Jamie Dimon runs the biggest bank in America. When Bloomberg asked him what he thinks of the AI bubble, he didn't talk about stock prices.
"I always said the best thing to do with any of these things is deal with it before it becomes a crisis," he said. If it does become one, "we will deal with it - it just will be in a much less pleasant way."
The line that caught Briefs Finance CEO Jaspreet Singh's attention came next: Dimon's big concern isn't the AI bubble. It's the credit crisis that could be building behind it.
"We can't borrow endlessly and spend endlessly," Dimon said. "There will be a point where the market will ask for more and more and more," and at some point that feeds into corporate debt and credit spreads.
Dimon's Real Worry Isn't the AI Bubble. It's the Borrowing Behind It.
What does "more and more and more" mean? Companies have been borrowing money like crazy to invest in AI.
As banks like JPMorgan Chase get more concerned about the AI bubble, interest rates go up and borrowing gets more expensive. The companies building AI still need to borrow more, so the cost keeps climbing.
Eventually it hits a breaking point. Companies that want to borrow can't afford to, but they need the money to keep growing, and that's where a credit crisis starts.
The pressure is already visible:
- The Chinese AI bubble popped, with AI stocks there falling more than 40%.
- South Korea's AI market fell in 2026 as well.
- In the US, Google stock crashed on concerns about its AI debt, and Intel crashed on the same worry.
Why a 1.5% Credit Spread Makes the AI Bubble Riskier Than It Looks
Start with a mortgage. A year ago you could borrow $400,000 at around 6.17%, which worked out to something like $2,400 a month.
Today that same $400,000 costs about 7.24%, or roughly $2,700 a month. Same house, same loan amount, $300 more every month because rates went up.
Now swap $400,000 for $400 million, which is closer to what a company borrows. The extra cost gets a lot bigger.
Dimon's concern isn't just where rates sit today. It's where the credit spread could go, and a credit spread is the premium a company has to pay to borrow because lending to it carries risk.
A Federal Reserve chart puts that spread at roughly 1.5% right now, which is very low. Low might sound like good news.
The last time spreads were this low was February 2007, just before the housing market burst. Before that it was January 2000, just before the dot-com bubble popped.
If the economy or the AI market starts to hurt, that spread can jump quickly. Borrowing gets much more expensive right when AI companies need access to a lot of money they don't have.
The Economy Runs on Credit, and AI Companies Are the Biggest Borrowers
Our economy runs on spending. The more money you spend, the more money somebody else makes.
Buy a burrito at Chipotle and you're supporting the economy. Put that burrito on your Amex after you've already spent your paycheck and you're supporting it even more, because now you're spending money you don't have.
That's a credit-based economy, not a money-based one. Spending depends on how much credit people and companies can qualify for, not just what's in their pockets.
Corporations are huge borrowers, and right now the biggest borrowers are big tech and AI companies. That money lands in somebody's pocket, whether it's contractors, employees, or the crews building the buildings.
A credit crisis means AI companies can't borrow as much. Less money flowing in means fewer jobs, fewer contracts, and less growth.
Jaspreet lays out five stages of a credit crisis:
- Boom. A new industry takes off and lending grows with it.
- Loosening. Banks see money to be made and relax their standards.
- Pricey loans. Banks see rising risk and raise interest rates.
- Cracks. Defaults and bankruptcies start showing up.
- Pullback. Banks stop lending.
Most investors panic through those stages, and the ones who understand them tend to find opportunities instead. Jaspreet and his analysts research those opportunities with ChatGPT and Claude, and he's showing exactly how in a free livestream on November 17.
Stage 1: Everyone Wants a Piece of the Boom
In the 1990s the new technology was the internet, and money poured into internet companies through debt and investment dollars because everybody wanted a piece of the action.
Today it's artificial intelligence, with a record amount of debt flowing into the industry. AI is changing the economy, every company wants to be the leader, and the fastest way to grow is to borrow as much as possible.
It isn't just the companies borrowing. Investors borrow to buy the stocks too.
China set a record for money borrowed to invest in AI stocks right before its AI bubble burst. South Korea saw record margin borrowing, which means buying stocks with money borrowed from a brokerage, poured into AI stocks before that bubble burst too.
Those investors didn't just lose their money. They still owe the debt they borrowed to buy the stocks, which makes it a double hit.
Stage 2: Lending Gets Loose, and Private Credit Fills the Gap
The clearest example of loose lending is the run-up to 2008. President Bush said housing was the best investment you could make, and banks did everything they could to get loans out the door:
- NINJA loans, short for no income, no job, no assets. You could sign for a $400,000 house with none of the three.
- Subprime loans. A 480 credit score couldn't qualify, so banks found a way to lend to a 550 score at a higher interest rate.
- 120% loan-to-value mortgages. Want a $400,000 house? Here's $480,000, so you have money to move in, renovate, and buy something new.
Everybody believed housing prices could never fall, so banks lent to anyone who could breathe and sign.
The AI bubble has seen nothing like 2008. But lending did loosen over the last several years, especially during the era of extremely low interest rates after the pandemic, in 2020, 2021, and 2022.
The catch is that businesses don't get 30-year fixed-rate mortgages. They get adjustable loans, generally three to five to seven years long.
In 2026 a big chunk of those 2020 and 2021 loans is readjusting, and a 3% loan from 2021 is resetting at 7%, 8%, or 9%. On hundreds of millions of dollars, that's a big jump in costs even for a company that isn't borrowing anything new.
What Is Private Credit, and Why Is It Bleeding?
The bigger worry here isn't banks like JPMorgan Chase, Wells Fargo, and Bank of America, which have their own concerns. It's private credit.
