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

What will happen to tech stocks after the U.S. midterm elections? A historical phenomenon possibly more worth watching than the S&P 500: Since its inception in 1999, $QQQ has experienced six midterm elections in 2002, 2006, 2010, 2014, 2018, and 2022. If we take the closing price on election day as the starting point and observe the performance over the following 12 months, all six times resulted in gains, a 100% historical win rate. $QQQ $NVDA $MSFT $GOOG $AMZN It is important to distinguish that the S&P 500 has had 19 midterm elections since 1950 with a full-year gain each time, but $QQQ was established in 1999, so data from the previous decades cannot be directly applied to it. Although there are only six samples, the results are still worth studying. Here are some typical examples: In 2002, the U.S. had just experienced the bursting of the internet bubble, and the Nasdaq 100 index fell about 37.6% that year. However, from the November midterm election to the same period in 2003, $QQQ rebounded over 30%, with tech stocks seeing a clear recovery. In 2018, the Federal Reserve continued raising interest rates, and U.S.-China trade tensions escalated. The S&P 500 and Nasdaq experienced sharp declines in December, but from the November 2018 midterm election to November 2019, $QQQ still gained about 18% cumulatively. In 2022, the Federal Reserve aggressively raised rates, and the Nasdaq 100 index fell about 33% for the year. Yet from the November 2022 midterm election to the same period in 2023, $QQQ rose over 30%. The AI investment boom, earnings recovery of tech giants, and market repricing of interest rate outlooks jointly drove that rebound. Why have tech stocks historically performed so strongly after midterm elections? One reason may be the reduction of political uncertainty. After the election, market expectations for future fiscal, tax, and regulatory policies become clearer, allowing investors to reassess corporate earnings and valuations. But more importantly, it may be the macroeconomic cycle. After 2002, tech stocks recovered from the bubble burst; in 2019, the Fed shifted to rate cuts; and in 2023, there was an AI investment boom. Each of these rallies had its own fundamental and liquidity background and cannot be simply attributed to the midterm elections. On November 3, 2026, the U.S. will hold midterm elections again. The current environment facing $QQQ is not exactly the same as the previous six times. AI infrastructure remains an important growth direction, with companies like $NVDA, $MSFT, $GOOG, and $AMZN continuing to invest in computing power. But tech stocks also face high valuations, capital expenditure returns, energy costs, and interest rate pressures. If Treasury yields gradually decline after the election, AI capital expenditures can translate into actual profits, and tech giants’ earnings expectations continue to grow, then $QQQ indeed has a chance to continue the historical pattern. Conversely, if corporate earnings expectations are significantly revised downward or AI investments start to drag on the free cash flow of large tech companies, even if political uncertainty fades, market gains cannot be guaranteed. One more easily overlooked point: a gain one year after the election does not mean the election day was the lowest point. 2018 is a typical example; after the November midterm election, the U.S. stock market still experienced a severe correction in December. Therefore, historical data supports maintaining a relatively positive outlook for the coming year but does not support going all-in unconditionally now, nor does it indicate that short-term corrections have ended. After the past six midterm elections, $QQQ’s price one year later was always higher than on election day. This is a historical pattern worth referencing, but whether it will hold for the seventh time in 2026 still depends on tech stock earnings growth, valuations, and liquidity. Being bullish on the tech sector long-term is not contradictory to controlling position sizes in the short term and waiting for more reasonable entry prices.

