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

299Following

666followers

Feed

$AVGO's single-quarter AI semiconductor revenue reached $16.7 billion, a year-over-year increase of 221%, yet CEO Chen Fuyang publicly warned on October 12 that the biggest constraint on future AI development may no longer be wafers or memory, but capital. He even compared some AI investors' behavior to gambling in a casino. $NVDA $MU $SOXX $QQQ These remarks deserve serious study. Broadcom is one of the most direct beneficiaries of AI infrastructure expansion, with custom ASICs and high-speed network switching chips benefiting from the massive capital expenditures of cloud providers. The company's latest quarterly revenue was $29.6 billion, with free cash flow of $13.7 billion. A company earning substantial profits from AI whose CEO begins to emphasize capital constraints indicates that there is a need to reassess the balance between strong industry demand and sustainable investment returns. In the past two years, the market mainly worried about insufficient GPUs, limited HBM capacity, advanced packaging constraints, and power shortages in data centers. As long as these bottlenecks exist, upstream suppliers can expect high growth. But if the biggest future limitation becomes funding, the situation changes. Building a large AI data center requires massive upfront capital to purchase GPUs, servers, network equipment, and to cover power and construction costs. Hardware vendors recognize revenue upon equipment delivery, but data center operators need to lease computing power continuously over several years to gradually recoup their investments. The problem is that while AI computing demand is growing rapidly, the long-term economic returns on computing power investments have not yet been fully validated. Especially in an environment where U.S. Treasury yields are high, increased financing costs directly compress project returns. If future computing power rental prices fall, equipment depreciates faster than expected, or customer revenue growth lags behind investment, some projects may need to adjust their construction schedules. Recent credit stress in the $CRWV bond market already indicates that investors are beginning to reassess this business model. However, this does not mean that cash-rich companies like Microsoft and Google face the same financing risks. I believe this is also an important background factor for the semiconductor sector's continued decline in October. Currently, there is no sufficient evidence that AI orders have fully deteriorated, but the valuation expansion following September's rise is being reexamined. Existing bulls continue to reduce risk exposure, and new funds lack sufficient willingness to support, so even with strong earnings from TSMC and ASML, the sector cannot stabilize. This does not mean the long-term AI bull market is over. Broadcom's latest financial report also proves that AI infrastructure demand is still growing rapidly. The question is whether future capital expenditures can maintain the current pace and whether these investments can translate into sufficient revenue and free cash flow. Next, I will focus on the earnings reports of Microsoft, Google, Meta, and Amazon, especially their capital expenditure guidance, cloud business growth, AI commercialization revenue, and free cash flow. If these companies can prove that new investments are generating reasonable returns, market concerns about AI capital expenditures may gradually ease. But if investments continue to grow significantly without synchronized cash returns, even if semiconductors rebound noticeably, they may again face valuation compression. I remain optimistic about AI's long-term development and tend to believe the current phase is a deep pullback within a bull market. However, in the short term, I will continue to manage positions with oscillating adjustments. For rebounds lacking sustained buying support, I prefer to reduce excessive risk exposure rather than chase gains directly. Future competition in the AI industry will not only depend on who can buy more GPUs but also on who can build infrastructure at lower capital costs and truly profit from these investments. This is the issue I believe the market needs to reprice in the next phase.

