Today's US Stock Market Summary: This Is Just a Healthy Shakeout, the AI Wave Is Only Halfway Through

Semiconductors indeed took a heavy hit today, with the AI infrastructure line broadly under pressure, and several stocks saw intraday declines of over 5%. Watching your account balance drop is never easy—it's a normal emotional reaction, not a lack of rationality or poor execution on your part. I feel the same way today as everyone else.

If you've been on the bull side during this period, it's likely been a tough time. The inner struggle and discomfort you're feeling right now are shared by many in the market—you're not fighting alone.

Let me state the conclusion first: This decline is essentially a natural pullback after too rapid a rise.

Take the Philadelphia Semiconductor Index (SOXX) as an example. A 15% to 25% correction in the middle of a strong rally is actually quite healthy. Individual stocks usually see even larger swings. This is a normal process of turnover and digestion within the trend, not the end of the trend. More importantly, if we view the entire AI industrial wave as a complete curve, my personal judgment is that we are probably only about halfway through, with a long road ahead.

Hold patiently, and let time do the compounding.

Buying on dips and holding patiently is the simplest approach, but it's also the easiest way to outsmart the smart money.

Over the past two years, AI has risen quickly, with several rounds of valuation corrections in between. I think that's perfectly normal—even inevitable. Almost every correction triggers market chatter about the AI trade being over, the bubble bursting, and this time being different. But looking back, these voices have appeared many times before, yet the AI industry hasn't stopped; it has continued to advance step by step to where it is today.

As of today, I haven't seen anything that would overturn my original judgment. Cloud vendors are still expanding AI capital expenditures, the pace of enterprise AI deployment is accelerating, commercialization by leading model companies like OpenAI and Anthropic is speeding up, token consumption for large models is growing rapidly, data centers are being built continuously, and demand for AI infrastructure components like storage, networking, optical interconnects, and advanced packaging hasn't suddenly disappeared because of the recent drop.

Let me give you an example: if you bought Bitcoin in 2019 or 2023 and held it according to the four-year cycle theory, you'd understand better.

Even in a bull market, it's normal for Bitcoin to experience several 30% to 40% pullbacks along the way. Altcoins are even more volatile, having gone through multiple 70% to 80% shakeouts. But if you truly held all the way to the end of that four-year cycle in 2025 and looked back, you'd find that the US stock market is actually much friendlier.

Holding US stocks in a bull market, you typically experience only two or three 10% corrections a year, and individual stock fluctuations are around 20% to 30%. But as long as the direction is right, it keeps moving upward, with solid fundamentals. Companies are making money for you even while you sleep—that feeling is very reassuring.

The fundamentals of the US stock market are still very good, and corporate profitability continues to strengthen. The US stock market is truly a market where, as long as you have enough patience, roughly understand the cycles, and know where you are, you can steadily make money by holding. That's because behind it are stable financial and legal systems, and the major tech giants cannot cut capital expenditures at this point (this year's confirmed CapEx is $725 billion, only increasing, not decreasing; next year is expected to be $1 trillion; beyond 2028 is unknown, but I expect the inflection point around the end of 2027).

This is still an arms race. The US and China are also continuously competing on the AI track, and true leading model companies like OpenAI and Anthropic haven't even gone public yet. There are still many stories to come.

My prediction for a full bear market is around end of 2027 to mid-2028. We're still in a bull market now. As long as we're in a bull market, stay calm, hold patiently, don't use leverage, try to buy common stock or LEAP calls expiring in January 2028, and add positions at key moving averages (I'll explain this later).

So, I hope everyone can let logic overcome emotion. Now, let me go through today's specific events in detail.

