwriting. Apollo’s Zelter told analysts to “focus on the $40 trillion” fixed-income replacement opportunity, as though the size of the addressable market had anything to do with whether investors could get their money back from a specific fund. Ares said the “majority of repurchase requests were made by a limited number of family offices and smaller institutions in select geographies who represent less than 1% of our over 20,000 shareholders” — an extraordinary sentence that manages to blame the sellers, minimize their significance, and imply that the 99% who did not redeem have made the smarter choice, all in a single breath. Blue Owl’s Lipschultz offered: “the trends we observed within Blue Owl’s credit portfolios remained strong and did not align with the headlines or investor fears.” That was fourteen days before halting.

The deflection targets vary — legacy assets, family offices, headlines, addressable markets — but the function is the same in every case: shrink the problem, externalize the cause, and imply that the people asking questions are the ones who are confused.

The third pattern is Inevitability Language, and it is the subtlest and most damning. Buried in prepared remarks or analyst Q&A on the same calls that contained the strongest reassurance language, each platform also introduced words and phrases that quietly pre-accepted failure. FSK said “we do believe some level of default is inevitable in a sub-investment grade portfolio” — true in the abstract, but when spoken on the same call where you are also telling investors the vast majority is performing, it functions as an advance excuse. Apollo’s Rowan, on the February 9 call, referenced “increased probability of outcomes outside of established lanes or an established playing field” — a hedge embedded inside a presentation full of record numbers and optimistic guidance. Then, on March 25, one day after gating ADS, Rowan appeared at Bloomberg Invest and said: “It was foreseeable. It was predictable.”

Foreseeable and predictable. But 43 days earlier, when his co-president was asked about it by Goldman Sachs, the answer was dismissive. Those two positions cannot both be true. Either the stress was visible in early February and investors were told otherwise, or it genuinely was not visible and Rowan’s post-gating framing is revisionist. Neither interpretation is flattering. Both are important for anyone trying to calibrate how much to trust management language going forward — at Apollo or anywhere else in private credit.

---

The Hard Numbers: FSK’s Seven-Quarter Deterioration

FSK provides the clearest deterioration arc because it is publicly traded, files quarterly with the SEC, and I analyzed seven consecutive earnings calls from Q4 2022 through Q4 2025 — enough to track the full trajectory from healthy to junk-rated. The numbers tell the story without embellishment:

|---|---|---|---|---|---|---|

| Q4 2022 | $24.89 | ~2.5% | ~2.0% | ~105% | $0.80 | $0.70 |

| Q3 2024 | $23.82 | Improving | Improving | ~110% | $0.74 | $0.70 |

| Q4 2024 | $23.55 | ~3.5% | ~2.5% | ~112% | $0.72 | $0.70 |

| Q1 2025 | $23.19 | ~4.0% | ~2.8% | ~115% | $0.68 | $0.64 |

| Q2 2025 | $22.04 | ~4.5% | ~3.0% | ~116% | $0.64 | $0.64 |

| Q3 2025 | $21.99 | 5.0% | 3.4% | ~116% | $0.60 | $0.64 |

| Q4 2025 | $20.89 | 5.5% | 5.1% | 122% | $0.48 | $0.57 |

The trajectory is unambiguous. Net asset value has declined 16% from its peak. Non-accruals at cost have more than doubled from approximately 2.5% to 5.5%. KKR-originated non-accruals — the loans that KKR itself selected and underwrote, not positions inherited from prior management — jumped from 3.4% to 5.1% in a single quarter, a 50% increase that demolishes the “legacy investments” narrative that Pietrzak relied on throughout the call.

Net leverage stands at 122% debt-to-equity. The regulatory ceiling is 125%. That is three percentage points of headroom. One bad quarter of asset markdowns erases it. If leverage breaches 125%, FSK faces a choice between forced asset sales in an illiquid market — which would crystallize losses and push NAV down further — and raising equity at a massive discount to book value, which dilutes existing shareholders. Neither option is attractive. Both are worse than what the transcript language suggests management is preparing investors for.