Semiconductor equipment stocks have fallen into a historically rare market condition, with some companies experiencing declines even more severe than the chip leaders. $LRCX has dropped for 10 consecutive days, with a cumulative pullback of about 22%; $KLAC also fell for 10 straight days, down about 17% cumulatively; $AMAT has declined for 9 consecutive days, with a total pullback of about 22%. These three companies almost monopolize the most important equipment technologies for advanced semiconductor manufacturing worldwide. $SOXX $TSM $ASML Here’s the question: TSMC just reported a net profit increase of nearly 69% year-over-year, ASML’s Q3 revenue reached €11.66 billion, and they even mentioned that some customers have started discussing equipment demand for 2029 and 2030. Long-term demand for advanced processes, HBM, and AI chips remains strong, so why are equipment stocks being sold off so fiercely? I believe the market is currently reassessing the capital expenditure growth rate beyond 2027. The business models of semiconductor equipment companies differ from those of chip design companies. $LRCX depends on demand for etching, thin film deposition, and cleaning equipment; $AMAT covers materials engineering and multiple key manufacturing steps; $KLAC controls advanced process inspection, metrology, and process control technologies. These companies have deep technological barriers, but their orders are still influenced by the capital expenditure cycles of wafer fabs. The market’s concern is that if AI data center financing costs continue to rise and some cloud computing companies slow their expansion, will upstream wafer fabs still need to increase equipment purchases at the previous pace? Even if orders remain strong in 2026, stocks may already be pricing in a slowdown in growth for 2027. There is another easily overlooked issue. These three companies have already experienced very significant gains. Even with the recent roughly 20% plunge, they still have considerable gains year-to-date. This suggests the market may be digesting previously overly optimistic earnings expectations rather than suddenly discovering these companies lack competitiveness. High interest rates, oil prices, AI financing pressure, and capital withdrawal after the short squeeze in September are all accelerating this process. However, I don’t think it’s necessary to be overly pessimistic about the entire equipment industry now. TSMC maintains its capital expenditure plan of $60 billion to $64 billion for 2026, and ASML even expects that planned capacity for 2027 and 2028 may not meet customer demand. This information at least indicates that long-term investment demand for advanced manufacturing has not collapsed comprehensively. Moreover, the competitive barriers of $KLAC, $LRCX, and $AMAT will not disappear just because of ten consecutive days of decline. If stock prices continue to fall but company orders, earnings expectations, and cash flow do not deteriorate significantly, I would seriously study the long-term valuations of these companies. That said, I remain cautious in the short term. After such extreme consecutive declines, equipment stocks could indeed see a noticeable technical rebound. But as long as U.S. Treasury yields, AI financing conditions, and the overall $SOXX downward trend do not improve, I tend to use rebounds lacking sustained buying support to reduce overly high positions. For those with ample cash, this adjustment is creating opportunities worth researching, but there is no need to go all-in at once. I remain optimistic about the long-term development of semiconductors and do not believe the bull market has definitively ended. The market is currently repricing growth expectations for the entire AI industry chain over the next few years. If advanced process demand continues to materialize and equipment stock valuations further return to reasonable ranges, this plunge may not be a bad thing for long-term investors. But two things must be distinguished: the technological barriers have not disappeared, but that does not mean previous stock prices were necessarily reasonable. I will continue to watch $KLAC, $LRCX, and $AMAT. The truly worthwhile entry point is when valuations fully reflect risks while the companies’ long-term competitive advantages remain intact.

