Tether Entered Private Credit When Defaults Hit a Five-Year High: A Structural Read of the StableFund

Hook

The number that matters is not the $3 billion target. It is 2.8%.

Blue Owl — one of the largest managers in the roughly $3 trillion private credit market — reported a 2.8% default rate in its most recent quarter, the highest in at least five years. Publicly traded private credit funds are carrying defaults not seen since 2021. Redemption pressure across the vehicle class is climbing. In May, the Financial Stability Board published a warning that private credit has never been stress-tested through a full economic downturn, and specifically flagged redemption-gated vehicles as a liquidity weak point.

Tether and Fasanara Capital announced StableFund inside that window.

Then the details, as reported publicly: a target raise of up to $3 billion. A $400 million anchor between the two firms. An evergreen vehicle. Short-duration, asset-backed lending across a 60-country fintech network covering SME financing, consumer credit, trade receivables and supply chain finance. Tether positioned as co-sponsor, originator and advisor. Fasanara as investment manager.

And then the silences, which are the actual story. No leverage figure. No fee structure. No subordination waterfall. No redemption terms. No statement of whether USDT is the funding asset, the collateral, or merely the accounting unit. No named decision-maker for final credit approval.

Seven inputs. Zero numbers. That is not a press-release problem. Under an evergreen structure with undisclosed gates, the absence of those numbers is the product. Code doesn't lie, but markets do — and so do offering memoranda written to be read only by people who already agreed.

The Three-Trillion-Dollar Room Tether Just Walked Into

Private credit is the largest quiet pool of capital in modern finance. It is roughly a $3 trillion asset class by most estimates, and it grew because banks retreated from middle-market lending after 2008 and because a decade of near-zero rates made yield scarce. The business model is simple in the abstract: raise capital from institutions, lend directly to companies that banks no longer want to underwrite, charge a floating spread over a reference rate, collect fees, and mark the book with models rather than market prints.

Four names dominate the top of that stack — Ares, Blackstone, Blue Owl and Golub. They have scale, brand, decades of workout experience, and the ability to warehouse risk across multiple vehicles. Fasanara Capital is not in that tier. It is a London-based asset manager with roughly $6 billion under management, a network that extends across more than 60 countries, and a specialization in fintech-originated and asset-backed credit. That last phrase is doing real work, and I will get to it later, because "asset-backed" and "safe" are not synonyms and never have been.

Tether's position on the other side of the table is more interesting than the market treated it. Tether is not merely a stablecoin issuer. According to the reporting, Tether already accounts for roughly 60% of a $23 billion crypto lending market, with something on the order of $13.5 billion in outstanding loans. That is a book of collateralized crypto credit that, in a bear market, has one structural characteristic most people ignore: its collateral is priced by the same order books that price its borrowers' solvency. When the market gaps, collateral and credit quality gap together.

So read the StableFund announcement as what it structurally is: a diversification of a lending franchise away from a single, highly correlated collateral base, into a nominally uncorrelated one. That is rational. It is also exactly what a risk committee would propose in a drawdown. I have watched enough of these proposals to know that the rational version and the executed version are frequently different documents.

I have a frame for this that comes from a failed trade. In 2020, during the DAI–USDC depeg, I ran a small arbitrage bot on Uniswap V2 with $500 of my own savings, tuning gas and pool weights off live block data. It executed 47 profitable trades in 72 hours for about $320 net — and then died on a reentrancy path I had never audited, because I had read the whitepaper and never the bytecode. The lesson was not "arbitrage is risky." The lesson was that the document that describes the system and the system that runs are two separate artifacts, and the second one is the only one that settles. StableFund is a document right now. It has not settled anything yet.

The Structure, Stripped

Strip the marketing language and you have three structural decisions that determine everything about how this fund behaves under stress.