Iran Rekindles the Strait of Hormuz

Technically, the day started out relatively stable, with chip stocks bouncing and the S&P steadily testing highs, until a major news broke in the afternoon Eastern Time—Iran fired on commercial oil tankers in the Strait of Hormuz again, this time targeting three vessels belonging to Qatar and Saudi Arabia. The trigger was these tankers switching to a US-backed Oman channel, prompting Iran's armed retaliation. Although no casualties were reported, the fragile ceasefire collapsed. The US quickly revoked the previously issued oil sales exemption for Iran. Upon the news, oil prices instantly surged over 3%, retaking the $70 mark, and the 10-year Treasury yield rapidly climbed above 4.5%. This sudden geopolitical risk directly interrupted the financial sector's rally, with major banks and regional bank stocks plunging before the close, ending at the day's lows. Trump had previously threatened that the US could completely destroy all of Iran's power plants and infrastructure in an afternoon. Coupled with the funeral of Iran's former Supreme Leader ending on July 9, high domestic sentiment, and Netanyahu flying to Washington next week to meet Trump on the Iran issue, this geopolitical tension is unlikely to cool off in the short term. I suspect the president has gone long on oil again. After 20 days of oversold conditions on the daily chart, a technical rebound is due. Short-term resistance levels: SMA 200 at 74, EMA 200 at 78.4, and the gap at 84. This is likely just a technical bounce, followed by further declines, as the VIX hasn't reacted much. In the short term, keep an eye on the XLE ETF.

Samsung's Earnings Letdown Triggers Semiconductor Plunge

Samsung Electronics reported historic quarterly earnings, with operating profit surging 19 times year-over-year. Quarterly profit even exceeded the full-year 2025 expectations, making it one of the most profitable companies globally.

But this stellar report turned into a liquidity trap in the market. One interpretation is that while Samsung's revenue met market expectations, it didn't meet the whispered expectations of some Wall Street funds—meaning a 100-point earnings report was anticipated as 120 points. Despite the fact that inventory was sold out due to surging demand and a 20% price increase was planned, a deeper concern emerged: Samsung's massive profitability might be squeezing the margins of other parts of the AI supply chain, similar to the logic when Apple's huge profits were seen as capturing supply chain gains. The result was that good news was taken as bad news. South Korea's KOSPI plunged, Samsung locally fell 7% to 9%, SK Hynix collapsed over 10%, and the selling quickly spread to US stocks. However, the Korean market has since stabilized. Watch whether EMA 50 at 7800 can hold. I think SK Hynix will break down, retest the downtrend line and EMA 20 at 2,400,000, and then continue to correct (Figure 1). The correction should be largely complete before the earnings reports of tech giants. There's still SK Hynix's earnings this month—with expectations lowered before earnings, they are likely to rise afterward.

Technical Analysis: S&P Relatively Resilient, Nasdaq and Small Caps Under More Pressure

The S&P 500 closed still holding the 5-day moving average, forming a doji for the day. Although less than 2% from the all-time high, it staged a failed breakout that attempted to break above the descending trendline from the highs but fell back below.

The range between 749 and 750 is the current key battleground between bulls and bears. If it can reclaim 750, it could push toward 752 and beyond. If it breaks below the intraday low (around 745) in the short term, it could further decline to 740 or even retest the June low near 723. On a finer technical level, the daily MACD remains in an upward bullish formation, indicating the channel to test the all-time high is still open. However, the 30-minute chart shows a double bearish divergence in MACD and RSI, suggesting a healthy 1-hour level pullback or consolidation in the short term, possibly 1-2 days before resuming the attack on overhead resistance.

The Nasdaq (Q) faces more obvious pressure. In the short term, while there is a bullish divergence in MACD and RSI, it hasn't been fully confirmed at the close. If the open prints a green histogram confirming the divergence, Q could mount a 1-hour level bounce to reclaim the 5-day moving average. However, there is heavy institutional selling pressure overhead, with 720 as a key resistance to watch.

The Russell 2000 small caps (IWM) have the worst technical picture among the three, confirming a daily-level top reversal signal. The daily chart shows an extremely rare triple MACD bearish divergence and quadruple RSI bearish divergence, indicating that the previous upward momentum is severely exhausted. The door is open for a test of the monthly lower band or even a weekly-level correction. IWM is already in an extreme high-risk zone beyond the 2026 annual expected range. Without a strong higher low, bears will likely maintain control.