$SOXX has fallen for nine consecutive days, and the semiconductor sector has been hit to a historically rare extent. But I still believe the long-term bull market for AI is not over, and there are still opportunities for a new major rally in the future. Look at March and July of this year, and review similar adjustments over the past few years, and you'll know there's no need to completely lose confidence now. $QQQ $SPY $NVDA $MU Let's first look at March this year. On March 30, $SOXX closed at $309.79, and the market was also in a highly pessimistic phase. By June 30, $SOXX had already risen to $640.76. In just three months, it increased by about 107%. Now look at July this year. On June 30, $SOXX was still around $640, but by July 29 it had dropped to $465, a pullback of over 27%. At that time, many investors were also worried about AI valuations, capital expenditures, and the sustainability of future growth. From July 29 to October 5, $SOXX rebounded from $465 to $589.51, an increase of about 26.8%. Although it did not fully recover the late June high, it proved that even after nearly a 30% crash, semiconductors still have considerable price recovery ability. Let's extend the timeline. In the first week of September 2024, $SOXX plunged 11.77%, but the following week it rebounded 9.28%. In September and October 2023, U.S. Treasury yields continued to rise, with the 10-year Treasury briefly hitting 5%, and semiconductors adjusted for two consecutive months. Then in November, $SOXX rose 16.09%, and continued to rise 12.96% in December. These market moves share a common feature: during adjustments, macro pressures and valuation concerns are amplified, but as long as corporate earnings continue to grow and capital flows back, stock prices have the chance to recover quickly. Back to October 2026. I have previously emphasized that the current situation is more like after the September rally ended, existing AI bulls are continuously exiting, incremental buying is insufficient, combined with valuation repricing triggered by U.S. Treasury yields, oil prices, and AI financing risks. But this does not mean the long-term growth logic of the entire AI industry has been destroyed. TSMC's net profit in Q3 grew nearly 69% year-over-year, and demand for advanced processes remains strong. Companies like $ASML, $NVDA, and $AVGO still hold extremely important technologies and customer resources in the industry chain. I believe AI computing, advanced processes, storage, and high-speed interconnect will still have long-term growth potential. This is why I have not turned completely bearish due to the October crash. Of course, I do not think this time will necessarily replicate the rebounds of March and July. Currently, U.S. Treasury yields remain high, oil prices increase inflationary pressure, the November midterm elections are approaching, and AI infrastructure financing issues have not been fully resolved. Especially this time lacks the obvious short-covering base seen in July, so capital inflows may take longer. Therefore, my judgment remains divided into two stages. In the short term, after consecutive extreme declines, a significant rebound is worth expecting, but the market may continue to adjust after the rebound. I tend to use rebounds lacking sustained buying support to reduce positions. In the long term, as long as AI demand continues to grow, corporate earnings keep materializing, and the credit environment does not deteriorate systemically, I remain optimistic that semiconductors will stage a strong rally again. From March to July, the market has already proven multiple times this year that sharp short-term declines do not necessarily mean the end of the long-term trend. But to participate in the next rally, the premise is not to be forced out due to leverage and position loss of control during this adjustment. For those holding quality stocks without leverage pressure, there is no need to be completely despairing just because of nine consecutive days of decline. Market recovery may take weeks or even months. Some stocks may not return to previous highs, but companies with true moats and sustainable profitability are still worth long-term study. I cannot guarantee when the next big bull market will come, but I will not deny the AI industry's opportunities in the coming years just because of the October pullback. The history of March and July this year is worth remembering: the market can fall very hard in a short time and can also quickly resume rising once expectations improve. Control risk in the short term, wait for the market to stabilize in the mid term, and continue to focus on quality companies in the long term. I still believe the bull market has a chance to return, but there is no need to bet full positions and leverage that it will come back tomorrow.

$SOXX has fallen for nine consecutive sessions, marking the longest losing streak since 2006. What’s most worth studying now is what historically happened after similar concentrated sell-offs of this magnitude in semiconductors. I found five market instances with valuable references. The results are clear: semiconductors often experience significant rebounds after extreme declines, but the timing and sustainability of these rebounds vary noticeably. $QQQ $SPY $NVDA $MU First, September 2024, the closest to this time. As of the week ending September 6: $SOXX dropped 11.77% $QQQ dropped 5.79% $SPY dropped 4.14% At that time, the market was worried about a slowdown in U.S. economic growth and overvaluation of tech stocks, leading to a concentrated sell-off in semiconductors. The following week: $SOXX rebounded 9.28% $QQQ rose 5.94% $SPY rose 4.01% Afterward, semiconductors continued to rise for two consecutive weeks by about 0.37% and 4.03%. This shows