It’s been a tough month for the market, especially tech stocks. Nvidia CEO Jensen Huang even mentioned the B-word in his post-earnings remarks this week, if only to decisively shoot down the idea of a bubble.
At present, worries about the huge amounts being spent on artificial intelligence are high enough that another great Nvidia quarter wasn’t enough to buoy stocks or dispel bubble talk. Stocks were broadly lower the day after it reported, and Friday’s bounce isn’t enough to erase recent losses.
Some might say that is a good thing: The market had gone up so quickly and smoothly after the tariff-related spring swoon that bears are worried stocks have gotten ahead of themselves and are far too pricey.
Gavekal Research’s Charles Gave argues that in one sense the current situation looks worse than the dot-com bubble. At that time a winner-take-all environment meant investors were acting rationally when buying a bunch of companies with relatively fixed costs, knowing one would eventually see sales explode and become dominant in their industry—bringing their stock along for the ride.
By contrast, today companies aren’t spending money up front and then sprinting to grow sales, but rather making a small number of sales in between investing huge amounts of capital with no apparent end point. Not to mention the U.S. doesn’t have enough electricity to power them.
“All this suggests that for AI companies, unlike dot-coms, the marginal cost of sales is relatively high, and the potential profit margin correspondingly thin,” Gave writes. So if 1999-2000 was a rational bubble, “2025 looks like an irrational bubble, and a capital-intensive irrational bubble to boot. In other words, it is a bad bubble.”
So to an extent, the market could be self-correcting on the same premise now, given all the focus that’s come on the tightknit AI economy. Even with stocks climbing Friday, the S&P 500 and Nasdaq Composite are still down for the week.
At the same time, there are reasons to think that this is a temporary setback for the market general, and tech specifically, even as the recent tech skepticism means the market is still on track for its worst month since March.
Wells Fargo analyst Ohsung Kwon put out his 2026 outlook, and predicts that the S&P 500 will end the year at 7800, nearly 20% higher than today. While stocks are expensive, earnings are growing as well, making those valuations look less outlandish.
Yet more to the point, he thinks AI stocks will be the big winners in the second half of the year, because no one in power can afford a bear market.
“Equities are now a bigger part of US household net worth than real estate and investment income tax could be as big as a quarter of government revenue,” he writes. “A K-shaped economy led by wealth effect means a bear market could trigger an economic downturn, which neither the Federal Reserve nor the Government can afford especially into midterms.”
The upshot is the Fed is likely to boost liquidity—the firm is forecasting it will purchase $25 billion in Treasuries a month beginning in April.
“Historically, ample liquidity has helped Tech more than other sectors,” he notes. “We expect increased speculation and risk-on as more liquidity gets injected to the system, resulting in a potential AI bubble trade.”
There’s the B-word again. But since there’s plenty of time before it’s likely to become a problem, it can wait until another day.
Write to Teresa Rivas at teresa.rivas@barrons.com