Net investment income per share has dropped 40% from its peak, falling from $0.80 to $0.48, and management guided Q1 2026 NII down another 6% to $0.45. The dividend has been cut twice. When an analyst asked about the timeline for recovery, Pietrzak said: “I don’t think this is a quarterly sort of discussion as we work on that evolution... I would caution to say that it’s not an overnight thing. It will take some time... we’re talking about things over kind of a longer period, which I would call sort of ‘26 and ‘27.” An 18 to 24-month workout. That is management telling you, in the clearest language available on a public earnings call, that the fund will be in a stressed state for at least the next two years.

Moody’s saw the same math. They downgraded FSK to junk on March 23-24, 2026.

FSK now trades at approximately a 50% discount to its stated NAV. The market is telling you that it does not believe the book value is real. When a publicly traded fund trades at half its stated asset value, the market is not pricing a temporary sentiment issue — it is pricing the possibility that the underlying loans are worth materially less than what the fund says they are.

FSK is to this cycle what New Century Financial was to the subprime crisis — not necessarily the largest exposure, but the entity where the deterioration is most advanced, most visible, and most difficult to reverse. New Century didn’t cause the 2008 crisis. It was the place where the crisis became undeniable. FSK, with its junk rating, its 122% leverage approaching the 125% regulatory ceiling, its KKR-originated non-accruals jumping 50% in a single quarter, its 18-24 month workout timeline guided by its own CEO on a public earnings call, and its stock trading at half its stated book value, occupies the same position in this cycle. My Monte Carlo simulation independently identified it as the number one cascade trigger, appearing in 49% of stress scenarios. Whether FSK survives in its current form is a question for KKR and its shareholders. What is not in question is that FSK is where the stress in private credit is most concentrated and most measurable.

---

The Broader BDC Stress Picture: Risk Scoring 29 Funds

FSK is not an isolated case, and this is where the analysis shifts from a single-fund deep dive to a market-wide risk assessment. I built a composite risk scoring framework that evaluates every publicly filing BDC across four financial dimensions: PIK exposure (weighted 30%), Level 3 illiquid asset concentration (30%), NAV discount to market price (25%), and cross-fund portfolio overlap (15%). Each fund receives a score from 0 to 100, calibrated against the historical performance of BDCs that ultimately failed or were absorbed — including Allied Capital and American Capital, both of which were acquired by Ares Capital after extended periods of stress. The framework identified six funds currently in SEVERE territory and thirteen in ELEVATED:

|---|---|---|---|---|

| FDUS | 85.7 | -- | 96 | SEVERE |

| GBDC | 79.9 | -15.4% | 268 | SEVERE |

| NMFC | 78.0 | -38.2% | 164 | SEVERE |

| FSK | 73.4 | -56.6% | 591 | SEVERE |

| OBDC | 71.9 | -26.2% | 108 | SEVERE |

| BBDC | 71.2 | -- | 240 | SEVERE |

| ARCC | 67.1 | -9.0% | 418 | ELEVATED |

| BXSL | 51.6 | -7.4% | -- | ELEVATED |

FSK has the highest PIK mention count in the dataset at 591 — nearly six times the count at OBDC, the Blue Owl subsidiary whose parent halted redemptions on February 19. PIK mentions in SEC filings are a proxy for the scale of a fund’s exposure to borrowers who are paying interest in IOUs rather than cash. The more a fund discusses PIK arrangements in its quarterly and annual filings, the more borrowers it has that cannot service their debt from operating cash flow. A PIK mention count of 591 means FSK is discussing payment-in-kind arrangements on nearly every page of its filing. That is not a handful of troubled names. That is a portfolio-wide phenomenon.

Across the 29-fund dataset, 636 companies are currently in PIK status. Of those, 24 appear in multiple fund portfolios simultaneously, meaning a single borrower default ripples across several funds at once. This is the interconnection that makes the system more fragile than it appears when you look at any individual fund in isolation. Private credit is marketed as uncorrelated — each fund underwrites its own deals, manages its own book, makes its own decisions. The reality is that 1,564 companies sit in two or more fund portfolios, and the 24 PIK companies in multiple funds create direct contagion pathways between funds that are supposed to be independent.