that even after a weekly plunge exceeding 10%, as long as macro expectations improve and selling pressure eases, the sector can quickly recover most of its losses. Second, August 2024, with more volatile adjustments. In the week of August 2, $SOXX fell 9.84%. It then rose 2.40% and 9.18% in the following two weeks respectively. Interestingly, although semiconductors experienced a clear rebound, they plunged again by 11.77% in the first week of September. This is the situation I have been emphasizing: a large rebound does not mean the correction is over. If one sees two consecutive weeks of gains and assumes all risks are gone, a new round of declines may still follow. Third, October 2023, with a macro environment worth referencing. At that time, the U.S. 10-year Treasury yield briefly touched 5%, and the market worried about the Fed maintaining high interest rates for a long time, putting continuous pressure on tech stock valuations. $SOXX fell 7.54% in September and continued to drop about 6.8% in October. But by November, as Treasury yields fell and market expectations for monetary policy improved, semiconductors showed significant recovery. $SOXX rose 16.09% in November and continued up 12.96% in December. Two consecutive months of correction did not end the subsequent uptrend. This history is very relevant now. High interest rates can continuously suppress valuations, but once rate expectations change, growth stock valuations can recover very quickly. Fourth, June 2022, a cautionary negative example. At that time, the Fed rapidly raised rates, continuously compressing tech stock valuations. $SOXX plunged 18.25% in June but rebounded 18.48% in July. Looking only at July’s movement, one could mistakenly think the market had bottomed. However, it fell again 8.59% in August and 11.05% in September. By November, semiconductors rose 16.51% again. The whole process is a typical downtrend: sharp drop, big rebound, another drop, followed by a strong recovery. So a strong rebound alone does not directly prove the market has entered a new uptrend. Fifth, March to June 2025. In March 2025, $SOXX fell 10.92%, continuing down 1.93% in April. This time, the market did not immediately recover after the first big drop month but continued to digest macro and policy risks. Until May, $SOXX rose 10.18%; in June, it further increased 17.03%. This indicates that corrections may last weeks or even months, but as long as the macro environment improves and earnings expectations stabilize, capital can still return to the semiconductor sector. Back to October 2026. This market movement is structurally similar in decline to September 2024, but I believe current macro pressures are more complex. The U.S. 10-year Treasury yield remains around 5%, oil prices are high, the November midterm elections are approaching, and AI infrastructure financing risks are being reassessed by the market. Meanwhile, valuation normalization after the September rise is still ongoing. From the market perspective, I tend to think the current situation mainly reflects existing AI longs exiting rather than the largest-scale concentrated short attack. Insufficient incremental buying has caused semiconductors to break down continuously. But this situation will not last forever. History from August and September 2024 and October 2023 shows that semiconductors can produce considerable rebounds after extreme corrections. Therefore, I remain optimistic about a clear technical rebound ahead and do not believe the long-term bull market has ended yet. However, I now prefer to reference August 2024 and 2022’s downside. After a rebound, another decline is entirely possible. If Treasury yields stay high, oil prices do not fall significantly, and AI credit market pressures persist, short-term oversold recoveries will not be enough to reverse the mid-term correction trend. Next, I will watch whether $SOXX can stop making new lows, break through the previous rebound high, and form higher lows on pullbacks. At the same time, $QQQ and $SPY need to stabilize, and semiconductor market breadth must improve. If these conditions do not appear, even a short-term rebound of 5% or 10% will still be considered a bounce within a correction. For those previously heavily invested, I prefer to use such unsupported rallies to reduce risk exposure. For those with ample cash, there is no need to rush to fully invest on the first rebound. History gives me a clear conclusion: After a 10% plunge in semiconductors, the next phase can quickly rebound close to 10%, or it may first consolidate, continue to probe lower, and then take a month or longer to recover. The current panic is not unprecedented, and a clear rebound ahead does not require a miracle. But rebounds and trend reversals must be judged separately. No need to be overly pessimistic short-term; mid-term still requires position control. As long as Treasury yields, oil prices, and capital structure do not improve substantially, I will not change my view of a choppy downtrend because of a few big bullish candles.

