Dow Near Record Highs, Nasdaq 100 Enters Correction—The Real Reason Why Only 'Semiconductors' Were Sold Off in the July 28 NY Market

The New York market on July 28, 2026, was an extremely rare session where opposite records were set within the same day.

The Dow Jones Industrial Average closed up 537.24 points (+1.03%) at 52,747.32, reaching record-high territory. 'Old-school blue-chip stocks' that were well-received in earnings reports pushed the index higher, with Coca-Cola up 5% and Sherwin-Williams up 8%. However, on the same day, the Nasdaq Composite fell 0.22% to 24,876.91, and it was reported that the Nasdaq 100 had entered a correction phase (down 10% from its recent high). The semiconductor ETF (SMH) fell over 3% for its fourth consecutive day of losses, while Micron and AMD both dropped over 8%. The S&P 500 ended almost exactly in the middle, unable to move, at 7,428.78, up 0.21%.

The trigger was not the United States, but Asia. South Korea's SK Hynix and Samsung Electronics plummeted over 14% and 13% respectively, and a circuit breaker was triggered for the KOSPI. The Nikkei Stock Average closed down 2,566 yen (-3.95%) at 62,364 yen, with Tokyo Electron and Advantest falling over 10%. Reports that a Chinese state-owned enterprise had begun manufacturing immersion DUV (deep ultraviolet) lithography equipment, combined with the presence of Chinese memory giant CXMT (ChangXin Memory Technologies), which had listed the previous day with a 470% gain, ignited the sell-off starting from Asia.

At this point, many commentary articles stop by saying, 'Caution regarding the AI bubble has spread.' However, there is one thing that cannot be explained by that alone. If the market had begun to doubt the future potential of AI itself, the S&P 500 should have fallen, the U.S. 10-year Treasury yield should have plummeted due to a flight to quality, and gold should have been bought. In reality, the 10-year Treasury yield moved little in the low 4.6% range, gold fell 1.25%, and the Dow remained near record highs. Capital did not flee the stock market; it simply moved 'within' the stock market.

So, what did investors sell? In conclusion, what was sold was not the demand for AI, but the 'number of years' that semiconductor manufacturers could continue to reap excess profits. Whether or not one understands this difference will significantly change how one approaches semiconductor stocks going forward and how one interprets tomorrow's Tokyo market.

Below, I will verify the causal relationships of this day based on the consistency of figures across multiple markets, and break down how the FOMC on July 29 and the Tokyo market might react tomorrow.

Background

Conclusion: The spark for this decline was not the U.S. economy or interest rates, but news of a type that had previously been difficult to use as daily stock price material: 'China's localization of technology.'

The U.S. market in July was caught between two opposing forces.

One is the cooling of inflation. The U.S. Consumer Price Index (CPI) for June, announced on July 14, fell 0.4% month-on-month and was 3.5% year-on-year, a significant drop from 4.2% in May. This was lower than the market forecast of 3.8%, and the 5.7% month-on-month decline in the energy index had an impact. Core CPI, excluding food and energy, was flat month-on-month and up 2.6% year-on-year.

The other is crude oil. Safety in the Strait of Hormuz was threatened by the conflict between the U.S. and Iran, and it was reported that crude oil briefly exceeded $100 per barrel in July. This is why the energy index of the CPI remains at a high level, up 15.7% year-on-year, with gasoline up 26.7%. In other words, the current decline in inflation was of a nature that 'could be reversed at any time depending on crude oil.'

As a result, the outlook (dot plot) from the June FOMC raised the median policy interest rate for the end of 2026 to 3.8% from 3.4% in March, reversing to content that suggests a rate hike rather than a rate cut. It is an unusual distribution, with 9 out of 18 members expecting at least one rate hike within the year. The policy interest rate remains held at 3.50–3.75%, and if a hold is decided on July 29, it will be the fifth consecutive meeting. For Chair Kevin Warsh, whose first meeting was in June, this is his second time at the helm.

