1999 Carnival Replayed? Wall Street Uses Internet Era Tactics to Tackle AI Bubble

In the face of the AI frenzy, some large investors are revisiting the old script of the 1990s, remaining vigilant about the risks of a bubble burst while trying to carve out a share of the feast.

As the U.S. stock market continues to hit new highs, the market capitalization of AI chip giant NVIDIA has soared to over $4 trillion, creating a frenzied atmosphere that makes professional investors uneasy. Francesco Sandrini, multi-asset head at Europe’s largest asset management company Amundi, candidly stated: “What we are doing now is exactly the strategy that worked from 1998 to 2000.” He pointed out signs of irrational exuberance on Wall Street, such as unusually active risk options trading related to major AI stocks.

This shift in strategy indicates that funds are gradually flowing out of giants like the "Mag7" in search of the next growth point within the AI ecosystem. Investors hope to gain returns by betting on relatively reasonably valued assets such as software, robotics, and Asian tech companies, which are seen as having high growth opportunities yet to be discovered by the market.

However, this "treasure hunt" on Wall Street is also accompanied by warnings and divisions. Some investors worry that history may repeat itself, fearing that the current data center construction boom could echo the overcapacity issues faced by the fiber optic cable industry in the past. Others believe that merely rotating investments within the AI sector is insufficient to hedge against risks and choose to use European assets or healthcare stocks to guard against the potential impact on the U.S. economy if the AI bubble bursts.

Learning from History: Rotation Strategy Outperformed the Internet Bubble

History provides a reference for today's investors. During the internet boom of the late 1990s, some hedge funds did not choose to short the market counter to the trend but successfully navigated the bubble through flexible rotation strategies.

A study by economists Markus Brunnermeir and Stefan Nagel showed that between 1998 and 2000, hedge funds achieved an average quarterly outperformance of about 4.5% by skillfully navigating the internet bubble and successfully avoiding the most severe declines after the bubble burst. Their core tactic was to sell high-priced internet stocks before they peaked and to shift profits into other stocks that had not yet caught the attention of ordinary investors.

“Even at the peak of the bubble in 2000, those who acted quickly could make substantial profits,” said Simon Edelsten, chief investment officer at Goshawk Asset Management, who believes the current market environment is very similar to that of 1999. Edelsten participated in telecom IPO projects in 1999. He expects the next phase of the AI boom to spread from companies like NVIDIA, Microsoft, and Alphabet to related industries.

The "Shovel Seller" Logic: Mining for Second-Tier Winners in the Supply Chain

In the current AI "gold rush," investors are practicing an age-old logic: When everyone is out to pan for gold, the safest business is to sell shovels and picks to the gold miners. This means that investors are trying to benefit indirectly from the wave of investing hundreds of billions of dollars in building AI data centers and procuring advanced chips from "super-scale companies" like Amazon, Microsoft, and Alphabet, rather than directly increasing their holdings in these giants.

Simon Edelsten favors IT consulting firms and Japanese robotics groups that can generate revenue from AI giants. Kevin Thozet, a member of Carmignac's investment committee, is reducing his holdings in "Mag7" stocks while building positions in Taiwan's Jaden Precision, which provides transfer boxes for AI chip manufacturers like TSMC. Becky Qin, a multi-asset manager at Fidelity International, sees uranium as her favored new AI trade, as the energy-intensive AI data centers may drive up demand for nuclear energy.

Bubble Alerts and Diversification Hedging

Although the rise of leading AI stocks like Microsoft, Amazon, and Alphabet is strongly supported by earnings, some investors still see "the elements of a bubble."

Arun Sai, a senior multi-asset strategist at Pictet Asset Management, warns that any new technological paradigm inevitably experiences excess behavior during its development. He is concerned that the frenzied construction of data centers may lead to overcapacity, reminiscent of the fiber optic cable boom in the telecom industry years ago. As a hedge, he is optimistic about Chinese stocks, believing that if the rapid development of AI technology in China weakens Wall Street's enthusiasm for AI, these stocks may benefit.

Oliver Blackbourn, a portfolio manager at Janus Henderson, stated that he is using European and healthcare assets to hedge his positions in U.S. tech stocks, in case a collapse in AI stocks drags down the entire U.S. economy.

He admitted that predicting how long the AI boom will last is nearly impossible, as determining the peak can usually only be done in hindsight. As he said:

"Before the bubble bursts, we are all in 1999."