$SOXX has fallen for nine consecutive days, with semiconductors continuously breaking support levels, and retail investor sentiment is very poor. However, from the perspective of the options market, I believe there is no need to completely lose confidence at this point. Although the current trading structure is fragile, it also has the conditions to amplify a rebound. $QQQ $SPY $NVDA $MU As of the close on October 16, $SOXX was at $493.39, with a Put/Call open interest ratio of about 1.86, indicating that put option holdings are significantly higher than call options. At the same time, third-party models estimate that the Gamma exposure has already entered negative territory. This set of data deserves serious study. A high put holding indicates that the market has many bearish or downside protection-related positions. However, this may include a large number of institutional insurance purchases on existing stocks and should not be directly interpreted as everyone shorting. Negative Gamma implies another type of risk. If market makers are overall in a negative Gamma state, they may need to continue selling the underlying assets to hedge Delta during declines, further amplifying the drop. The lower the price falls, the more apparent the selling pressure from some hedging trades may become. But this mechanism works both ways. Once the market starts to rise, the same hedging structure may force market makers to buy the underlying assets, further amplifying the rebound. So don’t think that after nine consecutive days of decline, the market only has the option to continue falling. The extreme moves in the past few days may include both the exit of existing long positions and the amplification of volatility caused by options hedging trades. The speed of price decline does not necessarily fully correspond to the deterioration of corporate fundamentals. And October 16 happens to be the monthly options expiration date. As some options expire, the original Delta and Gamma risk exposures will change, and market makers will need to readjust their hedging positions. The new options position structure will gradually form next week, and short-term price behavior may change accordingly. Of course, options expiration does not mean selling pressure will necessarily disappear, nor does it guarantee an immediate market rebound. But at least for now, the already formed speed of decline cannot be directly extrapolated to every future trading day. I still have expectations for a subsequent technical rebound. Nine consecutive days of selling have put semiconductors in an extremely weak state. As long as the speed of existing capital outflows slows, new buying begins to absorb, combined with some put position closures and hedging adjustments, it is entirely possible to form a considerable rebound. However, I also do not believe there is enough evidence yet to prove that the market is about to experience a large-scale short squeeze like in July. What is truly worth observing next is whether put protection demand begins to decline, whether implied volatility falls, and whether the rise can be continuously supported by spot trading volume and market breadth. If these indicators improve simultaneously, the sustainability of the rebound will be more promising. My previous judgment has not changed. This October adjustment looks more like a continued exit of existing AI longs, combined with a re-pricing after prior valuation expansion. U.S. Treasury bonds, oil prices, and AI financing pressures have not yet been resolved, so even if a significant rebound occurs, a secondary pullback still needs to be guarded against. Therefore, there is no need to be overly pessimistic now, but also no reason to use high leverage and short-term options to bet on a rebound next week. The options market can amplify short-term rallies but cannot guarantee any stock will return to its original price. For those already holding quality underlying stocks, as long as fundamentals have not clearly deteriorated and positions are within tolerable ranges, there is no need to hastily abandon long-term judgments just because of several days of panic selling. Those with heavy positions should prioritize risk control. If a strong rebound lacking sustained buying support occurs later, they can still use the rise to reduce exposure. I still tend to believe the bull market is not over, but the adjustment may need more time to digest. From the options structure perspective, the current high volatility is not only released downward. After selling pressure weakens, price recovery may also be rapid. The most important thing now is to retain enough capital and patience. There is no need to be completely despairing after nine consecutive days of decline, nor to bet all positions to prove you can catch the bottom. It is uncertain when the rebound will happen, but the volatility risk already accumulated in the market warrants considering both upward and downward directions simultaneously.

