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

692followers

Feed

Risk-free guaranteed profit? After the U.S. midterm elections, has the U.S. stock market really never fallen in the following year? An extremely exaggerated historical statistic: from 1950 to 2022, the U.S. experienced 19 midterm elections in total, and the S&P 500 index achieved positive returns in the 12 months following each election, with a 100% success rate and an average increase of about 15%. $SPY $QQQ $SOXX This is not speculation but a pattern statistically derived by Capital Group based on historical data. The statistical scope is from the midterm election day to one year later, comparing the changes in the S&P 500 price index, excluding dividend reinvestment. All 19 times showed gains, even including years after economic recessions, inflation crises, and U.S. stock bear markets. Some typical examples: In 2002, the U.S. had just experienced the burst of the internet bubble, and the S&P 500 fell more than 23% for the year, but it rose about 15% in the year following the midterm election. In 2018, the Federal Reserve continued raising interest rates, and U.S.-China trade tensions escalated; the S&P 500 fell 6.2% that year, but from the midterm election to November 2019, the index still rose over 11%. In 2022, the Federal Reserve aggressively raised rates, the U.S. stock market entered a bear market, and the S&P 500 fell about 19.4% for the year, but it rose about 14%–15% in the 12 months after the midterm election. Why does this pattern form? One possible explanation is the fading of political uncertainty. Before the midterm elections, the market needs to constantly reassess tax, fiscal spending, regulatory, and trade policies. After the election results are confirmed, some policy risks become easier to price. If Congress forms a divided government, major legislative advances may be constrained, so the market does not have to continuously pay a risk premium for drastic policy changes. Another reason is that midterm election years often come with greater volatility. According to historical statistics, since 1950, the average maximum drawdown of the S&P 500 within midterm election years reached about 17.5%. In some years, the market had already undergone significant adjustments before the election and then rebounded as earnings and risk appetite recovered. But there is an issue that cannot be ignored: all 19 past instances of gains do not guarantee continued gains after 2026. Expanding the timeframe to 1930, the 12 months following the 1930 and 1938 midterm elections both saw negative returns. Moreover, even if the final year ended with gains, there could have been very sharp declines in between. The next U.S. midterm election will be on November 3, 2026. The current U.S. stock market faces not only political uncertainty but also high interest rates, energy prices, AI capital expenditures, and tech stock valuation issues. It is reasonable to maintain a relatively positive outlook for the coming year, but one should not blindly go all-in just because of a 100% historical success rate in the short term. If inflationary pressures ease after the election, U.S. Treasury yields fall, and corporate earnings continue to grow—especially if earnings expectations for semiconductor companies like $NVDA, $MU, and $AVGO remain stable—then the historical pattern may be validated again. Conversely, if credit risk expands and earnings expectations are revised downward, even if political uncertainty fades, it will not prevent the market from re-pricing valuations lower. The fact that all 19 instances since 1950 showed gains is indeed noteworthy. But this statistic is more suitable to help us understand the market environment for the coming year rather than to predict daily ups and downs after the election. What truly determines investment returns remains the purchase price, corporate earnings, and position management.

Semiconductors plunged sharply at the market open during the day, then suddenly staged a V-shaped rebound intraday, and after the market closed on October 20 Eastern Time, they collectively surged again. $QQQ rose about 0.7% after hours, with some semiconductor-related stocks up about 3%. The strength of this rebound has clearly exceeded the scope of a typical technical recovery. $NVDA $MU $SNDK $AVGO $SOXX The direct catalyst came from $SMCI. The company suddenly raised its revenue forecast for Q1 of fiscal 2027 from $14.5–15.5 billion to $18.5–19.5 billion, with a new midpoint guidance of $19 billion, which is 25% higher than the market expectation of about $15.2 billion. More importantly, the gross margin forecast was also raised from 10.4%–10.8% to 12%–12.5%. After the announcement, $SMCI surged about 12% in after-hours trading. Why is this earnings forecast important? Because in the past two weeks, semiconductors have been continuously plunging, and the market has been repeatedly trading on concerns about excessive AI capital expenditure, rising financing costs for data centers, and potential future order slowdowns. But now, a core AI server supplier suddenly tells the market that not only is order demand stronger than expected, but the product mix and gross margin are also improving. The previously overly pessimistic expectations are now being challenged by real operational data. There are also changes at the macro level. During the session on the 20th, crude oil noticeably fell back, news of diplomatic contacts between the US and Iran eased some geopolitical concerns, and US Treasury yields also declined. However, there is currently no evidence that the two sides have formally reached a ceasefire agreement, so the entire after-hours rally cannot be attributed to the end of the war. However, $SMCI's revenue growth does not mean that the financing risks across the entire AI industry chain have disappeared. Strong server orders indicate that demand still exists, but whether customers can continue to pay, whether data centers can deliver on time, and whether computing power investments will ultimately generate sufficient returns still need to be verified. I had previously tended to interpret the rallies on the 19th and 20th as oversold rebounds rather than trend reversals. Now it must be acknowledged that this after-hours surge indeed increases the possibility of a reversal, especially as semiconductors begin to outperform the broader market again. But after-hours trading liquidity is limited and not enough to confirm the end of the entire downtrend. Going forward, if $SOXX can continue to break through previous resistance, and $NVDA, $MU, and $AVGO form a continuous upward trend again, while tech stock earnings expectations remain stable, then the previous cautious judgment will need to be adjusted. There is currently no reason to conclude that the AI bubble has burst just because of consecutive sharp declines, nor is there a need to immediately go all-in chasing the sudden after-hours surge. The truly valuable signal is whether this rise can transition from an after-hours sentiment reaction into sustained capital support during regular trading hours.

