The implication every time is the same. The Fed was slow. Read the signal late. Waited for the smooth middle of the distribution to confirm what the tails were already screaming.

Tuesday at the Senate hearing he asked for the job. Called for “regime change in the way the Fed conducts policy.”

Then he told you what regime change looked like to him.

Trimmed means. Look through tariffs. AI productivity to carry the disinflation forecast.

Every item bends the same direction. Smoother read. Slower reaction. Lower bar to cut.

Request Access

What the Dallas Fed actually says

First, the setup.

Trimmed mean PCE is an inflation measure that ranks every item in the consumer basket each month, throws out a fixed percentage from the top and the bottom, and averages what is left. The idea is to filter outliers and reveal the underlying trend. Core PCE, on the other side, is the Fed’s main inflation gauge. It strips out food and energy because those bounce around. Both filters. Different numbers in different regimes.

The Dallas Fed published a paper on April 16 whose conclusion creates an immediate problem for Warsh’s preferred framework.

Trimmed mean PCE lagged core and headline in 1974. Lagged again in 2021. The trimmed mean cuts the top and bottom of the price change distribution. When that distribution skews positive, the measure strips more signal off the top than the bottom. What comes out is a downwardly biased read on the underlying trend.

The Dallas Fed says it in plain language. When skewness flips from negative to positive, the trimmed mean can mislead.

The distribution is skewing positive right now. Tariffs in core goods are doing the work. Non-housing core services have shown some acceleration in early 2026 and are becoming a larger source of upside outliers. The Dallas Fed flagged the combination explicitly. They added a second problem on top. Gradual price increases from tariffs may not register as outliers at all. The trimmed mean does not strip them out. It absorbs them slowly into the trend.

The institution that builds the index is warning that the index can understate inflation risk in this regime. Warsh is leaning on it anyway.

Look at the gap

Trimmed mean PCE is running at 2.3% through February. Core PCE is at 3.0%.

That is a 70 basis point gap that did not exist before September of last year. The two measures tracked each other tightly from 2023 through September 2025. Then they diverged. Core kept rising. Trimmed mean kept falling.

The gap is not noise. It is the asymmetric trimming the Dallas Fed warned about, showing up in real time. Tariff pass-through in core goods is creating positive skew. Abrupt moves get trimmed out. Gradual moves bleed in slowly. The measure is weak as an early-warning signal in this regime.

Bessent already moved

Scott Bessent was publicly calling for rate cuts not long ago. He recently acknowledged near-term cuts are inappropriate.

The question is whether Warsh pivots before or after he is confirmed. The Banking Committee votes April 29. Powell’s term ends May 15. That is the window.

The front end is not pricing this asymmetry. The base case for next week’s FOMC is hold, soften forward guidance, two 25 basis point cuts later this year. That same consensus also concedes core PCE is stuck at 3 percent.

The market is pricing both. That path requires energy fading, labor softening, and core rolling toward 2 percent on schedule. Without those, one side gives. Whichever way Warsh leans tells you which side gives.

The AI productivity argument is showing up in credit as inflation

Warsh has leaned on AI productivity as part of the case for a lower-rate path, even though he tempered the claim at the hearing.

Look at the leveraged finance pipeline.

Roughly $13 billion of data center high yield issuance hit in April alone. That is around 40% of total April high yield supply. Core Scientific brought a $3.3 billion deal at 99.25 OID. Edged Compute brought $1.3 billion. The BB-quality HY bid and IG crossover demand absorbed all of it without a hiccup.

The capex that promises disinflation is inflationary on the way in. The capex is happening now. The power demand is here. The labor competition for the workers building the infrastructure is too. The productivity gains are years out.

Warsh is using a multi-year disinflation narrative to justify a near-term rate cut framework. That timing mismatch is the problem.

The case for dovish-Warsh, and why it fails

For the dovish Warsh trade to work, most of these have to break the right way within twelve months.

AI productivity hits in six to twelve months, not three to five years. Tariffs prove one-off rather than persistent. Energy normalizes by Q3 with the Strait reopening cleanly. Non-housing core services rolls over rather than accelerating further.

That is a parlay, not a base case. Several require resolutions that are not in motion. The front-end base case requires the parlay to hit. That is a favorable path through several unresolved shocks, not a clean base case. The market is pricing it as if it were.

What this means for credit

Read more

That’s the only question worth asking and management keeps answering a different one. DOGE. Tariffs. Budget pressures. Blame it on the cycle and offer a $2bn buyback program to make shareholders happy.

The equity market isn’t buying it and for good reason. Here’s Gartner’s Magic Quadrant, the product that built the brand. November 2025, generative AI model providers. IBM ahead of Anthropic. That’s all you probably need to know.

The stock is down more than 70% from its highs including 38% this year alone. Roughly $20bn of market cap vaporized. All due to multiple re-rating. Two separate single-day moves north of 20% after earnings.

Meanwhile, the bonds have barely budged. The 2028s at 140 over, the 2031s at 150, the 2035s at 195 bps. Held by sleepy long-only IG investors running a screen that still shows sub-1x net leverage, 70%+ FCF conversion, and an IG rating. A healthy business on the spreadsheet.

But is it really?

The Business Model

Gartner (NYSE: IT) is a ~$6.5bn revenue enterprise technology research and advisory firm, headquartered in Stamford, CT, with ~21,000 employees globally.

The business model works like this: pay an annual fee per seat, somewhere between $20,000 for basic research access and $80,000-plus for senior advisory, and whoever holds that seat can read Gartner reports, call a Gartner analyst, tap into peer networks, and attend Gartner conferences.

You know the magic quadrants every company references? That’s them.

The product is basically credentialed human expertise on technology decisions, originally aimed at IT leaders and, since the 2017 CEB acquisition, extended into HR, finance, legal, procurement, marketing, and supply chain.

The company operates under three segments, each with different economics.

• Insights: the subscription research engine. ~77% gross margins, ~90% of gross profit. ~14,000 enterprise clients across GTS (tech vendors) and GBS (functional leaders outside IT).

• Conferences: ~50 events a year, anchored by Symposium/Xpo. Blended ~45% gross margins. Revenue from attendee tickets plus sponsor fees from vendors paying to reach the audience.

• Consulting: small, project-based, contracting. ~27% gross margins. Contract optimization, sourcing, and benchmarking work sold into existing Insights accounts.

Contract Value (CV), the annualized book of active subscriptions, closed FY’25 at roughly $5.2bn. Clients pay annually and upfront, which coupled with minimal capex (<2% of sales), translates into robust FCF generation (70%+ FCF conversion).

What's Been Happening

The last 18 months at Gartner have been rough. CV growth, the metric the whole business ladders up to, has decelerated every quarter for four straight quarters. FY’24 exited at 8%. Q1’25 printed 5%. Q2 came in at 3%. Q4 at 1%, well below what consensus had modeled. Four consecutive misses against a progressively lower bar.

The equity has repriced accordingly. IT is down more than 70% from its FY’24 high of ~$584, with most of the damage concentrated in two single-day drops of 20%-plus following Q2’25 and Q4’25 earnings.

The multiple expansion that drove the stock through 2022-2024 has fully unwound and now hovers around ~7x NTM EBITDA.

Management’s explanation has been consistent and cyclical:

• DOGE: federal contract cancellations running at dollar retention near 50%, with the bulk of the renewal cohort concentrated in FY’25.

• Tariffs: 35-40% of the client base sits in tariff-exposed industries where capex cycles have stalled.

• Macro: cost-cutting across the enterprise base has stretched sales cycles and pushed existing accounts to trim seats at renewal.

The numbers beneath the CV headline are worse than the headline. GTS wallet retention, which measures how much existing clients spend at renewal versus the prior year, has fallen from 108.8% in FY’21 to 97.5% in FY’25. A book that used to quietly expand inside every account is now, in aggregate, shrinking inside them.

Inside that number, the small tech vendor cohort is the single largest drag. Gartner has called out small tech vendors as the primary driver of client count decline for multiple quarters in a row, with retention in that cohort running in the low-to-mid 70s.The deterioration started in FY’22 and has been uninterrupted since. DOGE and tariffs are FY’25 problems.

Management’s response has been to double down on capital return rather than change the business. FY’25 buybacks hit nearly $2bn against ~$1.2bn of FCF, the largest annual repurchase in company history, funded partly by drawing down cash and using incremental debt capacity. Another $1.2bn of authorization remains.

On recent calls, management has pointed to AskGartner adoption and analyst productivity gains as evidence the product transformation is working and CV will reaccelerate through FY’26-27.

The salesforce situation makes the setup worse. Quota-bearing headcount (QBH, the seat-selling reps that carry the renewal and new-logo targets) was reduced through FY’25, which pulled down average rep tenure at the same time the book turned.

Gartner reps take three to four years to ramp. Even assuming a cyclical recovery in FY’26, the salesforce that would need to capitalize on it is less tenured than the one that navigated prior cycle troughs. Any rebound will be shallower than historical comparisons suggest. Unsurprisingly, salesperson productivity has declined in lock-step.

Here’s what the bond market is saying.

Read more

This one is very clean. Non-traded BDCs are facing the biggest redemption wave of their existence. They’re also sitting on $140 billion of broadly syndicated loans. BSLs are liquid, marked to observable prices, and can be sold in days. So that’s where the BDCs go to meet redemptions. Forced selling hits the tape, BSL prices dislocate in specific names for technical reasons, and public credit takes the hit.

Indiscriminate selling. The kind of thesis that reads well in a strategist note and gets repeated at every credit roundtable for six months running.

The question is whether it’s actually showing up in prices.

So I pulled the data.

The Mechanics Are Worth Walking Through

The entire thesis rests on BDCs being forced sellers of BSLs, and before getting to whether they are, it’s worth laying out what a non-traded BDC actually does when a tender window opens. Most commentary skips the part where a BDC’s funding hierarchy absorbs redemption pressure long before it ever reaches the Level 2 bucket, and that’s where the thesis tends to overreach.

A non-traded BDC facing a tender window has to meet redemption requests within 90 days, capped at 5% of NAV under standard repurchase policies. The cash comes from six places, roughly in order of cost.

The first source is organic portfolio cash flow. Interest payments, loan maturities, and prepayments generate meaningful liquidity every quarter.

OTIC disclosed $1.3bn in liquidity against a $6bn book. OCIC reported $11.3bn. Industry-wide portfolio cash flow has historically run around 8% per quarter, comfortably above the 5% redemption cap. That historical average may understate current conditions though, given rising PIK, slower prepayments, and extension activity.

Cash on hand comes next, generally 2-5% of NAV at most vehicles.

After that, revolver capacity. Most BDCs carry significant undrawn capacity on their credit facilities, but each draw pushes leverage toward the 1.0-1.25x internal cap these vehicles operate under. Breaching the cap carries real consequences: ratings pressure, tighter covenants on the next refinancing, reduced NII as interest expense climbs.

New debt issuance sits fourth in the hierarchy and it’s still functioning for the larger names. GSCRED and OBDC both printed notes this week at spreads 30-50 bps wider than where they would have cleared six months ago, but the window is open for BDCs with scale and rated paper.

CLO warehousing is the fifth lever. Private credit CLO issuance is running near record pace this year, roughly $9.5bn YTD, with BDCs retaining meaningful portions of junior tranches across the industry. That extends leverage one level deeper inside the securitization in a way that doesn’t surface cleanly in the headline debt-to-equity ratio, but it generates cash.

Level 2 BSL sales come sixth. This is where the contagion thesis lives.

Beneath that sit Level 3 sales, which is the ugly option because secondary bids for private credit paper sit below par and selling at 85 or 90 cents means booking losses that flow into NAV. Suspending redemptions is the break-glass last resort.

Nobody’s forecfully selling yet. At least not meaningfully below par.

The implication of the hierarchy is that BSL selling only begins once the upstream levers are substantially exhausted. If portfolio cash flow, revolvers, debt issuance, and CLO warehousing are still absorbing the redemption pressure, the transmission into public markets simply doesn’t activate. The question of whether the mechanism is live becomes, more precisely, a question of whether Q1 pushed those upstream levers far enough.

Q1 Put Real Pressure On Those Levers

The entire non-traded BDC industry either capped or stretched at the 5% redemption limit in the same quarter, which is a different situation from the episodic pressure individual vehicles have faced in prior cycles.

Back-leverage terms have tightened meaningfully across the major banks, with spreads widening 50-150 bps over the past six months. JPMorgan, Goldman, and Barclays are exercising markdown rights on collateral that most sponsors assumed would never actually be used in practice.

Fundraising has stalled alongside all of this. Direct lending raised $10.7bn in Q1 against $27.7bn in Q4, which removes the flywheel dynamic where new AUM historically funded distributions in the background while the portfolio matured.

None of the top four levers have exhausted. But they’re all under real pressure simultaneously, and the pressure is compounding rather than dissipating. If redemption requests persist through this year and next at anywhere near the Q1 pace, the buffers thin out and the marginal cost of each successive draw rises.

Which shifts the question from whether BSL selling could eventually happen to what it would look like when the relief valve opens. That turns on what BDCs are actually holding in the Level 2 bucket.

The Level 2 Bucket Isn’t A Random Slice Of The BSL Market

BDCs don’t accumulate Level 2 positions randomly. They hold BSLs that overlap with their private origination footprint, which skews the bucket toward the same sectors where those BDCs lend privately. Software, business services, and healthcare services are disproportionately represented in the Level 2 book relative to their weight in the Leveraged Loan Index, and the concentration is meaningful enough to matter when the selling starts.

The problem with that concentration is that it’s already adverse. These are the sectors where CLO managers have been pulling back on new exposure for several quarters, for all the AI-disruption reasons the market has been working through since the fall. Which means the marginal BSL seller (a BDC meeting redemptions) meets the marginal BSL buyer (a CLO that’s already trying to reduce exposure to the same sectors) in precisely the credits where both sides are most stressed.

Prices gap rather than drift when the buy and sell sides are concentrated in the same names, and a BDC balance sheet problem becomes a BSL market problem at that point. From there the situation compounds, because the BSL market doesn’t absorb concentrated selling passively. It has a buyer base with hard structural constraints that can shift from absorber to amplifier faster than most of the contagion commentary has acknowledged.

Why CLOs Can Flip From Absorber To Amplifier

CLOs hold roughly 70% of the BSL market, which makes them the structural buyer, but the extent to which they absorb forced selling or amplify it depends heavily on how constrained their balance sheets are when the selling starts. CLO managers don’t buy unlimited quantities at any price. Overcollateralization (“OC”) tests require the collateral pool to maintain minimum par value relative to the debt tranches, CCC baskets cap exposure to distressed credits, and weighted average price tests constrain overall portfolio composition.

A CLO that buys a stressed loan at 85 cents and holds to maturity is generally fine on expected P&L. But if that loan gets downgraded or other loans in the portfolio get marked down, the OC test can fail, which triggers cash diversion away from equity holders and toward paying down senior debt. Managers running tight OC cushions can’t absorb forced selling even when they’d like to, and when the selling concentrates in specific sectors where CLO exposure is already elevated, the constraint binds faster than it would in a more diversified selling wave.

The cascade from there is fairly predictable.

Imagine BDCs sell BSLs into weakening CLO demand. Prices widen in the concentrated sectors. OC tests tighten. CLO managers stop adding and start rebalancing out of the weakest names. BSL widens further. New CLO issuance slows because the arbitrage math breaks at wider spreads, and CLO warehouses that were loaded in Q4 2025 at tighter levels get stuck holding paper that can’t price cleanly into a deal.

The knock-on effect reaches past secondary prices. Private credit borrowers who were planning to refinance their PC debt into BSL, which could be a natural release valve for stressed PC loans throughout this cycle, find the venue either closed or too expensive. They get pushed back into PC-internal solutions: amend-and-extend transactions, PIK toggles, liability management exercises. Which increases PC stress, which increases BDC redemption pressure, which forces more BSL selling.

Private credit stress transmits to public credit through the BDC BSL sleeve, and public credit weakness reflects back into private credit by closing the BSL refinancing channel. The loop feeds itself once it starts, which is why the question of whether it’s started matters more than the question of whether the mechanism could exist in principle.

The Thesis Is Clean. The Question Is Whether The Data Agrees.

The argument above sounds good on paper. The mechanism is coherent. The funding hierarchy is real. The redemption pressure is real. The sector concentration is real. The CLO feedback dynamics are real.