Then, news flowed in from a completely different source. On July 27, Chinese memory giant CXMT gathered record popularity on its first day of listing, rising 470% and leaping into the ranks of the largest listed companies in China by market capitalization. Furthermore, it was reported that a Chinese state-owned enterprise had begun manufacturing immersion DUV lithography equipment, which had been almost exclusively produced by the Dutch company ASML. With macro tensions continuing, materials that shake the very structure of the sector overlapped—this was the stage setting for the 28th.

Facts

Conclusion: Within a single day, opposite records were simultaneously established: 'The Dow is near record highs, while the Nasdaq 100 has entered a correction.'

I will list the figures and what each of them means.

NY Dow: Up 537.24 points (+1.03%) to 52,747.32. The content of the rise includes Sherwin-Williams up 8% and Coca-Cola up 5%. The former is paint, the latter is beverages, and both are unrelated to AI. Coca-Cola exceeded market expectations for both sales and profit and raised its full-year outlook. The rise in the Dow was also a confirmation that 'real demand outside of AI is solid.'
S&P 500: Up 0.21% to 7,428.78. The fact that the market-cap-weighted index was almost flat means that gains and losses canceled each other out within the index.
Nasdaq Composite: Down 0.22% to 24,876.91. It is reported that the Nasdaq 100 has entered a correction phase.
Semiconductors: SMH down over 3% for the fourth consecutive day of losses. Micron and AMD down over 8%. In Asia, SK Hynix was down over 14%, Samsung Electronics down over 13%, and a circuit breaker was triggered for the KOSPI after a drop of over 9%. The Nikkei Stock Average was down 2,566.27 yen (-3.95%) to 62,364.92 yen, with Tokyo Electron and Advantest down over 10%, and Kioxia hitting its daily limit low.
Crude Oil: WTI down 4% to $79.26 per barrel, Brent down 4.8% to $84.09, both at their lowest levels in about a week. It was reported that Iranian Foreign Minister Araghchi held successive phone talks with the foreign ministers of Saudi Arabia and Oman, with the security of the Strait of Hormuz as the agenda. This is the shedding of the geopolitical premium.
U.S. 10-Year Treasury Yield: Around 4.62–4.64%. This is a straightforward reaction where lower crude oil prices lowered inflation expectations, and bonds were bought (yields fell).
Gold: Down 1.25% to around $4,026 per troy ounce. The strong dollar and waiting for the FOMC acted as a drag.
Dollar-Yen: Around 163.75 yen, almost flat.
Of these, the point that will be decisively effective in later analysis is that 'none of the 10-year Treasury yield, gold, or crude oil showed risk-off movements.'

Market Reaction

Conclusion: What investors sold was not the demand for AI, but the 'remaining years' during which semiconductor manufacturers can secure excess profits.

First, let's start by process of elimination. In a typical risk-off scenario, three things occur: a broad decline in stocks, a rise in bonds (lower yields), and a rise in gold. How was the 28th? Stocks were far from a broad decline, with the Dow near record highs; the lower yield was explained by lower crude oil prices; and gold actually fell. In other words, the price movements of this day did not meet a single requirement for macro risk aversion. Capital did not leave the stock market; it simply moved sideways from semiconductors to consumer staples and capital goods.

So what happened? What was effective here was the 'type' of news reported. China's in-house production of DUV lithography equipment and the rise of CXMT are not stories about a decrease in demand for AI chips. With demand as a premise, it is a story that rewrites 'who can take the profits generated from that demand, and for how long.'

Semiconductors, especially memory and manufacturing equipment, are 'rent (excess profit) industries' that generate abnormal profit margins during boom periods due to high barriers to entry. And the stock prices of rent industries are determined not by textbook P/E ratios or next-term EPS, but by 'how many more years that rent will last.' The length of the period during which high profitability continues is the dominant variable, not the level of profit, which is the numerator.

