How Much Longer Can the AI Bubble Last? Inside Wall Street's Great Debate

Dow Jones10-06

The fun parts about blowing a bubble are seeing how big it can get -- and how long it takes to pop. The same appears to be true for investors witnessing tech stocks' increasingly stretched valuations, thanks to the artificial-intelligence revolution.

The market knows the AI bubble has to burst eventually, but the question is how much longer it can inflate before the chewing gum splatters on traders' faces. Timing exactly when a bubble will burst, however, is never easy.

Nvidia, the world's most valuable company, now has a market value of $4.5 trillion. Meanwhile, Sam Altman's OpenAI recently raised a round of financing that values the ChatGPT creator at $500 billion, making it the world's most valuable private company, topping Elon Musk's SpaceX.

The Technology Select Sector SPDR exchange-traded fund -- which owns Nvidia, Microsoft, Apple, Broadcom, and Palantir as its top five holdings -- now trades about 28 times next year's earnings estimates. That's up from around 21 times back in late 2022, when OpenAI's ChatGPT had just come on the scene. The question for analysts and investors is whether these valuations are near a peak, or if there is still room to grow.

There is a growing "this time is different camp" on Wall Street. Tech bulls maintain that the AI enthusiasm of 2025 isn't like the internet bubble of those irrationally exuberant late 1990s.

There is a lot of merit to that argument. For one, Nvidia and the so-called hyperscalers spending big on AI -- cloud companies like Microsoft, Amazon.com, Meta Platforms, and Alphabet -- have strong earnings growth and lofty profit margins to support their pricey valuations. And as long as big tech companies continue to spend on AI, that could keep the stock market's bull run going a while longer.

"Why is the AI story so pervasive? It's the sheer amount of money the hyperscalers have thrown at it," said Bob Savage, head of markets macro strategy at BNY, in an interview with Barron's. "Investment spending lifted second-quarter economic growth. We don't see that slowing."

There's also the fact that stocks' current rally has broadened out in recent months, which could keep the party going for both tech and the overall market. Small-cap stocks are taking part. Other sectors that have ties to AI, most notably utilities and certain real estate stocks, are getting a big boost too. And growing hopes for more interest-rate cuts from the Federal Reserve are a big part of the latest market surge too, with industrials, financials, and consumer discretionary stocks getting a lift.

Evercore ISI chief equity and quantitative strategist Julian Emanuel said in a report Friday that the "ongoing bull market -- which has broadened beyond tech, unlike in 1999 -- is central" to his firm's expectation that the S&P 500 could hit 7,750 by the end of 2026. That's 15% higher than current levels around 6,730.

And valuations, while not necessarily cheap, are actually much more reasonable now for big tech stocks than they were a quarter of a century ago. The Nasdaq 100 index, which has the Magnificent Seven along with Broadcom, Netflix, Palantir, Advanced Micro Devices, and ASML as top weightings, is trading at 28 times 2026 earnings estimates, compared with a forward price-to-earnings ratio of 47 back in February 2000.

Still, others worry that the giddiness about AI is too reminiscent of the late 1990s dot-com froth. In other words, we've seen this movie before, and it might not end well.

"There is growing hesitation to meaningfully chase the AI trade," said Tom Essaye, author of The Sevens Report, wrote on Friday. "The risk that we see the AI-narrative challenged in the weeks/months ahead is rising and has the potential to spark a meaningful profit-taking pullback in tech and the broader equity market."

Others are even more blunt about the possibility of the AI bubble suddenly popping.

"It feels like the risk/reward of deploying capital into AI plays, especially at a time when other, less richly valued, and less crowded, markets also have the wind in their sails, is simply not attractive," said Louis-Vincent Gave, founding partner and CEO of Gavekal Research, in a report Friday. "The odds of the AI bubble deflating in the coming quarters now seems high."

He added that "the bigger uncertainty is the impact an AI face-plant would have on the broader markets."

But there is some good news on that front. International stocks in developed markets such as Germany, South Korea and Japan -- as well as emerging market bonds in Latin America -- have been solid performers, offering alternatives for investors to consider if they exit the AI trade. Investors are also putting money into gold, silver and other precious metals as the dollar has weakened.

"Suddenly, there are other options in which to deploy capital," Gave said. So even if the AI bubble does burst, the rest of the market doesn't necessarily have to implode along with big tech.

It's also worth noting that bubbles can take a very long time to pop. Former Federal Reserve Chair Alan Greenspan gave his famous "irrational exuberance" speech about stocks on Dec. 5, 1996. But the Nasdaq Composite didn't hit its dot-com peak until March 2000.

At the request of the copyright holder, you need to log in to view this content

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

  • Lancelotz
    10-06
    Lancelotz
    There is no AI bubble. Bubble means over valuation without people using or paying for the service. Most AI compute are in high demand - meaning real companies or people are paying for the service. You can listen to hyscalers earnings call and you know demand far exceed supply now. Give it another 18 months perhaps to re-look at this "bubble". 
Leave a comment
1
1