---

Recovery Rates: The Silent Killer

The default rate gets the headlines. The recovery rate is doing the real damage, and almost nobody is talking about it.

| Period | Average Recovery Rate |

|---|---|

| 25-year middle market average | 63% |

| Recent 12-month realized | 41% |

| PIK/distressed companies | ~30% |

| Stress case (Envision Healthcare) | 20% |

When a borrower defaults and you recover 63 cents on the dollar, the loss is manageable — painful for the specific position but absorbable at the fund level. When you recover 41 cents, the arithmetic changes entirely. That 22-percentage-point collapse in recovery rates means that every default hurts roughly 50% more than historical models assume. A fund that was underwritten to withstand a 6% default rate with 63% recovery finds itself in trouble at a 6% default rate with 41% recovery because the loss-given-default has increased by more than half.

Why are recovery rates falling? The answer is leverage. The average private credit loan today is structured at higher leverage multiples than it was a decade ago. When a leveraged company fails, the debt stack is deeper, the enterprise value available to distribute among creditors is lower, and first-lien lenders recover a smaller percentage of their principal. The 25-year recovery average of 63% was built during an era when leverage was lower and lenders had more cushion beneath them. That era ended around 2020-2021, when the flood of capital into private credit pushed leverage higher and covenant protections thinner. The recovery rates we are seeing now — 41% realized over the most recent twelve months, and as low as 20% in severe cases like Envision Healthcare — are likely the new normal for the duration of this cycle.

---

The Monte Carlo: Independently Estimating What Nobody Else Is Measuring

This is the core of the analysis. The gating scoreboard tells you what already happened. The FSK deterioration tells you the trajectory of the weakest major fund. The PIK count tells you how many borrowers are not paying cash interest. But none of that answers the question that matters most: how bad can this actually get for the $3.5 trillion private credit market as a whole, and what is the probability that it becomes systemic?

To answer that, I built a Monte Carlo cascade simulation — 10,000 iterations modeling default, recovery, and cross-fund contagion across the full market. The simulation is calibrated entirely from publicly available data. I did not assume my way to a conclusion.

The calibration inputs are drawn from current market data and historical analogs. Default rates are modeled at 6% for a normal environment (the current Fitch trailing twelve-month rate), 12% for a recessionary environment (consistent with the 2008 leveraged loan peak of approximately 11%), and 20% for a crisis environment that includes concentration shocks and cascading failures. Recovery rates are modeled using a Beta distribution with means of 50% in normal conditions, 35% in recession, and 25% in crisis — all below the 25-year average of 63%, reflecting the leverage-driven compression described above. Companies currently in PIK status receive a 3.5x multiplier to their base default probability, derived from the empirical observation that 15-25% of PIK companies default within 24 months versus a base rate of approximately 6%. When a fund’s cumulative losses exceed 15% of its AUM in the simulation, it enters stress mode and begins forced selling at a 20% discount to fair value, which propagates mark-to-market losses to overlapping funds through the 378 contagion edges in my cross-fund portfolio map.

| Metric | Value |

|---|---|

| Median loss across 10,000 simulations | $395 billion |

| 95th percentile loss | $914 billion |

| Probability of losses exceeding $500 billion | 24.9% |

| Fund identified as #1 cascade trigger | FSK (appears in 49% of stress scenarios) |

| Average number of funds entering stress per simulation | 4.2 |

The most likely outcome — carrying roughly 50% probability — is between $200 billion and $400 billion in cumulative losses across the private credit market. That is severe for the sector. It means multiple funds gate or restructure, BDC stocks decline 30-50% from current levels, and several platforms that entered this cycle overleveraged and overexposed to software and healthcare roll-ups do not survive in their current form. But it stays contained within private credit. Banks survive. Pension funds take marks in their alternatives allocations but do not face solvency questions. This is a private credit recession, not a financial crisis.