If the mechanism is live, the evidence should show up in BSL prices for the specific credits that BDCs are most likely to be selling. The concentrated sectors should be widening faster than the aggregate index. Names that appear across multiple BDC Level 2 books should be underperforming otherwise-comparable names that don’t.

That’s testable.

So I picked a non-traded BDC with meaningful BSL exposure, pulled up its BSLs, and priced each one at three specific points: December 31, 2025 as year-end, March 3, 2026 as the leveraged loan market low, and April 20, 2026 as current. Then I compared the movement of those names against the aggregate leveraged loan index and the technology subindex, to see whether the BDC-held names were moving with the broader market or showing concentrated selling pressure beyond what the index reflected.

Here’s what I found.

Read more

Going into March, Avis was a deteriorating but not yet broken credit. Net leverage around 7.5x. Negative FCF in 2025 of ~$700mm. Bonds widening but not distressed. A 2027 maturity and a management team in the middle of a turnaround that hadn’t yet shown results.

The 4Q print made it worse. EBITDA missed by $100mm+. DPU blew out to $338 against a sub-$300 guide. A $518mm impairment on the EV fleet. The 2026 guide came in $150mm below where anyone was modeling. The business had not stabilized and the credit trajectory was clearly lower, not higher.

That was the setup 8 weeks ago. Coming out of that print, the consensus framing was straightforward. Avis would be in trouble if things didn’t turn, and the turn wasn’t visible in the numbers.

The stock is now up 300%+ from the March lows. The bonds have tightened materially. Management has filed an ATM shelf that’s about to take out most of the 2027 maturity wall through an equity raise. None of that was in the consensus credit model 8 weeks ago.

Worth walking through how we got here, because the answer isn’t a fundamental improvement.

Request Access

What actually happened

In late February, Pentwater Capital disclosed a 10% stake in Avis. An April Form 4 showed the direct stake had grown to 22%, plus synthetic exposure through cash-settled total return swaps. Combined with SRS Investment Management’s 49% position, two holders controlled over 70% of the shares directly, and more than 100% once the synthetics were layered in.

Short interest at 54% of the float. Section 16(b) kept the sellers on the sidelines. The squeeze math was visible in the filings by late March.

The stock roughly doubled from late February through early April on the ownership math alone, moving from around $97 to $190 before Iran escalated on April 7. Everything Avis did beyond the rental car tailwind was the squeeze.

A few things about the Iran conflict worth unpacking.

First, the domestic travel substitution. Americans who had planned European trips are expected to start rebooking domestically, which is supportive for airport rental car demand. Avis’s US business captures that directly. Not enormous in dollar terms, but directionally positive for 2Q and 3Q rental days and RPD.

Second, the used car price channel. Crude up meaningfully pushes input costs higher on new vehicles, which supports used vehicle residuals and therefore Manheim. The March Manheim rental risk index was already up 6.2% y/y, the best since July 2023. Iran escalated that trend. For Avis, higher residuals flow through DPU, which is the single biggest line item on the rental car P&L. A sustained Manheim tailwind matters more than the rental day volume.

Third, the macro backdrop. The Iran ceasefire last week opened the HY primary market to its biggest week since September 2020. Stressed paper got bid aggressively. The window for a stressed issuer to execute an opportunistic equity raise is as open right now as it’s ever been.

Add it up and Iran is unambiguously good for Avis on the fundamentals. Rental days up modestly, DPU down modestly, risk appetite up meaningfully. But the fundamental improvement is maybe $100-200mm of EBITDA if the ceasefire holds, against 2026 guidance of $800mm-$1bn. Useful but not enough to fix a credit with 7.5x leverage and a 2027 wall.

The pattern is mechanical. Shorted company with concentrated ownership sees a technical squeeze. Iran adds a geopolitical tailwind that makes the equity move more durable. Retail attention compounds it. Market cap expands faster than fundamentals justify.

Management, paying attention and not being stupid, files an ATM shelf and drips equity into strength. A $500mm raise at current prices deleverages the company by 0.5-0.7x and takes out most of the 2027 wall. The recap happens quietly while the bonds follow along. From signal to potential recap, about 60 days.

What the credit profession actually misses

For credit investors, focusing on the left tail is absolutely a rational mindset. You’re buying something with capped upside, so most of the work is figuring out what can go wrong. Recovery math, covenant protection, refinancing risk. That’s the job, and for most credits it’s the right job.

Where the bias shows up is the right-tail probability, and it gets ignored more systematically the closer a bond trades to par. At 40 or 50, distressed investors are already running right-tail scenarios. They have to. That’s the job at that dollar price. At 85 and above, with the credit stressed but not yet distressed, the analyst is still running the near-par playbook. Downside scenarios get modeled. The right tail gets underweighted.

The mechanism isn’t new. AMC, Carvana, and Hertz each went through a version of it. Stressed public company sees an equity catalyst, management issues into strength through an ATM or follow-on, and the proceeds deleverage the balance sheet. The bondholders benefit directly. By the time these outcomes got priced, the bonds were already distressed. Nobody wrote the recap up as a scenario when those names were still stressed but not broken. The framework didn’t include “management opportunistically issues equity into a squeeze” as a path worth weighting until the bonds had already cratered.

Not theoretical zero weight. Actual zero weight. Sell-side recovery work on AMC unsecured in mid-2020 clustered well below par. Avis is different because the recap is happening while the bonds are still near par. Nobody wrote the ATM recap up as a scenario going into the 4Q print. Eight weeks later the bonds are pricing differently, and they never had to go through distress first.

Equity leads, bonds follow

There’s a broader observation sitting underneath the Avis case.

Credit used to be the smart money. The discipline was built on doing careful work on the cap structure, the docs, and the recovery waterfall, and getting paid for that work because equity people weren’t doing it. Bonds moved first. Equity followed.

For many public credits, that has inverted. Equity moves first. Bonds follow. The variables determining the outcome are often equity-market variables. Float, short interest, insider lockups, retail flows, management willingness to issue into strength. The credit analyst modeling DPU and leverage is doing real work, but the work isn’t sufficient when the outcome is being set in equity markets.

AI datacenters are the cleanest current example outside the Avis pattern. The subsector of HY is massively outperforming the broader index YTD. CoreWeave’s 9% ‘31s are up almost 4 points in two weeks on no meaningful change in credit fundamentals. What changed is the equity narrative around AI capex and hyperscaler counterparties. The credit followed the equity.

That inversion doesn’t typically apply to IG or most of performing HY credit. But for a subset of public credits, equity is the leading indicator and credit is the price-taker. That’s a meaningful change in how information moves between markets, and it has consequences for where analytical effort actually produces edge.

The market this week

The same split shows up in this week’s tape. A primary market running hot for risk-on paper, left-tail names still grinding lower, and the biggest single-name moves sitting in distressed credits that got repriced on non-fundamental catalysts.

HY primary printed $16.5bn across 16 tranches. Biggest week in the last 2 years. Meridian Arc’s $5.7bn AI data center deal was the largest AI-driven HY print on record. CCCs led gains. 2026 supply on pace to be up 25% year-over-year. The Iran ceasefire opened a window and stressed paper got bid aggressively.

The primary pipe is telling you risk appetite is back. Meridian Arc clearing at 6.25% secured. CoreWeave printing a $1bn add-on on top of last week’s deal. SoftBank tapping for OpenAI funding. Sealed Air pricing $1.85bn in 8.25% ‘33s. LifePoint Health upsizing to $1.5bn at 7%. These are issuers accessing a market that was half-shut a month ago.

Meanwhile, high yield rallied 66bps last week with B spreads tightening through the week. At 266bps, HY spreads are at their 10th percentile over the last year. Inflows also totaled ~$3bn into HY funds, the largest since June.

Underneath the rally, the 3-month bifurcation shows the entire picture. The top movers are stressed names getting repriced higher on technical or strategic catalysts. NFE up 20 points. Radiate up 13. Community Health up 9. Organon up almost 9 on the Sun Pharma bid.

The bottom movers are different stories entirely. US LBM down 24 on housing-market stress. CNR down 16. ION down 21 on AI displacement fears. CDK down 18 on the same. Software and housing, two sectors with the most distressed paper. See below for a few articles I’ve written on these topics.

What to do with this

You don’t need a dedicated right-tail strategy. Most of you running liquid credit books already have exposure to these types of situations, whether you’ve thought about them this way or not. The point is narrower.

When you’re looking at liquid bonds in a stressed public name, stop systematically underweighting the right tail, especially when the bonds are still closer to par than to distressed. That’s where the framework defaults to downside-only and the right tail quietly falls out.

And if it wasn’t obvious by now, equity-price moves can be a real driver of bond volatility regardless of fundamentals, both on the way up and the way down.

On the way up, that looks like what’s happening at Avis. A squeeze, an ATM, a recap that wasn’t in anyone’s model. On the way down, it’s the mirror: an equity narrative cracking before the credit metrics do, bonds repricing lower because the story changed rather than because the business did.

The AI datacenter names running on hyperscaler momentum are exposed to the same dynamic in reverse if the AI capex narrative shifts. Same mechanism, just a different direction. The credit analyst modeling Avis’s DPU and leverage going into the 4Q print was doing the right work. The analysis wasn’t wrong. It just wasn’t complete for a name where the equity-market variables were going to decide the outcome.

Being confidently short on the back of it was the mistake. The same kind of mistake is available on the long side in names where the equity is doing the lifting and the credit is along for the ride.

Weekly Bulletin Board

1) Wharton Online partnership

I recently partnered with Wharton Online and Wall Street Prep on their Restructuring and Distressed Investing Certificate Program. Relevant timing given the current market backdrop in software and building products. The program covers capital structure, bankruptcy mechanics, and distressed investing, led by practitioners at firms like Silver Point and Ropes & Gray. Code JUNKBOND gets you $300 off, with another $200 off if you enroll by May 11. Starts June 8.

If you liked this, you might also like:

JunkBondInvestor is an independent financial research publication providing general market commentary, corporate credit analysis, and educational content. The information provided is for informational purposes only and does not constitute investment advice, financial planning, or personalized recommendations of any kind. Nothing published constitutes a solicitation, offer, or recommendation to buy, sell, or hold any securities, nor does it guarantee any financial outcome. Any credit opinions or outlooks expressed are solely the independent views of the authors and are intended to reflect general credit trends, not specific investment recommendations. Investing in high-yield bonds, distressed debt, and leveraged loans carries significant risk including potential loss of principal. JunkBondInvestor operates under the Publisher’s Exemption of the Investment Advisers Act of 1940 and applicable state laws. This issue includes sponsored content and partnership links. JunkBondInvestor receives compensation from sponsors for featured placements and may earn referral fees from partner programs. All sponsored content is clearly marked. Editorial views and recommendations are independent and not influenced by sponsor relationships.

Amazon signed the deal on Monday. $11.6 billion. A 117% premium to where the stock traded before the rumors started. A 5x+ bagger from its 2-year lows.

So the trade worked. Now what? Who’s next?

That’s the only question that matters when a deal like this prints. Globalstar has spectrum, a functioning D2D network with Apple, and an independent corporate structure. Amazon wanted all three badly enough to pay 88x EBITDA.

So you run down the list. Viasat. AST. Iridium. NextNav. They’ve all rallied. Everybody with a Bloomberg terminal figured out the same thing at the same time. Satellite spectrum is an asset that can’t just be printed, and the buyers are lining up.

One name kept coming back. Iridium. Global L-band operator. Spectrum block sitting adjacent to the one Amazon just bought. Independent. Publicly traded. The only other player in the world who can claim globally harmonized L-band authorized for safety-of-life services.

I started writing this two days ago. IRDM was at $36. The stock is $43 as I’m writing this. Up 14% yesterday. Up 20% over two sessions. A 52-week high. Every investor who saw the headline ran the same math and the easy entry is gone.

So here’s the real question. Is this a buy? Or is this post 3 days too late?

What Amazon Actually Paid For

Before getting to Iridium, a quick decomposition of the Globalstar deal, because the comp is what the math hinges on.

At ~$12 billion of enterprise value, Globalstar was acquired at roughly 88x current EBITDA. Nobody pays 88x for this operating business. Amazon paid for spectrum and the Apple relationship. The question is how to allocate the $12 billion across the assets in the package.

Globalstar holds three distinct spectrum assets:

• 8.725 MHz of L-band for satellite uplink;

• 16.5 MHz of S-band for satellite downlink; and

• A 11.5 MHz subset of the S-band that’s been authorized for terrestrial use and standardized by 3GPP as Band 53 for LTE and 5G.

Band 53 is the most strategically interesting piece. It’s a rare block of carrier-grade mid-band spectrum not owned by any wireless operator, deployable in 12 countries, with Qualcomm chipset support, Nokia and Airspan radios, and a proprietary 5G RAN platform targeting industrial and defense applications.

Globalstar’s own Analyst Day framed Band 53 against U.S. terrestrial spectrum precedents of $0.21 to $3.16 per MHz-POP, with management pointing to a valuation range of $0.50 to $2.00 per MHz-POP on 10 billion-plus global MHz-POPs. That math implies $5 billion on the low end, $10 billion-plus on the high end.

Worth being honest that these numbers are aspirational. Band 53 has one commercial deployment today (a retailer’s micro-fulfillment center). Anchoring to global MHz-POPs treats footprints that aren’t generating revenue as if they were.

Read more

Howard Marks wrote 6,000 words about how private credit lost its mind. Goldman mailed a shareholder letter that read like a recruitment brochure. Blackstone closed a $10 billion opportunistic fund the same week Moody’s downgraded the sector. Cliffwater published a defense. Bain went on Barron’s to say there’s no bubble.

All in the same week. Not one of them disinterested.

None of them are lying, either. That’s the trick.

Every piece of public commentary was a fundraising document dressed up as market analysis. And the 5% redemption gate, the mechanism trapping billions inside non-traded BDCs right now, is what lets these narratives run at the same time without anyone having to be wrong.

The Gate Creates a Narrative Market

A large subset of private credit managers are now processing redemption queues that will take 2-4 years to clear. Blue Owl’s technology fund saw 40.7% of investors request their capital back in Q1. The larger credit fund saw 21.9%. Barings saw 11.3%. All fulfilled at the 5% cap.

These gated managers aren’t frozen. They have revolvers, CLO warehouse lines, and portfolio cash flow from loan maturities and interest payments. OTIC told investors it has $1.3 billion in liquidity across cash, undrawn debt, and other liquid assets against a $6 billion portfolio.

But the constraint is real even if the capital is technically available. When your debt costs are rising, your hurdle rate on new originations goes up, which means you need wider spreads to hit the same return targets, which means you’re less competitive on deals where an ungated fund can bid tighter.

Boards and risk committees in defensive mode tend to slow deployment even when the liquidity is there. The result isn’t that gated funds stop lending. It’s that they lend more cautiously, at higher internal hurdle rates, with less leverage flexibility, while ungated funds and closed-end vehicles are deploying aggressively into the same pipeline with a structural cost-of-capital advantage.

The queues aren’t shrinking either. Moody’s spelled out the feedback loop in its sector downgrade: investors who were prorated in Q1 will resubmit in Q2 for their remaining balance. Investors who weren’t planning to redeem see the gates go up and submit preemptive requests to get in line. The queue grows even as it’s being processed. At 5% per quarter against a 40% backlog, the math alone takes 8 quarters to clear, and that assumes zero new requests.

That vacuum is what everyone else is positioning around. The public commentary sounds like market analysis when it really functions as a sales pitch.

Goldman Telling You Exactly What’s Happening

Goldman Sachs Private Credit Corp. (GSCRED) was the only major non-traded BDC to come in under the 5% quarterly redemption cap (4.999% to be exact). The shareholder letter didn’t just report this. It built an investment thesis around it.

The fund said the retreat of retail capital will produce a more lender-friendly environment, reversing the supply-demand imbalance that benefited borrowers in recent years. Wider spreads. Stronger documentation.

Why did Goldman avoid the gate? Institutional LPs accounted for more than 80% of their platform’s capital. Long-duration commitments from pensions and sovereign wealth funds that aren’t subject to the same redemption dynamics as retail vehicles. The fund didn’t outperform on credit quality. It outperformed on LP composition.

The bond market is reading the same signals. GSCRED sold $750 million of notes on Tuesday, a day after OBDC sold $400 million of bonds, the first BDC offerings since February. The GS deal priced at 255bps over, about 30 bps tighter than initial price talk. The Blue Owl notes priced at 270bps with a 6.5% yield, rated Baa2/BBB-.

