Investment Rating - The report maintains a "Hold" rating for Micron Technology (MU US) with a target price raised to $361.00 [1][3][13] Core Views - Concerns arise over the rapid increase in storage prices potentially weakening downstream demand, with server-grade DDR5 modules rising over 600% and consumer-grade DDR5 modules increasing by 250-400% from early 2025 to February 2026 [1][2] - Despite short-term concerns, there is potential for further stock price increases as market expectations adjust, particularly with anticipated growth in AI applications and capital expenditures from major cloud service providers (CSPs) expected to exceed $710 billion in 2026, reflecting a 61% year-over-year growth [2][10] - Micron's revenue and net profit estimates for fiscal years 2026-2027 have been significantly increased, reflecting the impact of rising storage prices and improved utilization rates, with net profit projections increasing by 211.6% to 559.0% [3][11] Financial Summary - Revenue projections for 2026 are estimated at $77.341 billion, with a year-over-year growth rate of 106.9%, and for 2027, revenue is projected at $101.608 billion, reflecting a growth rate of 31.4% [9][20] - The estimated earnings per share (EPS) for 2026 is $34.70, a significant increase from $10.42, and for 2027, the EPS is projected at $46.56, up from $6.29 [5][11] - The report indicates a substantial increase in gross profit margins, with a projected gross margin of 67.3% for 2026, compared to 39.8% in 2025 [20] Valuation - The valuation approach uses a sum-of-the-parts (SOTP) method, estimating the DRAM segment at $352.684 billion and the NAND segment at $20.487 billion, leading to a total equity value of approximately $394.939 billion [13][14] - The target price of $361.00 is based on a price-to-earnings (P/E) ratio of 10 for the DRAM segment and 5 for the NAND segment, with net cash of $21.721 billion included in the valuation [13][14]

Core Insights - NVIDIA is set to unveil a groundbreaking chip at the GTC conference in mid-March, which is expected to integrate Groq's LPU technology for a new inference product [1][4] - The shift in global computing demand is moving from training to inference, with predictions indicating that by 2026, inference will account for two-thirds of all AI computing power [3] - The new chip is anticipated to enhance decoding efficiency, addressing the limitations of current GPU architectures in handling large model parameters [5][6] Group 1: Chip Development and Technology - The upcoming chip is likely to be a new inference chip system that incorporates Groq's LPU technology, marking a significant integration of external architecture into NVIDIA's core AI computing product line [4] - The Groq LPU is designed specifically for inference acceleration, utilizing SRAM for model parameter storage, which offers significantly higher memory bandwidth compared to traditional GPU architectures [6] - NVIDIA may adopt a 3D stacking approach similar to AMD's V-Cache technology, integrating LPU units directly on top of GPU cores to enhance performance [7][8] Group 2: Market Trends and Predictions - The market is expected to see the emergence of specialized inference chips worth billions, which will be deployed in data centers and enterprise servers, with some chips potentially having power consumption comparable to general AI chips [3] - The industry is witnessing a trend where advanced manufacturing processes are becoming increasingly critical, with a focus on achieving high interconnect density and energy efficiency in chip designs [10] - There is a potential risk for domestic packaging and testing companies to be pushed out of the high-end market as the value of advanced chips concentrates on front-end manufacturing and advanced packaging [10]