But there is a one-in-four chance that losses exceed $500 billion, and at that level, the dynamics change qualitatively. The simulation shows that cascade mechanics — where one fund’s forced selling marks down overlapping portfolios at other funds, triggering their own covenant breaches and forced selling — add 30-40% to total losses compared to a no-cascade scenario. The system is more interconnected than it appears from the outside. When I mapped portfolio overlaps across all 29 funds, I found 1,564 companies sitting in two or more fund portfolios, with 530 companies in three or more. A single large borrower default does not hit one fund — it hits three, or five, or eight simultaneously. And when one of those funds is forced to sell at a discount, the marks on that borrower’s loans change across every fund that holds the same name.

One in four is not a tail risk you ignore. It is a scenario you prepare for.

---

The PIK Time Bomb

There are 636 companies currently in PIK status across the BDC complex, and this number is the single most important datapoint in this entire analysis. PIK — payment in kind — means the borrower is not paying cash interest on its loans. Instead, the interest is being added to the outstanding loan balance. The lender books income it has not collected in cash. The borrower’s debt grows. The lender’s stated portfolio yield remains high even as the actual cash coming in the door declines.

Historical data from two decades of BDC performance shows that 15-25% of companies that enter PIK status ultimately default within 24 months. That conversion rate is approximately 3.5 times the base default rate — PIK is not a benign accommodation extended to a borrower going through a temporary rough patch. It is a leading indicator of distress, and the historical data on this point is unambiguous.

The current PIK vintage entered status primarily in 2024 and 2025. The 24-month conversion clock on those companies runs out in Q3 and Q4 of 2026. At a 20% conversion rate — the midpoint of the historical range — approximately 127 of those 636 companies will default. At the current compressed recovery rate of 41 cents on the dollar, those defaults translate into real, material losses concentrated in the Q3 and Q4 2026 filing season.

This is not a prediction based on a view about the economy or interest rates or AI disruption. It is arithmetic applied to historical conversion rates. The companies are already in PIK. The clock is already running. The only question is whether the conversion rate lands at the low end of the historical range (15%, meaning roughly 95 defaults) or the high end (25%, meaning roughly 159 defaults). Either number produces meaningful losses across a sector that is already gating redemptions and trading at deep discounts to stated book value.

---

Could This Take Down the Economy? And Do We Have the Tools to Stop It If It Does?

This is the question the piece has been building toward, and it deserves a direct answer rather than an implied one.

The most likely outcome — carrying roughly 50% probability in my simulation — is that private credit goes through a painful but contained repricing. Defaults rise, several funds restructure or merge, BDC stocks take another leg down, and the sector emerges smaller and more disciplined on the other side. That is not a financial crisis. It is a credit cycle doing what credit cycles do, and it would be healthy for a market that grew from $500 billion to $3.5 trillion in less than a decade without ever being seriously tested.

But the 25% tail is not 1%. One in four is a probability that demands preparation, not dismissal. And the reason it demands preparation is not just the size of the potential losses — it is the fact that the tools the government used the last time around do not apply to this market.

In 2008, the entities at the center of the crisis were banks. Banks are regulated by the Federal Reserve. The Fed can provide emergency liquidity through the discount window. The Treasury, under TARP, injected capital directly into bank balance sheets. The FDIC guaranteed deposits. These tools — lender of last resort, direct capital injection, deposit insurance — were designed for a banking crisis, and they worked for a banking crisis.

Private credit is not a banking crisis. The entities at the center of this stress are business development companies, non-traded funds, and alternative asset platforms. They are regulated by the SEC, not the Fed. There is no discount window for a BDC. There is no TARP mechanism that allows Treasury to inject capital into a non-traded fund. There is no deposit insurance equivalent for an investor who put money into Apollo Debt Solutions or Ares Strategic Income Fund. When a BDC gates, the investors are on their own. The fund decides how much to return and when. There is no backstop.

