AI bubble alarm sounds! Investors restart "survival strategies" from the internet bubble era

According to the Zhitong Finance APP, due to concerns about the bursting of the artificial intelligence (AI) bubble while simultaneously holding a cautious stance on shorting the sector, some investors have restarted a classic strategy from the internet bubble era of the 1990s—pulling out of overhyped AI stocks and shifting towards potential "next wave winners." This strategy helped some investors successfully avoid the market crash decades ago.

As U.S. stocks continue to hit new highs, the AI chip manufacturer NVIDIA (NVDA.US), which has benefited the most from this AI boom, has seen its market value surpass $4 trillion. Meanwhile, professional investors are looking for investment methods that can participate in the bull market while avoiding excessive risks. Some investors are recalling the internet boom of the 1990s—when the bubble spread from startups to the telecommunications and technology sectors, hedge funds profited by pulling out of overvalued popular stocks before the crash and shifting to companies with still significant upside potential.

Francesco Sandrini, head of the multi-asset department and chief investment officer for Italy at Amundi, Europe’s largest asset management company, stated: "What we are doing now is exactly what worked from 1998 to 2000." He pointed out that there are signs of irrational exuberance on Wall Street, such as unusually active high-risk options trading around large AI stock prices. However, he also indicated that he expects this new tech enthusiasm to continue and hopes to profit by betting on "reasonably valued assets that could become the next wave of rising stars."

Francesco Sandrini noted that the core of this strategy is "to find the highest growth opportunities that the market has yet to discover," including software companies, the robotics industry, and the Asian technology sector. Other investors also plan to gradually exit from the "Magnificent Seven" of U.S. stocks but still hope to maintain a diversified layout in the AI field.

Flexible Operations

Simon Edelsten, who was responsible for telecommunications industry IPOs at Dresdner Kleinwort Benson in London in 1999 and is now the chief investment officer at Goshawk Asset Management, stated: "The likelihood of this AI boom ultimately bursting is very high because too many companies are spending trillions of dollars to compete for a market that does not yet exist." He expects the next phase of the AI boom to spread from NVIDIA and companies like Microsoft and Alphabet to related industries.

Historically, correctly grasping the various stages of a bubble has been a way to profit without taking the risk of "topping out" too early. A study by economists Markus Brunnermeier and Stefan Nagel showed that most hedge funds did not directly short related stocks during the internet bubble but instead operated flexibly, outperforming the market by about 4.5% on average each quarter from 1998 to 2000, successfully avoiding the most severe downturns. They timely sold overvalued internet stocks and reinvested the profits into other companies that had not yet attracted attention from mainstream investors Simon Edelsten stated, "Even during the peak of the 2000 bubble, agile responders could still make substantial profits." He added that the current market environment is quite similar to that of 1999. Simon Edelsten pointed out that he is optimistic about IT consulting firms and Japanese robotics companies—these companies are expected to gain new revenue from AI giants, which aligns with the typical path of a "gold rush." "When someone is mining for gold, (you should) go buy from the local hardware store selling shovels."

Stay in the AI field, but don't take too much risk

Investors are also trying to benefit from the trillions of dollars being invested in AI data centers and advanced chips by so-called "hyperscalers" like Amazon, Microsoft, and Alphabet, while trying to avoid directly increasing their risk exposure to these companies.

Becky Qin, a multi-asset fund manager at Fidelity International, stated that uranium is her favored new AI-themed investment target, as the energy-intensive AI data centers may heavily rely on nuclear power.

Kevin Thozet, a member of the investment committee at asset management firm Carmignac, is gradually cashing in on profits from the "seven giants" of the U.S. stock market while increasing his holdings in Gudeng Precision, a Taiwanese company that provides wafer carriers for AI chip manufacturers, including TSMC.

Asset managers are also concerned that the data center construction boom could lead to overcapacity—similar to the fiber optic cable expansion bubble in the telecommunications industry years ago. Arun Sai, a senior multi-asset strategist at Pictet Asset Management, pointed out, "In any new technological paradigm, we cannot avoid experiencing phases of excess." Despite the strong current performance of major AI companies like Microsoft, Amazon, and Alphabet, he still believes there are "signs of a bubble" in the market and tends to view Chinese stocks as a hedging tool.

However, some investors do not agree with this "relative value" strategy to cope with potential AI risks. Oliver Blackbourn, a portfolio manager at Janus Henderson, stated that he is hedging his U.S. tech stock positions by holding European and healthcare assets to guard against a potential collapse of AI stocks dragging down the U.S. economy. He noted that no one can accurately predict how long the AI frenzy will last—because "the peak of a bubble can only be judged in hindsight," and "before the bubble bursts, we are still in 1999."