The interesting detail on the Blue Owl deal: PIMCO bought the entire $400 million offering. This is the same PIMCO whose president said last month that the firm was staying away from “pretty bad loans” being sold by funds meeting redemption requests.1 Public caution about private credit loan quality in one breath, $400 million deployed into BDC unsecured debt in the next.

Either way, the deal only exists at these terms because of the stress PIMCO was publicly warning about. Another actor whose public commentary and capital deployment tell slightly different stories.

The Marks Memo Is a Fundraising Document

Howard Marks’ “What’s Going on in Private Credit?” was widely circulated last week. The diagnosis was historically grounded and broadly accurate: direct lending grew too fast, attracted too many inexperienced managers, and produced a goldrush where standards collapsed. Software concentration became the sector’s biggest risk because of how the capital stack evolved, not by accident.

The part worth reading more carefully is the pivot.

After 4,000 words building the case for why the market is broken, Marks shifts to explaining why Oaktree isn’t. Private credit is well under half of their performing credit assets. Direct lending is less than half of their private credit book. Software exposure is “extremely small on an absolute basis and relative to peers.” 80% of their private credit capital is institutional. The memo builds the case for an industry in trouble, then presents Oaktree as the exception.

Oaktree/Brookfield runs opportunistic credit funds that deploy into exactly the kind of stress the memo describes. The memo creates the fear. The fund deploys into the fear. That mechanism is as old as alternative asset management, and it doesn’t make the analysis wrong. But keep it in mind when the memo gets passed around the desk as though it were written by a neutral party.

There’s a detail the memo doesn’t mention. Oaktree Specialty Lending (OCSL) has a negative individual outlook from Moody’s. Marks wrote 6,000 words about the fragility of everyone else’s direct lending book while his own publicly traded BDC is on negative watch by Moody’s.

The Opportunistic Funds Tell the Same Story From the Other Direction

Moody's downgraded the BDC sector last week. The same week, the largest opportunistic credit fundraises in a decade closed.

Blackstone Capital Opportunities Fund V closed at over $10 billion. Ares Special Opportunities Fund III at $9.8 billion. Guggenheim Private Debt Fund IV at $8.4 billion. Adams Street raised $7.5 billion for its 3rd private credit vehicle, more than double its prior vintage. Record numbers, raised in the worst sentiment environment for private credit in a decade.

The Adams Street raise is particularly telling. The fundraise took roughly a year, meaning it stretched across Liberation Day tariffs, the SaaSpocalypse, the Iran war, and the entire BDC redemption wave. Bill Sacher, Adams Street’s head of private credit, said institutional investors are “a little less prone to headlines and a little bit more informed about what’s going on with the underlying investments.” That’s someone who just raised $7.5 billion telling you the BDC crisis didn’t matter to his LP base. About 40% of the commitments came from outside the US.

The pitch for every one of these vehicles is built on the same foundation: direct lending funds are gated, software loans are repricing, spreads are 100 bps wider than they were 18 months ago, and this is the best deployment vintage since 2009. The pitch is probably right. New-issue direct lending spreads have moved from SOFR + 450-475 to SOFR + 550-575. Covenant-heavy volume is up. The data supports the thesis.

What makes it interesting is the circularity. Blackstone runs BCRED, which is under gate pressure. Blackstone also just closed a $10 billion opportunistic credit fund to deploy into the opportunity BCRED’s stress is helping create. Ares runs non-traded BDCs with elevated redemptions. Ares also just closed a $9.8 billion special opportunities fund. These firms aren’t picking a side. The gated fund generates the stress narrative. The opportunistic fund monetizes it. Both vehicles sit on the same platform, managed by teams that share research and deal flow, marketed to different LP bases with different stories about the same market.

The Defense Is Equally Positioned

Cliffwater’s “Ten Misconceptions About Private Debt” published a detailed rebuttal to the negative private credit narrative last week.

They argue credit health metrics remain below historical stress averages: non-accruals at 1.3% versus a 2.1% long-term average, credit losses at 0.64% against a 1.0% average. Portfolio cash flow has historically averaged 8.4% per quarter, well above the 5% redemption cap. Private debt valuations aren’t smoothed, and Cliffwater’s research shows private loan price drops actually exceed subsequent realized losses, meaning marks are conservative.

Their most provocative number: if every single software borrower in the private debt universe defaulted over five years with zero recovery, portfolio-level total returns would still come in around 4.5% annualized. That matches the current investment grade bond yield. The carry eats the losses even in a doomsday scenario.

Cliffwater also created the Cliffwater Direct Lending Index, the main benchmark for the asset class, and manages interval funds invested in private debt. The defense of the asset class is also a defense of the franchise.

Bain went a similar direction, telling Barron’s that private credit isn’t a bubble and software isn’t the real issue, disclosed 10-15% software exposure, and noted that Bain doesn’t have a retail problem because it prioritizes institutional and ultra-high-net-worth clients. The interview reads as reassurance directed at existing and prospective LPs wondering whether Bain has the same vulnerabilities as Blue Owl.

The Banks Want You to Notice They’re Fine

Jamie Dimon’s annual letter gave private credit the backhanded compliment treatment. He said the market “probably does not present a systemic risk,” pointing to its $1.8 trillion size relative to much larger asset classes. Then he spent several paragraphs cataloguing everything that could go wrong: weakening standards, generous add-backs, weak covenants, growing PIK usage, aggressive private ratings, poor transparency, and the likelihood that retail investors will seek legal remedy if things deteriorate.

The framing sounds like concern. Read the context.

JPMorgan disclosed $50 billion in private credit exposure this week. Wells Fargo reported $36 billion. Citigroup reported $22 billion. Dimon said on the analyst call he’s “not particularly concerned” because banks sit behind substantial loss cushions. Wells Fargo showed a 40% loss buffer with the ability to adjust margins dynamically if collateral performance deteriorates.

Compare that to the BDC liability side. Moody’s flagged that BCRED’s five-year notes widened from ~180 bps to the upper 200s. Unsecured debt issuance is still possible for the largest BDCs but increasingly uneconomical, pushing them toward secured borrowing and all the structural subordination that entails. The banks are funding at tighter spreads with more flexible terms. The BDCs are watching their cost of capital rise in real time.

The banks aren’t worried about contagion. They’re establishing themselves as the safer, better-capitalized alternative. If retail money is leaving private credit, the banks want to be where it lands next. If private credit lenders are pulling back from software and middle market lending, the banks want to fill the gap. Matt Levine made the point cleanly in his Bloomberg column yesterday: you could imagine banks gloating, and some of them basically are.

What Nobody Is Incentivized to Say

The private credit market may be going through exactly the mid-cycle repricing the asset class needs. Spreads are wider. Documentation is tighter. Leverage is lower on new deals. The capital that compressed terms and enabled sloppy underwriting is trapped behind gates or heading for the exits. By the standards credit investors actually care about, the lending environment is better than it was 18 months ago.

The biggest beneficiaries of that improvement will be the same firms whose retail fundraising excesses created the problem.

The managers who raised too much retail capital and deployed it into software loans at S+450 are gated now. Bad for their current funds. But the gate is creating the wider spreads and stronger docs that will make the next vintage significantly better. The worse this cycle gets for 2021-2024 vintages, the better it gets for 2026-2028 deployment. The same platforms will raise those funds. They’ll point to the improved environment as evidence they’ve learned their lesson and they’re now deploying into the best opportunity set in a decade.

Every piece of public commentary in the past week is laying the groundwork for that pitch.

Marks is positioning Oaktree as the disciplined manager who saw it coming. Goldman is positioning as the institutional-quality platform that avoided the retail mess. The opportunistic funds are positioning their deployment vintage as the post-correction opportunity. The banks are positioning for market share recapture. Cliffwater is positioning the asset class itself as resilient and misunderstood.

The gate is what makes all of these positions coexist. It prevents the stress from becoming a liquidation event. That’s the function everyone talks about. The function nobody mentions is that the gate bridges vintage cycles. It holds the current portfolio together long enough for the next fund to launch. It lets the stress generate the track record dispersion that makes “manager selection matters” a credible pitch. Boom, correction, recovery, new fund.

Next time you read a shareholder letter or an investor memo about private credit, ask one question. What is this person raising, and how does this story help them raise it?

If you liked this, you might also like:

Disclosure: The information provided is based on publicly available information and is for informational purposes only. While every effort has been made to ensure the accuracy of the information, the author cannot guarantee its completeness or reliability. This content should not be considered investment advice. Any investment decisions made based on the information provided are at your own risk. It is essential to conduct your own research and consult a qualified financial advisor before making any investment decisions. Investing involves risks, and past performance is not indicative of future results. By using this information, you acknowledge that you are responsible for your own decisions and release the author from any liability. Always seek professional advice tailored to your financial situation and objectives.

Meanwhile, another sector is breaking down just as badly, for unrelated reasons, and getting a fraction of the coverage.

Building products is the 2nd worst performer across high yield and leveraged loans YTD. $20+ billion of distressed debt, with private issuers that still haven’t reported Q4. When those numbers land, the picture gets worse. Four simultaneous headwinds hitting capital structures that were built for a completely different environment. The mainstream financial press has largely looked past it.

The loudest story in leveraged credit right now has most people looking in one direction. This week’s post is about what’s happening in the other one.

Request Access

The Setup

The housing market broke in 2022 and the repair timeline keeps getting pushed out.

Rates went from 3% to 7% in twelve months. Existing home sales collapsed. The homeowners sitting on sub-3% mortgages stopped moving. Transaction volume, which drives demand for everything that goes into a home, fell off a cliff.

That much is well understood.

What’s less understood is the sequence of what happened next and why the pain is arriving now rather than two years ago.

When rates spiked, new construction didn’t stop immediately. Builders had backlogs. Units started in 2021 and 2022 kept completing through 2023 and into 2024. Building products manufacturers watched their order books and concluded the cycle was holding. Revenue held up. EBITDA held up. Sponsors told lenders the worst was behind them. It wasn’t.

The completions backlog cleared in late 2024. When it did there was nothing behind it. Starts had been running below replacement for two years.

Volume fell sharply and the manufacturers had no buffer. Cost structures built for boom-cycle volumes. Debt loads underwritten to boom-cycle EBITDA. Neither of those things existed anymore.

The data today confirms the damage is deepening rather than stabilizing. Only 3% of the housing stock is currently turning over versus 4-5% in a normal environment. That has happened only 4 times in 50 years. In February, there were 630,000 more home sellers than buyers in the United States, the largest gap in records going back to 2013. The market is not clearing.

Home Depot’s CFO called the housing recovery a 2027-plus story at the JPM conference last week. KB Home recently cut its full year guidance. This is a public homebuilder putting hard numbers on the completions cliff that building products credit markets have been pricing for months.

~70% of outstanding mortgages carry a rate below 5%. Every one of those homeowners faces a moving penalty measured in hundreds of dollars per month in higher payments if they transact. This doesn’t resolve until either rates fall meaningfully or home prices correct enough to offset the payment gap.

The operating leverage math makes it worse than the revenue figures suggest. A 10% revenue decline doesn’t produce a 10% EBITDA decline for these companies. Fixed cost bases built for 2021-2022 volumes mean a modest revenue move translates into a catastrophic margin move. The damage is disproportionate to the revenue move because the cost structure doesn’t flex.

That’s just the operational side of things. Then the Iran conflict layered an entirely new set of problems on top of an already impaired setup.

Crude oil is up over 50% year-over-year and sitting at nearly $100 per barrel. For building products manufacturers this isn’t an abstract macro variable. Asphalt, PVC, resins, petrochemicals. All of it moves with oil. Polypropylene was up 20% in March alone. Aluminum now up 45% year-over-year. These are the direct inputs for windows, doors, insulation, and building envelope products. Earnings will begin to reflect input cost pressure regardless of what diplomats agreed to.

Aluminum tariffs compound the problem on a separate track with no clean resolution. The tariff tailwind that was supposed to help domestic manufacturers by shutting out Chinese competition never materialized because demand collapsed simultaneously. They got the cost inflation from tariffs on their own inputs. They never got the pricing umbrella on the revenue side.

Four simultaneous headwinds, each with a different resolution timeline, none of which the companies control. Suppressed housing demand. Unaffordable new construction economics. Oil-driven input cost inflation. Aluminum tariffs. It would be a difficult environment for a normally capitalized sector.

Building products is not normally capitalized.

The Debt Structure Trap

Many of these companies were taken private or recapitalized during the 2020-2022 window, and the capital structures reflect the assumptions of that moment. Housing demand was booming. Rates were near zero. PE multiples were expanding. Leverage at 8-10x made sense when debt cost 4% and end markets were growing.

That window closed but the debt didn’t go with it.

Those floating rate loans priced at SOFR plus 400bps that cost roughly 4.5% all-in when SOFR was near zero now cost closer to 9%. Cash interest expense has effectively doubled without any change in the principal balance. Coverage ratios have deteriorated not only because EBITDA collapsed but because debt service also did damage. The cov-lite structures that gave sponsors flexibility on the way up now give lenders limited visibility on the way down. And the maturity walls that felt distant in 2021 are now a 2027-2029 problem that is no longer theoretical.

The refinancing math doesn’t work for the most levered names. A company running at 8-10x leverage with interest coverage <1x cannot access the loan market at any reasonable cost. The HY bond market is closed to credits with those metrics. The path forward is either an operational recovery that organically deleverages before the maturity wall arrives, an LME such as a distressed exchange, or an outright restructuring. In the current operating environment, organic recovery is the hardest path to underwrite.

Why It’s Flying Under the Radar

If the data is this clear, the obvious question is why this sector isn’t generating the same attention as software.

The answer comes down to how these companies are financed and who owns them. Most of the stressed names are PE-owned, financed primarily with leveraged loans, and not subject to regular public disclosure requirements. They report to lenders behind invite-only datarooms. When deterioration begins, it shows up first in secondary loan prices as informed holders sell. By the time it surfaces as a public narrative, the stress has often been building for months.

Software names generate headlines because they have visible equity comp points that move markets. Building products names don’t have that information infrastructure. The stress is real. The signal is just harder to see.

There’s also a selection bias in how people read sector-level credit data. The HY bond index sample of building products companies skews toward the larger, better-capitalized names that had access to the public bond market in the first place. The loan market sample includes the aggressive 2020-2022 LBO transactions done at peak valuations with maximum leverage. When you look at HY bond performance for the sector, you’re largely looking at the survivors. The loan market is where the bodies are.

There is also a technical dimension worth understanding. ~75% of BSLs are held inside CLOs. When those CLOs move past their reinvestment periods they become forced sellers rather than natural holders. Par 1L prices in some of these names are a technical artifact of CLO demand, not a fundamental signal about enterprise value.

I wrote about one of the public names in this sector in detail last year, before the deterioration became this visible. Check it out here.

What the Market Is Saying

This week’s broader HY rally makes the building products situation easier to miss, not easier to dismiss.

HY spreads tightened roughly 21bps on the week to 278bps, driven almost entirely by the ceasefire relief trade. The index returned +0.90% on the week and is now positive on the year at +0.80%. The index is averaging a functioning market and a deteriorating one into a single number that flatters the former and obscures the latter.

Primary market activity in the sector has been essentially nonexistent. Last week, CoreWeave printed $1.75bn at the tight end of talk. Univision printed $1.5bn. Building products names have been absent from the new issue calendar this year entirely. Two primary markets are running in parallel right now, and building products is firmly on the wrong side of that divide.

What to Watch

The spring selling season over the next four to six weeks is the verdict.

Monday’s existing home sales print for March will be backward-looking, reflecting contracts signed in January and February before the Iran conflict and before mortgage rates moved higher. Even a decent print tells you nothing about current reality. The real verdict arrives in April and May data which won’t land for weeks. By which point several private issuers will have reported, restructuring conversations will have advanced, and the window for refinancing will have narrowed further.

The current bull case requires a specific sequence:

• The ceasefire holds.

• Oil normalizes by mid-year.

• The Fed acknowledges slowing growth and signals cuts.

• Mortgage rates ease.

• Spring data surprises.

The names that survive to that environment with balance sheets intact become compelling. Fixed rate or hedged debt, manageable leverage, revenue weighted toward R&R rather than pure new construction. That profile exists in this sector. It is not the profile of the names currently generating the most attention.

For the most distressed names the question was never whether housing recovers. Housing always recovers. The lock-in effect is a coiled spring. Pent-up demand is real. Demographics still favor household formation. The structural case for an eventual recovery is not in dispute.