This is not a criticism of Treasury Secretary Bessent or incoming Fed leadership — it is an observation about the architecture. The regulatory tools were built for the last crisis. Private credit grew up outside the perimeter those tools were designed to cover. If the 25% tail scenario materializes and cascading losses propagate through the 378 contagion edges connecting these funds, the question will not be whether policymakers want to intervene. It will be whether they have the legal authority and the mechanical infrastructure to do so. As of today, the honest answer is that the playbook does not exist.

The $718 billion in pension fund allocations to private credit makes this everyone’s problem, not just the investors who bought non-traded BDCs. State pension systems — teachers, firefighters, municipal workers — have been increasing their private credit allocations for years on the premise that these were safe, income-generating, diversified investments. FSK’s South Carolina joint venture, backed by a $39 billion state pension, just increased its commitment. If the tail scenario plays out, the people affected will not be family offices in select geographies. They will be retirees who never heard of a business development company and did not choose to have their pension allocated to one.

And here is the political problem that makes this different from 2008, and potentially worse. In 2008, the government bailed out the banks. It was enormously unpopular — Occupy Wall Street was the direct consequence — but the argument was that the banking system was too interconnected to let fail, and the alternative was a complete collapse of the payments and credit infrastructure that the entire economy runs on. The public hated it, but there was a logic to it, and there were tools to execute it.

Now imagine the tail scenario arrives in private credit. Losses cascade. Pension funds take marks. The gatings spread. Who gets bailed out? Apollo Global Management, a firm that just reported $5.9 billion in earnings? KKR? Ares? Blackstone? The idea that Main Street taxpayers would accept a rescue package for the largest private equity firms on the planet — firms whose executives are among the wealthiest people in finance — is politically inconceivable. It would make the TARP backlash look quaint. No elected official would survive voting for it.

But the alternative is not clean either. If you let it burn — if you say these are sophisticated investors who knew the risks, and the private equity industry can absorb its own losses — then the pension funds burn too. The teachers in South Carolina burn. The firefighters in whichever state allocated 12% of their retirement fund to private credit burn. The retirees who never made the decision and cannot absorb the loss are the ones left holding it. Have you ever driven on the interstate and seen someone who started a fire in their backyard to burn a pile of trash, except the wind picked up and now it is a brush fire headed for the neighbor’s property? That is the risk of letting the private credit sector sort itself out without a plan. The fire may have started on Wall Street’s property, but the wind does not respect property lines, and $718 billion in pension exposure means the neighbors are already downwind.

So who should be getting ahead of this? I do not have a definitive answer, and I am not sure anyone does, which is itself part of the problem. But I have some thoughts on where to start.

The Financial Stability Oversight Council — FSOC — exists for exactly this purpose. It was created after 2008 to identify systemic risks before they metastasize. FSOC has the authority to formally assess whether private credit constitutes a systemic risk and to designate specific platforms as systemically important financial institutions, which would bring them under enhanced Federal Reserve oversight. In fact, FSOC met yesterday — March 25, 2026, the day after the Apollo and Ares gatings — and the readout confirms they “addressed recent developments in the private credit sector” as part of their quarterly financial stability briefing. They also voted to publish revised guidance on the process for designating nonbank financial companies, the very mechanism that could bring BDC platforms under enhanced oversight. So the apparatus is aware and the tools are being discussed. But there is a significant difference between monitoring developments in a quarterly briefing and conducting a formal systemic risk assessment with the urgency that seven gatings in 33 days warrants. The former is routine. The latter requires a decision that the situation has escalated beyond routine.

SEC Chairman Atkins oversees the regulatory framework for BDCs. The SEC could require enhanced stress test disclosures from any BDC above a size threshold — the way the Fed requires annual stress tests for large banks. Right now, BDCs self-report their risk metrics, and as the transcripts in this piece demonstrate, management characterizations of those metrics can diverge significantly from reality. Mandatory, standardized stress testing would give investors and regulators a common fact base rather than relying on each platform’s chosen narrative.