$LRCX Lam Research's fiscal year 2026 revenue reached $23.2 billion, a 26% year-over-year increase, with net profit exceeding $7.2 billion. However, many focus only on storage companies like $MU and $SNDK, overlooking the companies that actually sell manufacturing equipment to wafer fabs. Regardless of where storage prices ultimately go, as long as wafer fabs continue to expand capacity, equipment suppliers can secure orders in advance. Lam Research is one of the world's leading semiconductor manufacturing equipment companies, with core businesses including etching, thin film deposition, wafer cleaning, and related equipment services. It is distinctly different from ASML: ASML mainly handles lithography, while Lam Research primarily participates in material processing and structural manufacturing after lithography. Etching can be understood as forming extremely precise structures on wafer materials according to design requirements; deposition involves forming thin films of different materials on the wafer surface. Advanced chips require these processes to be repeated multiple times, and as chip structures become increasingly complex, process demands continue to rise. One of Lam Research's greatest advantages comes from 3D NAND manufacturing. Traditional NAND mainly relies on planar structures, while 3D NAND increases the number of stacked layers to enhance the data storage capacity per unit area. The higher the number of stacked layers, the more difficult it is to process the deep holes and complex structures inside the wafer, which demands higher precision, uniformity, and stability from etching equipment. This means NAND manufacturers may need to purchase more advanced etching and deposition equipment even without significantly increasing wafer output, as long as they continue to upgrade stacking technology. Therefore, Lam Research can benefit from both capacity expansion and manufacturing technology upgrades. $MU, Samsung, SK Hynix, SanDisk, and Kioxia's NAND manufacturing systems are all part of this industry chain. However, the specific procurement amounts still depend on each wafer fab's capital expenditure plans and cannot be directly inferred from NAND price increases. Besides NAND, DRAM and HBM are also important growth areas. HBM requires advanced DRAM manufacturing, chip stacking, and packaging technologies. As product complexity increases, the demand for equipment in some deposition, etching, and cleaning processes will also rise. Lam Research also participates in advanced logic chip manufacturing. GAA transistors, advanced interconnects, and some advanced packaging processes impose new requirements on material deposition and precision etching. Therefore, it is not a company solely reliant on the storage industry's prosperity. According to the fiscal year 2026 annual report, Lam Research's full-year revenue was $23.233 billion, a 26% year-over-year increase, with a GAAP gross margin of 50.5%, operating margin of 35.3%, and net profit of $7.265 billion. The recently announced June 2026 quarter revenue reached $6.722 billion, a 15.1% quarter-over-quarter increase, with GAAP gross margin further improving to 51.7%. Equipment sales revenue was about $4.25 billion, and customer support and other related revenue was about $2.47 billion. This service revenue is equally important. After wafer fabs purchase equipment, they require ongoing maintenance, parts replacement, and process upgrades. The more equipment installed, the larger the subsequent service business revenue base usually is, which can also somewhat mitigate fluctuations in new equipment procurement cycles. However, Lam Research's profit growth does not mean cash flow grows in tandem. Fiscal year 2026 operating cash flow was about $5.86 billion, lower than the previous year's $6.17 billion; after deducting about $970 million in capital expenditures, free cash flow was about $4.89 billion. While profits hit new highs, cash conversion still needs to be observed. Another risk comes from the China market. In fiscal year 2026, China contributed about 34% of revenue, with a 26% share in the June quarter. This includes various customers with factories in China and does not equate to all revenue coming from Chinese domestic companies. Changes in U.S. export controls may still directly impact its equipment sales. In terms of competition, companies like $AMAT Applied Materials and Tokyo Electron also participate in the etching and deposition equipment market. Although Lam Research has strong technical capabilities, it does not hold the exclusive supplier position in the EUV field that ASML enjoys. In the long term, Lam Research's opportunities are not just about rising storage prices but also about the increasing complexity of chip manufacturing, which may continuously increase the equipment and process value required per wafer. However, the equipment industry often has a more cyclical investment pattern than end-chip demand. If Samsung, SK Hynix, Micron, or foundry companies delay capacity expansion, equipment procurement may be quickly affected. When researching $LRCX, it is important to closely track global wafer manufacturing equipment spending, DRAM and NAND capital expenditures, advanced process orders, service revenue, and free cash flow. It has strong technical advantages and profitability but still cannot escape the semiconductor industry's capital expenditure cycles.