Decision one: evergreen versus closed-end. StableFund uses an evergreen structure. That means it can raise and deploy continuously, with no preset wind-down date. Practitioners like evergreen because it smooths fundraising and avoids forced sales at bad moments. The trade-off is that an evergreen vehicle never has a terminal date at which the book must be realized at a market-clearing price. In a closed-end fund, the wind-down is a valuation discipline mechanism — eventually, you find out what the assets are worth, because you have to sell them. In an evergreen fund, that moment can be deferred indefinitely, and the liquidity you offer investors has to be manufactured from cash flow and subscriptions rather than from an auction.

This is not a theory. It is the mechanical reason interval funds, nontraded REITs and evergreen private credit vehicles have historically shown smoother NAV charts than their listed counterparts. The smoothness is a feature of the redemption mechanic, not evidence that the assets are less volatile. When break-the-buck pressure finally arrived at nontraded REITs in 2022 and 2023, the response was to use the gates that always existed in the fine print — quarterly caps, pro-rata proration, and in some cases a suspension of repurchases. Nobody defaulted. But a lot of investors discovered that "quarterly liquidity" meant something narrower than they assumed.

If 2.8% is the market's current default print, then the relevant question for an evergreen lender is not "what is the default rate" but "at what point does the redemption queue exceed the cash generation, and what happens then."

Decision two: who originates, who manages, who advises. Per the reporting, Tether is co-sponsor, asset originator, and advisor. Fasanara is the investment manager. The article on which this analysis is based explicitly flags that the final authority over credit approval is not stated. That single omission is larger than it looks. In a lending operation, the credit decision is the business. Origination feeds the pipeline, management services the book, but somebody signs the loan. If that somebody is not named, then you cannot know whose credit culture governs — a manager with a fintech and asset-backed bias, or an originator with a stablecoin treasury and a growth mandate.

Decision three: what USDT actually does here. This is the one I keep circling. The reporting does not establish whether USDT is the funding currency of the loans, the collateral posted, the settlement rail between originators and the fund, or simply a reporting unit. Those are four different funds with four different risk profiles. If USDT is the funding currency, the fund carries stablecoin redemption risk alongside credit risk. If USDT is the collateral, you have a mark-to-market loop between a $1 peg and a loan book. If USDT is just the rail, the fund is a conventional credit vehicle with an unusual settlement layer — interesting but not structurally novel.

I cannot price a fund when I do not know which of those three it is. Neither can an institutional allocator. Which raises the question of who is actually subscribing at the $3 billion target.

The Carry Trade Nobody Printed

Here is the arithmetic the announcement politely declines to publish, and it is the reason a stablecoin issuer — not a bank — is the natural entrant to this trade.

USDT is a zero-yield liability to its holders. That is the defining feature of the instrument. Tether earns on the reserve backing it and pays nothing to the people holding the token. Its cost of funds, at the margin, is close to zero in nominal terms, minus operations, attestation and compliance. That is the lowest funding cost in modern finance for anything that behaves like a deposit.

Short-duration asset-backed lending, in the segments Fasanara targets, generally prices at a floating spread over a reference rate. Even with policy rates off their peak, that spread is meaningfully positive in absolute terms. The gross carry on zero-cost funding against a floating-rate, short-duration credit book is, on paper, extraordinary. Banks cannot match it, because banks pay something for deposits and carry capital charges against the exposure. Money market funds cannot match it, because they are constrained to near-risk-free assets.

So the pitch writes itself: take the cheapest liability in the system, match it to short-dated self-liquidating assets, and capture the spread. It is a textbook carry trade. And a carry trade has exactly one failure mode — the spread is real until the credit losses exceed it, and then it is leverage in disguise.

This is where volatility enters the conversation. Volatility is just unpriced risk — and in private credit, most of the risk is unpriced because most of the book is unmarked. Private credit assets are not marked to a live market. They are marked to models, to comparable transactions, and to the manager's judgment, reviewed by an auditor on a schedule. That is not fraud. It is a known and disclosed feature of the asset class. But it means the reported NAV of the book is a smoothed estimate of an unknowable number, and the true volatility only reveals itself when a borrower actually stops paying or a credit is restructured at a discount.