Added to this was a broker report transmitted from South Korea stating that 'memory prices will peak in 2027.' This is not a story about moving next-term EPS by 1–2%. It is a story about pulling the end point of the high-profit period closer than the market had assumed. In terms of discounted cash flow (DCF), the plateau portion of the cash flow is shaved off by several years. This explains why SK Hynix and Micron moved 8–14% in a single day. The reason individual stocks move by double digits on a day when the index only moves 0.2% is that the variable that moved was not 'this term's numbers' but the 'lifespan of the business model.'

Simultaneously incorporated was the 'emergence of new suppliers.' CXMT's 470% gain on its first day of listing was accepted as proof that China's domestic capital market can supply large-scale funds to the memory industry. The in-house production of lithography equipment adds the condition that 'capital investment can be continued even under sanctions.' If the number of suppliers increases, the lifespan of the rent shortens further.

From here on is my own hypothesis.

The market is currently processing this series of news as 'commoditization by China = bearish for tech as a whole.' However, the fact that semiconductors and manufacturing equipment are becoming commoditized and prices are falling is, for the buying side—hyperscalers, software companies using AI, and general business companies incorporating AI into their operations—a simple reduction in cost. It is merely a profit pool allocation problem of whether the upstream (the side selling chips) or the downstream (the side using AI) takes the fruit of the same AI.

The price movements on the 28th may indicate the early stages of that allocation shift. The fact that capital left semiconductors but did not leave the stock market as a whole, and that stocks like Coca-Cola, which symbolize downstream real demand, were actually bought, does not contradict this hypothesis. If this view is correct, the focus going forward will shift from the blanket question of 'will AI-related stocks go up or down' to the selective question of 'which layer of the AI value chain are you standing in.' Of course, this is one hypothesis, and I cannot deny the possibility that a slowdown in demand itself will be confirmed later.

Future Outlook

Conclusion: The immediate protagonist is the FOMC on July 29, but the point of contention this time is not 'will there be a rate cut' but 'will they move to a rate hike.'

The policy interest rate will be announced at 2:00 PM U.S. Eastern Time on July 29 (3:00 AM on the 30th, Japan time), and Chair Warsh's press conference will follow 30 minutes later. A hold is considered likely, but it is reported that the futures market has priced in a probability of a rate hike at the July meeting of about one-third, and nearly 80% for the September meeting. In mid-July, when crude oil surged, the probability of a rate hike within the year rose to 38%, and it has retreated somewhat due to the drop in crude oil on the 28th. In other words, the current interest rate outlook has effectively become a dependent variable of crude oil prices.

There is one point to keep in mind. It is reported that Chair Warsh has expressed a policy of narrowing forward guidance, and because the amount of information in the statement will decrease, the market will be more likely to overreact to his choice of words at the press conference. Fluctuations after the event passes may be larger than before.

The bullish scenario is a development where, in addition to a hold, a cautious tone is shown based on the fact that core CPI has fallen to 2.6% year-on-year. If rate hike expectations recede and crude oil prices continue to fall, combined with lower interest rates, the shift of funds to areas other than high-tech will progress further.

The bearish scenario is a development where inflation concerns return due to a re-rise in crude oil or a resurgence of the situation in the Strait of Hormuz. If the pricing in of a September rate hike approaches 100%, growth stocks with high valuations will face a double headwind of rising interest rates and intensifying competition.

Regarding semiconductors themselves, there are three materials to assess over the next few weeks. First, the order trends of manufacturing equipment companies; second, the spot prices of DRAM and NAND; and third, follow-up reports on the actual yields and mass production capabilities of Chinese players. This sharp decline is merely a reaction to 'future news,' and verification based on actual data is yet to come. It can be said that this is a phase where one should avoid making definitive judgments.

Consideration of the Impact on the Japanese Market on the Day

Conclusion: The Tokyo market will move not based on 'did the NY Dow go up' but on 'what went up in NY.' Since it has already largely priced this in on the 28th, there is both room for a rebound and a risk of the issue being rehashed.

What is important in thinking about the Tokyo market on July 29 is that Japanese stocks have already priced in this news with a significant decline of -3.95% on the 28th. In terms of sequence, the structure is that Japan was sold first, and then the U.S. market returned a selective reaction of 'Dow up, semiconductors down.'