The question is whether the recovery arrives before the maturity clock runs out.

The restructuring happens before the recovery arrives. The businesses survive in some form. The capital structures don’t. And by the time the housing market finally thaws and transaction volume returns and completions pick back up, the names that make it to that environment will be reorganized entities with cleaned-up balance sheets, new ownership, and no memory of what the debt cost their predecessors.

That’s the tragedy of the timing. Playing out in credit markets, in BSLs, high yield, and private credit, in secondary prices that most people aren’t watching.

The Names

Here are the specific names I’m tracking in building products right now, organized by where they sit on the distress spectrum, with current prices and yields.

This isn’t a recommendation to buy or sell anything. If the spring selling season data disappoints and the private issuer earnings confirm what the public name downgrades are already signaling, the names below are where the restructuring conversations start.

The other side of that trade is also worth understanding. When the housing market does recover, the names that survive to that environment in their current capital structures will have significant convexity. Bonds trading at 20-40 cents don’t need perfection. They need the business to still exist and the recovery to arrive before the maturity clock runs out.

Read more

The hyperscalers are spending $600 billion+ on data center CapEx this year. The picks-and-shovels trade is crowded and priced. You know the names. The easy money is gone.

Meanwhile, a fracking contractor just tripled in the last 6 months.

Liberty Energy. Completion services. Cyclical, capital-intensive work that trades at 3-4x EBITDA because that’s what the market thinks it’s worth, and usually the market is right.

Then someone at Liberty looked at what they actually do. Run massive high-horsepower equipment in demanding field environments. Manage power under pressure, literally. And they made a call: those same skills could power an AI data center. Not metaphorically. Actually power it, with dedicated natural gas generation built and operated by people who already know how to keep equipment running where utility contractors would fail.

The question now: is this a real infrastructure pivot? Or a frac contractor that got ahead of itself?

What This Company Does

The frac business

Liberty Energy is the 2nd-largest fracking company in North America, with ~20% market share and ~40 active fleets operating across every major U.S. shale basin. When an O&G producer wants to extract hydrocarbons from shale rock, it needs to pump fluid into the earth at extremely high pressure to crack the rock open. Liberty shows up with the equipment, the crews, and the expertise to do it. The oil company owns the land and the oil. Liberty provides the service.

Revenue runs at roughly $100-120mm per active fleet annualized, implying a $4bn revenue base at peak activity levels. Costs are predominantly variable: natural gas or diesel fuel to run the pumping equipment, maintenance on a fleet of heavy machinery, labor for field crews, and proppant which is the sand pumped into the cracks to hold them open. Proppant largely passes through to the customer so it inflates the revenue line without contributing meaningfully to margin.

The margin profile is highly cyclical. At the 2023 peak, with WTI above $75 and frac supply tight, EBITDA margins ran 25-26%. At the current trough they are running 13-15%. Mid-cycle normalized is probably 18-20%.

When activity recovers and fleets reactivate, fixed costs get absorbed across a larger revenue base and incremental EBITDA margins can run 40-60% on the way up. EBITDA per active fleet went from $4mm annualized in 2020 to $31mm in 2023 on a relatively modest increase in fleet count.

Over 8 years as a public company, ’s median EV/EBITDA has been roughly 3-4x regardless of execution quality. The ceiling is structural, tied to the cyclicality of the underlying business, and it is the core strategic problem the second business is trying to solve.

The power business

Fracking is the old business. The new business is power.

If you’re unfamiliar with the current industry setup, the current AI buildout is increasingly creating a power crisis. A single large AI data center can require 100-500MW of continuous power and the U.S. grid was not built for this. Interconnection queues now stretch more than 4 years in most major markets. Traditional utility-scale power development takes 7 to 10 years from conception.

Hyperscalers cannot wait.

The solution many are turning to is behind-the-meter (BTM) power. Rather than connecting to the public grid, a data center installs its own dedicated generation on-site. The generator sits physically behind the utility meter, supplying power directly to the facility without touching the grid. No interconnection queue. No transmission constraints. Power available in months rather than years.

The economics are ~$300k of EBITDA per MW per year on $1.5mm per MW of all-in capex, implying roughly a 5-year cash payback and high-teens unlevered returns. At 1 GW fully deployed, that is $300mm of annual EBITDA. At the 3 GW, management is targeting by 2029, it is $900mm.

Margins at steady state run somewhere around 40-50% EBITDA. Once an asset is built and contracted, revenue is predictable and the primary variable cost is fuel, which in many contract structures passes through to the customer.

In early 2026, the commercial proof points arrived. A 1 GW power development agreement with Vantage Data Centers, including a firm reservation for 400MW of 2027 capacity. Then a 330MW preliminary ESA with an unnamed Texas developer (although this was recently cancelled). Management raised its capacity target to 3 GW deployed by 2029, up from 1 GW six months prior.

Why BTM Has More Duration Than It Looks

The bull case on BTM power rests on two arguments that are related but distinct, and it matters which one you believe because they have different shelf lives.

Read more

The gates go up, redemptions get managed, the 5% quarterly cap does its job. Apollo’s Jim Zelter went on Bloomberg last week and said as much. Properly structured. Properly disclosed. The rules were clear.

He said this right before Blue Owl announced 40.7% redemption requests.

The whole structure rests on one assumption: that the loans mature when they’re supposed to.

The people saying everything is fine are correct about today. The gates are holding. But the match-funding argument gets tested in 2028, when the software maturities arrive and the borrowers who can’t refinance need more time. If the redemption queue is still running when that happens, the liquidity mismatch isn’t a sentiment story anymore.

And just to be clear about what we’re actually talking about. This is about one specific product, the non-traded BDC, sold through wealth channels to retail and HNW investors who were told they were getting private credit returns with semi-liquid access. Institutional direct lending is a different business. Publicly traded BDCs are a different conversation.

Why the Timing of These Loans Matters More Than People Think

When people say a BDC is match-funded, they mean the timing works.

Loans mature, cash comes in, investors who want out get paid, the vehicle continues. The liabilities are gated because the assets aren’t liquid on demand. You accept quarterly liquidity with a 5% cap because the yield compensates you, and the whole structure balances because the loans actually mature on schedule and produce the cash flows the model assumed.

But if the loans don’t mature on schedule, cash doesn’t arrive on schedule. The vehicle has to fund redemptions from somewhere else: new investor capital, bank lines, CLO proceeds, secondary loan sales. Each of those sources has a cost and a limit. I talked about hierarchy of redemption pressures last week.

Right now, each of those limits is being tested simultaneously.

New investor capital, the cleanest source, is slowing. Inflows turned negative in Q1 for the first time. Bank lines are tightening. JPMorgan is restricting lending to some private credit funds after marking down loan values in their portfolios. Meanwhile, CLO issuance is running near record pace partly to fill that gap.

$9.5 billion in CLOs issued so far this year. Near the 1Q’24 record. People in the industry call it capital stack diversification. It’s also a funding gap getting papered over with securitization.

Notably, BDCs are retaining the junior tranches of these CLOs. The equity and mezz pieces that absorb first loss. Citigroup estimated total BDC exposure at roughly $12 billion, about a third of all junior capital in the private credit CLO market. The leverage inside the BDC gets extended one level further inside the securitization, in a way that doesn’t surface cleanly in the headline debt-to-equity ratio.

The secondary market is now asking those same assets to clear at those same marks (they won’t).

More private credit loans for sale right now than at any point in this market’s short history. Goldman, Morgan Stanley, JPMorgan, Jefferies, all running processes. Asking prices near par. Bids not meeting them. One direct lender put it simply: buyers aren’t there in a systematic way at par, and the bid-ask spread remains very wide in reality.

That’s the only independent price signal available on loans otherwise valued by the people who own them. Right now that signal and the quarterly NAV are not saying the same thing.

All of that matters, but only because it tells you the same thing: the market is already rejecting the marks. The core problem is what happens when the maturities arrive and the borrowers can’t refinance.

The BREIT Comparison and Why It Breaks Down Here

The industry’s response to all of this has been consistent: the gate is working as designed, the structure is sound, BREIT proved the model.

It’s worth taking that argument seriously because it’s not stupid.

BREIT gated in 2022. Blackstone held firm. Investors asked to withdraw roughly 20% of shares at the peak. The fund fulfilled 100% of requests over a fourteen-month period. The assets recovered, sentiment normalized, and the vehicle survived. The industry has been banking on that outcome ever since. It became the template. It became the answer to every skeptic.

The argument being made right now, explicitly in industry whitepapers and implicitly in every shareholder letter, is that non-traded BDC redemptions are market-driven rather than performance-driven. Sentiment is the problem, not credit quality. The underlying loans are healthy. The gates buy time for sentiment to normalize, just like they did for BREIT.

That argument works if the underlying assets are actually fine.

BREIT’s assets were buildings. Stabilized real estate. Class A industrial, multifamily, things that exist independent of the capital structure sitting on top of them. When Blackstone said trust us, the assets are fine, they were making a claim that could eventually be tested against observable transactions in the same markets. The gate bought time for sentiment to catch up to fundamentals, and the fundamentals turned out to be intact.

Non-traded BDC assets are middle market loans. Many of them to software companies. These are businesses that went to private credit because the syndicated market passed. The spread you earned was the market’s way of quantifying the risk you were absorbing. When a non-traded BDC says the portfolio is resilient, they’re making a claim that’s much harder to verify, on assets valued by the same manager that originated them, held them, and has every incentive to keep them at par.

The BREIT playbook assumed the problem was sentiment and sentiment heals. The assumption in that entire framework is that redemptions are market-driven, not performance-driven.

The BDC question is whether the software maturity wall is the scenario where that assumption breaks down. If the 2027 to 2029 maturities don’t refinance cleanly, redemptions stop being a sentiment problem and start being a cash flow problem. Those are different situations with different endings, and running them together in shareholder letters is doing investors a disservice.

The Software Maturity Wall Is Coming

We’ve covered this multiple times now. The private credit boom in software lending peaked roughly 2021 to 2023 by most estimates. Sponsor-owned software businesses, many taken private at elevated multiples, borrowed heavily from direct lenders at spreads that reflected both the illiquidity premium and the credit risk the syndicated market didn’t want to absorb.

Those maturities start coming due in 2028 in a major way

And the default backdrop they’re maturing into is not benign. Morgan Stanley estimates the private credit default rate reaches 20% in the current cycle, approaching the sector’s historical stressed peak in 2009.

Some of those businesses will be fine. Revenue held up, the model worked, they refinance cleanly. That happens. Probably more often than the bears want to admit, and it’s worth saying that a performing software business with stable cash flows and a credible growth story gets refinanced. The match-funding argument works. The gate was never really tested.

Some will be genuinely impaired. The business deteriorated, the sponsor is deeply underwater, there’s no equity value left to protect. In those cases lenders take the keys. They stop waiting for a refinancing that was never coming and run the business for cash instead. Lenders in that position often do reasonably well. First lien, recurring revenue, a software business that still functions even if the growth story is gone. Operating it for distributions beats waiting for a par recovery that isn’t there.

Medallia is probably heading there. Thoma Bravo called the $6.4 billion take-private a “mistake.” The paper is being shopped in the secondary market right now. Nobody is offering par. The lenders who end up owning it may not have a disaster on their hands. They have a software business. They’ll run it for cash.

The clearly impaired credits get resolved. That’s not where the structural problem lives.

The Middle Is Where It Gets Complicated

The problem cases are the ones that aren’t clearly fine and aren’t clearly impaired.

Not good enough to refinance cleanly into the 2027 market. Not bad enough that the sponsor hands over the keys. A business that’s technically performing, still generating revenue, sponsor still holding out hope for an exit, but rates are higher than the original underwriting assumed, growth has slowed, and the syndicated market isn’t going to touch it at a price that works for the capital structure.

That borrower and that lender sit down and have a conversation. The lender can mark the loan down and take the loss. Or extend the maturity, collect an amendment fee, and give the business another eighteen months.

Think about what marking down actually means in this context. It’s not an accounting entry. If NAV goes down, more investors head for the exit, the fund sells assets into a market that won’t pay par. Nobody in that seat marks down if they can avoid it.

So you extend. The fee gets collected. The loan stays at par. The quarterly letter reports a resilient portfolio.

Some of those extensions will work out. The business stabilizes, a refinancing happens in 2028 or 2029, the lender gets paid. The extend decision was correct. This is the base case the managers are implicitly assuming when they say the portfolio is resilient.

Some won’t. The extension buys time the business uses to get worse rather than better. Now the lender is holding a more impaired credit at the same par value, two years later, in a vehicle that has been funding redemptions from bank lines and CLO proceeds the whole time.

Non-accruals stay low throughout all of this. An extended loan is technically performing. A loan restructured to pay interest in kind rather than cash is not in default. The statistic is accurate and misleading at the same time, which is a recurring feature of how private credit health gets reported.

The match-funding argument doesn’t require every borderline credit to go bad. It just requires enough of them to extend rather than mature that the cash flow timing stops working the way the model assumed.

What the Redemption Queue Is Actually Telling You

Last quarter investors asked to pull 40.7% of Blue Owl’s tech-focused BDC, OCIC. Nearly 22% from the flagship. Both funds enforced the 5% cap. More than $4 billion is now sitting in the queue.

OCIC has $11.3 billion in liquidity. OTIC has $1.3 billion. The funds can meet the 5% cap for now, and probably for a while. It’s worth highlighting that the near-term liquidity is real, the gates are holding, and the managers saying they’re well-positioned to meet upcoming tenders are not wrong about today.

The real problem is that this gets ugly on a timeline longer than one quarter, which is exactly why people keep hand-waving it away.

A 5% quarterly cap on a fund facing 40% redemption requests means the queue takes two years to clear under static assumptions, with no additional requests. Two years during which investors at the back of the line are holding a vehicle that’s paying out the more motivated investors first, while the 2027 and 2028 software maturities are arriving and being evaluated one by one.

Corbin Capital’s John Cocke put it plainly: if outflows hold at 5% per quarter for a year, it becomes “an existential question of the vehicle.” Not a prediction. A description of the math.

Moody’s cut the sector outlook to negative this week. $14 billion in Q1 redemption requests. The proration psychology is now in motion: investors watching peers get gated and deciding to queue before the window closes further. Once that dynamic sets in it’s genuinely hard to reverse.

The industry’s own scenario analysis, modeled on BREIT, assumes requests peak around 20% for a couple of quarters and then normalize over three to five quarters. That’s the optimistic path. It assumes the underlying credits are healthy, sentiment is the driver, and time solves the problem.

The Question Nobody Is Asking

Last week the Department of Labor proposed rules that would move private credit into 401(k) retirement portfolios. Scott Bessent called it the beginning of a Golden Age.

The week after Moody’s cut the sector to negative. After $14 billion in Q1 redemption requests. After Blue Owl disclosed nearly half of one fund’s investors trying to leave.

Nobody in Washington reads the filings.

Doug Ostrover said two weeks ago, about the mismatch between what investors expected and what the structure actually provides: “Between us, and the advisers who sell our products, I don’t think we made it clear enough.”

Here’s the question the industry’s BREIT framework doesn’t answer. BREIT worked because the assets were fine and time solved the problem. The entire non-traded BDC defense right now rests on the same assumption: redemptions are sentiment-driven, credit quality is intact, give it a few quarters.

But what if the redemption cycle and the maturity wall arrive at the same time?

The queue is running now. The 2027 maturities start in roughly eighteen months. If a meaningful portion of the borderline credits can’t refinance at that point, the lenders extend. The NAV holds. The queue keeps running. New capital stays away because nobody wants to invest in a vehicle that’s prorating redemptions. The vehicle is now funding a persistent queue from a shrinking pool of maturing assets, bank lines that are tightening, and CLO proceeds that are getting more expensive.

This is going to be a slow grind. The kind that doesn’t make headlines until it’s already well advanced. The type where the quarterly letters keep saying resilient right up until they don’t.

Centerbridge’s co-founder said yesterday the current moment is a genuine entry point for experienced investors with real liquidity and long time horizons. He’s right. The distressed opportunity in private credit over the next three years is going to be significant, precisely because the A&E cycle creates the conditions for it.

But that opportunity accrues to the people on the outside looking in.

The people on the inside, in the queue, waiting on cash flows from loans that may or may not mature on the schedule the model assumed, are in a different position entirely. They’re in something that doesn’t have a clean historical precedent, because the last time a product like this was stress-tested, the assets were buildings and not software businesses facing an AI disruption cycle with a maturity wall coming due.

