Written by: thiigth
Howard Marks, a Wall Street veteran and founder of Oaktree Capital, recently made a thought-provoking statement. He essentially said that if this AI frenzy doesn't ultimately turn into a classic bubble burst, it will be the only exception in human financial history.
The problem is that the vast majority of people are focusing on the wrong things.
We're still debating whether Nvidia's stock price is too high, or who will become the next Cisco. Everyone's glued to the candlestick charts, trying to find the slightest clue to a crash. However, the real eye of the storm isn't in the bustling stock exchange, but in that silent, hidden corner—the credit market—where life and death are decided.
This is not a math problem about the price-to-earnings ratio (P/E), but a high-stakes gamble built on massive debt.
01 The Disappearing "Cash Cow"
Our biggest misconception in this story is that tech giants are incredibly wealthy. We assume they're using their surplus earnings to leisurely reshape the world with AI.
But the numbers in the ledger reveal a completely different reality.
Let's rewind to before ChatGPT ignited the world. Back then, Microsoft was the most reliable "cash cow" on the planet. Its cash reserves exceeded its debt by 30%, making it incredibly stable. But just a few years later, in order to feed the insatiable AI, the situation has fundamentally reversed: now, Microsoft's debt exceeds its cash reserves by 20%.
Meta (the former Facebook) is even more extreme. In 2022, Zuckerberg had three times his debt; and by the last quarter, his debt had surpassed his cash reserves by 15%.
Amazon, which has always been fond of high leverage, now has 50% more debt than cash. As for Oracle, the software empire that once made money hand over fist, it now not only has a tight cash flow but is also burdened with a mountain of debt.
In just four years, Silicon Valley giants have changed. In pursuit of AI, the "holy grail" whose realization is uncertain, and to avoid being left behind in the next industrial revolution, they have not only emptied their savings of many years but also staked their future income, signing massive IOUs.
02 "Real Estate Projects" That Cannot Be Abandoned
This isn't just about money; it's a story about "no way out."
Looking back at the 2008 subprime mortgage crisis, when housing prices collapsed and people were broke, heavily indebted homeowners made a painful but rational decision: they handed their keys back to the bank and walked away. This is personal default—leaving the mess for the bank while life goes on.
But in the game of AI, there is no option to "throw the key to the bank".
Microsoft, Oracle, and Meta are now acting less like homeowners and more like developers who have to get the job done. When they signed multi-billion dollar chip contracts and drove the first pile into the server farm on the wasteland, they signed a "pledge of allegiance."
If credit tightens tomorrow and Oracle can't borrow money anymore, can it point to its half-built data center and say, "I'm done"?
cannot.
If these unfinished products are not completed, they will not only be worthless, but also face huge penalties for breach of contract, and it will mean the instant collapse of core competitiveness.
So even if there's a pit of fire ahead, they have to grit their teeth and keep burning money.
This is a classic debt trap: regardless of whether AI is profitable now, you can't stop. If you stop, the previous billions will instantly vanish into nothing.
To continue on this path, a continuous influx of credit is needed.
This is why the AI bubble is said to be more terrifying than the stock market bubble. A stock market crash (like in 2000) only means that everyone's wealth shrinks and life becomes a little tighter; but if credit collapses, causing these giants to be unable to borrow money to keep their infrastructure running, that's cardiac arrest.
03 Fear is the highest form of leverage.
Why would these exceptionally intelligent CEOs collectively put themselves in this corner?
Because of fear.
This is not merely greed, but a deep-seated existential anxiety. At the crossroads of technological change, inaction means death. If you dare to sit on the shore and watch, you are destined to miss the next Age of Exploration. To stay at the table, the giants have no choice but to take on debt and bet.
Thus, we witnessed this enormous two-way gamble:
If AI succeeds: It could take 6 to 10 years. But during those 10 years, companies will be burdened with heavy interest payments, and money that could have been used to develop new products will have to be used to pay off debts. This will make once agile tech giants stumble and falter.
If credit collapses first: Before AI truly makes big money, if banks feel the risks are too high and tighten their purse strings, then this business loop built on lending will snap instantly, like a taut string.
The current tech industry is like pouring gasoline on a smoldering fire. On the surface, we only see the stock market frenzy, but beneath the surface, the accumulated debt fuel has seeped into every inch of the soil.
04 The End
Most investors think, "I'm smart enough to get out before the bubble bursts."
This is a classic survivor's illusion. Everyone fantasizes about cashing out at the peak, buying government bonds, and then sunbathing at the beach.
But history never follows this script.
It's like taking an escalator upstairs: smooth, comfortable, and even making you drowsy; but going downstairs often means you're jumping straight into the elevator shaft. When the heavy door of the credit cycle suddenly closes, everyone is crammed into a crowded track—that is, into those few core tech stocks—and the exit will be instantly blocked.
So stop focusing solely on whether Nvidia's stock price goes up or down tomorrow. The real risk lies in whether the credit foundation supporting this prosperity remains solid.
When Microsoft starts operating like a highly leveraged real estate developer, it's time to wake up: this is no longer just about technological dreams; it's a brutal story about finding a glimmer of hope in a quagmire of debt.
This article is compiled from Eurodollar University podcast content and is for reference only. It does not constitute investment advice.
