Broadcom drags down the stock prices of Micron, SK Hynix, and Samsung, with Wall Street titans issuing the latest warning: The AI bubble is approaching the eve of the 1929 Great Depression
The chip sector is currently the well - deserved hot target in the entire market. However, the sharp rise in its stock prices has also increased market concerns.
On the 4th local time in the United States, U.S. chip stocks such as Broadcom tumbled sharply. Just one day before, Ray Dalio, the founder of Bridgewater Associates, the world's largest hedge fund, pointed out in a media interview that the artificial intelligence track is booming, but there are already signs of a bubble, and such bubbles will ultimately burst.
Broadcom Triggers Sell - off
After rising for several consecutive days, U.S. chip stocks have entered an adjustment phase.
Broadcom (AVGO) closed down 12.59% on June 4th, recording its largest decline in 16 months, and its market value evaporated by more than $285 billion. According to an analysis by the Financial Times, this sharp decline also made Broadcom the fourth - largest company in U.S. history in terms of single - day market value decline.
The trigger for the stock price decline was Broadcom's announced chip sales forecast. According to Reuters, the company expects its artificial intelligence chip sales in the third quarter to be $16 billion, lower than analysts' expectations. In addition, Broadcom's reaffirmation rather than an increase in its sales forecast of $100 billion in artificial intelligence revenue for the fiscal year 2027 also disappointed investors.
Broadcom is an important player in the artificial intelligence infrastructure market, competing with companies such as NVIDIA and AMD (Advanced Micro Devices) to provide semiconductors required by technology companies for training and running AI models. In the past year, Broadcom announced a list of customers for its custom AI chips, including Google, Meta, OpenAI, and Anthropic. However, investors are also questioning whether large technology companies can continue to significantly increase their capital expenditures.
Broadcom's performance forecast has also triggered a sell - off of chip stocks by investors. The stock prices of semiconductor companies including Micron Technology and Qualcomm were generally under pressure on Thursday. Micron Technology tumbled 7.74%, AMD fell 3.56%, and Qualcomm dropped 2.62%.
Currently, although the market remains optimistic about the high - growth rate of the AI industry, once the performance fails to meet expectations, the stock prices will also be affected.
As U.S. chip stocks declined, after the Asian market opened on June 5th, South Korean chip giants tumbled across the board. SK Hynix plunged more than 9%, and Samsung Electronics dropped more than 7%. South Korea's KOSPI index plummeted more than 6% at one point, triggering the circuit - breaker mechanism, and program trading was suspended for 5 minutes. As of the close, SK Hynix closed down 9.92%, and Samsung Electronics closed down 6.40%.
Wall Street Bigwigs Issue Warnings
Dalio warned in a Bloomberg interview on Wednesday that the artificial intelligence track is booming, but there are already signs of a bubble, and such bubbles will ultimately burst.
"Every major technological change will inevitably give rise to a bubble. The reason behind this is that it is difficult for enterprises to accurately control the scale of investment: either they invest heavily to seize market share, or they invest insufficiently and ultimately cede the market," Dalio said.
Driven by the strong demand for high - bandwidth chips in artificial intelligence data centers, chip - manufacturing enterprises have become the hottest targets on Wall Street, driving U.S. stocks to repeatedly hit record highs. This sharp rise in stock prices has also sparked a market debate: Is the current market overheated? Are investors blindly entering the market?
Dalio is worried about the profitability of AI enterprises. He believes that when the market enters the investment realization period, the bubble will burst.
He pointed out that many investors have a cognitive misunderstanding: if they are optimistic about a new technology, they directly buy related stocks. However, the technological prospects and the valuation of stocks are two different things, and an excessively high stock price itself contains risks. In addition, it is necessary to distinguish between book wealth and actual income. Book wealth can easily be created out of thin air through inflated valuations. For example, if a company is valued at $1 billion and only raises $50 million in funds, the founder can become a billionaire through shareholding. However, this book wealth cannot be directly used for consumption, and to realize it, assets must be sold.
He said, "The process of the bubble bursting is the process of the shrinkage of book wealth realization. Although artificial intelligence itself is a high - value and high - quality technology, the current AI - dominated market is following the path of bubble evolution."
He said that the bursting of the bubble is almost an inevitable result. According to him, he has a set of quantitative indicators for monitoring bubbles, covering multiple data such as retail investors' positions and market sentiment. The current bubble level of the AI sector is approaching that before the Great Depression in 1929 and the peak of the Internet bubble in 2000, but it has not reached the extreme value.
Dalio believes that the bubble has two stages: inflation and puncture, and the core inducement for puncture is always the concentrated realization demand. In history, the Japanese real - estate bubble, the U.S. stock market crash in 1929, and the burst of the Internet bubble in 2000 were all cases where monetary tightening forced the market to sell assets concentratedly to realize cash. No matter how good the technology is, it cannot change this cyclical logic. The only variable is how long the bubble can last. To accurately time the market, one needs to judge the degree of bubble inflation and closely monitor the trigger signals for concentrated realization.
In addition to Dalio, Michael Burry, a legendary Wall Street investor and the prototype of the character in the movie "The Big Short", has also issued another "warning" about the artificial intelligence bubble.
Well - known technology blogger EdZitron said in a Bloomberg interview recently that Anthropic and OpenAI are dangerous and unsustainable companies and should not go public. The AI bubble is a scam, and retail investors are the targets to be harvested. AI has no return on investment. It is completely different from AWS (Amazon Web Services) or Uber, and there is no recovery story after the bubble bursts.
Michael Burry reposted EdZitron's tweet about the interview on social media and said, "It's worth listening to." Michael himself has also frequently warned recently that the current AI mania is similar to the Internet bubble period.
However, some institutions are optimistic about the prospects of artificial intelligence. In its mid - 2026 outlook report released recently, JPMorgan Chase said that the AI super - cycle has just begun.
The report said that the current mainstream view in the market around the AI super - cycle has become overly pessimistic. This technology is being implemented on a large scale at the consumer and enterprise levels, spurring record - high investments in the power and infrastructure sectors, bringing tangible improvements in production efficiency, and is expected to reshape the patterns of healthcare, education, population problems, and debt sustainability. However, Wall Street is bearish on it. This phenomenon reflects more the shortcoming of the market's pricing logic in the face of structural changes than doubts about the AI technology itself.
This article is from the WeChat official account “China News Service - New Economy” (ID: jwview). The authors are Wei Wei and Luo Kun. It is published by 36Kr with authorization.