Investors in fear of an AI bubble have reverted to dot-com-era defenses. What are they buying?

Investors who are wary of the growing hype around AI but don't want to be left out of the rally have reverted to the safety net strategy from the dot-com era of 25 years ago. They are avoiding the key names of the AI boom, but are picking related companies and technologies and looking for new potential breakout areas. What do they advise?

Some market participants draw parallels between the current market situation and the Internet boom of the 1990s, which started with startups and spread to the telecom and IT sectors, but ended in a crisis. Back then, hedge funds were able to capitalize on the wave by timely exiting overvalued securities before their peak and switching to those that still had growth potential, Reuters writes.

"We are doing what worked in 1998-2000," says Francesco Sandrini, head of multi-asset strategies and chief investment officer in Italy at Amundi, Europe's largest asset manager. He notes signs of "irrational exuberance" on Wall Street - for example, the frenzied trading in risky options tied to shares of leading AI companies.

Still, Sandrini expects the new wave of tech enthusiasm to continue and expects to lock in profits by investing in assets with fairer valuations that could be the next beneficiaries of growth. He said Amundi's strategy is to look for "the most promising areas that the market has not yet noticed," including companies in software, robotics and Asia's technology sector.

"The likelihood of this AI boom turning out to be a bubble is very high, as companies are spending trillions competing for the same market that doesn't actually exist yet," says Simon Edelsten, chief investment officer at Goshawk Asset Management. He expects the next wave of AI hype to extend beyond Nvidia, Microsoft and Alphabet to related industries.

A study by economists Markus Brunnermeier and Stephan Nagel found that during the dotcom bubble, hedge funds mostly did not bet against it, but were able to outperform the market by about 4.5 percent per quarter in 1998-2000 while avoiding the most painful downturn, Reuters wrote. They sold expensive Internet stocks in time to funnel profits into other areas, the agency explained.

"Even in 2000, when the market peaked, nimble players could make good profits," Edelsten added.

- Goshawk's Edelsten favors IT consultants and Japanese robotics companies, which he said could benefit from a boom in investment from artificial intelligence giants. He compared what's happening to the classic gold rush market timeline: "When someone finds gold, it's worth buying the local hardware store, where prospectors will buy all their shovels."

- Investors are also trying to capitalize on the trillions of dollars that the so-called hyperscalers - Amazon, Microsoft and Alphabet - are pouring into developing data centers and manufacturing advanced chips, but have avoided investing directly in their stocks, Reuters writes. According to Fidelity International multi-asset portfolio manager Becky Tsing, her main new bet on AI growth has been uranium: data centers, which consume enormous amounts of energy, could drive demand for nuclear power.

- Kevin Tose, a member of the investment committee at management firm Carmignac, is taking profits on stocks from the Magnificent Seven and building up a position in Taiwan's Gudeng Precision, which makes chip shipping containers for AI chip makers such as TSMC.

- Arun Sai, senior multi-asset strategist at Pictet Asset Management, points out the rapid construction of data centers can turn into overcapacity - similar to what happened during the fiber optic network boom in the telecom industry. "In any new technology cycle, the path from point A to point B inevitably goes through redundancy," says Sy. He is betting on the Chinese market as a defensive tool, believing that China's rapid technology progress could dampen Wall Street's enthusiasm.

- Oliver Blackbourn, portfolio manager at Janus Henderson, said he is hedging his positions in US technology stocks by adding European and medical assets to his portfolio in case the collapse of AI-related securities drags down the entire US economy. According to him, it is impossible to predict how long the AI frenzy will last: it is usually possible to determine the moment of peak only retrospectively.

"We're living in 1999 - until the bubble bursts," summarized Blackburn.

This article was AI-translated and verified by a human editor