Sector Rotation: Funds Haven't Left, They're Rotating

Despite the semiconductor sector's clear pressure, the market didn't see a complete crash. Funds show a distinct rotation pattern. Communication services (Google, Meta), healthcare, consumer defensive stocks—non-AI core sectors—mostly closed green. Energy performed best due to the oil price surge, at one point rising over 2.6%. While hardware was being shaken out, software stocks rose against the trend. This rapid sector rotation makes the index look stable but sharply increases the difficulty of individual stock operations—chasing and cutting losses easily leads to losses on both sides.

Noteworthy Individual Stock Moves

Tesla continued to weaken today, at one point falling over 4%. If it breaks below the key psychological level of $400, technically there's risk of a further decline to the $390 support level. However, it will likely perform well before earnings. In July, keep an eye on short covering before MAGS earnings.

Boeing was relatively resilient, activating its fourth final assembly line for the 737 Max to boost production. It's currently forming a higher-low structure, with potential to challenge the $243 to $255 resistance zone in the coming weeks.

Energy stocks rallied across the board on the oil surge, with Occidental Petroleum jumping nearly 6%, and Chevron and ExxonMobil also showing strength.

The XLV ETF (Healthcare) rose 1.54% today.

Nvidia Became a Safe Haven for Funds

In the bloodbath of the semiconductor sector, Nvidia stood alone. It opened lower and fell over 2% intraday but quickly reversed to positive, forming a stark contrast to the overall sector. It was used by funds as a tool to hedge systemic sector risk.

Memory chip stocks also showed notable resilience. Institutions are clearly buying the dip on names like SanDisk, Micron, the Korean ETF, and AMD. SanDisk, despite a 36% pullback from its highs, quickly bounced after touching the 55-day moving average, forming nearly a long-legged doji on the daily chart, firmly holding the 61.8% Fibonacci retracement level, indicating solid buying support at the bottom.

The three semiconductor equipment giants (Applied Materials, KLA, and ASML) all precisely retraced to their respective 55-day moving averages. For funds that missed the previous rally, this is exactly the entry point that technicians have been waiting for. It's highly probable that all of them will see a bounce.

Short-term, the technical breakdown of the semiconductor index will be painful, but the fundamentals haven't been falsified (I'll analyze with a weekly chart, Figure 2).

The weekly SOXX shows a bearish engulfing candle. Since there are many semiconductor company earnings in July and August, and earnings are likely to be good, expectations are being lowered before earnings, so they should rise after. I think the weekly SOXX will form a bear flag and consolidate upward until sometime in August, then due to midterm election seasonality, it will decline again to September/October. The final target trend line for SOXX in September/October is around 500, roughly the weekly EMA 20. I'm not trying to create panic; the chart pattern is what it is. I need to give you corresponding strategies. My strategy is to add positions in batches: for example, add SK Hynix at EMA 50 and EMA 100, and eventually switch all memory holdings to SK Hynix. For optics, add Lite, Cohr, and NOK at EMA 200 and SMA 200, and GLW at EMA 100. But also consider sub-sectors like optics; the mid-level correction for optics may end in mid-to-late July, followed by the optics earnings season. When expectations are this low, earnings often perform well. Just a slight beat will cause shorts to be routed by longs.

Summary

We've discussed a lot today about geopolitics, earnings analysis, and technical patterns. Ultimately, it all comes down to answering one question: can you still hold now? My answer hasn't changed. The short-term is bearish, accounts are suffering—these are facts. But these two facts cannot overturn the larger fact that the AI industrial cycle is not over.

If you believe in a multi-year industrial trend, then the past two days of volatility are essentially just a small noise on that curve. What truly determines whether you make money in the end is not whether you precisely dodged every pullback, but whether you were shaken out by short-term panic before the direction truly turned bad. As long as we're still in a bull market, knowing this will keep you from panicking too much.

Hold patiently, and let time do the compounding.