$ASML ASML's net profit for the third quarter approached €5 billion, with a gross margin of 57.5%. Management even stated that, based on the current planned capacity, it may not be able to fully meet customer demand in 2027 and 2028. Global chip manufacturers are all expanding production, but only a very few companies truly master advanced lithography technology. What exactly does ASML rely on to make money, and why does it hold such a strong position in the industry? ASML is a Dutch semiconductor equipment manufacturer, mainly providing lithography equipment for wafer fabs. Chip manufacturing involves many processes such as lithography, etching, deposition, and ion implantation, with lithography responsible for transferring the designed circuit patterns onto the wafer. Without sufficiently advanced lithography technology, wafer fabs find it difficult to achieve large-scale manufacturing of advanced chips. ASML mainly operates two types of lithography equipment: DUV and EUV. DUV is widely used in multiple process steps of mature and advanced nodes, while EUV is the key equipment for manufacturing advanced logic chips. Especially in the EUV field, ASML is currently the only equipment manufacturer to achieve commercial mass production supply worldwide. This means that no matter how advanced the chips developed by $NVDA, $AMD, or $AVGO are, as long as the final manufacturing process requires EUV, they must rely on the wafer manufacturing systems supported by ASML equipment. $TSM, Intel, and Samsung are all important users of advanced lithography technology. According to the Q3 2026 financial report released on October 14, ASML's quarterly revenue reached €11.662 billion, net profit €4.96 billion, with a gross margin of 57.5%. During the quarter, 113 new lithography systems were sold, while equipment maintenance, upgrades, and service revenue reached €3.289 billion. This part of the installed equipment service business is easily overlooked. Lithography machines are expensive and have a long service life; after wafer fabs purchase them, they require long-term maintenance, upgrades, and efficiency improvements. Therefore, ASML not only makes money from selling new equipment but also continuously generates revenue from equipment already installed in wafer fabs worldwide. ASML's technological barriers are not only in the equipment itself. EUV requires extremely complex light sources, precision mirrors, motion control systems, and software coordination, with optical system accuracy reaching very high levels. This capability depends on long-term R&D investment and the collective accumulation of the entire supply chain, which other equipment manufacturers cannot quickly replicate just by investing money. The most important technological upgrade in the next phase is High-NA EUV. Traditional EUV systems have a numerical aperture of 0.33, while the new generation High-NA increases it to 0.55, which can further improve lithography resolution and reduce some multi-patterning steps required by advanced processes. However, High-NA equipment costs more, and whether customers adopt it on a large scale depends on actual yield, capacity, and manufacturing costs. AI is an important driver of ASML's growth. Increased chip demand from $NVDA, $AMD, and $AVGO drives wafer fabs like TSMC to invest in advanced capacity; the development of HBM and advanced DRAM also increases demand for related lithography equipment. Management expects full-year 2026 revenue to reach €44 billion to €45 billion, with a gross margin of about 56%. Some customers have even expressed interest in equipment purchases for 2029 and 2030, but purchase intentions cannot be directly counted as formal orders. ASML's biggest risk comes from the semiconductor capital expenditure cycle. Wafer fabs typically concentrate equipment purchases during industry upturns; if AI demand growth slows or customers face overcapacity, equipment orders may quickly slow down. Export restrictions on the Chinese market will also affect sales of some products. In the long term, ASML's strongest advantage is its irreplaceable technology. However, a monopoly position does not mean the stock price only rises without falling. If the market assigns an excessively high valuation in advance, even if the company continues to grow in the coming years, it may still experience valuation compression. Researching $ASML requires continuous observation of EUV orders, equipment delivery cycles, High-NA adoption progress, installed equipment service revenue, and global wafer fab capital expenditures. As long as advanced process nodes continue to upgrade, ASML remains one of the most difficult companies to replace in the entire semiconductor industry chain.