The SEC could also mandate disclosure of cross-fund portfolio overlaps. The contagion graph I built — 378 edges connecting 29 funds through 1,564 shared borrowers — was assembled from public filings, but it required reading and cross-referencing hundreds of SEC documents that no retail investor and few institutional investors would ever assemble on their own. If the SEC required BDCs to disclose their top portfolio overlaps with other funds in a standardized format, the interconnection risk that currently hides in plain sight would become visible to every investor before they commit capital. The information is public. The assembly of it into a usable picture is not. That is a gap the SEC can close with a rule change, not an act of Congress.

For non-traded funds specifically — the ones that just gated — the most practical reform would be minimum liquidity buffers. If you market quarterly liquidity to retail investors, you should be required to hold enough liquid assets to meet a reasonable redemption scenario. The fact that seven platforms marketed 5% quarterly redemptions and then could not meet requests of 9-15% means the liquidity promise was structurally unsound from the beginning. A minimum buffer requirement — say, liquid assets sufficient to cover 15% of NAV in any given quarter — would either force funds to hold more cash or stop marketing liquidity they cannot deliver. Either outcome protects investors.

The right response is not panic. It is planning. A 25% probability event with no existing containment framework is a gap that should be closed before the event arrives, not after. The data to assess the risk exists — it is in the SEC filings, in the contagion graph, in the PIK conversion arithmetic. What does not yet exist is the institutional will to look at it clearly. I built this analysis with publicly available data. Everything in this piece — the risk scores, the contagion edges, the PIK counts, the Monte Carlo calibration — can be independently verified by anyone with access to EDGAR and a spreadsheet. The information is not hidden. It is ignored.

---

The Apollo Arc: When the Largest Platform in Private Credit Gated Its Own Fund

Apollo Global Management is the most important case study in this analysis — not because it is the weakest fund (FSK holds that distinction), but because it is arguably the strongest and most sophisticated platform in private credit. Apollo built the machine exactly as it said it would. It reached $938 billion in AUM. It crossed $300 billion in origination. It posted record earnings. And then it gated its own flagship non-traded BDC 43 days after its co-president called the stress “pebbles.” If Apollo is gating, the question is no longer whether private credit has a problem. The question is how deep the problem runs. I read through every available Apollo earnings call and investor presentation from Q2 2021 through Q4 2025, including the October 2021 Investor Day — a five-hour marathon that laid out the strategic vision that Apollo has since largely executed.

In October 2021, Marc Rowan stood on stage and set the theme for the next five years: “Consistency over heroism.” Jim Belardi, Athene’s CEO, told the room: “The reality is we just don’t lose much money in either a recession or a deep recession.” Rowan reinforced the message in the broadest possible terms: “No one decision in the context of this business is consequential. This is a fully diversified business.” He said Apollo would reach $1 trillion in AUM. He said origination would grow from $80 billion to $150 billion annually. He said fee-related earnings would compound at 18% annually for five years.

By almost every metric, Apollo delivered. AUM reached $938 billion. Origination crossed $300 billion — double the original target. FRE hit $2.5 billion for 2025, up 23% year over year. They built the machine exactly as promised. The question the transcripts raise is not whether Apollo built a great business — it did. The question is what happens to a $938 billion platform when the machine’s own products cannot return investors’ capital on schedule.

On February 9, 2026, forty-three days before gating, the language on Apollo’s Q4 2025 earnings call reached peak confidence. Rowan called the business “firing on all cylinders.” He reported record earnings. He said the firm was “eating our own cooking” and used the phrase “principal’s mindset” eight times. On ADS specifically — the $25 billion non-traded BDC that would gate 43 days later — Zelter was unequivocal: “Approximately 8% return, lowest leverage, top of the capital structure, large company, no PIK. I assure you, ADS is on offense.”

When Goldman’s Alex Blostein pressed directly on rising redemptions in the non-traded BDC space, Zelter delivered the line that, in my assessment, will define this cycle: “So much of the conversation in the last several weeks has been about a quick reset of prices in the public equity market and then the impact to non-investment grade software lending, which has been particularly focused well by the non-traded BDCs... I urge all of you on this phone call, focus on the $40 trillion. That’s the opportunity set... I know it’s great headlines about the software bump in the road, but there’s a bigger, bigger picture and don’t miss that.”