The gate held for BREIT because the problem was sentiment.

The question worth asking right now is whether the problem is something else entirely. And whether by the time we know the answer, the queue has been running long enough that the match-funding argument has already quietly failed.

If you liked this, you might also like:

Disclosure: The information provided is based on publicly available information and is for informational purposes only. While every effort has been made to ensure the accuracy of the information, the author cannot guarantee its completeness or reliability. This content should not be considered investment advice. Any investment decisions made based on the information provided are at your own risk. It is essential to conduct your own research and consult a qualified financial advisor before making any investment decisions. Investing involves risks, and past performance is not indicative of future results. By using this information, you acknowledge that you are responsible for your own decisions and release the author from any liability. Always seek professional advice tailored to your financial situation and objectives.

Request Access

Everyone’s trying to figure out the right move in software credit right now.

Here’s an idea: maybe the best trade is to actually do nothing.

You already know the setup. The 2028 software maturity wall. The AI disruption overhang. The terminal value problem we’ve been talking about the last couple of months.

The question was always whether the restructuring wave would be consensual or coercive. It won’t be. The lawyers are already circling and salivating.

So what’s next? If the sponsor plays nice, you get an A&E. The more cynical view, and the more likely one based on everything we’ve seen over the last three years, is that you’re about to get your collateral stripped.

Which raises the question nobody is asking: if that’s where this is going, is the holdout actually the better trade?

Here’s how it goes. The large holders negotiate themselves into new senior paper. They get their deal, they get their collateral back. Meanwhile if you hold out, and just the right number of other lenders hold out with you, the stub remaining after the LME might be small enough that the company just pays you back at par with cash on hand.

No negotiation or haircut. You waited and got paid.

At least, that’s the theory. Here’s what it looks like in practice.

How the Playbook Works

Bloomberg reported this week on Vibrantz Technologies, a paint-additives company owned by American Securities. The owner wouldn’t share restructuring details until lenders signed NDAs barring any communication with peers. Once lenders were siloed, the sponsor cut different deals with different groups.

I mean, c’mon!

The larger creditors got first-out paper. Smaller lenders were approached after the structure was already set, offered something worse, on a take-it-or-leave-it basis with 90% of the lender base already signed.

Scott Greenberg described it aptly:

“Weaponizing NDAs and side deals to prevent lenders from coordinating pushes the limits of the collective action provisions in credit docs that are supposed to require coordination.”

Supposed to.

You probably already have a view on the recovery math. But the Fitch data is worth stating explicitly because it’s starker than most people realize. Companies that went through an LME before eventually filing recovered ~41 cents on first lien. Companies that skipped the LME and went straight to Chapter 11 recovered closer to 68 cents. You’d think that would slow things down. It hasn’t.

A Harvard Law paper from earlier this year found that a majority of coercively restructured firms end up filing anyway. The LME buys time. It doesn’t fix the balance sheet. And it leaves certain lenders materially worse off than a straight filing would have.

And it’s not evenly distributed.

The reason this matters now is timing.

The 2021 vintage is the single largest entry class in the loan index by every measure, with a higher concentration of B3-or-lower borrowers than any vintage on record. 25% of outstanding software loans mature in 2027-28, vs. 21% for the broader index. A meaningful share of those issuers have already defaulted or completed an LME and the wall is still two years out.

The Trade

A lot of these companies are still cash flowing. The AI disruption thesis is real but it plays out over years. Operational deterioration is slow. Debt service is current. The maturity is the problem, not the business.

If the LME is designed to subordinate you, strip your covenants, and leave you in a worse structure, maybe the better trade is to not participate at all. Sit in the 70s on a 2028 maturity. Clip your coupon. The stub maturity that rings post-LME might produce a better outcome than fighting over collateral in a deteriorating negotiation while the sponsor extracts whatever value remains.

The holdout only works if the company doesn’t actually need to file. That’s the whole underwrite. You either believe the business survives to the maturity or you don’t.

But the nuance is the holdout only works as a minority strategy. If enough of the lender base sits on their hands, the maturity wall doesn’t get addressed, the company can’t refinance, and you’ve created the exact outcome you were trying to avoid. A room full of holdouts is just a default waiting to happen. And frankly, that might actually be the better approach if the Fitch data is any indication. It’s basically what’s playing out at Medallia.

It does mean however the actual research required might not be just identifying the right names. It’s understanding who else is in the capital structure and how much they own.

A dispersed lender base with no dominant holder gives the holdout room to work. A cap table where 2 or 3 large funds control majority economics probably doesn’t, because those holders have enough leverage to drive a negotiated outcome on their own timeline regardless of what you do.

Most people are spending their time trying to identify the right names. The smarter work right now is mapping the lender base. Both pieces of the analysis matter.

One more thing before moving on. This is predominantly a BSL and high yield issue. You can’t run the Vibrantz playbook on a unitranche. A single direct lender holding a $500 million club deal doesn’t get siloed behind an NDA the same way a dispersed syndicated lender base does.

Private credit has its own version of the vintage stress problem but the mechanism is different. That market will see its own restructuring wave but the form it takes is a separate question. I’ll be getting into this in Private Credit Unfiltered, my weekly private credit series. Check out the 1st edition below:

What the Pricing Is Telling You

Software loans gained just over 1% in March, best month in nearly a year. Still down close to 6% YTD. The bounce got attention. The structural problem underneath it didn’t.

The bottom chart tells that story clearly. The software sub-index sits at 89 cents vs. 95 for the broad loan index. The divergence opened sharply in early 2026, partially recovered in March, and remains wide. The stabilization is real. It’s not a resolution.

Loan yields came in around 8.75% this week, ~125bps above high-yield bond yields. The one-year average gap runs closer to 66bps and it normally runs in HY’s favor. Loans are floating rate, senior secured, shorter duration. They should yield less than bonds. Right now they yield substantially more.

Part of that gap is explainable. Forced retail outflows hitting loans with no equivalent on the HY side. CLO technical dynamics. Software concentration in the loan index that simply doesn’t exist in HY to the same degree. Some of this is structural and probably persists. The inversion alone doesn’t tell you loans are cheap.

For the names where the business is functional and the problem is the capital structure, the holdout might work precisely because the LME process is designed to give par to someone else. The large holders accumulate, negotiate their first-out paper, and move on. The smaller holder who didn’t participate, sitting in a 2029 maturity on a company still generating cash, collects coupon and waits.

Not every name or even most names. A specific subset the market is currently pricing like all the others.

What the Week Said

HY spreads tightened roughly 30bps this week. Index yields fell from 7.68% to 7.27%. Equities rallied hard, SPX up nearly 4% from the lows. VIX down too. The 10-year drifted back toward 4.3%. The market recalibrated from pricing a potential hike to none.

Brent crude barely moved.

That’s the disconnect worth sitting with. The commodity at the center of the conflict didn’t confirm what every other asset class was pricing. Two to three week ceasefire timelines have been floated repeatedly over the past month. The equity market tests into them, reverses on headlines, and tests again. This week felt like another iteration of that pattern rather than a genuine reassessment.

Selling in March was historically elevated and the trading community walked into April carrying a considerable short base. A lot of the week’s price action reflected that positioning getting squeezed, not a fundamental change in the outlook.

On Iran, the honest answer is nobody knows. Polymarket is pricing a 61% chance of ground invasion before 2027. Trump is threatening to bomb power plants. Iran rejected the Hormuz ultimatum and says the strait reopens only when war damages are fully compensated. Kuwait’s oil infrastructure got hit again over the weekend.

It does feel like we’re closer to peak negative sentiment than peak escalation risk, but I hold that view loosely. The short base is real, the positioning data supports a squeeze, and neither side has taken the next truly catastrophic step yet. That could mean we’re approaching an off-ramp. It could also mean the worst is still ahead. The market is betting on the former. The headlines are not confirming it.

What’s locked in either way is the economic damage. Inflation is going higher. Gas above $4 a gallon for the first time since 2022. Supply chains disrupted. A ceasefire tomorrow doesn’t undo any of that.

HY primary was quiet. March closed as the lightest month for HY supply since late 2025 and the lowest March total in three years. Four consecutive weeks of outflows above $1 billion each.

In the secondary, top movers included PARA/WBD, NFE, and UNIT, which rallied on M&A speculation. The bottom of the table was more telling: CNR hitting record lows, ION under pressure from the same AI overhang weighing on software credit broadly, and XRX debt weak.

Where This Leaves You

The LME wave is likely coming and most people in this market already know it.

The Vibrantz situation is an example of how it works. Fragment the lender base, assemble majority control behind an NDA, deliver a structure to the room that was negotiated without them. For most names in the distressed software cohort that process plays out the way it always does.

Be honest about what the trade actually is though.

If you’re an original lender who bought at par in 2021, the holdout framing looks different than it does on paper. You’re not choosing between a 41 cent LME recovery and par at maturity. You’re choosing between crystallizing a loss now or carrying a mark-to-market impairment for two years on the hope that a maturity rings clean.

That depends on your fund structure, your investor base, your ability to absorb the interim pain.

The holdout as described here is primarily a secondary market entry point. Buying in the 70s on a name where the business still works and the maturity is the problem. Most people aren’t talking about it that way.

The other thing most people aren’t doing is the lender base work.

A dispersed cap table with no dominant holder gives the holdout room to operate. A concentrated one doesn’t. You need to know who’s in the room before you decide whether to show up.

Which names does the sponsor actually need to restructure and which ones just need time. Most people are working on the first question. The second one is where the edge actually is.

If you liked this, you might also like:

JunkBondInvestor is an independent financial research publication providing general market commentary, corporate credit analysis, and educational content. The information provided is for informational purposes only and does not constitute investment advice, financial planning, or personalized recommendations of any kind. Nothing published constitutes a solicitation, offer, or recommendation to buy, sell, or hold any securities, nor does it guarantee any financial outcome. Any credit opinions or outlooks expressed are solely the independent views of the authors and are intended to reflect general credit trends, not specific investment recommendations. Investing in high-yield bonds, distressed debt, and leveraged loans carries significant risk including potential loss of principal. JunkBondInvestor operates under the Publisher’s Exemption of the Investment Advisers Act of 1940 and applicable state laws.

Today it owns ~2% of SpaceX.

That is not a typo. In the span of 12 months, a leveraged spectrum holding company with a wall of near-term debt maturities and a half-built 5G network bleeding cash has transformed into one of the most interesting equity stories in the market. The inflection point was the SpaceX deal: $8.5 billion in cash and $8.5 billion in SpaceX equity in exchange for spectrum licences that EchoStar had been forced to sell anyway.

The question investors are now asking is simple.

SpaceX is reportedly targeting a $2 trillion-plus valuation in an IPO that could happen in weeks. At those levels, EchoStar’s stake is worth over $40 billion. The stock trades at a ~$37 billion market cap. So are you buying SpaceX at a discount through EchoStar, or has the market already figured that out?

That back-of-the-envelope math is exactly what most people are doing. The problem is that EchoStar is genuinely complicated. Most of what is being posted online is missing those details.

There was a Restructuring Support Agreement signed last month that changes the DISH DBS capital structure in ways that are not obvious from the headlines. There are intercompany loans and a waterfall of pre-committed AT&T proceeds that reduce what actually flows to holdco equity. There are tax considerations that most analyses ignore entirely. And there are operating businesses in various states of distress, including one with a debt maturity four months away that the bond market is pricing at 75 cents on the dollar.

Maybe none of it matters. Maybe the SpaceX narrative is powerful enough to override everything else. But I think you should know what you are buying before you decide that.

This piece is not an attempt to debate whether $2 trillion is the right number for SpaceX. That deserves its own analysis. The goal here is narrower: to work through the full sum-of-the-parts and figure out whether you are buying EchoStar at a meaningful enough discount to make it worth owning indirectly, vs. simply waiting and buying SpaceX when it goes public.

That is the question worth answering.

How We Got Here

You can find the full history elsewhere, so I will keep this brief.

EchoStar started life as a pay-TV company. DISH Network was for decades one of the two major satellite television providers in the United States, competing directly with DirecTV for the roughly 20 million American households that subscribed to satellite TV. It was a good business for a long time.

As that business began its secular decline, Charlie Ergen (co-founder/chairman) made a bet that spectrum would become the most valuable resource in wireless. Through a series of FCC auctions starting in the 2000s, DISH accumulated one of the largest spectrum portfolios in the country, spanning AWS-3, AWS-4, 600MHz, 3.45GHz, and several other bands. The thesis was that eventually the major carriers would need that spectrum badly enough to pay up for it.

The problem was the FCC had other ideas. It attached aggressive buildout requirements to the licences, forcing EchoStar to actually construct a 5G network rather than simply sit on the spectrum. The company spent years and billions of dollars building that network, acquired Boost Mobile from T-Mobile in 2020 to have a retail wireless presence. By 2025, it was burning cash at an alarming rate with a going concern disclosure and over $20 billion of debt.

Then the FCC came calling again, this time threatening to reclaim licences that EchoStar had not fully built out. Rather than fight a prolonged regulatory battle, Ergen did what he has always done: turned a problem into a transaction.

The spectrum sale to AT&T in August 2025 and the deal with SpaceX in September 2025 resolved the FCC dispute, retired the most pressing debt maturities, and in the process handed EchoStar a stake in the most valuable private company in the world.

In the process, the stock has more than 5x’d, rallying from the teens to nearly $130 as the narrative shifted from near-bankruptcy to SpaceX proxy.

Sum of the Parts

The stock has re-rated dramatically to reflect the new reality, but whether it has re-rated enough, or too much, requires actually working through the numbers. So let us do that.

The way I think about EchoStar is in five buckets. The first is contracted cash: signed, binding deals where the amounts are fixed. The second is the three operating businesses. The third is the remaining spectrum. The fourth is the holdco deductions, which reduce what ultimately flows to equity holders. The last and most important is the SpaceX equity stake.

Starting with the easiest.

Bucket 1: Contracted Cash ($33.2 billion)

EchoStar has signed two deals that together guarantee $33.2 billion of cash inflows regardless of what happens to the operating businesses or the SpaceX stake.

The first is the AT&T spectrum sale: $22.65 billion for the 600MHz and 3.45GHz licences. Signed in August 2025, pending FCC approval, expected to close mid-2026. The largest spectrum transaction in US history.

The second is the SpaceX deal, signed in September 2025: $8.5 billion in cash for the AWS-4 and H-Block spectrum plus international S-band rights. On top of that, SpaceX contractually committed to fund $2 billion of interest payments on EchoStar’s outstanding debt through November 2027, effectively meaning SpaceX is servicing part of EchoStar’s balance sheet for the next 18 months. That brings the SpaceX cash package to $10.5 billion.

Combined: $33.2 billion of contracted, signed, committed cash. Not contingent on any valuation, not dependent on finding a buyer, not subject to negotiation. The only remaining step on the AT&T deal is FCC sign-off, which is widely expected given the current regulatory environment.

Read more

The quarterly report looked fine. The NAV hadn’t moved. Redemptions were being honored and the people rushing for the exit seemed like they were overreacting to headlines.

So you stayed.

Rational, patient, exactly the kind of investor the pitch deck described when they sold you this thing.

And while you were being patient, every liquid asset in your portfolio walked out the door with the people who panicked.

That’s the part nobody is writing about.

I’m starting a new weekly series on private credit. There’s a lot of noise out there right now, some of it alarmist, some of it too rosy. I’ll try to cut through it.

Everyone Is Writing About the Wrong Story

The coverage so far has focused on the managers. Who honored redemptions. Who gated. Who wrote the check to buy goodwill.

The headlines treat this as a story about institutional character. That is not your story. Your story is what happened to the portfolio while all of that was playing out.

Redemptions at non-traded BDCs are rising sharply and gross inflows are slowing. At a few funds, outflows have already exceeded inflows. If that becomes widespread, the manager effectively has three levers.

Draw on credit facilities and issue debt. Some are doing exactly that. But re-leveraging a deteriorating collateral pool has limits, and the banks providing that leverage are paying close attention. JPMorgan has started independently revaluing private credit collateral used as loan security.

Find new equity inflows to replace what is leaving. That window is closing. Gross inflows slowed materially in the back half of 2025, and net flows across the private BDC cohort dropped sharply quarter over quarter.

Which makes asset sales the key pressure point if leverage and replacement inflows are insufficient.