Market Overview - Global markets showed positive momentum with A-shares rising, driven by technology stocks and easing concerns over the Strait of Hormuz blockade[3] - U.S. stocks experienced volatility, with the Dow Jones down 0.07% and the S&P 500 down 0.21%, while the Nasdaq remained flat[10] Oil Market Dynamics - Oil prices saw a significant drop of 20% due to misleading reports about U.S. naval escorts, but recovered to a final drop of 11.94%[28] - Iraq's oil production cuts and attacks on UAE refineries contributed to ongoing supply concerns in the Middle East[4] Fixed Income Market - U.S. Treasury yields rose by 5-8 basis points amid a lackluster auction for 3-year bonds, with the 10-year yield reaching 4.16%[5] - Investment-grade bond spreads narrowed by 3-7 basis points in Asia, reflecting improved sentiment following easing geopolitical tensions[32] Trade and Export Performance - China's exports in the first two months of the year increased by approximately 17%, driven by strong demand in the semiconductor and automotive sectors[6] - The adjustment of export tax rebates and a low base from the previous year contributed to this growth[6] Stock Market Highlights - Hong Kong's Hang Seng Index rose by 2.17%, with technology and AI sectors leading the gains, particularly driven by the success of OpenClaw[12] - The A-share market saw a significant increase, with the Shanghai Composite Index up 0.65% and the ChiNext Index up 3.04%[15] Key Corporate Developments - Amazon announced plans to issue at least $37 billion in bonds, marking a record for single-day issuance[32] - TSMC reported a 30% year-on-year revenue increase in the first two months, although it fell short of expectations[10] Currency Movements - The U.S. dollar index fell by 0.4%, while the dollar strengthened against the Chinese yuan, trading at 6.868[26] - The euro traded at 1.161 against the dollar, reflecting a slight decline of 0.2%[26]

Macro Commentary - China's exports showed significant growth of 21.8% in the first two months of 2026, driven by a later Chinese New Year, low base effects, and an upturn in the global semiconductor industry [2] - Exports to the US saw a notable reduction in decline for the first time since Q2 2025, while imports rebounded significantly due to increased demand for AI-related materials [2] - The forecast for China's export growth is expected to slow slightly from 5.2% in 2025 to 4% in 2026, while imports are projected to recover from -0.5% to 2% [2] Market Performance - The Hang Seng Index closed at 25,960, up 2.17% for the day and 1.29% year-to-date [2] - The Shenzhen Component Index rose by 1.84% year-to-date, reflecting a strong performance in the Chinese stock market [2] - The US markets showed mixed results, with the S&P 500 and Dow Jones experiencing slight declines of 0.21% and 0.07% respectively [2] Company Commentary: NIO - NIO reported a 76% year-on-year revenue increase in Q4 2025, reaching 34.7 billion yuan, with a gross margin improvement of 3.6 percentage points to 17.5% [5] - The company achieved a non-GAAP operating profit of 1.25 billion yuan, exceeding previous forecasts, and recorded its first quarterly net profit of 122 million yuan [5] - Despite the positive results, NIO faces challenges in 2026 due to market competition, AI race, and fluctuations in component prices, leading to a slight reduction in sales forecasts [5][6] Profit Forecast and Valuation - Based on a sales assumption of 460,000 units for 2026, with five major models accounting for 60%, the forecasted gross margin is expected to rise to 16.3%, with net losses narrowing to 3.8 billion yuan [6] - NIO's brand value is recognized as a competitive advantage, but the costs associated with brand building may hinder overall profitability compared to peers [6] - The target price for NIO shares has been adjusted to $6 in the US and HK$47 in Hong Kong, based on a price-to-sales ratio of 0.8x for FY26E [6]

Core Viewpoint - The AI industry is still in its early development stage, requiring trillions of dollars in ongoing investment to fully realize its potential, despite current investments in the industry amounting to hundreds of billions of dollars [3]. Group 1: AI Five-Layer Architecture - The AI "Five-Layer Architecture" is defined as a "five-layer cake," consisting of energy, chips, infrastructure, models, and applications, with each layer supporting and driving the others [4]. - The energy layer is identified as the fundamental principle of AI infrastructure, emphasizing that real-time generated intelligence requires real-time generated power, making energy supply a critical bottleneck for AI's scalable development [4]. - The chip layer serves as the physical foundation for computing power, with advancements in chip technology directly influencing the speed of AI expansion and the reduction of intelligence costs, although current chip technology is struggling to keep pace with the explosive growth in AI computing demands [4]. Group 2: Infrastructure and Model Layers - The infrastructure layer, referred to as the "AI factory," includes essential components such as land, power delivery, cooling systems, and network coordination, with a significant global push to build chip manufacturing plants, supercomputer factories, and AI factories, marking one of the largest infrastructure developments in human history [5]. - The model layer encompasses various types of information processing, with the potential of AI models still largely untapped; open-source models play a crucial role in accelerating the demand for underlying infrastructure, chips, and energy [7]. Group 3: Application Layer and Employment Impact - The application layer is where AI generates economic value, with a wide range of applications such as drug discovery platforms and autonomous vehicles; the innovation space in this layer remains vast, with predictions that traditional software forms may evolve into a new paradigm of AI Agents [7]. - Contrary to fears of job loss due to AI, it is believed that AI will create numerous new job opportunities, particularly in skilled labor sectors needed for infrastructure development, filling significant labor gaps in various industries [8].

