On September 29, 2025, an auto-parts maker called First Brands filed for bankruptcy. It owed roughly $5 billion in senior loans, and about $2.1 billion of that sat inside CLOs, spread across 69 different fund managers.

A CLO, or collateralized loan obligation, is a machine that buys hundreds of loans made to risky companies and repackages them into new bonds. Some of those new bonds are rated triple-A, the same grade as U.S. Treasuries. The reader may own one through a pension or insurance policy.

The machine we are drawing is how that repackaging works. It takes loans that could default and turns most of them into debt that almost never does. The trick is a strict order of who gets paid, and a switch that changes that order under stress.

The First Brands failure was a live test. A borrower inside thousands of these funds went bankrupt, and there were questions of fraud, with reports that some collateral had been pledged more than once and billions had simply gone missing.

Yet the CLOs mostly shrugged. First Brands was about 0.2% of total CLO collateral, and analysts called the exposure manageable, with comfortable room left in the funds' protective tests. The machine did the job it was built for: contain one bad loan.

That is the CLO market at roughly $1 trillion in the U.S., buying most of the leveraged loans that fund private-equity buyouts. The question the failure raised is not what happens to one loan. It is what happens when many go bad at once.

An email announcing a new 401(k) investment option is coming to 70 million inboxes. Most will delete it – and miss what Larry Benedict calls the decade's biggest opportunity. He's naming the one ticker to own before it arrives – free. Watch here.

Start with the raw material. A CLO holds a few hundred loans made to companies rated below investment grade, the kind of borrowers that carry real default risk. On their own, these are junk. Pooled together, their fates stop moving in lockstep.

Now stack the claims. The CLO issues its own bonds in layers, called tranches. The top layer is rated AAA and gets paid first. Below it sit lower-rated layers, and at the very bottom is the equity, which gets paid last and takes the first loss.

Cash flows down this stack like water down steps. Interest from the loans pays the AAA layer first, then the next, and so on down to the equity. Losses fill up from the bottom. The equity is wiped out before the layer above it takes a scratch, and so on up the stack.

The load-bearing part is a switch called the overcollateralization test, or OC test. It constantly checks that the loans are still worth comfortably more than the senior bonds. As long as they are, cash keeps flowing to the bottom. That is the machine in normal weather.

When too many loans default or get downgraded, the test fails. The switch flips. Cash that would have gone to the equity and lower layers is seized and used to pay down the AAA layer instead. The bottom is starved to keep the top whole. That diversion is why the AAA tranche has never defaulted since 1994.

The first joint is who sits where. Insurance companies and pension funds hold the safe AAA layer. Hedge funds and private-credit vehicles hold the risky equity at the bottom. The whole design moves loss away from the top and onto the bottom on purpose. When the OC switch flips, the equity holder is the one who stops getting paid.

The second joint is downgrades, not just defaults. The switch can trip on ratings alone. When a wave of loans is cut to the lowest rung, the test can fail even before any loan misses a payment. A single event like First Brands, which set off billions in downgrades in one month, tightens every fund holding that name at once.

The third joint is what the pool cannot see. First Brands allegedly hid debt off its balance sheet and may have pledged the same collateral twice. A CLO can be perfectly diversified across names and still be blind to a borrower lying about what it owes. The machine manages the risk it can measure, not the fraud it cannot.

What persists: the structure, which is genuinely tough. Pooling spreads the risk, the waterfall ranks the claims, and the OC switch actively defends the top. Through 2020 and through First Brands, the senior layers held, and the AAA tranche's clean record since 1994 is real, not marketing.

What can break: the layers below the top, quietly and first. The protection for the AAA holder is paid for by the equity and mezzanine holders, who absorb the diversion when the switch flips. In a broad downturn, the bottom of the stack can be gutted while the headlines still call the machine safe.

This machine connects to two others the reader's money touches. It is the largest buyer in the leveraged loan market, the debt behind private-equity buyouts, and it feeds the insurance and pension portfolios that hold its AAA bonds. When the loans sour, the strain runs from the buyout all the way to the retirement account.

The next reading arrives with the fall wave of CLO trustee reports and the fallout from First Brands and similar failures, which will show whether the recent defaults stay contained to a few names or start tripping overcollateralization switches across the market.