$SOXX has fallen for 9 consecutive trading days, and many people have lost confidence in this round of the market. However, from the perspective of candlesticks and market structure, I believe there is no need to be overly pessimistic right now. The short-term downtrend is indeed very obvious, but continuous declines do not mean it can only keep falling in the future. $QQQ $SPY $NVDA $MU Let's first look at the candlesticks of this decline. On October 6, $SOXX closed at $589.45, and by October 16 it had dropped to $493.39, a cumulative decline of about 16.3% over nine trading days. In just two weeks, semiconductors have hardly seen any sustained rebound, with highs and lows continuously moving lower. Especially on Friday, the intraday low touched $492.83, and it finally closed at $493.39, almost at the day's lowest point. This closing structure indicates that sellers maintained a clear advantage until the end of trading. There is not enough basis to declare a bottom now. But there is a detail worth noting. On October 15, $SOXX's intraday low was $495.78, then it rebounded to $517.12 at one point, finally closing at $514.71. Although it still fell compared to the previous trading day, the intraday candlestick shows a clear low-level support, and the closing price was higher than the opening price. This at least indicates that the market is not completely without buying interest. However, Friday's new round of selling broke the previous rebound structure again, showing that buyers are not yet strong enough to reverse the trend. I think everyone should understand one thing now: the downtrend can continue, but the speed of decline will not always remain the same. After consecutive large bearish candles, the market may first experience volatility contraction and sideways movement at low levels, or it may directly have a quick rebound. Some short sellers taking profits, weakening selling pressure, and funds rebuilding positions can all drive price recovery. This is why I still expect a strong rebound to appear later. From a technical perspective, I will focus on whether $SOXX can stop making new lows and climb back to the $510–$520 range, and whether the pullback after the rebound can form a higher low. If it stops making new lows for several consecutive trading days and then breaks through the previous short-term rebound high, with more semiconductor component stocks strengthening simultaneously, the short-term market structure may begin to improve. $QQQ and $SPY are also very important. If these two major indices can stabilize simultaneously, the foundation for semiconductor rebounds will be more solid. Of course, even if these signals appear, I will not immediately conclude that the mid-term correction is over. US Treasury yields, oil prices, and AI financing pressures still exist, and the market could continue to oscillate downward after a technical rebound. Therefore, I still recommend that investors with heavy positions use rebounds lacking sustained buying support to appropriately reduce risk, rather than rushing to leverage up to recover losses. But I also want to remind everyone not to assume that after nine consecutive bearish candles, there will never be bullish candles again. The current candlesticks do look bad, and the market has not confirmed a bottom. But short-term price structure deterioration and the end of the entire bull market are two completely different judgments. I still tend to believe that this is a deep correction within a bull market. As long as the long-term profitability logic of companies is not fundamentally broken, the market still has room for revaluation and recovery in the future. You don't need to guess every day whether there will be a 10% rebound tomorrow, nor should you completely abandon your long-term investment plan because of one or two weeks of decline. If you have leverage, first control risk; if you hold quality stocks, re-examine the fundamentals; if you have cash, be patient and wait. Candlesticks reflect the current strength of buyers and sellers, not the final judgment of a company's value over the next ten years. The current advantage of sellers does not mean buyers will never return. The market has not given a bottom signal yet, but there is also no reason to believe this round of decline will never end. Be patient, keep your position within what you can bear. Wait for the market to truly stop falling and for trend repair, then reassess opportunities.

