While everyone is watching AI, debt is preparing its next exit. Somewhere else, as usual.
Every doctor knows it: if you don’t treat the cause, the pain doesn’t go away. It moves. Global finance seems to have forgotten this. Or rather, it has decided not to remember.
2000, 2008, and now
I lived through the turn of the millennium. We counted down the seconds to a computer bug that never came. The real shock arrived a few months later: the dot-com bubble burst, and the Nasdaq lost nearly 80% in just over two years.
Today, the parallel is tempting. Very expensive US markets, a handful of AI-related stocks pulling the whole index, a visible disconnect between the stock market and the real economy.
But there is a major difference. In 2000, the market was paying for thin air: companies with no revenue. Today, it pays a high price for something solid: giants that earn real profits. What should worry us is not the price of the champions. It is what happens while we are watching them.
The subprime is being built elsewhere
In 2008, nobody was watching mortgages to fragile households. Today, while every eye is fixed on AI, three fires are smouldering in the blind spot.
Private credit. Investment funds lend directly to companies, outside traditional banks. Around 1.7 trillion dollars outstanding, far less closely monitored than a bank balance sheet. We know the mountain exists; we can hardly see what is inside it.
Subprime auto loans. Americans owe around 1.7 trillion dollars on their cars, part of it high-risk. Defaults in this segment are at their highest level in more than thirty years. It isn’t the bomb; it is the canary in the coal mine: when people stop paying for their car, they will soon stop paying for the rest.
Commercial real estate. Remote work has emptied the offices: about one in five is vacant in the United States. And nearly 900 billion dollars of commercial real estate loans mature in 2026, each one a refinancing to negotiate in a market that has changed.
The triple penalty
The owner of an office building takes three blows. He loses his rents. His property loses value. And when it is time to refinance, the bank finds that the building is no longer worth what is owed on it, and refuses to lend again.
The debt, meanwhile, hasn’t moved by a single cent.
When hundreds of owners go through this in the same year, the losses travel up to the lenders: regional banks, private credit funds. And since all of them borrow from the big banks, the whole web moves when you pull on one knot.
The risk hasn’t disappeared, it has moved house
After 2008, banks were reined in. Logical. But risky credit did what water does: it found the crack. It flowed towards shadow finance, where the regulator sees less clearly.
We secured the lit room. The risk settled in the room next door, in the dark. It changed address, not nature.
The disease is the engine
Why is the source never treated? Because the source is the fuel. When a bank grants a loan, it creates money with a simple accounting entry. Money is born from debt. The system therefore needs ever more debt to keep running.
Truly curing the disease would mean accepting less growth, for years. No politician, no central banker, no voter wants that. So at every crisis, we choose the painkiller over the operation. Gains stay private, losses become public, and we start again. The system calls this stability.
A little debt is not poison: it makes investment and building possible. The poison is the dose, and the habit of postponing the moment of truth.
The same mechanism, at the scale of a shop
This is not just a Wall Street affair. The same logic plays out in any business. Rising revenue can hide a melting margin, exactly as a stock index hides the building that is cracking.
We watch the show: the shop window, the footfall, the busy Saturday. The mechanics are elsewhere: in cash flow, in stock gathering dust, in payment terms that keep stretching.
A shop rarely dies from a lack of customers. It dies from having treated the symptom, one more promotion, one more credit line, without ever touching the cause.
Build earthquake-proof
Nobody can date a crisis. We can, however, know that it will come back. It is not an accident; it is an appointment.
So we don’t predict the earthquake; we build earthquake-proof. Little dependence on refinancing. Solid cash reserves. And dry powder: not cash to hide behind, but ammunition to act when others are afraid.
Because in a crisis, good assets don’t disappear. They change hands. They pass from those who must sell to those who can buy.
Discipline in peacetime pays for freedom in a crisis. And clear-sightedness begins when you look away from what everyone else is watching.
Sources for figures: Mortgage Bankers Association (2026 maturities), Moody’s Analytics (office vacancy), Fitch Ratings and the Federal Reserve Bank of New York (auto loans), market estimates for private credit. Data available in 2026, rounded.
This text is a personal reflection, not investment advice.