Pebbles and boulders. The non-traded BDC market — the exact product category that would gate 43 days later — was dismissed as a distraction from the real story.

On March 24, Apollo capped ADS redemptions at 5% of net asset value, filling 45% of the $1.2 billion in requests from investors who wanted their money back. The fund that was “on offense” returned less than half the money investors asked for. Bloomberg reported that Apollo had to double a credit line to $1 billion and sign a new $500 million facility to manage liquidity — actions consistent with a fund that was not, in fact, managing pebbles.

On March 25, one day after gating, Rowan appeared at Bloomberg Invest in New York and offered this assessment: “It was foreseeable. It was predictable.”

Foreseeable and predictable. But 43 days earlier, it was pebbles.

And then, this morning — as I am writing this — Bloomberg reported that Zelter appeared at a conference in Melbourne and offered a third framing. The stress was no longer pebbles. It was no longer foreseeable and predictable. It was now a communication failure: “Certain distribution channels in certain parts of the globe may not have fully communicated the risks.” The blame has migrated from the market (pebbles, not worth worrying about), to management’s own foresight (foreseeable, we saw it coming), to the sales team (distribution channels didn’t explain it properly). Three positions in 43 days from the same person about the same event. Blue Owl co-CEO Doug Ostrover, appearing alongside Zelter, echoed the framing: “Between us, and the advisers who sell our products, I don’t think we made it clear enough.” This from the executive whose firm said “lots and lots of liquidity” fourteen days before halting redemptions.

I have read and re-read the transcripts, the post-gating remarks, and now the Melbourne comments. They are irreconcilable. The deflection keeps moving outward — from the market, to hindsight, to the distribution teams — but the underlying fact does not change: seven platforms gated $159 billion in 33 days, and on the last earnings call before each one broke, management told investors the opposite of what was about to happen.

---

The Risk I Might Be Wrong About FSK

There is a scenario where FSK’s trajectory reverses, and intellectual honesty requires me to lay it out clearly. Finian O’Shea — the Wells Fargo analyst widely considered the axe on BDC coverage — asked the key question on FSK’s Q4 2025 earnings call: “Do you ever think about a grand bargain — lower fees plus a KKR balance sheet injection?” He also noted that Blackstone had restructured its fee arrangement with BXSL and “the stock was fine,” implying that FSK could do the same.

If KKR announced a meaningful capital injection, a fee waiver or restructuring, and a strategic commitment to rebuilding the platform, FSK’s stock could rally 20-30% and the bearish thesis would face a significant challenge.

My read of the transcripts suggests this outcome is more unlikely than likely. Pietrzak’s response to O’Shea’s question was to cite FSK’s historical IRR of 9.1% since inception and say “we’ve been happy with the partnership.” He did not address the structural question O’Shea was actually asking, which was: FSK is shrinking, permanently stuck below book value, and the current fee structure creates a misalignment between KKR’s economics and shareholder outcomes — at what point does something have to give? When the top analyst in the space asks the existential question and management pivots to a backward-looking metric, that is a tell. It is the equivalent of answering a question about the future by talking about the past.

KKR has also been growing its own direct lending platform aggressively — Apollo’s ADS, Ares’s ASIF, and Blackstone’s BCRED are all direct competitors to FSK, and KKR has its own products in the space. The incentive to inject billions of dollars of balance sheet capital into a subscale, junk-rated vehicle with a declining NAV when you could deploy that capital in your own proprietary platform is, at best, unclear.

But it is a risk. And if you are making investment decisions based on a bearish view of FSK, it is the primary risk to that thesis, and you should size your position accordingly.

---

The Forward Risk Framework: What to Monitor and When

I am not making a prediction. Predictions in credit markets are less useful than monitoring frameworks, because the variables that determine whether stress stays contained or cascades are observable in real time if you know where to look. What follows is the framework I am using to track whether the private credit stress that produced seven gatings in 33 days is a painful but contained sector repricing, or the early stages of something that bleeds into the broader financial system.