In October, the US stock market showed a set of very exaggerated data: $LLY rose 9.32%, $BRK.B rose 6.30%, $AAPL rose 3.98%, but $MU plummeted 19.92%, and $SKHY fell even more by 29.46%. In just half a month, for companies all at the trillion-dollar level, the stock price performance differed by nearly 39 percentage points. $SOXX $QQQ $NVDA I compiled data from October 1 to 16 for 15 US-listed companies that currently or at the beginning of October reached a trillion-dollar market value, and the results were even more polarized than expected. The 7 companies that rose: $LLY Eli Lilly: +9.32% $BRK.B Berkshire Hathaway: +6.30% $AAPL Apple: +3.98% $MSFT Microsoft: +3.17% $AMZN Amazon: +0.71% $GOOGL Google: +0.69% $TSLA Tesla: +0.17% The 8 companies that fell: $SPCX SpaceX: -1.38% $META Meta: -2.29% $AVGO Broadcom: -4.33% $NVDA Nvidia: -7.25% $TSM TSMC: -10.15% $AMD AMD: -14.13% $MU Micron: -19.92% $SKHY SK Hynix: -29.46% There is a very obvious phenomenon here. Since October, the worst-performing large companies are almost all concentrated in the AI semiconductor industry chain. $SOXX fell from $576.33 to $493.39 during the same period, a cumulative drop of 14.39%. But at the same time, companies like Eli Lilly, Berkshire Hathaway, and Apple still rose. This indicates a very clear sector divergence in this round of adjustments, and it is not that the enterprise value of the entire US stock market is declining synchronously. I believe this is largely related to the market's reassessment of AI investment returns. In recent years, AI semiconductors have attracted a lot of capital. The market was willing to pay high valuations for GPU demand in the coming years, HBM capacity expansion, advanced process technology, and data center construction. But when US Treasury yields rise, AI financing costs increase, and the credit risk of some data center companies is questioned, these long-term growth expectations begin to be repriced. Especially for storage stocks that had previously surged. The declines of $MU and $SKHY significantly exceed those of $NVDA, $AVGO, and $TSM, indicating that the market prices risk differently for various semiconductor companies. The storage industry itself is clearly cyclical. Even with strong current demand for HBM and DRAM, investors worry about future supply expansion, price cycle reversals, and slowing profit growth. For companies like $NVDA and $AVGO, the market is also reexamining the sustainability of their customers' massive capital expenditures in the coming years. However, I do not think that these declines alone can confirm that the AI industry's fundamentals have fully deteriorated. Price drops can come from lower profit expectations, valuation compression, deleveraging, or institutions reducing risk exposure. Different causes correspond to completely different subsequent investment opportunities. Meanwhile, the relative strength of Eli Lilly, Berkshire Hathaway, and Apple also shows that investors have not sold off all large companies simultaneously. Recently, traditional sectors like consumer goods, healthcare, and finance have also shown clear relative strength. This aligns with the market's risk appetite shifting from the previously crowded AI trades to other sectors. But based on stock price rises and falls alone, we cannot directly conclude that funds selling semiconductors have all flowed into these companies. I have always been a long-term AI bull and still tend to believe this is a deep correction within a bull market. But I also repeatedly emphasize that being optimistic about long-term development does not mean ignoring the risks that have already occurred in the short term. $SOXX has continuously broken support levels, and storage stocks have cumulatively fallen nearly 20% to even 30%. Such a trend should not be simply responded to with "long-term optimism." If positions are too high or leverage too large, one should reduce holdings when necessary. Even if a significant rebound follows, it is necessary to observe whether the price structure, market breadth, and macro pressures truly improve. For investors with ample cash, it is worth seriously studying companies whose competitive advantages have not obviously deteriorated but whose valuations have already fallen significantly. But a 20% drop does not mean cheap, nor does it mean the bottom has been reached. This October large-cap stock performance list actually reveals a very noteworthy change: The market has not stopped pricing quality companies but is reconsidering which growth is worth continuing to pay a high premium for. I remain optimistic about AI's long-term development, but going forward, stock selection may be far more important than the past phase where simply buying AI concepts could make money. What truly deserves attention is, after this round of valuation compression, which companies can continue to deliver profits, generate free cash flow, and maintain their competitive advantages.

