Many people are asking whether we’re in the middle of an AI bubble, and the answer, in my opinion, is probably yes. The much harder question is what to do about it.
A recent analysis by Jason Furman, a Harvard economist and former Chair of the Council of Economic Advisers under President Obama, calculated that 92 percent of economic growth in the first half of 2025 was related to data-center buildout for AI.
Furman further showed that without the surge in tech-infrastructure investment, the annualized GDP growth rate would have been just 0.1 percent—far below the realized 1.6 percent.
Jeff Bezos put it another way. At a recent tech conference, he argued that the AI industry is in a “kind of industrial bubble.” But, he added, industrial bubbles “aren’t as bad—and can even be good—because when the dust settles and you see who the winners are, societies benefit from those inventions. That’s what’s going to happen here, too.”
Naturally, when clients hear the word bubble, they get concerned and want to know what to do about it. I do the same thing.
If you want to trim your tech exposure, that’s easy enough—we can do it a dozen different ways and still maintain market exposure. The harder question is whether you should trim your tech exposure.
Who wouldn’t want to identify a bubble, sell those assets, and sidestep the problem? Unfortunately, it’s not that easy.
I think the tech bubble and burst of the late 1990s and early 2000s is instructive—not only because many of us remember it firsthand, but because there are clear parallels to today.
On December 5, 1996, then–Fed Chair Alan Greenspan first uttered the phrase “irrational exuberance” in reference to stock market valuations. In fairness, he didn’t exactly say the market valuation was irrationally exuberant, but rather asked how we could ever know for sure when prices had become so.
Many commentators at the time took his remark as a warning that valuations were unsustainably high—and so did I. But anyone who sold out based on that speech would likely have been disappointed in the years that followed, even though Greenspan turned out to be right.
An investor who sold the S&P 500 the day after Greenspan’s speech would have missed an incredible run. Between his speech and the market peak on March 24, 2000, the S&P 500 rose 116.3 percent, or 27.0 percent annualized. Treasury bills gained just 17.5 percent, or 5.1 percent annualized.
That run-up turned out to be quite a cushion for the losses that followed.
After the peak on March 24, 2000, the S&P 500 fell by -47.4 percent, or -22.3 percent annualized. That period included the bursting of the tech bubble, 9/11, and the recession that followed. During that time, Treasury bills rose 10.5 percent, or 4.0 percent annualized.
Now, let’s look at the returns from start to finish—from Greenspan’s remarks to the bottom of the market.
From the speech to the market bottom, S&P 500 investors earned 13.8 percent total, or 2.3 percent per year. That’s not great, obviously, and T-bills did better, making 19.8 percent total, or 4.6 percent annualized.
It was a wild six years, and the results weren’t outstanding, but simply buying and holding the S&P 500 wasn’t the worst outcome (being overweight in tech stocks was).
And if you maintained a balanced, globally diversified portfolio of stocks and bonds, rebalanced as stocks rose and again when they fell, it worked out reasonably well – certainly within long-term market expectations, since we know there are bad times too.
The lesson isn’t that bubbles don’t exist—they do, and we may be in one now. It’s that they can inflate for far longer than anyone expects, and calling the top can be just as costly as missing it.
As Bezos said, industrial bubbles often leave behind real progress. He should know: Amazon stock fell by -94.4 percent from its peak in December 1999 to its trough in September 2001. Still, it survived and did fabulously for investors and consumers alike.
Of course, I have no idea when the current AI enthusiasm will end — it could be tomorrow or five years from now. What I do know is that it’s hard for investors to stay the course when markets are falling — but it can be just as hard when they’re rising, as well.