The second quarter of 2026, April through June, is the decision quarter. This is when 10-Q filings drop for the first quarter, and the single most important data point in those filings is PIK counts by fund. If PIK counts jump 20% or more across the top ten BDCs, the “idiosyncratic and isolated” narrative that Arougheti introduced in November — and that every other platform has echoed since — dies a public, data-driven death. Q2 also brings the second gating test, as quarterly tender windows reopen in April. If redemption requests are higher than they were in Q1 — above the 9-12% range we saw in March — platforms that returned 45 cents on the dollar last quarter face an impossible choice: gate harder and risk trapping investors in a cycle of escalating withdrawal attempts, or fill more and drain the liquidity that keeps the fund operational. Enterprise software earnings in April and May — not the hyperscalers, but the mid-cap and legacy names that private credit actually lends to, companies like those in the IGV basket and below — will set the AI disruption narrative for the quarter. If guidance cuts at these companies reference AI-driven automation reducing customer spend, seat compression, or contract downsizing, the private credit thesis around software exposure hardens further. These are the borrowers sitting in BDC portfolios. Their earnings are the leading indicator, not Google’s.

The third quarter, July through September, is when the PIK conversion wave arrives. The 24-month clock on 2024-vintage PIK toggles starts converting. Borrowers who were granted PIK status in 2024 either resume paying cash interest or get reclassified as non-accruals. Historical conversion rates suggest 15-25% will not make the transition. If the Federal Reserve has not cut rates by July, the refinancing window that PIK companies need to roll their debt stays shut, and the clock runs out. This is the quarter when the first large wave of non-accrual reclassifications should appear in BDC filings.

By the fourth quarter, we will know which scenario we are in. Either the sector has stabilized — credit metrics in the Q2 and Q3 filings showed improvement, redemption requests normalized, and the gating wave of March 2026 becomes a case study rather than a prologue — or the deterioration has continued, additional platforms have gated, and the question has shifted from “is private credit stressed?” to “who holds the exposure?”

The single most important number to watch between now and then is the Q2 2026 10-Q PIK count across the top ten BDCs. If it jumps 20% or more from current levels, the probability of the tail scenario in my Monte Carlo — losses exceeding $500 billion — shifts from 25% toward 40% or higher.

---

*Eric Jackson is the founder of EMJX, EMJ Capital, and EventHorizonIQ, a structural intelligence platform that tracks 130+ real-time sensors across credit, equity, crypto, and cross-asset regimes — including cross-fund contagion modeling, Monte Carlo stress simulation, and regime detection. His TonalityIQ system has analyzed thousands of earnings transcripts across hundreds of companies. This private credit analysis was generated using TonalityIQ applied to 38 transcripts across four platforms spanning 2021-2026, combined with a 10,000-iteration Monte Carlo cascade simulation and cross-fund contagion mapping of 29 BDCs.*

*Previous long calls: Carvana at $15 before the 32x run, Opendoor at $0.73, IREN at $9, CIFR at $3. He is also long CVNA, OPEN, IREN, CIFR. EventHorizonIQ’s regime detection system identified the BTC regime shift to STRESS on January 26, 2026, eleven days before the February crash. Eric and EMJ Capital are short FSK and OWL through put options and are financially incentivized for those stocks to decline. EMJ Capital may short other BDC-related names in the future. Positions can change at any time. Do your own work.*

*For the full data behind this analysis — 29-fund risk scores, the contagion graph with 378 fund-pair edges and 1,564 overlapping portfolio companies, Monte Carlo distributions, and quarterly TonalityIQ pattern updates — visit eventhorizoniq.com. To learn more about EMJ Capital, visit emjcapital.ltd. To learn more about EMJX, visit emjx.ai.*

*Track record: eventhorizoniq.com/scoreboard*

*This is not investment advice. This is structural analysis. Do your own work.*

EventHorizonIQ

Discussion about this post

Ready for more?