$TSLA's adjusted EPS for Q3 exceeded expectations, but revenue basically stagnated, and operating profit dropped 56% year-over-year. Free cash flow was significantly better than expected, mainly due to capital expenditures being lower than market forecasts. This earnings report reflects some recovery in automotive business profit margins, but overall profitability remains under pressure. After the U.S. market close on October 21, Tesla released its Q3 2026 earnings report. Revenue was $28.046 billion, roughly flat year-over-year, below Tesla's official analyst consensus of $28.341 billion. Adjusted EPS was $0.48, higher than the expected $0.45 but lower than last year's $0.50. GAAP EPS was $0.22, compared to $0.39 in the same period last year. Operating profit was $718 million, down 56% year-over-year; operating margin was 2.6%, compared to 5.8% last year. R&D expenses reached $2.761 billion, up 69% year-over-year, mainly related to investments in AI, autonomous driving, and robotics projects. The automotive business showed some improvement. Q3 deliveries were 486,532 vehicles, down 2% year-over-year but above prior market expectations. Excluding regulatory credits, automotive gross margin reached 18.1%, up from 16.3% in Q2. Product mix adjustments, improved regional sales mix, and growth in FSD revenue supported the margin. However, automotive revenue still declined about 3% year-over-year, and regulatory credit revenue fell from $417 million last year to $113 million. The single-quarter gross margin recovery does not yet prove that vehicle sales and profitability have re-entered a sustained growth phase. Cash flow was the biggest divergence from market expectations this quarter. Free cash flow in Q3 was $1.53 billion, while Tesla's official analyst consensus was negative $4.506 billion. But actual capital expenditures were only $4.73 billion, below the prior market expectation of $7.771 billion. The improvement in free cash flow cannot be fully attributed to operational efficiency gains. Lower-than-expected capital expenditures also played a significant role. If investments in AI infrastructure, Robotaxi, and Optimus accelerate again in the future, cash flow may remain under pressure. The energy storage business also needs to distinguish between shipment volume and actual revenue. Q3 energy storage deployments were 13.7 GWh, up about 10% year-over-year, but energy business revenue declined about 16% year-over-year to $2.881 billion. Deployment growth did not translate into revenue growth. In contrast, services and other business revenue reached $4.605 billion, up about 33% year-over-year, becoming the faster-growing part of the revenue structure this quarter. In AI and autonomous driving, FSD active subscriptions reached 1.74 million, up 67% year-over-year; this statistic includes both one-time purchases and monthly subscriptions. Cybercab has started participating in Austin Robotaxi operations, but this business is still in early commercialization and has not yet become a major profit source for the company. From a valuation perspective, $TSLA still carries high expectations for AI, autonomous driving, and robotics commercialization, but current profits mainly rely on automotive, energy, and services businesses. Improvements in automotive gross margin, FSD user growth, and better-than-expected cash flow can support short-term market sentiment; however, revenue stagnation, a sharp year-over-year decline in operating profit, and ongoing capital expenditure needs for sustained growth still limit the extent to which this earnings report improves long-term profit expectations. Going forward, key focuses include whether automotive gross margin can continue to rise, the scale of Robotaxi commercialization, actual delivery of Optimus, and the impact of future capital expenditures on free cash flow. The operational quality of this earnings report improved compared to Q2, but there is still insufficient evidence that overall profits have re-entered a sustained growth phase. Short-term stock price may rise due to EPS and cash flow beating expectations, but whether long-term valuation can be supported ultimately depends on whether new businesses can contribute sufficient scale of actual profits.