The FSB's May note put a specific marker on one of these mechanisms: rising use of payment-in-kind provisions. PIK lets a borrower defer cash interest and roll it into the loan balance. From the lender's income statement, the interest is still being recognized. From the lender's cash flow statement, it is not. When PIK usage rises across a book, reported yields stay flat while cash collection quietly deteriorates, and you only find out in the workout. This is not a niche footnote. It is one of the ways a credit book looks healthy in the quarter before it doesn't.

If Tether is funding a floating-rate book with a zero-cost liability, the carry is enormous and the tail is entirely in credit. If Tether is also taking subordinated exposure — first loss — in exchange for that carry, then the carry is not a spread. It is an insurance premium, and the risk on the other side lives in the reserve that backs USDT.

Seven Inputs, Zero Numbers

Let me be precise about what is missing, because vagueness here is how people lose money.

Leverage. Not disclosed. This is the single most important input for any credit fund, because leverage converts a modest default rate into a wipeout of equity. A 2.8% default rate on an unlevered short-duration book is manageable with recoveries. The same default rate on a 3x levered structure with first-loss exposure is a solvency event. Without the number, the 2.8% figure is uninterpretable in this context.

Subordination. Not disclosed. Who sits first in the waterfall? If Tether's $400 million is junior to third-party institutional capital, Tether is taking material first-loss risk, and the question becomes whether that risk is segregated from the reserve backstopping USDT. If the $400 million is pari passu, the risk is shared. If it is senior, then Tether's participation is largely a credit-enhancement story for other investors, and the press framing as a bold expansion is mostly marketing.

Redemption terms. Not disclosed. Frequency, notice period, gates, side pockets, suspension rights. The FSB specifically flagged redemption vehicles. When a fund offers periodic liquidity against assets that cannot be sold quickly, the redemption right is a promise, and the gate is the tool that lets the manager break it. I do not object to gates. I object to buying into a structure without knowing where the gate is.

Fees. Not disclosed. Management fee, incentive fee, hurdle, catch-up, origination fees at the asset level, servicing fees, and whether any of them are waived for affiliates. Fee structure is also a signal — it tells you who is being paid for what, and affiliate fee flows are where sponsor-aligned structures quietly transfer value.

USDT's role. Not clarified. Covered above. It determines whether this is a stablecoin story or a credit story wearing a stablecoin costume.

Final credit authority. Not stated. Covered above. It determines whose underwriting standard applies.

Jurisdiction and legal form. Not stated. Evergreen funds are commonly domiciled in jurisdictions chosen for regulatory efficiency rather than transparency. That is routine. It still matters for enforcement, tax, and which regulator has visibility.

Seven inputs. Zero numbers. In my 2025 hackathon work simulating compliance checks on a DeFi lending protocol, the finding that ended up mattering most was not a bug in the code — it was that three governance functions had effective control concentrated in a path the documentation never described. The same discipline applies here. In a fund, the documentation is the code. If the documentation does not describe who holds the keys, the keys are held somewhere the buyers were not shown.

The Evergreen Problem

There is a structural mismatch at the heart of every evergreen credit vehicle, and StableFund inherits it by construction.

The assets Fasanara targets are short-duration and self-liquidating: trade receivables, supply chain finance, SME working capital, consumer credit. In theory, that means the book turns over quickly — money goes out, money comes back with interest in weeks or months. Good liquidity profile, low duration, low interest-rate sensitivity. That is the sales pitch, and it is not dishonest.

But short duration on the asset side does not eliminate mismatch on the liability side. If the fund offers redemption on a short cycle, investors can queue faster than receivables can be collected, because collection is a legal process with counterparties in 60 countries and redemption is a button. The mitigation is gates and notice periods. The problem is that gates are most likely to be used exactly when everyone wants out, which is when the reputational cost is highest and the temptation to keep them tight the first time is strongest.

The deeper issue is that an evergreen vehicle with no terminal date has no natural moment at which the assets must be marked to reality. A closed-end private credit fund that winds down in year seven forces a reckoning. An evergreen fund can roll forever, re-marking gently, writing down one loan at a time, and reporting a stable NAV through the whole process. The NAV is not a lie. It is a smoothed estimate that has never been tested by an auction.