Core Insights - The article discusses NVIDIA CEO Jensen Huang's concept of the "Five-Layer Cake" in AI, which includes Energy, Chips, Infrastructure, Models, and Applications [2][3] - Huang emphasizes that the largest scale of AI infrastructure construction has just begun, with thousands of billions already invested and the need for infrastructure worth tens of thousands of billions [3] - AI factories are being built at an unprecedented scale, requiring a vast workforce including electricians, plumbers, and network technicians [3][4] Summary by Sections Energy Layer - Energy is identified as a core element that fundamentally determines the production scale limit of intelligence [4] Chip Layer - Chips are being redesigned for efficiency, which is crucial for the scalability of AI [4] Infrastructure Layer - The construction of AI factories and related infrastructure is ongoing, with significant investments in land, power supply, cooling systems, and network communication [3][4] Model Layer - AI models have crossed a significant threshold, with improved performance and reduced hallucination phenomena, enabling large-scale applications that create real economic value [3] Application Layer - The application layer is accelerating due to the widespread availability of advanced models, which in turn drives demand for training, infrastructure, chips, and energy [3][4]

Group 1 - The core view of the report indicates that the A-share market is expected to maintain a trend of upward fluctuations, with a strategy of opportunistic low-positioning being favored [2][3][10] - The report highlights that the recent geopolitical tensions in the Middle East have influenced global markets, leading to a rebound in the A-share market, particularly in technology stocks [2][9] - It is noted that while oil and gas sectors may experience short-term spikes during geopolitical conflicts, these gains are often not sustainable, and a cautious approach is recommended [2][10] Group 2 - The report emphasizes the importance of focusing on high-quality stocks in the oil and gas and chemical sectors that have stable dividends and strong performance certainty [3][10] - It suggests that investors should also consider technology growth stocks that are less affected by oil price fluctuations, such as AI computing, semiconductors, and humanoid robots [3][10] - The report identifies specific sectors that showed strong performance, including optical communication modules, semiconductors, and communication equipment, while noting the volatility in oil and gas stocks [4][6][8]

Group 1: 和黄医药 - The company is expected to return to a growth trajectory in 2H25, with a significant rebound in tumor product revenue, increasing by 16% compared to 1H25 [1] - Sales of furmonertinib in overseas markets are projected to rise by 22% year-on-year and 25% quarter-on-quarter in 2H25 [1] - The company has reduced R&D and SG&A expenses by 30% and 9% respectively in 2025, achieving a net profit of $457 million, indicating sustainable profitability [1] Group 2: ATTC Platform - The ATTC platform is anticipated to have a breakthrough year, with HMPL-A251 and HMPL-A580 entering global Phase I trials, and HMPL-830 expected to submit IND applications in 2H26 [2] - Short-term catalysts include the results of global Phase III SAFFRON and Chinese Phase III SANOVO studies for savolitinib, and NDA submissions for various indications [2] - Revenue forecasts for 2026-28 have been raised, with a target price adjustment to HKD 36.60, maintaining a buy rating [2] Group 3: 宁德时代 - The company reported strong annual performance, with revenue of RMB 423.7 billion in 2025, a year-on-year increase of 17%, and a net profit growth of 42.3% to RMB 72.2 billion [3] - The gross margin and net margin for 2025 were 26.3% and 17.0% respectively, with a notable increase in Q4 gross margin to 28.2% [3] - The company’s battery production capacity reached 772 GWh in 2025, with a capacity utilization rate of 96.9%, significantly up from 76.3% in 2024 [4] Group 4: Industry Outlook - The global smartphone and PC shipment volumes are expected to decline by 11% in 2026, indicating a challenging market environment [8] - The semiconductor sector remains promising, with TSMC reporting a 22% year-on-year revenue growth in February 2026, suggesting resilience in the industry [9] - The technology sector is experiencing increased volatility, but the fundamentals of AI remain unchanged, providing potential investment opportunities [7]