The U.S. midterm elections are just around the corner! I've discovered a very interesting historical pattern on Wall Street: in the past six midterm election years, the healthcare ETF $XLV has risen every time from late July to the end of November, with an average gain of 8.6%! With the U.S. midterms coming up on November 3 this year, I think it's necessary to revisit this sector that has long been overshadowed by AI, semiconductors, and storage. First, what exactly is $XLV? It is a U.S. healthcare industry ETF under State Street, mainly investing in large pharmaceutical, biotech, health insurance, medical device, and life sciences companies within the S&P 500. The fund size is about $43.5 billion, with a management fee of only 0.08%. The trailing twelve months (TTM) PE ratio of its tracked index is about 30.61x, the forward one-year PE is about 20.16x, and the expected EPS annual growth rate over the next 3–5 years is approximately 12.13%. Its core holdings include: $LLY Eli Lilly: 15.31%, focusing on weight loss drugs, diabetes, and innovative medicines. $JNJ Johnson & Johnson: 10.11%, innovative drugs and medical devices. $ABBV AbbVie: 7.87%, immunology and oncology drugs. $MRK Merck: 5.75%, oncology drugs and vaccines. $UNH UnitedHealth: 5.45%, health insurance and medical services. Additionally, there are healthcare giants like $TMO, $AMGN, $GILD, $ABT, and $PFE, with the top ten holdings accounting for about 60.4%. By industry distribution, pharmaceuticals make up 36.68%, biotechnology 19.83%, healthcare services 17.22%, medical devices 14.94%, and life sciences tools 10.63%. So $XLV is not simply an innovative drug ETF; it packages together the larger publicly traded companies across the entire U.S. healthcare industry chain. What really interests me is its historical performance around midterm elections. According to statistics from the six midterm elections in 2002, 2006, 2010, 2014, 2018, and 2022, $XLV achieved positive returns every time from July 25 to November 30 in the election years, with an average increase of 8.6%. For example, it rose 13.32% in 2002, and even during the 2022 U.S. stock bear market, it still gained 6.95% in that period. Of course, this is just a specific time window. Looking at the entire election year, $XLV actually declined twice in the six midterm elections. So we cannot take a 100% historical window win rate as a guarantee of gains in 2026. But why does the healthcare sector often perform well around midterms? I believe there are three main reasons. First, policy uncertainty. The U.S. healthcare industry heavily depends on Medicare, Medicaid, drug pricing, and government regulation. Before elections, the market needs to repeatedly assess policy risks. After the election, as the future Congress power structure becomes clearer, some policy uncertainties may fade, and the market may readjust healthcare company valuations. Second, defensive characteristics. When the economy is weak, tech companies can delay data center construction, and enterprises can reduce server purchases, but people's demand for essential medicines, health insurance, and medical services does not suddenly disappear. So when Wall Street worries about economic growth, interest rates, or tech stock valuations, the healthcare sector may receive some defensive allocation funds. Third, and what I think makes 2026 especially worth studying, is capital rotation. This year, AI, semiconductors, optical communications, and storage sectors have seen astonishing rallies, with stocks like $MU, $SNDK, and $LITE attracting large amounts of capital. As some tech stock valuations and positions become increasingly crowded, institutions may reconsider asset allocation by year-end, shifting some funds toward industries that previously lagged but have relatively stable earnings. However, this does not mean capital will necessarily flow entirely from AI to healthcare. True confirmation of rotation requires seeing $XLV consistently outperform $SPY, along with improving earnings expectations for healthcare companies, rather than assuming large-scale institutional buying based on a single day’s rise. Looking at real catalysts for 2026. $LLY’s weight loss and diabetes businesses remain highly watched growth areas; big pharma’s new drug R&D, oncology treatments, and medical device demand also provide profit sources independent of AI. Meanwhile, the health insurance sector is showing clear differentiation—for example, $HUM recently benefited from improved Medicare Advantage star ratings, significantly boosting its stock price. From a technical perspective, $XLV closed at $170.81 on October 9, up 1.58% in one day. I will focus on the breakout around $172–173, support in the $168–170 range, and whether there is more obvious buying near $165–166. But there are two risks that cannot be ignored. On one hand, $XLV’s forward PE of about 20x is not extremely cheap, especially with $LLY alone accounting for over 15% weight; competition in weight loss drugs and valuation changes could impact the entire ETF. On the other hand, U.S. drug price negotiations, Medicare payment policies, patent expirations, and medical cost reimbursements remain real risks; the election’s end does not mean these issues disappear. So my judgment is that $XLV is worth attention not just because it rose in the specific windows of the past six midterms, but because it combines healthcare demand resilience, policy expectation changes, independent profit drivers, and potential capital rotation logic. If in the coming weeks $XLV can break out of its recent consolidation on strong volume, continue to outperform $SPY, and the earnings outlook for healthcare leaders keeps improving, then this sector long overshadowed by AI’s spotlight deserves Wall Street’s renewed research. In recent years, AI has been the market’s biggest star, but the next phase of capital attention may not be limited to AI. The truly worthwhile opportunities are always those where fundamentals are improving but the market has not yet fully priced them in.