Liquidity is the only truth. Price is an opinion until somebody is forced to trade. That is why I trust auction-based marks over model-based marks, why I trust deposit and withdrawal flows over stated AUM, and why the most important number in the StableFund announcement is not $3 billion or $400 million but the one nobody has published: the redemption terms.

What "Asset-Backed" Actually Means in Sixty Countries

The strongest structural defense of this fund is its asset selection. StableFund is reportedly targeting short-duration, asset-supported lending — not long-dated direct corporate loans of the kind that dominate the stress reports. That distinction is real. Direct lending to a mid-market company for seven years at a floating spread has term risk, refinancing risk and enterprise-value risk. Factoring a trade receivable that self-liquidates in 60 days has credit risk, operational risk and fraud risk, but not much duration risk.

There is a precedent that every analyst should keep in the front of their mind when hearing "supply chain finance," and it is Greensill. Greensill Capital built a large, apparently low-risk book of supply chain receivables and reverse factoring, packaged it for institutional funders, and collapsed in 2021 — dragging down funds that had trusted the structure and the markings. The failure mode was not duration. It was the integrity of the underlying receivables and the concentration of a single obligor hidden inside a diversified-looking book.

This is the frailty of asset-backed lending in a 60-country fintech network. The assets are genuinely diversified across geographies, which reduces correlation. They are simultaneously genuinely difficult to underwrite at the margin, because each jurisdiction has different collateral law, different collection practice, different fraud exposure, and different enforceability. A short-dated book is easy to mark when the marks come from a servicing system the fund controls. It is hard to mark when the question is whether the receivable exists, whether the same invoice was financed twice, and whether the counterparty is who the onboarding file says it is.

None of this is an accusation. Fasanara's entire franchise is built on operating this kind of book, and that is a genuine competence. The point is that "asset-backed" is a claim about the pool's liquidity, not about its credit quality, and the two are frequently confused by audiences who hear the phrase and stop listening. In a 60-country SME and consumer book, the actual risk drivers are origination fraud, servicing capability, and the correlation between borrowers who are all dependent on the same fintech lenders and the same macro cycle.

Short-duration, asset-backed, diversified. Three traits that each reduce one risk and leave another untouched.

The Subordination Question and the USDT Transmission Channel

Here is the scenario I have been unable to rule out, and it is the reason I would treat this announcement as a risk-management event rather than a growth event until the terms are published.

Suppose Tether's $400 million is subordinated. Suppose it takes first loss. Suppose the fund is levered beyond what has been disclosed. Now trace the transmission path: USDT reserve backing includes or is associated with the capital in this fund; the fund's lending book deteriorates as private credit defaults rise; the fund draws on Tether's tranche; the loss is absorbed at the sponsor level; the reserve backstopping USDT absorbs the hit; USDT's credit standing shifts from functionally risk-free to merely well-capitalized.

That chain is not inevitable. It is conditional on the subordination structure, which is exactly the thing that has not been disclosed. But it is the only channel through which this announcement can matter to the price of USDT, and it deserves to be stated plainly instead of buried under expansion narratives.

The mitigating factors are real. Tether's reserve portfolio — per its quarterly attestations, historically dominated by short-dated Treasuries and repo — generates substantial income, which is a buffer. Tether's operating cash flow is enormous. Fasanara is a competent credit manager. And the underlying assets are short-dated and self-liquidating, which limits the loss given default versus a long-dated book.

The aggravating factors are also real. Tether already carries $13.5 billion in crypto-backed loans, a book whose collateral value correlates directly with the market cycles that stress USDT demand. StableFund adds a second, differently-shaped credit book to the same balance sheet, one whose marks are less visible and whose losses would arrive more slowly. A hedge fund running two uncorrelated credit books is diversification. A stablecoin issuer running two credit books — one of them unmarked — is a different animal, because the liability side of the balance sheet is a $1 peg that people hold as if it were cash.