Private credit grew out of a gap in the market. A $200 million loan for a risky AI startup was too big for a regional bank and too small for the biggest banks, so companies that size struggled to find a lender.
Hedge funds built an industry to fill it. Firms like BlackRock and Blackstone lent to these middle-market companies, especially tech and software businesses that used to be cash-flow machines, and they borrowed from investors to keep the lending going.
Today those private credit firms are bleeding. BlackRock, Blackstone, and Apollo have all halted investors from pulling money out, because the companies they lent to aren't paying the loans back and letting investors leave would collapse the funds.
Higher interest rates pushed up the burden on borrowers, and many can't keep paying, because while a few key tech companies make a lot of money, plenty don't. A traditional software company that didn't invest in AI fast enough may not be growing fast enough to compete.
Dimon referenced 2008 directly. He said this feels like 2005, 2006, and 2007, when there was a lot of money to be made and people took whatever loans they could because they thought they'd get rich.
Stage 3: Loans Get Pricey, and the AI Bubble Meets Higher Rates
Dimon said JPMorgan is no longer being so loose with its lending requirements. In his words, they don't have to chase that.
Higher interest rates have historically been the needle that pops bubbles. In the 1990s, valuations skyrocketed as money flooded into dot-com companies, the Federal Reserve raised rates to cool the economy, and the bubble burst.
Rates have climbed a lot over the last several years, partly because of the Fed and partly because of higher Treasury yields. A Treasury yield is the interest rate you earn when you lend money to the US government.
Yields rose over the last 12 months not because of the Fed but because investors and countries around the world grew less interested in lending to the US. They had worries about the dollar and inflation, and the government had to offer higher yields to bring them back.
Dimon points to a second reason loans are getting pricey, which he calls increased competition for capital. A few years ago investors had extra cash and savings sitting around, and those savings aren't there the way they were.
Now companies are fighting for dollars, and more banks are deciding they'd rather not lend at all, so the ones that do charge more to justify the risk. Sticky inflation adds to it, because a bank earning only interest on a ten-year loan wants protection from inflation eating its profit.
That creates two problems for companies:
- A company that needs money to keep growing, or even to keep operating, now pays a lot more for it.
- A company trying to refinance loans from years ago faces significantly higher rates, assuming a bank will refinance it at all now that valuations have fallen.
Picture a company that thought it was worth $100 million three years ago and is valued at $80 million today. The value it expected to grow shrank instead, and the debt on top of it got more expensive.
Stage 4: Cracks Show Up as Defaults, From China to Google
Every credit crisis shows cracks the same way: more defaults, then more bankruptcies.
In the dot-com era, banks lent millions to internet companies that were just an idea, with no revenue and no profit. Rates rose in the late 1990s, loans got pricey, and hundreds of internet companies defaulted in 2001.
In 2008, mortgage rates rose, adjustable-rate mortgages reset, home prices fell, and foreclosures flooded in.
In 2026 the cracks look like this:
- China. As the AI bubble burst, investors who borrowed to buy AI stocks declared default.
- South Korea. Investors who bought AI stocks on margin had to declare default too.
- United States. The bubble hasn't burst, and some don't believe there is one. But Google saw a lot of pain after pouring all of its free cash flow into AI while parts of its business shrank.
- Intel. The company, which is taking investment from the US government, had one of its worst days on concerns it won't see the AI growth investors expected.
Dimon expects the next cracks to show up in the data center industry.
Ray Dalio, founder of Bridgewater, the largest hedge fund in the world, is seeing the same thing. He posted that "the market is now beginning to tighten, as investors are beginning to require larger concessions," which means lenders want more protection and higher interest before they hand over money.
Stage 5: The Pullback Is the Actual Credit Crisis
If the cracks spread, banks pull back. They tighten restrictions and make it much harder to borrow.
That's the credit crisis, because spending in our economy isn't just the money in your pocket, it's the credit you can qualify for. When banks stop lending, money stops entering the economy as credit, and spending stops with it.
After the dot-com bubble burst, banks stopped lending to internet companies because they couldn't tell the good ones from the bad. After 2008, NINJA loans, subprime loans, and 120% mortgages disappeared, and you needed a 30% down payment, a great credit score, a great income, and great assets just to be considered.
What Investors Can Do Before the AI Bubble Question Gets Answered
There's a chance the AI bubble bursts, the market crashes, the economy hurts, and it all turns into a great buying opportunity. There's also a chance it never bursts and markets keep climbing.
Jaspreet's starting point is what he calls ABB, or always be buying. Set up a system to buy the broad market every week, two weeks, or month, and keep buying whether markets are up, down, or sideways, for years and decades.
If that's not enough to hit your goals, he lays out three options:
- Retire later. Work until 72 instead of 65 and let the money compound longer.
- Invest more. Go from $500 a month to $1,000 by cutting expenses or picking up a second job.
- Earn better returns, which comes with more risk.
Better returns come from two places. The first is what Jaspreet calls POOP: panic leads to overselling, which leads to opportunity, which leads to profit.
Buying through the 2020 pandemic crash, the 2008 crash, or the 2000 dot-com bust meant buying almost any investment on sale. The problem is that nobody knows when the next crash comes, only that one will.
The second is market shifts: identifying where money is moving and investing there, not where it was.
Both take research, and Jaspreet recommends doing it with AI tools like ChatGPT and Claude, because it's almost like having a Wall Street analyst at your fingertips. Those tools are only as good as the person prompting them, which is exactly what his free livestream on November 17 covers, down to the ten prompts his analysts use.
Dimon has already started tightening at JPMorgan, and he made clear that doesn't make the rest of the banking industry safe. If this does become a crisis, it will get dealt with "in a much less pleasant way."
