The Shadow Bank That Ate the Loan Business: Private Credit's $2 Trillion Question
85% of buyout loans now come from funds, not banks — a $2 trillion market with no runs, no market prices, and three new warning reports from the world's financial watchdogs. How a century-old banking function moved outside the rulebook while nobody was watching.
By MyAudioBooks.ai ·
Listen free: The Shadow Bank That Ate the Loan Business: Private Credit's $2 Trillion Question
In the space of fifteen years, without most people noticing, one of the oldest functions of banking moved house. When a private-equity firm buys a company today, the loan that finances the purchase — the big leveraged loan at the heart of the deal — comes, about eighty-five percent of the time, not from a bank but from a fund: a private-credit manager, pooling money from pensions and insurers and endowments, lending directly to the borrower with no bank anywhere in the middle. The market those funds form now stands at roughly two trillion dollars under management, by the most conservative estimates, with outstanding loan volumes approaching two and a half trillion and projections reaching three and a half trillion or more by the end of the decade. A business that barely existed in 2010 — a corner of the lending world once dismissively called shadow banking — now writes the majority of the checks in American corporate buyouts, employs the yield that millions of retirees depend on, and has begun, this year, to attract a kind of attention from the world's financial watchdogs that it has never had before.
Between April and September of this year, the three most important financial-stability institutions on Earth each published a formal assessment of private credit. In April, the International Monetary Fund's Global Financial Stability Report flagged the market's early fault lines — defaults rising from a low base, and a worrying increase in stress-driven payment-in-kind arrangements. In May, the Financial Stability Board, the body that coordinates the world's regulators, published an entire report on vulnerabilities in private credit: its deepening interconnections with banks, insurers, and private equity; the credit quality of its borrowers; the liquidity mismatches hiding in some of its vehicles; and above all, its opacity — the fact that nobody outside the funds knows what the loans are worth. And this month, the Bank for International Settlements devoted a feature of its Quarterly Review to the market's newest and largest frontier: the financing of the digital economy, where technology companies now absorb roughly forty-four percent of all private-credit lending, much of it bound for the same data centers and artificial-intelligence buildout this channel documented last week.
This is the story of how a two-trillion-dollar lending market grew up outside the banking system, why the official sector has suddenly started writing reports about it, and what the reports do and do not tell us about whether the machine holds under the weight it has taken on so quickly.
At My Audio Books dot A I, you can create your own audiobooks from prompts, turn your documents into audio, all with one subscription, and store your items in your own personal library.
To understand what moved, you have to understand the loan that moved it, and its name is the leveraged loan. When a private-equity firm buys a company, it pays partly with its investors' money and mostly with borrowed money — debt loaded onto the acquired company itself, at interest rates floating above the market, sized at five, six, or sometimes even seven times the company's annual earnings. For forty years, those loans were made by banks, and the banks did not keep them: a lead bank would underwrite the loan and then sell pieces of it to dozens of other institutions in a process called syndication — the spreading of one loan across many balance sheets, so that no single bank carried the risk alone. The syndicated leveraged-loan market was one of the great invisible machines of corporate finance: trillions of dollars, hundreds of deals a year, priced daily, traded constantly, regulated heavily because banks did it. Then came the financial crisis, and the regulators, quite reasonably, made that machine expensive to run. Capital requirements rose. Leveraged-lending guidance tightened. The banks pulled back from the riskiest end of the business — and into the space they vacated stepped the funds. The first of them were small and scrappy: business development companies, specialty lenders, the odd credit arm of a private-equity house, making the midsized loans the banks no longer wanted. The borrowers discovered the product was better. The managers discovered the margins were better. The institutional investors discovered the yields were better. And the managers grew from specialists into giants — the largest private-credit platforms now manage hundreds of billions each, employ former bank lending teams by the floor, and compete head-to-head for the very largest buyouts, the billion-dollar club loans that once only a syndicate of global banks could write. The market did not just take the banks' leftovers. It grew past the banks' prime rib, and then it grew past the banks.
The pitch, from the fund managers, was and remains genuinely compelling. To the borrower, a private-credit fund offers what a bank syndicate cannot: certainty and speed. One counterparty, one negotiation, terms agreed in weeks rather than the months a syndicate takes to assemble, no risk of the deal falling apart in the market window between signing and closing. The product that made it possible is called the unitranche — a single, blended loan that replaces the old layered structure of senior and junior debt with one agreement, one rate, one lender group, simpler for everyone. To the investor — the pension fund hunting yield in a world of low rates — the funds offer floating-rate returns several points above comparable public debt, secured by the borrower's assets, diversified across hundreds of loans. And to the regulators of the twenty-teens, the whole arrangement looked, at first, like a feature rather than a bug: risk migrating out of the deposit-taking banking system into investment funds whose investors had knowingly signed up for it. What could be safer than risk owned voluntarily?