Goldman Sachs estimates the retail private credit sector could shed between $45 and $70 billion in assets over the next two years, reversing growth that took the sector from $34 billion in 2021 to $222 billion by end of last year. The question that determines what happens to your capital is not which manager honored redemptions. It is which assets they sold to do it.

How You Fund a Redemption

You don’t get to sell whatever you want when you’re under redemption pressure. You sell what has a bid.

Cash goes first. Then the liquid stuff (broadly syndicated loans if any), the ones with observable prices and real secondary market depth. Then the cleaner names, the credits a sophisticated buyer will underwrite quickly, close to par, without demanding a steep discount.

What’s left after that process is not random.

Heavily leveraged paper from 2021 and 2022, underwritten at 7-9x when rates were zero. PIK toggles. Names that are one bad quarter away from non-accrual. Software loans (which are under-represented) with AI disruption hanging over them.

The assets that remain are more likely to be the ones that didn’t find a bid. That is why they are still in your fund.

A handful of the largest managers have additional options. Blackstone injected $400 million of firm and employee capital to satisfy BCRED requests without touching the portfolio. Exceptional moves, available only to managers with institutional capital structures and balance sheets large enough to absorb the cost.

Most managers running non-traded BDCs are not Blackstone. Honored redemptions need to be funded somehow. The mix of how they get funded determines what you are left holding.

The Self-Selection Mechanism

The portfolio may deteriorate for two reasons at once: original underwriting quality, and a redemption process that selectively strips out the easier-to-sell assets first.

Every redemption cycle removes the assets with the most liquidity and buyer interest, leaving behind the ones with neither. Nothing in the disclosed data tells you the composition has shifted. But it has.

Here is what makes this particularly insidious. The headline credit metrics look fine.

Non-accruals across publicly traded BDCs have actually been falling. Loans are still “performing.” Borrowers are still paying, or PIK-ing with lender consent, but nothing has formally tripped.

That’s how private credit works. A loan doesn’t go non-accrual until there’s an actual payment default or a formal determination that collection is doubtful. Until then, it accrues. Sponsors and lenders have every incentive to amend, extend, and restructure their way around that moment. The reported metrics reflect what has already broken. They don’t reflect what is quietly deteriorating underneath.

The PIK data is where that deterioration becomes visible before it has to.

Lincoln International, which values more than 7,000 private credit companies each quarter, found that bad PIK, meaning PIK interest that was NOT part of the original loan structure but is now being utilized, sat at 6.4% of private loans in Q4 2025. Up from 6.1% the prior quarter. Up from 2.5% in late 2021 when Lincoln began tracking it. More than half of all PIK loans now feature this dynamic.

PIK loans are unlikely to leave in a curated portfolio sale. They stay, compounding deferred interest, increasing the face value of the obligation while the enterprise value underneath may be going the other direction. The gap between carrying value and recoverable value grows quietly. It doesn’t show up in NAV until it has to.

The Secondary Market Won’t Save You

Some managers are trying to solve this by selling loan portfolios in bulk. Package the assets, find a sophisticated buyer, sidestep the liquidation hierarchy.

The problem is who the buyer is.

HarbourVest, Coller, Pantheon, the secondary arms of the big private equity shops. These institutions do this for a living. They know what a portfolio assembled under redemption pressure looks like. They know the clean assets cleared first. They price what’s left accordingly.

A manager trying to raise liquidity under a redemption deadline is not negotiating from strength.

Blue Owl’s $1.4 billion sale at 99.7 cents was celebrated as proof the system worked. Look more carefully at who bought it. CalPERS. OMERS. BCI. Blue Owl’s own Kuvare insurance affiliate. Four institutions evaluating a pre-selected slice in a negotiated process. That is the best possible execution on the most sellable assets in the book. For anything beyond that kind of curated pool, buyers will probably not pay par.

The next portfolio sale is likely harder. The one after that harder still.

When a secondary transaction clears at a meaningful discount, it becomes the first real observable market price for assets that previously existed only at manager-determined marks. Once that price is observable, every manager running a similar book has a problem. Which means the NAV you are looking at today may not be the one you are looking at next quarter.

The NAV Illusion

On the surface, nothing looks wrong. Stable marks. Assets sold at or near par. The manager’s letter references portfolio resilience.

But the composition of what remains has quietly worsened, and the NAV may reflect that only slowly, unevenly, or after a forcing event.

We already have real examples of how fast marks can move when a stressed book gets forced into the open. At BlackRock TCP Capital Corp, writedowns in Q4 2025 reduced NAV by 19% in a single quarter. Moody’s downgraded FS KKR Capital Corp to Ba1, one level into junk, citing non-accruals rising to 5.5% of total investments. This doesn’t mean every fund is next. It proves the speed of adjustment, when it comes, can be severe.

The default trajectory is moving. Morningstar DBRS recorded a 78% year-over-year increase in private credit default events in 2025. Distressed exchange transactions now account for 94% of all downgrades to default or selective default. Not hard payment defaults. Restructurings where lenders accept worse terms to avoid triggering a formal event. Extend maturities. Accept PIK. Loosen covenants.

And the assets that didn’t find a secondary bid, the software loans and PIK positions the sophisticated buyers passed on, are the same ones staring down a specific refinancing wall.

Over $160 billion in private credit software loans mature by 2029. These need to refinance in an AI-disrupted environment on timelines now measurable in months.

The mask slips when a secondary transaction clears at a visible discount, a forced sale implies the marks are wrong, or non-accruals start climbing because the cleaner names have already been sold.

When any of those happen, the loop tightens.

NAV moves. Queue gets longer. More assets sold. Portfolio concentrates further into the names with the fewest bids. A liquidity problem creating the conditions for a credit problem that doesn’t yet exist in the reported data. The investor in the middle of it, watching a stable NAV, may have no way of knowing how far along this already is.

You Are Holding What’s Left

Most of the coverage treats the redeeming investor as the protagonist. The one who read the headlines and got in line. That investor is out. Their problem is over.

You stayed. Maybe the quarterly report looked fine. Maybe you believed the NAV. Maybe you requested a redemption, got a fraction of what you asked for, and remained in the fund involuntarily. Either way you are now holding a concentrated position in the assets that survived the selection process. The clean paper is likely gone.

You are not holding private credit. You are holding what’s left of it.

The quarterly report next month won’t tell you that. The NAV will look roughly the same. The manager’s letter will sound the same. The non-accrual rate will look stable because the names approaching stress are still accruing, still marked at cost, still waiting for the quarter that forces the issue.

The instrument panel can look fine longer than it should.

The Structure Was Always the Problem

The non-traded BDC was designed to solve a distribution problem for asset managers, not a liquidity problem for investors.

Quarterly redemption windows against 5-7 year loan maturities was never going to hold under sustained stress. The industry watched this play out in non-traded REITs in 2022. Same structure, same redemption caps, same manager-determined NAVs, same retail investors discovering what semi-liquid actually meant when everyone reached for the exit at once.

Among BDCs with aggregate NAV over $1 billion, redemptions have already risen 217% quarter over quarter. Sixth Street said it directly: if the non-traded REIT segment is any guide, this capital flow dynamic will last multiple years.

Then the industry built the identical structure in private credit anyway, with more capital and a broader retail distribution network than the REIT market ever had.

Apollo moving toward daily NAVs is the right direction. Blue Owl eliminating the tender offer mechanism on OBDC II is a confession that the structure couldn’t hold, packaged as an orderly solution.

The managers who use this moment to genuinely redesign the product will own the next cycle of retail private credit inflows. The ones who rebuild the same structure will face this again, at larger scale, with less goodwill.

What’s Behind the Gate

The gate is not the story. The gate is the moment you find out what’s behind it.

What’s behind it is a portfolio that survived the redemption process for one reason. Not because the manager protected it. Not because the underlying credits are strong. Because the assets with bids left first.

You didn’t panic. You stayed. You trusted the quarterly report and the stable NAV and the manager’s letter.

None of those are the price at which the remaining portfolio clears today.

Private credit is not broken. The non-traded retail vehicle is. And the investors still inside it are holding something that looks exactly like what they bought, right up until the quarter it doesn’t.

Disclosure: The information provided is based on publicly available information and is for informational purposes only. While every effort has been made to ensure the accuracy of the information, the author cannot guarantee its completeness or reliability. This content should not be considered investment advice. Any investment decisions made based on the information provided are at your own risk. It is essential to conduct your own research and consult a qualified financial advisor before making any investment decisions. Investing involves risks, and past performance is not indicative of future results. By using this information, you acknowledge that you are responsible for your own decisions and release the author from any liability. Always seek professional advice tailored to your financial situation and objectives.

You were watching the order book.

Fifty billion dollars in orders. Priced at the tight end of talk. Three bond tranches, all trading up on the break. Every desk with a story about how oversubscribed the deal was. Oil above $100, 10-year at eight-month highs, and the market took it down without flinching. You were telling yourself this was a sign of health.

Meanwhile the secondary market was bleeding. Quietly. Right next to you. Same screen.

CCC spreads blowing through 700bps. HY funds posting their largest weekly outflow of the year. Recent new issues fading. Bids drifting to new lows while the primary market remained hot.

You missed it. Most people did. Which one you noticed says more about your positioning than about the market.

What $50 Billion in Orders Actually Means

EA bulls felt vindicated. Fifty billion dollars. Five hundred accounts. The $2.875 billion of 7.25% secured notes hit 102 in the secondary. The $2.5 billion of 8.75% unsecureds up to 103. Investors were calling the pricing cheap for the rating.

The more skeptical read is that $50 billion in orders is not proof of health. It’s a sign of cash that has nowhere else to go and a market conditioned by fifteen years of buy-the-dip reflexes.

500 accounts chasing the same paper because rates are high enough to make the carry look attractive and everyone figures someone else stress-tested the macro assumptions underneath the deal. The biggest LBO financing since the GFC pricing tight in the same week that CCC spreads hit nine-month wides and HY funds posted their largest weekly outflow of the year.

Worth stepping back on what’s actually driving primary volume.

This is not a broad market reopening. EA, Nexstar, Sealed Air. M&A and LBO financings are carrying the entire market right now. Refi activity has collapsed. Strip out the acquisition financings and the primary market looks a lot quieter than the headline numbers suggest.

The appearance of health is concentrated in a handful of large transactions, not a broad return of risk appetite.

The Secondary Market Nobody Is Talking About

CCC spreads blew through 700bps last week, widest since June. BB yields climbed to a 10-month high. The Bloomberg US Corporate High Yield Total Return Index has returned -1.97% month to date. YTD returns are negative too. HY funds posted their largest weekly outflow of the year.

Recent deals are backsliding. FICO’s 6.25% notes at 97. PENN’s 6.75% bonds at new lows.

What’s also worth naming explicitly is what’s driving all of it.

Credit right now is behaving more as an output of oil and rates than anything happening at the issuer level. Spreads, yields, and risk appetite are being set by energy prices and Treasury moves first. Company-specific stories are layered on top. The macro is overriding the micro in a way that rarely persists but is very much the reality right now.

The credits that are actually moving are moving on events. Hertz jumped as travelers looked for alternatives to flying amid the TSA staffing crisis / hours-long airport security lines. JetBlue popped four points on M&A speculation.

Beneath the index level, the sector picture is increasingly divergent and it maps almost directly onto Iran exposure.

Energy is the one area of leveraged credit where fundamentals and geopolitics are pointing in the same direction. Improving credit profiles, genuinely under-owned, with a tailwind the market keeps treating as temporary. New deals pricing tight and trading up. One of the few places where primary and secondary are in agreement.

Everything with hydrocarbon input costs is moving the other way. Autos, building products, specialty chemicals, packaging. A 30% oil move doesn’t stay in the energy sector. It runs through every manufacturer with petroleum-based inputs and energy-intensive processes.

Building products and homebuilders had a weak earnings season with more issuers still to report. The sector was already dealing with soft housing and no policy support before Iran added an input cost layer on top. PE-backed structures built for a rate environment that never materialized are now facing a commodity shock on top. Price discovery in that sector is not done.

Consumer-facing credits are quietly getting harder to underwrite. Gas price expectations just hit their highest level since 2022. University of Michigan consumer sentiment slipped to a three-month low. Lower-income cohorts absorbing higher full-price costs while energy takes a bigger share of their wallet is a difficult backdrop for retailers, restaurants, and anything else levered to discretionary spending.

The sectors holding up are the ones with real assets, contracted cash flows, or a direct benefit from the energy price spike. Everything else is seeing a commodity shock, a rate shock, and a consumer confidence shock simultaneously. That combination was not in any of the underwriting models.

Why Iran Doesn’t Just Walk Back

The market keeps trying to trade this like a financial shock.

Financial shocks have circuit breakers. Rate cuts, liquidity injections, a policy reversal, the right statement at the right moment. You reverse the policy and the shock reverses with it. That’s why every tariff headline last year generated violent two-way price action. The underlying economy was intact. The damage was a decision, and decisions can be undecided.

Physical shocks don’t work that way. Infrastructure damage doesn’t reverse on a ceasefire. The supply disruption outlasts the conflict itself regardless of what the diplomatic calendar produces. That’s been true since the first week and it hasn’t changed.

The oil futures curve has Brent in the $80s by August. Maybe that’s right.

But the Strait of Hormuz problem remains unsolved. Neither side has a clean exit. The escalation trap keeps ratcheting. The Houthis entering with risk to Bab al-Mandeb adds a second chokepoint and a third actor with its own incentive structure entirely.

Private Credit: The Ratchet Accelerates

The macro pressure is hitting an asset class that was already under strain before the conflict started.

Investors have sought to pull roughly $13 billion from over a dozen private credit funds so far this quarter. More than $4.6 billion is sitting behind withdrawal limits. Apollo, Ares, and BlackRock all capped at 5% after requests came in higher than that. All “normal course” as per the docs as they like to say.

Meanwhile, Cliffwater paid out 7% for the first time and now has a negative outlook from S&P. FS KKR got cut to junk by Moody’s (no surprise). Oaktree had to bring in Brookfield’s own balance sheet capital just to meet redemptions in full.

The structure was always the problem. Retail investors were sold quarterly liquidity against assets that take years to mature. Works fine until redemptions exceed inflows. All at the same time. And that’s BEFORE any major credit concerns.

At that point managers sell their best liquid assets first. The clean BSLs at or near par. The NAV stays roughly stable on paper because those sales happened at par. But the quality of the collateral behind the NAV quietly deteriorated with every redemption cycle.

The remaining portfolio becomes progressively more concentrated in the marks-at-model positions. The software loans, the PIK borrowers, the credits nobody wanted to buy. The fund looks fine until it doesn’t, and when it doesn’t the markdown is sudden because it’s been accumulating for quarters.

That selling also lands directly in the broadly syndicated secondary market as incremental supply at exactly the moment the secondary is already soft.

There’s a lot more to say here and a weekly recap isn’t the right format for it. Rather than jamming a nuanced private credit take, I’m launching a dedicated series on private credit starting later this week.

The redemption dynamics, the mark-to-model problem, the CLO feedback loop, the vintage exposure. All of it deserves its own treatment. Keep an eye out for that.

How I’m Thinking About Positioning

Read more

Revenue was growing triple digits. The stock was near all-time highs. The board looked at all of that and decided to rehire Wally Liaw, the co-founder they had pushed out 5 years earlier after an SEC accounting investigation, and put him back on as a director. With full knowledge of the 2018 history. Nobody forgot. They just decided it didn’t matter anymore.

Three years later Liaw was indicted by the SDNY.

On March 19th, federal prosecutors unsealed charges against Liaw and two others. The alleged scheme: conspiring to divert $2.5 billion in Nvidia-equipped servers to China through falsified documentation. To defeat auditors (allegedly), serial numbers were physically removed from the hardware and reaffixed using a hair dryer. SMCI was not named as a corporate defendant.

The stock fell 33% overnight. The converts dropped 10 to 20 points. The 2028s now trade at 87 cents, the 2029s at 81, the 2030 zero coupon at 73. The 2029s yield ~11% on a company carrying over $4 billion of cash, $10 billion of inventory, and net leverage of ~0.5x. On those numbers, the bonds look like an opportunity, right?

It’s not so obvious. This is the 5th governance episode in 8 years at a company that has consistently looked at its own history and decided it did not apply anymore. The balance sheet is real, sure. The question is what is standing behind it.

Not the First Time

To understand why this indictment is different from a typical governance shock, you need to know what came before it.

In 2018, the SEC charged with widespread accounting violations, including premature revenue recognition and improperly reduced accrued liabilities. Wally Liaw resigned under pressure. The stock spent two years in the penalty box before the AI server cycle rescued it. In 2020, the SEC charges were settled.

