The market's advance-decline ratio remains at 1:4, with the broad-based decline in individual stocks showing no signs of improvement. On one side is the frenzied technology sector, and on the other are all other sectors. The market's trading volume has shrunk to 2.8 trillion, further validating the logic I discussed yesterday—this is currently a market of existing capital seesawing between strong and weak sectors.
The sectors mentioned yesterday, such as MLCC, coal, optical optoelectronics, metal materials, and semiconductor materials, continued their collective strength today. However, the performance of other sectors is difficult to describe, especially with traditional consumption sectors continuing to see capital outflows, and the battery sector facing a similar situation. Selling these traditional industries to chase technology has become a consensus to a certain degree.
AI has a five-layer cake: energy, chips, infrastructure (mega data centers), models, and applications (value creation, new product and service development and implementation). Currently, the focus is mainly on the first three layers. Some aspects within the infrastructure layer haven't fully played out yet, and the backend applications haven't been significantly speculated on either. Looking at this as a major cycle, this round of AI market activity is not over.
Furthermore, based on the logic that stock prices peak about six months before an industry cycle peaks, the peak of the industry cycle for many AI hardware components is still far from appearing. It's fair to say this round of AI activity far exceeds any cycle seen in recent decades.
The situation in overseas markets is similar. In the U.S. stock market, only AI and tech are rising, while other sectors are languishing. Particularly, the old-guard consumption sector has fallen 6.5% over the past two weeks, while the S&P 500 rose 3% and the U.S. information technology sector surged 13%. In just two weeks, U.S. consumption stocks have underperformed U.S. tech stocks by 20%.
Similarly, Bitcoin has fallen 50% from its highs, and gold has been steadily declining. Some attribute this to rising U.S. Treasury yields, but a more realistic reason is that liquidity is being sucked away by AI. Why can AI absorb most of the money? Because its gains are simply too exaggerated, with no signs of stopping. If blindly buying AI and other tech stocks can yield 6% in a single day, who would want to hold products with an expected annual return of only 6%?
This is reminiscent of a new term recently coined in South Korean society: "벼락거지" (byeorak-geoji), meaning "suddenly impoverished by lightning." It describes the feeling of working hard every day but watching those around you make a decade's worth of salary in a single market move, making you feel as if struck by lightning and instantly becoming relatively "poor." This extreme FOMO (Fear Of Missing Out) sentiment is driving everyone, eyes red, into the stock market, and the pace of buying is accelerating.
Despite the market frenzy, I believe not everyone in the market is bullish on AI-related themes. The bubble hasn't reached a manic stage yet. Historically, major market rallies often end only when the stock market breaks out of its usual circles and attracts a massive influx of new, outside participants.
However, one thing worth sharing is that even a well-known antique appraisal blogger has started talking about optical module CPO. This suggests the market's rise is indeed beginning to break into wider circles. Have you noticed more people around you recently sharing about stock trading? If people who normally don't discuss stocks have started talking about them, please leave a comment.
Having said all that, let's discuss a few noteworthy developments related to capital markets. The first is the potential for a Japanese Yen interest rate hike. In late May, an American scholar pointed out that "the Yen has fallen below the Turkish Lira, becoming the world's weakest currency." Cooling expectations for the end of the U.S.-Israel-Iran conflict pushed the international crude oil benchmark WTI futures price above $90 per barrel. Coupled with weak domestic manufacturing (especially the automotive industry being caught up and surpassed), the market expects Japan's trade deficit to widen further, leading to trades selling Yen and buying U.S. dollars.
There was a recent discussion about exchanging money to go work in Japan, only to find that the money earned, when converted back to Renminbi, is worth less and less, with purchasing power declining rapidly as wages fail to keep up with prices. This potential rate hike might be akin to drinking poison to quench thirst, unlikely to achieve a real effect. However, as the Yen is a global carry trade currency, a rate hike could cause some market volatility.
The second development is the pre-market decline in the U.S. semiconductor sector, primarily influenced by Broadcom Inc.'s earnings. After the market closed yesterday, Broadcom Inc. reported its Q2 FY2026 results. Net revenue exceeded $22 billion for the first time in a single quarter, slightly above expectations, representing a year-over-year increase of about 48%. Revenue from the semiconductor solutions business grew nearly 80% year-over-year. Non-GAAP adjusted earnings per share (EPS) increased over 50% year-over-year, both stronger than Wall Street expectations.
The company forecasted Q3 revenue of approximately $29.4 billion, about 2.8% higher than analyst expectations but not as impressive as before. The Q3 AI semiconductor revenue guidance is $16 billion, a more than twofold year-over-year increase, yet still about 7% below the average analyst estimate. This indicates the market has piled significant expectations onto AI itself. It's manageable if one or two companies fall short, but widespread earnings misses would be concerning. Whether subsequent companies can withstand earnings verification will be key to the continuation of this market trend.