The movement of futures is suggestive. It is reported that Nikkei 225 futures (September 2026 contract) gradually raised their level in night trading on the 28th, at 62,410 yen at midnight on the 29th (Japan time) (up 10 yen from the previous day) and 62,850 yen at 2:00 AM (up 450 yen). It is a structure where the firmness of the NY Dow spilled over while the Nasdaq decline suppressed the upside, and it is a level with more of a color of autonomous rebound than a significant rebound.

By sector, there is a possibility that polarization will be brought in as is. Those likely to face headwinds are semiconductor manufacturing equipment such as Tokyo Electron, Advantest, Disco, and SCREEN; memory such as Kioxia; and SoftBank Group, which has a high weight of related investments. Since these are also high-priced stocks in the Nikkei Average, it is necessary to be aware that the impact on the index is likely to be greater than actual corporate performance.

On the other hand, tailwind materials are also in place. The dollar-yen is at a weak yen level in the high 163 yen range, and the profitability of export companies itself has not deteriorated. The 4% drop in crude oil is a clear positive for industries with high fuel cost ratios such as electric power/gas, shipping, air transport, and chemicals. Given that what was bought in the NY market were consumer staples and capital goods, it is conceivable that funds will head toward domestic demand defensives such as food and retail, and capital goods such as electric wires and heavy electrical equipment in Tokyo as well.

If one takes an optimistic view, bad news was digested all at once on the 28th, and if the decline in high-priced stocks runs its course, the index will regain its composure. If one takes a pessimistic view, China's localization of technology is not a story that will be solved overnight, and there is a possibility that it will be repeatedly rehashed as a structural theme for Japanese equipment stocks. In addition, with the FOMC in the early hours of the 30th (Japan time) ahead, it is also an environment where it is difficult to actively buy up. It would be realistic to judge the direction after seeing the movement of high-priced stocks after the opening and the semiconductor stocks in Korea and Taiwan during Asian hours.

Implications for Japanese Individual Investors

Conclusion: What you should check now is not 'whether to buy or sell AI stocks,' but whether your portfolio is making the same single bet across Japan and the U.S.

1. Check if 'Japan-U.S. diversification' is truly diversification.

High-priced stocks in the Nikkei Average (semiconductor manufacturing equipment, SoftBank Group), stocks constituting the U.S. SOX index, and large-cap tech included in the S&P 500 and all-country stock indices are effectively investments in almost the same theme. On the 28th, Japan, Korea, Taiwan, and the U.S. moved in the same direction for the same reason. Even if you separate countries, it is not diversification if there is only one risk factor. I recommend rearranging your holdings and the contents of your investment trusts not by country, but by 'theme.'

2. Check if you are double-dipping on the benefits of a weak yen.

The level of 163 yen to the dollar is affecting both U.S. stock investment trusts without currency hedging and Japanese export companies at the same time. Conversely, in a phase where the yen strengthens, these two will turn disadvantageous at the same time. If you break down how much of your current unrealized profit is due to corporate strength and how much is due to exchange rates, your next judgment will become much easier.

3. Use the FOMC in the early hours of the 30th (Japan time) as material for confirmation, not action.

What actually moves in response to the statement and press conference is crude oil and the dollar-yen, rather than the policy interest rate itself. If these two move, sector rotation in Japanese stocks will change from the next day onward. As long as forward guidance is narrowed, this is not a phase to expand positions or use leverage in advance.

This article is for informational purposes only and does not recommend the buying or selling of any specific financial product. Please make investment decisions at your own responsibility.

Sources

CNBC
Bloomberg
Reuters
Yahoo Finance
TheStreet
Proactive Investors
Morningstar
CBS News
FXStreet
Investing.com
Trading Economics
Nihon Keizai Shimbun
Kabutan
Minkabu
Kabushiki Shimbun
U.S. Bureau of Labor Statistics (BLS)
U.S. Federal Reserve Board (FRB)