In 2024, Hindenburg accused SMCI of channel stuffing. SMCI’s own filings disclosed family-linked entities tied to CEO Charles Liang accounting for 4.3% of cost of sales. A civil lawsuit named Liaw for improper revenue recognition.

In October, E&Y resigned as auditor, stating it could no longer rely on management’s or the Audit Committee’s representations. Three months before the current indictment, SMCI’s 10-Q disclosed four outstanding material weakness items in internal controls. Not historical. Outstanding.

A company with a governance incident you can underwrite: model the liability, estimate the fine, apply a discount, and move on.

What the evidence describes is an institution that tolerates escalating misconduct, and that distinction matters because the legal liability becomes harder to bound, supplier relationships harder to underwrite, and regulators who feel misled tend to look harder at what they missed.

The Balance Sheet

On the surface the balance sheet looks fine. SMCI has approximately $4.9 billion of total debt against $4.1 billion of cash, leaving net debt of roughly $800 million. The revolving credit facilities, $2 billion from JP Morgan and $1.8 billion from CTBC, are entirely undrawn. Net leverage sits at ~0.5x. By any conventional measure this looks like a company with options.

Read more

You’ve seen this before.

Geopolitical shock, spreads widen, everyone holds their position and waits. It worked after Ukraine. It worked after every Middle East flareup in recent memory. The institutional muscle memory is so deep that most investors aren’t even making a conscious decision anymore.

They’re just holding. Assuming the premium fades because it always has.

Maybe it does this time too.

But the mechanism that allowed the market to look past prior conflicts was a specific one: the underlying economic damage was containable enough that pricing it out felt rational within weeks. Prior conflicts didn’t close a chokepoint carrying nearly 20% of global oil supply. They didn’t leave years-long damage at one of the world’s most important LNG export hubs.

The geopolitical risk premium faded because the physical reality underneath it eventually stabilized.

That’s the assumption that’s harder to make here. And it’s the one current spread levels are still relying on.

The Three Conditions

The geopolitical dip trade works when three things are true simultaneously.

Central banks can ride to the rescue. Shock hits, growth slows, the Fed cuts, liquidity floods the system, assets recover. This is the mechanical engine behind every successful dip-buy since 1987. Shock happens, Fed responds, dip gets bought.

Investors have dry powder. Buying dips requires capital on the sidelines ready to deploy. Institutional cash is sitting at lows. The ammunition that funded every dip-buy of the past decade was systematically deployed into a market that went essentially straight up from October 2022 through January 2026.

The shock is demand-driven, not supply-driven. A demand shock hits growth and pushes inflation down. That gives the Fed room to cut. Supply shocks work the opposite way. They hit growth and push inflation up simultaneously. The Fed cannot ease its way out of stagflation. Cutting into a supply shock makes the inflation problem worse. The central bank is paralyzed precisely when the market most needs it to act.

Right now all three conditions are inverted. Seven FOMC participants already see zero cuts this year. Markets are pricing a chance of a hike by year-end, a complete reversal from the 2.5 cuts priced three weeks ago. Dry powder is near multi-year lows. And Iran is a supply shock.

Rates First, Spreads Second

So if this isn’t a normal dip, what is it actually?

The easiest mistake right now is assuming credit is selling off because credit fundamentals suddenly cracked. That’s only part of the story.

This is a rates selloff first, a spread selloff second, and a credit event only after that. Treasury yields have backed up sharply, rate-cut expectations have been pushed further out, and the market is repricing toward higher-for-longer just as Iran makes the inflation backdrop harder to dismiss.

HY yields are back near nine-month wides but a meaningful part of that move is coming from the risk-free rate, not from a collapse in credit quality.

That distinction matters because it tells you what kind of selloff this actually is. If this were a classic credit panic you would expect more obvious distress in the lower-quality tail and more indiscriminate spread widening. Instead you are seeing a macro tightening event working its way through leveraged credit.

Spreads are wider but they are not anywhere near recession pricing. The market is repricing assumptions around inflation, rates, and funding conditions before it is repricing broad insolvency.

It also explains why BBs are not behaving defensively. In a normal risk-off move investors hide in quality. In a rates-driven selloff BBs give you more duration and less spread cushion. They get hit because they are not paying you enough to absorb the rate move.

But the more uncomfortable setup is one level down. Single-B credits don’t need perfection. They just need the world to stay good enough. Margins can’t weaken too much, financing windows can’t stay difficult too long, interest costs can’t remain punitive quarter after quarter.

A slow-burn shock erodes exactly those assumptions one quarter at a time without triggering the kind of visible distress that forces a repricing. The real danger may not be the credits everyone already knows are broken. It may be the ones still priced as if normal refinancing conditions are only temporarily delayed.

The Comment Worth Sitting With

Before the credit mechanics, one piece of context that matters.

Treasury Secretary Bessent said publicly last week that Kharg Island could eventually become a US asset. 90% of Iran’s oil exports transit Kharg. Whether he meant it literally or was sending a signal is almost beside the point. Once it’s said publicly, Iran’s ability to negotiate a face-saving exit narrows considerably. The incentive to sustain asymmetric disruption rather than deal increases. That’s a probability shift on conflict duration that spread levels haven’t moved to reflect.

And regardless of how the conflict resolves, some of the damage is already permanent. Qatar confirmed multi-year damage to Ras Laffan. 15 million tonnes of LNG annually offline for 3 to 5 years. A ceasefire tomorrow doesn’t rebuild that infrastructure. The damage already done locks in supply tightness for years regardless of the diplomatic calendar.

The Commodity Transmission Nobody Priced

Oil up 30% since Iran escalated doesn’t stay in the energy sector. It moves through the cost structure of every manufacturer with hydrocarbon exposure. Leveraged credit is loaded with them.

Start with autos. The auto supplier book runs on petroleum-based raw materials and energy-intensive manufacturing processes. Margin pressure that was theoretical six months ago is now arithmetic. Sponsors underwrote these businesses at commodity prices that no longer exist. Pass-through mechanisms that looked adequate in a stable environment are proving inadequate in a spike. The backend of the capital structure is repricing to reflect that.

Building products is a different problem but the same pattern. Soft housing, no policy support, PE-backed structures built for a rate environment that never materialized. The sector had a weak earnings season and private issuers haven’t fully reported yet. Price discovery isn’t done. The names already trading poorly are probably the honest ones.

Chemicals and packaging are getting compressed from both sides simultaneously. Rising resin and energy costs on the margin side, potential feedstock disruption from supply chain complications on the input side. That’s a squeeze with two blades in a sector carrying leverage that was sized for neither.

The through-line across all three is the same. Commodity shock plus leverage plus rate sensitivity plus zero margin for error. That formula was always going to find its victims. It’s finding them now.

There is also a second-order effect that hasn’t fully shown up yet. If the disruption lingers, the first real damage may not appear in earnings. It may appear in working capital.

Companies carry more inventory because they don’t trust transit times. Freight and insurance costs rise. Suppliers tighten terms. Customers pay more slowly because they’re managing the same uncertainty on their end.

EBITDA can look fine for a quarter or two while cash gets trapped in the plumbing. That’s how an operating problem becomes a credit problem. Not with one dramatic miss but with weaker cash conversion, more revolver draws, and less covenant cushion before the headline numbers tell you anything.

The Primary Market Is Open. That’s the Problem.

Only one HY deal priced last week. One.

This week’s pipeline is $15+ billion as larger M&A financings come forward (e.g., EA, Nexstar). That tells you capital markets are still functioning. It also tells you issuers are facing a much worse buyer base than they were a few weeks ago.

Fragile markets are often more dangerous than shut ones. A shut market is obvious. A fragile one encourages issuers to believe financing is still available right up until it isn’t. Deals can still get done. They just need more concession, more structure, and better timing.

How $15+ billion clears into a market with fresh redemption pressure and a deteriorating technical backdrop will be more informative than any spread quote this week.

How I’m Thinking About Positioning

Read more

Blue Owl earned most of them. The non-traded BDC situation was mishandled. Retail flows, which were the whole growth story, are broken in ways that do not fix themselves in a quarter or two. You cannot scare financial advisors and then expect them to keep selling your product to clients. That is not how trust works.

The stock is off 60%+ from its highs. The market has been saying something for a while now.

But at some point price matters. At some point the bad news is in the stock and the question shifts from what went wrong to what happens next. The question right now is whether Blue Owl is at that point, or whether the next shoe is still in the air.

In this piece I am going to walk through the valuation, the scenarios, and whether there is actually a trade here.

How Alternative Asset Managers Are Valued

Before getting to the numbers, it helps to understand how these businesses are actually valued. Alt managers are not like normal companies and the standard metrics mislead. The framework matters because the wrong lens produces the wrong answer.

Read more

No mark-to-market swings. No margin calls or forced selling. Just senior secured loans to PE-backed companies, locked up for years, generating steady income while public markets lost their minds.

Blue Owl told that story better than anyone.

$307 billion in AUM. $42 billion raised in 2025 alone. A stock that nearly tripled off its IPO.

Then OBDC II halted redemptions. A proposed merger that would have forced retail investors to effectively absorb losses. Short interest exploded. The stock crashed 60% from its peak. And the question private credit has been dodging for a decade landed directly on Blue Owl: are the marks on these loans actually real?

The stock is now at $9.11. It trades at a ~60% discount to Blackstone, a 50% discount to Ares, and below the multiple the entire alt manager sector hit at its worst point during COVID.

Nobody in the peer group is cheaper. It isn’t even close.

The marks on the software portfolio are generous. The retail channel takes years to heal, not quarters. The dividend is being paid above earnings. Management is already running behind its own five-year targets after twelve months.

But Blue Owl is three businesses sharing a balance sheet and a stock price.

The credit franchise deserves every bit of skepticism it’s getting. What it doesn’t necessarily deserve is to drag down a GP Stakes business earning permanent fees from owning pieces of other asset managers, and a digital infrastructure operation leasing data centers to Meta, Microsoft, and Amazon on 15-year contracts.

Neither of those cares whether some software company in the loan book can service its debt.

Distrust doesn’t stay contained, though. When your largest business is under a cloud, it slows fundraising everywhere, compresses the multiple on everything, makes the whole harder to separate than the parts.

This is Part 1 of two. Here I discuss what Blue Owl actually is, and the five debates that determine whether the discount is earned. Part 2 does the math, stress-tests the multiple across three scenarios, and answers whether you should own it here.

Three Businesses, One Stock Price

Blue Owl is not one business. It is three, and the market is pricing all of them as if they share the same problem.

The first is credit. The second buys minority stakes in PE firms. The third owns data centers leased to the biggest technology companies in the world. Understanding why the discount may be too large, or too small, starts here. Here’s a quick summary:

Before getting into what is broken, it helps to understand how the machine works. Alt managers make money by charging fees on other people’s capital. The fee stream is the business. What separates the great firms from the average ones is permanence: capital that is locked up for years cannot be pulled in a bad quarter, which means the fee revenue is closer to an annuity than a subscription.

Blue Owl built its model around that principle. Roughly 95% of revenues come from recurring management fees, FRE margins run at ~58%, and at year-end 2025 FRE earnings were running at $1.5 billion annually. The current problem is concentrated in one of those three segments.

Direct Lending / Credit (~61% of FRE Revenues)

When a PE firm buys a company, it usually needs to borrow money to finance the deal. Historically that meant going to a bank, which would syndicate the loan to other investors. Blue Owl cuts out the middleman. It goes directly to the PE sponsor, offers to fund the entire loan itself, and charges a premium for the speed and certainty of execution that a single lender can provide.

The borrowers are PE-backed companies, mostly in software, technology, healthcare, and business services. The loans are first-lien senior secured, meaning Blue Owl sits at the top of the capital structure and gets paid before anyone else if something goes wrong.

The fee model charges ~1.5% per year on the total assets it manages, regardless of performance. That fee is collected from four separate vehicles. OBDC and OTF are publicly traded; anyone can buy or sell shares on a stock exchange, and the market prices them daily.

OCIC and OTIC are non-traded, sold exclusively through financial advisors to retail investors. There is no exchange. If an OCIC investor wants out, they submit a redemption request and the fund is only required to honor up to 5% of its total assets per quarter. That structure is the permanent capital mechanic: even under stress, the fee base drains slowly rather than all at once.

That same structure is precisely what created the current problem. OBDC II, a fifth vehicle that is now being wound down, replaced its quarterly redemption process with return-of-capital distributions funded by asset sales. Investors who wanted liquidity discovered the exit was smaller than they were led to believe.

The resulting overhang, combined with a surge in redemption requests at OTIC, means the credit fee base is being drawn down at a pace Blue Owl cannot control. Atalaya, acquired in 2024, adds a different flavor of credit: consumer loans, equipment leases, and specialty finance with shorter duration and hard collateral rather than corporate debt.

GP Strategic Capital (~24% of FRE)

This business has nothing to do with lending. Blue Owl buys small minority stakes, typically 10-15%, in private equity management companies themselves. Not their funds. The firm. The entity that earns management fees and carried interest in perpetuity, regardless of which fund it is currently raising.

Think of it this way: if you owned 10% of KKR’s fee-earning engine, you would collect a share of every management fee KKR charges on every fund it ever runs. That is the GP Stakes model. The PE firm gets liquidity and a strategic partner. Blue Owl gets a permanent, compounding revenue stream with no credit exposure, no mark-to-market risk, and no connection to what is happening in the BDC market.

The portfolio spans roughly 40 managers across five fund vintages, including Clearlake, Stonepeak, Veritas, and KPS. The track record is real: HPS was acquired by BlackRock in late 2024 at a significant premium, and CVC and Bridgepoint are publicly traded comps.

Fund VI is the current fundraise, targeting $13 billion. Existing fee streams are unaffected by whether it closes on time. What is at stake is whether the fee base grows.

Real Assets / IPI (~15% of FRE)

The Real Assets segment spans net lease, real estate credit, and digital infrastructure, with $80.6 billion in AUM across the three strategies. The piece most relevant to the current debate is IPI Partners, the digital infrastructure business acquired in January 2025.

IPI Partners builds and owns data centers, then leases them to hyperscalers on long-term contracts. The tenants are Meta, Oracle, Microsoft, and Amazon. The leases run 15 years with financial penalties for early exit.

The confusion in the market stems from headlines about data center financing stress in early 2026. Those stories were about speculative AI compute companies borrowing in public debt markets. Blue Owl’s counterparties are IG hyperscalers who have already committed to the capex and need the physical infrastructure to execute it.

Five Debates, One Question: Is the Discount Earned?

The first three debates are about deterioration already underway in credit. The last two are about whether the rest of the platform can grow through it.

Debate #1: Are the BDC outflows a temporary shock or permanent damage to the wealth channel?

Read more

The capital structure says you win. The relationship says otherwise.

Not in every deal. But in enough of them that the current conversation is incomplete in ways that matter.

The logic is intuitive. Private credit sits senior in the stack. Equity is first loss. If the business deteriorates far enough, the sponsor gets wiped and the lender takes the keys with value to spare. That is the theory.

It is also incomplete. Senior secured does not mean unimpaired. It means you get paid before the equity. In a real workout, that gap between theory and recovery can swallow a lot of NAV.

And the math ignores the room.

Lenders don’t run companies. The deal team that got comfortable after five GLG expert calls, a sponsor-provided McKinsey report, and a 20% haircut to the sponsor model is not suddenly equipped to manage a 500-person global manufacturing business through a workout. That is not a knock. It is just reality. Underwriting a business and operating one are different skills. The credit community has exactly zero of the latter.

Sure, management runs the business and the incentive plan gets reset. But management ran the day to day. The sponsor ran the strategy. A reset MIP does not replace that. The best people in the building figure that out first. They are also the ones with the most options.

So what actually happens when a deal goes sideways?

Take the keys, or negotiate. Almost every time, they negotiate. Blackstone knows this. KKR knows this. Apollo knows this. The threat of taking the keys is not a threat to the sponsor. It is a threat to the lender’s time on every other position in the fund.

Here is the hypothetical conversation happening in workout rooms that nobody is writing about.

Blackstone sits across from a direct lender on a bad vintage deal. Nobody saw the tariff exposure, or the AI disruption, or whatever the problem turned out to be. Blackstone says: we can hand you the keys today. Or we can extend the maturity, restructure the debt, roll some equity into a new structure, give you upside participation on the recovery. We stay, we operate it, we try to get you out whole.