What could be safer is exactly the question the twenty twenty-six reports are now asking, because the arrangement's virtues came with architecture attached, and the architecture has never been tested. Start with the most fundamental difference: a bank holds deposits it must return on demand, and so it is regulated to the teeth — capital ratios, liquidity rules, examiner visits, stress tests. A private-credit fund holds its investors' money for years under lock-up terms, which means it cannot suffer a bank run — the genuine stability advantage — but it also means it operates largely outside the bank rulebook. Its loans are not marked to a market price, because there is no market; they are valued by the fund itself, quarterly, using models, and the FSB's report states the consequence in the flat language regulators reserve for things that worry them: valuation opacity. Nobody outside the fund knows what the loans are worth until the fund tells them. The second architectural fact is interconnection: the banks did not leave the building, they just moved upstairs. They lend to the private-credit funds themselves, they finance the funds' portfolio companies, they sell the funds hedging and services, and the insurance companies that buy the funds' products are themselves entangled with bank balance sheets — so that the risk that migrated out of the regulated system is wired back into it, along pathways the FSB describes as deepening and difficult to map. It is worth pausing on the structural comparison everyone in the industry quietly makes and nobody likes, because it cuts both ways. The last great credit machine to fail was built on originate-to-distribute: make the loan, sell the loan, hold nothing, care about nothing but the fee. The private-credit model is the opposite — originate-to-hold: make the loan, keep the loan, eat the losses yourself. On paper, that is a vastly better incentive structure, and it is a large part of why the default record is so clean. But hold-to-maturity has its own shadow: the incentive to keep a loan alive past its natural life, to amend and extend rather than realize the loss, to PIK the interest rather than admit the borrower cannot pay. The last crisis was a crisis of people selling loans they did not understand. The next one, if it comes, will be a crisis of people holding loans they cannot sell — which is a different disease, requiring different medicine, and the medicine has never been tested either.
The third fact is the newest: the market's center of gravity has shifted to the technology buildout, the data centers and AI infrastructure of the last two years, which means the loans are increasingly concentrated in a single, spectacularly capital-hungry thesis — the same thesis, note, that now dominates the equity markets as well. The machine that funds the AI boom and the machine that funds the buyout boom are becoming the same machine.
At My Audio Books dot A I, you can listen to this story and thousands of others that explore the hidden science and mechanics behind the headlines.
Two more pieces of the architecture deserve their own light. The first is the insurance companies, because they are where the FSB's interconnection worry gets concrete. Life insurers — the companies that hold the retirement promises of millions of households — have become some of the largest buyers of private-credit product, drawn by the same yield hunger as everyone else, and some have gone further, partnering with or buying the credit managers outright. When an insurer's annuity book is backed by unmarked loans to leveraged companies, the risk that was supposed to have left the regulated household-finance system has re-entered it through the back door, wearing a yield wrapper.
The second is the newest frontier of all: the retail investor. The industry's next growth spurt is the semiliquid evergreen fund — vehicles that let ordinary investors put money into private credit with the promise of periodic withdrawals, quarterly liquidity on an illiquid asset class. The official reports note the mismatch with care: a fund that holds ten-year loans and offers quarterly exits is making a promise it can only keep while inflows exceed outflows, and that promise is currently about a fifth of the market. The bulls read that as containment. The historians read it as a fraction with room to grow.
Now the early-warning data, because the reports are not writing about abstractions. Default rates in direct lending remain low — around two percent, by the market's own trackers, and the bulls cite that number constantly — but the shape of the stress matters more than its current size, and the shape is visible in the acronym the IMF flagged: PIK, payment-in-kind, the arrangement in which a borrower that cannot make its cash interest payment pays the lender in more debt instead, adding the interest to the loan balance rather than defaulting. PIK is not inherently sinister — some deals are structured with PIK options from the start — but stress-driven PIK, the kind a borrower takes because it has no other way to pay, is the classic mechanism by which a credit problem hides itself: the loan stays current on paper while the borrower's actual condition deteriorates underneath. The IMF's April report found exactly that pattern emerging — defaults rising from a low base, stress-driven PIK increasing — and the industry's own surveys concede the direction: managers report expectations of flat or lower returns amid rising competition and defaults, and the newest funds are raising money on projections of returns lower than the track records that made them famous.