I don't predict. I react. So I will state the trigger rather than the forecast: if the terms of this fund are published, and they show Tether subordinated, and they show leverage above a modest level, then USDT is no longer a pure reserve-backed instrument and should be valued accordingly. Until then, the correct posture is uncertainty, not comfort.

The FSB Said It First

The timing of this announcement relative to the FSB's May warning is not coincidental, and the source reporting appropriately led with it.

The FSB's specific concerns about private credit cut directly across StableFund's design: weakening borrower quality, elevated leverage, valuation opacity, growing interconnection with banks and insurers, rising PIK usage, increasing defaults, and liquidity risk in redemption-enabled vehicles. That last item is StableFund's structure. The second-to-last overlaps its market. Two of the seven concerns map onto its balance sheet and one maps onto its investor-facing mechanism.

When a global standard-setter publishes a fragility list and a product launches inside it months later, the correct read is not that the product is illegal. It is that the product will be examined by regulators with a pre-written list of questions, and that its terms will eventually be discovered whether or not the sponsor discloses them voluntarily.

There is a second regulatory vector that the source reporting does not address but that any reader should hold in mind. USDT operates under a patchwork of rules across jurisdictions, and the direction of travel in major markets is toward more reserve transparency, more reporting, and more constraints on where reserve assets can be deployed. A stablecoin issuer that puts reserve-adjacent capital into a private credit fund is creating a disclosure problem for itself in regimes that increasingly care about what backs the token and how correlated the backing is to the issuer's other bets. Fund shares themselves, evaluated under a standard investment-contract analysis, would almost certainly be securities, which puts the offering squarely in private-placement territory with its own investor-sophistication and disclosure obligations. None of that is fatal. All of it is why the missing terms matter more than the announced numbers.

Crypto-Native Lending Versus This

It is worth being precise about what StableFund is not, because the reflexive comparison in crypto is to on-chain credit protocols.

Aave and its peers are not comparable. On-chain lending is overcollateralized, marked continuously, liquidated by code, and visible to anyone with a block explorer. Its risk is transparent even when it is ugly. Maple and the institutional on-chain credit platforms are closer in spirit — they intermediate real credit through a blockchain-adjacent structure — but they remain, at least in their on-chain form, more legible than a private fund, because the exposure and the repayment flow touch a chain.

StableFund, as described, is a conventional private credit vehicle with a stablecoin issuer as sponsor and a stablecoin as a possible settlement and funding layer. That combination is genuinely novel at scale, and it is the reason the announcement matters. It is novel because it puts the cheapest liability in crypto against the least-marked asset class in traditional finance. If that structure works, it becomes a template: stablecoins as the funding backbone for real-world credit, with the blockchain relegated to the settlement rail where it belongs.

If it fails, it also becomes a template — the cautionary one, cited every time a stablecoin issuer proposes to deploy reserves into a real-world credit book. That asymmetry is worth naming. Infrastructure outlasts innovation; if this becomes infrastructure, the details will eventually get standardized and disclosed, and the structure will be replicated by every issuer with a large reserve and a low cost of funds. If it becomes a cautionary tale, the details will also get standardized, through regulation rather than markets.

What I Would Actually Monitor

When I built a monitoring harness for the GBTC premium and discount ahead of the 2024 ETF approval, I processed more than ten thousand hourly snapshots just to establish a baseline for a trade that lasted a few weeks. The lesson was that monitoring only becomes useful when you define the signal before the event, so you cannot rationalize afterward. Here is how I would define the signal for StableFund.

One: track USDT supply as a proxy for stablecoin demand. A sustained increase in supply alongside no disclosed deployment data would suggest the fund is a demand narrative rather than a demand source. A static supply with a growing fund would suggest the fund is funded from existing balances, which is a different and more interesting signal.

Two: track Tether's attestation cadence and reserve composition for any appearance of new categories of exposure. The reserve is the only public window into where the money is. If a new line item appears that plausibly corresponds to fund exposure, that is your first real disclosure, and it will arrive through accounting before it arrives through a press release.