$AMD's data center revenue in Q2 grew 107% year-over-year, but its gross margin remains significantly lower than $NVDA's. Both benefit from the AI computing boom, so why is there such a large gap in profitability between the two companies? Does AMD really have a chance to become the next Nvidia? AMD is not purely an AI GPU company. It mainly operates three major businesses: data center, client and gaming, and embedded, with core products including EPYC server CPUs, Instinct AI GPUs, Ryzen PC processors, Radeon graphics cards, and embedded computing chips. According to the Q2 2026 financial report, AMD's revenue reached $11.536 billion, up 50% year-over-year; GAAP net profit was $2.297 billion, with a gross margin of 54%. Data center revenue was $6.718 billion, up 107% year-over-year, accounting for about 58% of total company revenue. Client and gaming business revenue was $3.841 billion, and embedded business revenue was $977 million. Here is an easily overlooked issue: data center revenue does not equal all coming from AI GPUs. AMD also has a very important EPYC server CPU business, which is one of its most stable sources of long-term competitive advantage. The server CPU market has long been dominated by $INTC, but AMD has gradually improved its competitiveness in the server market through the Zen architecture, process upgrades, and core count advantages. Cloud computing, databases, virtualization, and AI servers all require CPUs; even if customers choose Nvidia GPUs, it does not mean the servers must use Nvidia CPUs. AMD's business with truly huge growth elasticity is the Instinct AI accelerator. Nvidia's biggest advantages are not only GPU performance but also the CUDA software ecosystem, NVLink interconnect, and a complete AI server system. AMD hopes to provide an alternative AI computing infrastructure through Instinct, the ROCm software platform, and the Helios rack-level system. Currently, AMD has established important partnerships with companies like OpenAI, Meta, Anthropic, Microsoft, and Oracle. OpenAI previously announced plans to deploy a total of 6GW of AMD GPU computing power, and Anthropic also announced a cooperation plan of up to 2GW. But it is necessary to distinguish between the scale of cooperation and the revenue already realized. These deployments often span multiple years, and actual shipments are affected by data center construction, power, product delivery, and customer funding arrangements, so the GW scale cannot be simply converted into current revenue. Another detail worth noting: AMD has provided OpenAI and Meta with conditional warrants to purchase up to 160 million shares of AMD stock each. These rights are linked to GPU procurement progress, stock price, and other conditions. As of June 27, 2026, no shares have vested under these warrants. They help bind major customers, but if a large number vest and are exercised in the future, they could also cause equity dilution. Compared to Nvidia, AMD's biggest current shortcoming remains its software ecosystem and system scale. CUDA has a long-accumulated developer base, frameworks, and optimization tools. Customers switching platforms must consider not only chip prices but also migration costs, stability, and actual inference efficiency. However, AMD does not need to completely replace Nvidia to achieve considerable growth. The AI computing market is large enough, some cloud providers want to reduce dependence on a single supplier, and customized inference and cost-sensitive tasks may also provide opportunities for AMD. From a profitability perspective, AMD's non-GAAP gross margin in Q2 was 56%, significantly lower than Nvidia's, but this gap cannot be entirely attributed to technological lag. The two companies have different product mixes, software pricing power, network system businesses, and revenue scales. If AMD can expand AI GPU sales and increase the proportion of high-margin businesses in its product mix, overall profit margins still have room for improvement. Risks are also obvious. AMD needs to continue investing in R&D, build a more mature software ecosystem, and compete with Nvidia for advanced packaging, HBM, and data center customer budgets. If the delivery of the next-generation Instinct products is delayed or actual performance and usage costs do not meet customer expectations, the growth priced in by the market may fail to materialize. The company previously expected Q3 2026 revenue of about $13 billion and a non-GAAP gross margin of about 56%. Future financial reports need to verify not only whether total revenue can exceed expectations but also how much of the data center revenue growth comes from AI GPUs, whether customer deployments proceed smoothly, and whether gross margins can continue to improve. $AMD's long-term opportunity lies in the combined growth of its CPU market competitive advantage and AI GPU business, but its stock price elasticity also implies higher expectation risk. It does not need to become a second Nvidia; what really matters is whether it can convert large-scale customer cooperation into sustained growth in revenue, profit, and free cash flow.