美股投资young

美股科技投资者 ai狂热者分享每日快讯. all in ai

300Following

694followers

Feed

Livestream

Feed

The current downturn in semiconductors is not over yet, but I still believe the AI industry chain will see a significant surge in the future. However, right now is far from a time to ignore risks. $SOXX $NVDA $MU $AVGO $TSM On October 22, $SOXX plummeted about 6.7% in a single day, falling from nearly $600 at the beginning of October to around $476, a pullback of nearly 20%. From October 19 to 21, the market tried to rebound but ultimately failed to hold. I have consistently emphasized that this rebound should not be taken as a reversal, so I chose to proactively reduce some of my stock positions while increasing Put protection. Currently, the short-term downtrend in semiconductors remains very evident. The rebound strength is insufficient, previous supports are continuously broken, and the sector’s decline is significantly greater than the Nasdaq. If $SOXX cannot firmly regain the $500–510 range soon, even if there are several consecutive days of gains afterward, it is more likely to be seen as a technical correction within a downtrend rather than the start of a new major upward wave. But a falling stock price does not mean AI demand has disappeared. $LRCX’s latest quarterly revenue reached $8.22 billion, a year-over-year increase of about 54%, with operating margins hitting a new high. The AI computing power, advanced processes, HBM, and storage demands behind $TSM, $NVDA, and $MU have not suddenly vanished due to this market sell-off. The market is now concerned about excessive AI capital expenditures, the speed of investment returns, and whether previously high valuations can be maintained. This adjustment essentially reflects the market’s re-pricing of the entire AI industry chain. Even if companies continue to grow, stock prices can still fall due to valuation compression. Fundamentals and stock prices never need to move in sync in the short term. My long-term view on semiconductors remains unchanged. Over the next few years, AI will gradually expand from model training to large-scale inference, enterprise applications, and intelligent agents, with continued growth potential in computing, storage, networking, and power demand. As long as these demands eventually translate into corporate profits, the next semiconductor rally could still be very strong. But before that, we must get through this adjustment. There is no shortage of companies worth holding long-term now; what is lacking is a sufficiently safe entry price. Keep cash on hand, don’t rush to bottom-fish, avoid excessive leverage, and gradually increase positions once the market stabilizes again. Being optimistic about AI’s development over the next few years does not mean you have to endure all the pullbacks until October 2026.

On October 23, when the South Korean stock market opens, it could be the toughest trading day for the entire Asian semiconductor and memory sectors. SK Hynix faces the risk of a single-day plunge of over 10%, and in extreme cases, the KOSPI could even see a drop exceeding 7% again. On October 22 Eastern Time, $SOXX plunged about 6.7% in a single day, with the entire U.S. semiconductor market almost indiscriminately sold off. $MU, $SNDK, $WDC, and $LRCX were all significantly impacted. The rebound attempts from October 19 to 21 were basically reversed, as the market lacked sufficient support. What’s more noteworthy is $WDC’s earnings report. Revenue reached $4.381 billion, a 55% year-over-year increase, with adjusted EPS at $4.64. The next quarter’s revenue guidance is around $5.1 billion. Such results, under normal market conditions, would be enough to support the market in reassessing the growth potential of the memory industry. But today, the selling pressure across the sector is so heavy that even a stock price rebound after the earnings release does not mean the market has stabilized. Now the pressure is about to transmit to Asia. On October 22, the South Korean market had already fallen 3.52%, with SK Hynix down 3.17%. After the Korean market closed, U.S. semiconductors plunged further, meaning that when the Korean market opens on October 23, it will need to digest the overnight U.S. stock risk again. Don’t think a 10% drop in SK Hynix or a 7% drop in KOSPI is impossible. On October 15, SK Hynix plunged 11.06% in a single day, and KOSPI fell 7.57%. When global capital focuses on reducing semiconductor risk exposure, even the core companies in the HBM industry cannot completely avoid sell-offs. This doesn’t mean October 23 will definitely repeat October 15, but the Korean market is indeed in a very dangerous position now. SK Hynix and Samsung have too much influence on KOSPI, and if foreign capital continues concentrated selling, the decline in the memory sector can easily evolve into passive index reduction and programmatic selling across the entire index. Long-term, I remain optimistic about SK Hynix, $MU, and the entire memory industry. The demand for HBM, DRAM, and enterprise-level storage from AI training, inference, and data centers has not disappeared just because stock prices have fallen. But in the short term, we must admit that the global semiconductor industry is still in a downward trend. No matter how good the earnings reports are, they may not stop capital outflows. What is most needed now is cash and patience, not leveraging to bet on a rebound the next day. Whether the South Korean market on October 23 can absorb the selling pressure from the U.S. stock market will be an important signal to judge whether this round of global semiconductor adjustment will continue to spread.

On October 22, the Nasdaq Composite Index closed down about 2.7%, and the Philadelphia Semiconductor Index plunged about 6.3%. $QQQ and $SOXX were once again heavily sold off, with tech stocks becoming the hardest hit sector in the market. From October 19 to 21, when the market tried to rebound, I repeatedly emphasized: this was a weak rebound, not a reversal. At that time, my strategy was clear: reduce some long stock positions while increasing Put protection. Looking back today, the timing of this risk control was indeed quite good. Why did I not recognize it as a reversal then? On October 16, $QQQ fell 2.5% with a trading volume close to 67 million shares. On the 19th it rose 1.21%, on the 20th it slightly fell 0.10%, and on the 21st it rose 1.05%, but the trading volumes for these three days were only about 37 million, 30.5 million, and 32.5 million shares respectively. The rebound lacked strength, volume did not show significant improvement, and the resistance around $730 was never effectively broken. The downtrend was far from being fully repaired. I have also consistently emphasized that the October decline may not have fully released risk yet. Earlier, long positions were overly crowded, semiconductor gains were too large, and both leverage and profit-taking needed to be digested. Whether new shorts have fully entered still requires actual position data to verify. But today's price action shows the market did not establish strong enough support during the previous rebound. So it is still not easy to declare a bottom now. If the market continues to break below previous lows, combined with AI financing pressure and controversies over capital expenditure returns from tech giants, a much larger sell-off is entirely possible. However, I will not immediately conclude that the U.S. stock market has entered a long-term bear market just because today's judgment was validated. Earnings from industry chain companies like $TSM and $LRCX remain strong, and next week’s earnings reports from $MSFT and $GOOG will further verify AI investment demand. The fundamental logic for a long-term bull market has not yet been completely overturned by this round of price declines. The recent reduction in positions and increase in Puts was not because of a long-term bearish view on U.S. stocks, but because the risk-reward ratio for the recent rise was no longer favorable. The same applies now. I won’t rush to go fully long just because of a one-day plunge, nor will I dismiss the entire AI industry due to short-term weakness. Since early October, I have repeatedly warned about pullback risks and reminded to reduce positions during the rebound; all these judgments were made before today’s plunge. Being bullish in the long term is one thing; being able to identify pullbacks early and protect principal is another. Going forward, I will continue to control risk and wait for the market to show real support before considering gradually increasing long stock positions.