Global DRAM and HBM supply is already tight, and Micron's Taiwan factory is preparing for a strike. The timing is quite sensitive; $MU has just experienced a significant pullback since October, and the market is still debating whether the memory industry is entering a new round of adjustment. Now, new uncertainties have emerged on the supply side. $MU $SNDK $NVDA $TSM According to a Reuters report on October 20, the union at Micron's Taoyuan factory in Taiwan plans to hold a one-day strike on October 22 due to employee dissatisfaction with the bonus and profit-sharing system. On the same day, Micron will also conduct labor-management mediation with another union at the Taichung factory. Micron employs about 15,000 people in Taiwan, with cumulative investments exceeding $50 billion. Taiwan is an important global base for its DRAM and HBM manufacturing, with the Taichung plant being especially critical, responsible for producing advanced memory products. It is important to distinguish that the Taoyuan union has announced the strike plan, but the Taichung union is still negotiating and has not decided whether to strike. There is currently no evidence that Micron's HBM production line has stopped, nor can we directly infer how much future output might decline. Why is this matter worth attention? Because the memory industry differs from ordinary chip design companies; DRAM and HBM production heavily rely on complex manufacturing processes and continuous, stable production line operation. Especially for advanced HBM, which not only requires DRAM wafers but also involves stacking, packaging, testing, and customer certification. Delays in any part can affect final delivery. However, a short-term strike does not necessarily lead to wafer scrapping or production line interruption. Semiconductor factories are highly automated, and companies may maintain critical equipment operation through shift arrangements and emergency personnel. The impact depends on the actual number of participants, involved processes, and duration. What deserves further study is the profit distribution issue in the memory industry. SK Hynix has previously agreed to allocate 10% of annual operating profit for profit-sharing bonuses, and Samsung's semiconductor division has established a special bonus mechanism linked to operating profit. Micron employees clearly hope to have a more explicit profit-sharing system amid the company's rapid profit growth. This indicates that after the memory industry's market conditions improve, besides shareholders and the company itself, employees also want to participate in profit growth distribution. But for investors, tight supply does not necessarily mean Micron's stock price will rise. If the strike scope is limited and has almost no impact on actual production, the market may quickly downplay the issue. If production continues to be disrupted, global DRAM supply may tighten further, supporting memory prices to some extent, but Micron itself may face delivery delays, additional costs, and customer relationship risks. Competitors Samsung and SK Hynix might gain relative advantages in certain product supply shortages. Additionally, $SNDK mainly focuses on NAND flash, which does not share the same DRAM and HBM supply risks as Micron, so labor disputes at Micron should not be directly interpreted as benefiting all memory stocks. The key focus next is the actual strike scope on October 22 and the mediation results at the Taichung factory. If labor and management reach an agreement quickly, the impact on the supply chain could be very limited. What is certain now is that AI storage demand remains strong, but manufacturing, labor costs, and capacity allocation issues on the supply side are becoming increasingly important. In the future, what determines memory companies' profitability is not just chip prices but also production efficiency, cost control, and order fulfillment capability.

China's three major tech giants suddenly placed chip orders exceeding $35 billion with NVIDIA, yet NVIDIA told customers it might only fulfill 5%–10% of the purchase demand. Interestingly, despite the exposure of such a massive potential order, $NVDA did not show a significant intraday rise today. $NVDA $BABA $TSM $AVGO According to an exclusive Reuters report on October 20, ByteDance, Alibaba, and Tencent collectively ordered over 3 million RTX PRO 5500 graphics cards. Based on a unit price of about $11,400 each, the total value exceeds $35 billion. Including other Chinese companies' purchase demands, the total order volume for this product in the Chinese market has surpassed 6 million units. However, NVIDIA expects to produce at most about 3 million units globally by 2027. Sources revealed that NVIDIA has informed some partners that the actual delivery volume may only be 5%–10% of the requested quantity. This situation reflects at least two issues. On one hand, China's AI computing power demand has not disappeared. Despite the U.S. continuing to restrict high-performance GPU exports, large Chinese internet companies are still willing to purchase NVIDIA products, even accepting data center alternatives with significantly weaker performance than the H200 and B300. The RTX PRO 5500 originally belongs to professional workstation graphics cards and is not a flagship accelerator specifically designed for large-scale AI data centers. But under high-end chip procurement restrictions, Chinese companies have started using it for AI inference. Compared to training massive models, inference focuses more on actual operating costs, deployment flexibility, and computing power utilization. This also explains why some Chinese companies showed limited interest in the RTX 6000D last year but are willing to purchase large quantities of similar-performance alternatives this year. High-end chip restrictions remain, but the computing power gap continues to widen. On the other hand, NVIDIA is already facing obvious capacity allocation issues. In 2027, the company needs to allocate more of TSMC's advanced process resources to the Rubin platform rather than indefinitely expanding RTX PRO 5500 production. For $TSM, this means the capacity competition between high-end GPUs and other chips may continue to intensify. However, $35 billion does not mean NVIDIA has locked in $35 billion in revenue. Orders have not been fully delivered, and final sales depend on actual capacity, contract execution, and U.S. export policies. The related purchase scale currently mainly comes from insiders cited by Reuters, not official announcements from the three companies. This may also be one reason why $NVDA's stock price did not rise significantly due to the news today. Wall Street is not lacking evidence of strong NVIDIA demand; the market now cares more about how much of this demand can convert into new revenue and how much profit growth the company can sustain. NVIDIA's current advantages remain obvious, but even the strongest demand must go through capacity and delivery to turn into profit. The potential $35 billion order is astonishing, but the actual amount that enters the financial report ultimately determines how much return shareholders receive.