And then comes the part that does not appear in any textbook:

We have significant upcoming deal flow. The kind you have been trying to get allocation on for two years. This was a bad vintage. We are prepared to treat it as one. But if you play hardball here, we will remember that when the post-dislocation cycle opens up. Maybe your competitor gets that allocation instead.

Tell me that conversation does not happen and the lender does not blink.

Direct lending is a repeat game with a very small number of counterparties. A lender who torches a relationship over one bad deal is making a locally optimal, globally irrational decision. The sponsors know the math. They are playing it.

This is not a universal law. It is a practical reality that shapes how a meaningful number of these situations actually resolve. And it depends entirely on the sponsor.

There are a lot of zombie PE firms out there right now. Funds that are not raising another flagship. Funds where the deal flow argument carries no weight because there is no deal flow. When the sponsor has no future leverage the negotiating dynamic collapses. And that actually makes the point sharper. The strategic value of a top-tier sponsor relationship becomes more important precisely because it is scarce. Lenders know who is going to be in business in five years. That calculus is in every conversation whether anyone says it out loud or not.

The lenders underwrote the business. The sponsors know how to run it. Those are two very different things. And in a stressed credit, only one of them matters.

That dynamic is playing out against a backdrop that makes the workout math harder across the entire credit complex. Not just in private credit.

The Workout Is the Easy Part

The dominant narrative is AI disruption layered on an Iran oil shock. Both real. Both getting all the analytical resources. Both obscuring something building in completely different parts of the credit universe.

Old-fashioned credit stress is broadening. Nothing to do with software or the Strait of Hormuz.

Start with non-prime consumer lending. Non-prime borrowers have the least buffer and crack first. When non-prime lenders are in covenant discussions with their own securitization lenders the question is not whether consumer stress exists. The question is how far up the income ladder it has already traveled. Further than the headlines suggest.

Building products is running a parallel script. Margin compression with no macro catalyst attached to it. Soft housing. No policy support. PE-backed structures leveraged for a rate environment that never materialized. Pure cyclical deterioration in a sector that has nothing to do with AI or geopolitical risk. Nobody is writing about it because there is no narrative hook. The stress is just there, quietly compounding.

Airlines are the same story with a more visible mechanism. Unhedged fuel exposure, no pricing power, and an oil shock doing what oil shocks do. This one does not get better while crude stays at $100.

Three sectors deteriorating simultaneously for completely different reasons. The market is calling this a sector rotation. It isn’t.

One more thing on Iran that is not getting airtime. The focus has been entirely on the oil price. Almost nobody is modeling what a sustained Hormuz disruption does to trade finance. Letters of credit. Shipping insurance. Working capital facilities for companies with Asian supply chains. These reprice quietly when underwriters decide a major global chokepoint is under active military threat.

The oil price move is visible and already in spreads. The trade finance tightening is not. If this disruption runs another 60 to 90 days the second-order effects start showing up in working capital lines and covenant calculations at companies where your credit model currently says nothing is wrong. By the time it is legible in the data it will already be in the price.

All of this matters for the workout dynamic described above. Every incremental deterioration in the macro backdrop narrows the sponsor’s negotiating options and widens the gap between what the lender can recover in theory and what they can recover in practice. The repeat game gets more expensive to lose every week.

The Market Is Solving for the Wrong Variable

HY BBs are down 1.35% month to date. CCCs are down 1.07%. In every risk-off episode in recent memory quality outperforms. Everyone learned that playbook. Everyone is currently executing it.

It is the wrong playbook.

BBs are the worst performing ratings bucket in high yield right now and the reason has nothing to do with credit quality. The 10-year moved ~25bps wider this month. BBs carry the most duration in high yield. The accounts that historically stabilize the BB tier in a risk-off move, crossover IG buyers, insurance companies, total return funds, are themselves forced sellers when rates spike this fast. The natural shock absorber for BBs in a risk-off environment becomes a source of supply in a rates-driven one.

This is a rates selloff dressed as a credit event. If you are treating 311 bps as a spread problem you are looking at the wrong number.

The mechanism that would normally correct this is unavailable. The Fed cannot cut into 3% inflation with oil at $100. It cannot signal hikes without accelerating the private credit unwind already building. The Fed put that has underwritten credit market risk-taking since 2008 is not suspended. It is logically unavailable given the current configuration of data.

The forward curve is still pricing ~1 cut for the full year. That residual expectation is the last piece of the Fed put not fully removed from market pricing. Watch Wednesday’s press conference. Not for what Powell says. For what he will not commit to. The silence is the signal.

The Primary Market as the Honest Signal

Moog priced $500 million at 5.50%. Defense and aerospace. Traded to 101 immediately. Kodiak Gas Services printed $1 billion at 5.875%. Natural gas infrastructure. Upsized. Clean.

FICO came at the midpoint of guidance and slipped to 99 on the break. Penn priced and immediately backtracked to 99. The deal cleared. The investor concern about profitability did not.

Then there is EA. The largest LBO on record is on deck for the primary market. When it launches it will be the most visible real-time test of investor appetite for large software-adjacent credits in the current environment. The primary marketing task will be convincing investors the company is not actually a software company. That sentence tells you everything about where the sector discount sits right now. Watch how it prices. Watch whether it upsizes or gets cut. The outcome will be more honest than any spread quote published this week.

EA is the exception that proves the rule. The rest of the primary market is not testing anything. It already knows the answer.

All but two deals this week were BB refis. Looks like health. The selection effect is so extreme it has inverted the signal entirely. Each month the primary market stays closed to lower-quality issuers more paper gets pushed into a narrower refinancing window at worse terms. The calm primary market is making the maturity wall problem worse by appearing to solve it.

By the time the compressed maturity wall becomes visible it will already be a negotiating problem rather than a refinancing problem. Those resolve differently. With worse outcomes for existing creditors.

While the leveraged credit primary market quietly closes its doors to lower-quality issuers the investment grade market just had one of the largest issuance weeks in its history. Those two facts are connected.

How I’m Positioning

Read more

Miami was fine.

The panels were polished. The networking was productive. The wifi barely worked. The elevators never came so you took the stairs. You did your 1x1s, worked the evening events, ate your boxed lunch in the overflow room as Jamie Dimon gave his fireside chat. Same as every year.

Except the mood wasn’t.

Software disruption dominated every conversation. Analysts cornering management teams about AI risk. The loan market coming off its worst monthly return since 2022. Iran hit over the weekend and suddenly you were managing two crises at once.

You showed up to talk about AI disruption. You left talking about oil and inflation. The dark joke nobody said out loud is that those two narratives run in direct opposition. AI is deflationary. Energy shocks are inflationary. You can’t hold both frameworks simultaneously and come out with a clean trade.

Then you checked your screens.

Oil up 30%. Payrolls -92,000. Private credit redemptions hitting records at multiple major managers while HY spreads stayed flat.

Something in that picture doesn’t add up. Several things, actually.

Here’s what didn’t make the agenda.

The Software Question Everyone Is Asking Wrong

The debate has hardened into two camps and both are missing the point.

Camp One: software credit is uninvestable until the terminal value question gets resolved. The tech loan bid at ~88 cents, AI disruption accelerating, the 2027/2028 maturity wall approaching. Stay away and wait for clarity.

Camp Two: the market is indiscriminately selling. “Mission-critical” platforms with embedded enterprise workflows, high switching costs, and proprietary data are not going to be replaced by a chatbot.

Both camps are right. Both camps are missing something that neither wants to say out loud:

The AI disruption trade is self-fulfilling in a way that changes the math entirely.

Spread widening at the sector level starves software companies of the capital they need to invest in AI. The companies most likely to successfully navigate the transition are precisely the ones that need continued capital access to build the defensive moat: new product development, engineering headcount, customer success investment, etc. By collapsing capital access indiscriminately across the entire sector, the market is selecting against the survivors. It is selecting for the impairment it is trying to price.

The software company that would have refinanced at par six months ago and used the proceeds to build AI-native product features is now facing a 200 bps new issue concession, a lender group demanding covenant resets, and a sponsor increasingly reluctant to inject equity into a narrative-impaired asset. The business that might have adapted is now less likely to adapt. The spread widening made the disruption thesis more true.

This is not a normal credit cycle dynamic. In a normal cycle, spread widening reflects deterioration that has already happened. Here, spread widening is causing deterioration that has not happened yet. The credit market is not observing the outcome. It is partially creating it.

Camp One is right that you cannot underwrite a mediocre software refinancing at par when the terminal value question is unresolved. The A&E math does not give you par. It gives you coupon plus a fee, which is 8-10% on an instrument you are underwriting as a 12-15% YTM. The pull-to-par is not coming.

But the self-fulfilling dynamic eventually creates a clearing event. The companies that survive the capital starvation phase will have demonstrated something the credit market cannot currently price: genuine AI resilience under stress. When the narrative breaks, the recovery in those names will be fast, violent, and concentrated.

The Fed Is Trapped and the Market Hasn’t Priced It

The self-fulfilling dynamic would eventually resolve itself if the macro environment were cooperating. It is not.

Before Iran, the Fed’s path was already narrowing. Core inflation stuck above target, break-evens drifting higher, a labor market that was softening but not breaking.

Then oil moved ~30% in a week and the conversation changed entirely. The forward curve is still pricing meaningful cuts by year-end. Payrolls at -92,000 would normally be the green light. With $90 oil and rising break-evens it is not. Cutting into that environment risks unanchoring expectations in a way that makes the problem structurally worse.

But staying on hold creates a different problem. Every LBO model underwritten assuming significantly lower SOFR by now is running a cash flow shortfall that compounds quarterly. The borrowers who need relief are precisely the ones the Fed cannot help without making the inflation problem worse.

The honest framing: the Fed has one rate for two economies.

Economy One is AI infrastructure, hyperscaler capex, and defense. Completely insensitive to rates. Spending hundreds of billions this year regardless of what Powell does.

Economy Two is everyone in leveraged finance who borrowed at 6-8x leverage assuming rates would normalize. Being slowly impaired by the same rates that are irrelevant to Economy One.

The “Fed put” that has underpinned credit markets since 2008 is functionally suspended. Credit spreads are still pricing it as active. That gap is where the risk is.

CLO Demand: The Invisible Systemic Risk

If the Fed cannot provide relief, the loan market is left depending on a single technical support that most investors are not watching carefully enough.

The leveraged loan market has effectively one marginal buyer right now. Not retail, which has seen billions in outflows YTD with February among the worst months in recent memory. Not hedge funds. CLOs.

Here is the mechanism that matters.

When a CLO fails its OC test because software loan prices have fallen, the instinct is to sell the problem assets. But most CLO structures calculate overcollateralization on par value, not market value. Selling a software loan at 87 cents takes a par hit that makes the OC test worse, not better. So managers do the opposite. They sell their best performing positions at or above par instead. The chemicals credit at par. The energy name above par. The industrial issuer that just reported strong earnings. The OC test gets fixed. The CLO looks clean on paper. A fundamentally sound credit in an unrelated sector is wider for no fundamental reason.

This happened in 2015-16. CLO managers sold their best non-energy assets to rebalance structures stressed by oil patch deterioration. Credits with nothing wrong widened for months. Almost nobody recognized the mechanism in real time. The investors who did cleaned up on the other side.

Software is 2016 energy. The collateral damage will show up in sectors where your credit work says nothing is wrong. That is not a reason to avoid those sectors. It is a reason to have the buy list ready when the widening arrives.

The deeper problem: the supply/demand gap in loans is being bridged almost entirely by CLO formation. Multiple tech deals have been pulled from primary since late January. As CLO managers become more selective about eligible collateral, the feedback loop from tighter CLO standards to wider primary pricing to more pulled deals becomes self-reinforcing. It does not require a crisis. It just requires the marginal CLO deal to stop making economic sense.

Primary Market as Live Confirmation

You do not need to model any of this. The primary market is showing you in real time.

Two markets are running on the same screen right now. Energy and infrastructure printing clean, upsizing, trading up. Software-adjacent situations shelved, delayed, or restructured before launch. The line between them is not rating or sector in the broad sense. It is whether your business model has a question mark attached to it that a credit investor cannot underwrite through to a confident answer at current prices.

The EA financing makes the point cleanly. The debt package for the largest LBO on record is shifting heavily toward bonds and away from loans. When banks underwriting a marquee buyout make that call, they are not making an academic observation. They are telling you directly that loan investor appetite for large software-adjacent credits is impaired at scale. Deal structure is more honest than spread quotes.

The fact that the EA pitch had to be made at all tells you where the market’s head is. Bankers have been spending their time assuring anchor investors that EA is not a software company in the disruptable sense. The IP in Madden NFL and The Sims is defensible. Which is a sentence that has never been uttered in a leveraged finance context before and yet here we are.

The pitch is reasonable and probably correct for EA specifically. But if you are running a $55B buyout and your primary marketing task is convincing investors your software company is not actually a software company, the sector discount is real and it is structural.

Private Credit: The Ratchet

The public market signals are clear. What makes this cycle harder to manage is that the stress is building simultaneously in a market where the signals are deliberately obscured.

This is not a liquidity problem. It is a trust problem. And trust problems do not respond to liquidity solutions.

When Blackstone injects firm and employee capital to avoid formally changing its tender offer terms, that is reputation management. Retail investors are not pulling money because they need liquidity. They are pulling because they do not believe the marks. Publicly traded BDC peers pricing at significant discounts to NAV suggests the institutional market agrees.

Here is what makes this structurally different from a normal credit cycle. Redemption pressure forces asset sales. Managers sell their best assets first because those are liquid and at par. The assets remaining are more concentrated in marks-at-model positions that have not been independently validated. The average NAV stays roughly stable. The actual quality of the collateral behind the NAV quietly deteriorates.

Each redemption cycle that forces par sales of the best credits leaves a portfolio incrementally more concentrated in the assets nobody wanted to buy. The ratchet does not require a crisis to operate. It just requires inflows to stay below outflows for a few more quarters.

What is actually inside those remaining portfolios is the part of this story that is not being told clearly enough. The 2020-2022 vintage MM PE playbook was built on one core assumption: labor costs are stable and manageable.

Read more

32% off reported NAV. A $17 stock trading at $11.82. Private credit, the hottest asset class of the last decade, suddenly democratized and available to anyone with a brokerage account.

You know the pitch. The market is wrong. Everyone is scared. You’re the smart money. You’re buying a dollar for 68 cents while the tourists panic.

But what if the market is not panicking? What if it has looked at this portfolio carefully and keeps arriving at the same conclusion?

OTF is a $14 billion portfolio of software loans, many tied to companies taken private at peak multiples during the free money era. Blue Owl manages it.

I went through the borrowers, reclassified the portfolio by what the companies actually do and when the deals were originated, then compared the marks against where similar paper is trading today.

The result was not comforting.

What OTF Actually Is

OTF is a $14.3 billion book of loans, almost entirely floating rate, to software and technology companies. Blue Owl manages it externally. The portfolio spans 199 companies across 39 industries, with 77% in 1L positions and avg. portfolio company EBITDA of $290 million. Leverage sits at 0.75x debt-to-equity, which is moderate by BDC standards.

Liquidity, cash plus undrawn capacity, is $2.3 billion. At the current stock price of $11.82, the regular dividend alone yields 11.8% annualized, 13.5% including the special dividend. On paper, it looks like a well-structured vehicle offering attractive income from a diversified credit book.

Two numbers from the filing are worth sitting with before accepting that framing. The first is the weighted average total yield on the portfolio, which fell from 10.9% in 4Q’24 to 9.6% by 4Q’25, a 130bps compression in a single year driven by SOFR moves.

The second is fair value as a percentage of principal on the debt book, which the filing reports at 99.2%. That number is not wrong. It is the number as of December 31, right before the tech sell-off. Both of those facts matter for what comes next.

Blue Owl’s own investor deck describes OTF as a “Defensively Positioned, Diversified Software Portfolio.” That phrase is worth sitting with for a moment, because both adjectives are doing considerable work.

Start with “diversified.” OTF is explicitly a technology lender, so meaningful tech concentration is expected and by design. The issue is not the concentration itself. It is that the filed sector breakdown actively obscures its extent, with consequences for how investors model a tech credit event. The filing breaks exposure across buckets like Systems Software, Application Software, Health Care Technology, and Diversified Financial Services. Systems Software and Application Software combined come to 31.5%. That figure creates an impression of balance that the underlying portfolio does not support.

Now look at what they left out.

Read more