Three: track private credit default and redemption data across the listed managers. If the 2.8% print is a leading edge rather than a peak, the fund's launch conditions are deteriorating into its deployment period, which is the worst sequence for a new book because early vintages are underwritten before the cycle is obvious.

Four: track FSB and major-regulator publications. Anything naming evergreen credit vehicles, PIK usage, or stablecoin reserve deployment is a direct input to this fund's operating environment.

Five, and this is the one most people skip: write the parser. I ran an LLM agent against my dashboard last year to filter news sentiment against on-chain whale movements, and found that AI-flagged sentiment aligned with subsequent price movement only about 12% of the time without human verification. I spent a weekend refining the filter and cut false positives by roughly 40%, but the failure mode never disappeared — the model could not distinguish a structural change from a headline about one. So I use the agent for speed and I read the primary documents myself for logic. Reading an offering document is a manual process. It is also the only process that works, because the whole game is noticing what is absent, and statistical models are trained to notice what is present.

Six: define the trigger and pre-commit. If terms are published and subordination is senior, downgrade the concern. If terms are published and subordination is junior with leverage above a threshold, upgrade. If terms are never published, treat the absence itself as the disclosure, because a sponsor with nothing to hide generally finds a way to say so.

Contrarian: You Are Watching the Wrong Ticker

Retail coverage of this announcement has framed it as a stablecoin story — Tether expanding, USDT gaining a new use case, the ecosystem growing. That framing is derived from the press release, which is the one document guaranteed to describe intent rather than mechanics.

The order flow says something different. Institutionally, a fund that raises up to $3 billion against a $400 million anchor is a fund that expects third-party capital to do the heavy lifting. The anchor is a credential, not the book. If the anchor were the book, the structure would look very different, and the terms would be public. So the relevant question is not "what does Tether want" — it is "what does the third-party allocator need in order to write a $2.6 billion check," and the answer to that is a full term sheet with leverage, waterfall, gates and fees. Which means those terms exist. They are simply not in front of you.

That asymmetry is the contrarian point. The market is reading a headline about expansion. The allocators, who are the only people whose money moves, are reading a private placement memorandum. When the terms eventually leak or get disclosed, the market will reprice the narrative against them — and the direction of that repricing depends on numbers retail has never seen.

There is a second blind spot. Retail observers consistently treat "asset-backed" as a risk adjective. It is a liquidity adjective. Trade receivables are liquid in the sense that they self-liquidate. They are not liquid in the sense that they are safe. Meanwhile the actual risk driver in this fund — subordination and leverage — is invisible in both the headline and the asset description.

The third blind spot is the correlation nobody models. Tether's existing crypto loan book and its new private credit book are nominally uncorrelated. But both are extending credit into a single macro environment. In a sustained risk-off regime, crypto borrowers face margin pressure at the same time that SME borrowers in a 60-country network face tightening credit and slowing trade. Two "uncorrelated" credit books on one balance sheet are only uncorrelated in the model. In a downturn, they can converge on the same capital base.

The retail read is a growth story. The institutional read is a term sheet. Only one of them settles.

Takeaway

Watch the disclosures, not the headlines. The three numbers that will decide whether this fund is a curiosity or a systemic event are leverage, subordination, and redemption terms — none of which have been published. Until they are, treat StableFund as a structure you cannot price, and treat unpriced structures the way you treat unmarked books: with distance, not conviction.

Two triggers are worth pre-committing to now. If Tether's exposure is subordinated and the fund is levered, then USDT's balance sheet carries private credit risk and its risk profile should be re-evaluated on that basis rather than on the reserve attestations alone. If the fund never publishes basic economic terms, then the absence of leverage, waterfall, gates and fees is itself the most informative disclosure in the entire announcement.

Efficiency is a feature, not a bug — and in a carry trade between the cheapest liability in finance and the least-marked asset class in finance, the efficiency is real right up until the moment the credit losses arrive. The only question that has ever mattered in that trade is who holds the first loss. That number is missing. Ask why.