Export Performance - In January-February 2026, China's export value increased by 21.8% year-on-year, up from 6.6% in December 2025, exceeding Bloomberg's consensus forecast of 7.2%[1] - The export growth was significantly boosted by seasonal factors due to the later timing of the 2026 Spring Festival, contributing approximately 7.4 percentage points to the year-on-year growth[2] - Integrated circuit exports surged by 72.6%, contributing 3.4 percentage points to overall export growth, while automotive exports rose by 67.1%, contributing 2.0 percentage points[3] Import Performance - Imports in January-February 2026 rose by 19.8% year-on-year, up from 5.7% in December 2025, also surpassing Bloomberg's expectation of 7.0%[4] - Notably, imports from Hong Kong increased by 333%, contributing approximately 1.7 percentage points to total imports, while imports from South Korea contributed 2.5 percentage points, reflecting strong semiconductor demand driven by AI investments[5] - Industrial metal raw materials saw an 8.9% year-on-year increase, contributing 1.9 percentage points to import growth[6] Trade Surplus - The trade surplus for January-February 2026 reached $213.6 billion, an increase of $42.7 billion year-on-year[7] - The overall trade dynamics indicate a robust performance in both exports and imports, with expectations of maintaining double-digit growth in the first quarter despite potential drag from the Spring Festival timing on March exports[8] Risks and Outlook - Potential risks include a slowdown in global AI investments and fluctuations in U.S. tariff policies, which could impact future trade dynamics[9] - Despite the anticipated drag on March exports, the overall outlook for the first quarter remains positive, supported by global AI investment acceleration and fiscal expansion abroad[10]

Market Performance - The Hang Seng Technology Index has experienced a maximum drawdown of 28% since October 2025, with a cumulative decline of 25.3% as of March 2, 2026[2][10]. - The index saw a peak cumulative increase of 58% from early 2025 until its peak in October 2025[2]. Key Drivers of Decline - The decline is attributed to two main factors: a visible AI revaluation and a hidden adjustment in consumer earnings expectations[2][3]. - The first phase of decline (October 2, 2025, to November 21, 2025) was driven by a 19.3% drop in valuation, while the second phase (January 14, 2026, to March 2, 2026) saw a 15.6% decline, primarily affecting internet consumption and comprehensive platforms[10][12]. Sector Analysis - The technology sector remains under pressure, with AI hardware showing resilience while AI software has largely digested its revaluation since July 2025[3][31]. - Consumer sectors, particularly consumer electronics and home appliances, are under significant pressure, with earnings expectations still needing stabilization[3][37]. Future Outlook - The Hang Seng Technology Index is projected to stabilize between 4,500 and 4,850 points, with a neutral scenario around 4,690 points[42]. - Key catalysts for recovery include the stabilization of consumer earnings, a peak in competitive pressures, and positive developments in AI applications by major firms[4][42]. Investment Sentiment - Southbound capital has significantly increased its allocation to the Hang Seng Technology Index since August 2025, indicating a "buy the dip" mentality[4][49]. - The latest sentiment index reading is at 58, which has not yet triggered a buy signal, suggesting caution in the market[4][47].