And one more layer belongs in the early-warning file, because it determines how loudly the warnings will eventually sound. The covenant protections that once forced struggling borrowers into early conversations with their lenders have thinned across the entire leveraged-loan era: covenant-lite structures — loans with minimal maintenance requirements, quarterly check-ins reduced to formalities — now dominate both the bank and private markets. The old covenants were the smoke detectors of corporate credit, tripped early and often. The new structures are smoke detectors with the batteries removed: problems surface later, and larger, than they used to, precisely when the cost of addressing them is highest.
The strongest case against the warnings — the bull case for private credit, made by its biggest managers and, quietly, by many of its borrowers — deserves to be heard at full strength, because it explains why the market keeps growing even as the warnings accumulate. The funds cannot be run; their investors are locked in for years, which makes the structure inherently more stable than a deposit bank. The loans are senior, secured, and diversified across hundreds of borrowers and dozens of industries; the two percent default rate, achieved through the rate shock of twenty twenty-two and the years since, is not a projection but a track record. The managers argue that their valuations, while model-based, are audited annually and tested constantly by the thousands of secondary trades and refinancings that occur in the loans every year — the market marks itself, they say, just not on a screen you can watch. The growth into AI infrastructure is, on this reading, simply the market financing the most important capital expenditure of the century, the way project finance once funded the railroads, the canals, and the power plants of earlier eras. And the regulators' own reports, the bulls point out, conclude that systemic risk is currently contained: the semiliquid retail vehicles that worried everyone are only about a fifth of the market, and the stress scenarios run by the official sector show the system absorbing moderate shocks. The machine works, the case concludes — the reports are the watchdogs doing their job, not finding a fire.
And the strongest case for concern is written in the last sentence of that defense: currently contained. Every financial-stability report ever written has said the risk was contained until it was not, and the specific signatures of this market — opacity, interconnection, concentration, and a credit culture that has never faced a real recession — are the signatures that have preceded every credit event in living memory. The opacity means the stress is invisible until it is large; the PIK mechanics mean it is actively being deferred; the interconnection means that when the losses surface, they will surface inside the banks and insurers the structure was supposed to protect; and the concentration means a single thesis — the AI buildout — now ties the loan market, the equity market, and the capital plans of the world's largest companies into one correlated bet. The regulators are not saying the machine is broken. They are saying they cannot see inside it, and that nobody can, and that the parts they can see are exactly the parts that failed somewhere else before. Three findings would disprove or confirm which reading of this machine is right, and each is observable. First, the default-and-PIK trajectory: if stress-driven PIK keeps climbing and realized defaults follow it upward through a genuine economic downturn, the hidden-loss thesis confirms itself in the only currency that settles credit arguments — losses. Second, the semiliquid vehicles: as retail investors flow into the new evergreen funds that promise periodic withdrawals, the first test of those withdrawal promises in a falling market will show whether liquidity mismatch is theoretical or real. Third, the AI lending book: if the data-center buildout's economics falter — if the tenants renegotiate, the power costs spike, or the models' revenue lags the capex — the forty-four percent of the market now tied to that thesis reprices at once, and the repricing will be the first true mark-to-market the industry has ever received.
It is worth saying what this article has not claimed. It has not claimed private credit is a fraud or a bubble; it is a real market performing a real function, with a track record that has so far survived everything thrown at it, and the article says so. It has not claimed the banks are safer; the post-crisis evidence that risk belonged outside the deposit system was real, and the migration had logic. It has not claimed the regulators are predicting a crash; their reports say, precisely, that risk is currently contained and that their concern is structural and prospective. And it has not claimed the AI buildout is unsound; it has claimed only that concentrating two trillion dollars of unmarked loans and the world's equity markets in the same physical thesis is a correlation nobody chose, and correlations are the things that turn sectoral problems into systemic ones.
Which returns to the eighty-five percent — the quiet statistic at the center of everything. A banking function that took a century to build moved, in fifteen years, to a new address, and it moved so smoothly that almost nobody outside the industry registered the change of management. The leveraged loan did not disappear; the banker's desk did, and with it went a century of accumulated oversight that nobody thought to pack. The reports on your shelf — the IMF's, the FSB's, the BIS's — are the sound of the new landlord being asked for the building plans, fifteen years after the tenants moved in, and answering, politely, that the plans are proprietary. The next downturn will inspect the building whether the plans are produced or not — downturns always do, that being their entire function in the system. The only question the reports really ask is whether anyone learns what the building is made of before the inspection arrives — and on the current evidence, the honest answer — for now, and uncomfortably for everyone asking — is: not yet.
At My Audio Books dot A I, you can create fiction, non-fiction, and turn your documents into audio, all stored in one place with a single subscription — plus get instant access to thousands of audiobooks and deep-dive investigations. Learn more today at My Audio Books dot A I.