美股频道更多独家策划、专家专栏，免费查阅>> 责任编辑：山上 2月26日，据韩联社，闪迪与SK 海力士周三在前者总部联合举办HBF规格标准化联盟启动会，这标志着 两大记忆体厂正式启动高频宽快闪记忆体(HBF)的全球标准化进程。SK海力士表示，已与闪迪携手合 作，开始将下一代储存技术HBF标准化，以强化其在AI芯片市场的地位。 HBF作为介于高频宽记忆体(HBM)和固态硬碟之间的新型储存阶层，旨在弥合HBM高效能与固态硬碟 大容量特性之间的差距，并满足AI推理场景对容量扩展性与能源效率的双重需求。业界普遍预测，以 HBF为关键零组件的整合式记忆体解决方案，市场需求将在2030年左右迎来全面扩张。 ...

Core Viewpoint - AMD's stock price plummeted despite reporting better-than-expected Q4 earnings, primarily due to disappointing Q1 revenue guidance and concerns over competition in the AI chip market [3][4]. Financial Performance - AMD reported Q4 revenue of $10.3 billion, a 34% year-over-year increase, surpassing the analyst consensus of $9.65 billion [4]. - Adjusted earnings per share (EPS) for the quarter were $1.53, exceeding analyst expectations of $1.32 by nearly 16% [4]. - The data center business grew 39% year-over-year to $5.38 billion, significantly above the analyst forecast of $4.97 billion [4]. Revenue Guidance - AMD's Q1 revenue guidance midpoint is approximately $9.8 billion, which, while above the consensus estimate of nearly $9.4 billion, falls short of the previous quarter's record revenue of $10.27 billion [3][4]. - The guidance disappointment is attributed to a lack of expected growth in AI spending, leading to investor disillusionment [3]. Market Competition - AMD faces increasing competition in the AI chip market, with major tech companies opting for custom AI chips, which raises concerns about AMD's ability to compete effectively [4][5]. - The demand for older MI308 chips in China generated $390 million in revenue, but sales are expected to drop to about $100 million in the current quarter, indicating waning demand for this aging product [5]. Analyst Perspectives - Analysts have mixed views on AMD's performance, with some noting that high expectations and unexpected revenue from China contributed to the stock's decline [6]. - AMD's CEO maintains an optimistic outlook, projecting AI revenue to reach hundreds of billions by 2027 and indicating strong demand for next-generation AI servers [6][7]. Investment Outlook - Despite the stock drop, some analysts view this as a potential buying opportunity, with target prices for AMD stock ranging from $280 to $300, suggesting a 40% to 50% upside from current levels [8]. - The market's reaction to AMD's earnings contrasts sharply with that of competitors like Supermicro, which saw a significant stock rebound after its earnings report [8]. Valuation Concerns - AMD is trading at a high forward P/E ratio of 40, indicating that the market has high expectations for flawless execution in chip releases and infrastructure launches [9].

相比较去年8月1595.88元的高点，寒武纪股价已经下跌了超30%。 | 今开 | 1255.00 | | 最高 | 1257.78 | | 成交量 | 11.92万手 | | --- | --- | --- | --- | --- | --- | --- | --- | | 昨收 | 1242.00 | | 最低 | 1070.00 | | 成交额 | 137.81亿 | | 换手率 | 2.85% | | 市盈(TTM) | 241.69 | | 总市值 | 4536亿 | | / / 0 -- | TO | 口口 | EE I/ | 三 | 天/ | EI/ | 古々 .. | 此前，1月31日晚间，寒武纪发布业绩预告，预计2025年度实现营收60亿元到70亿元，同比增长410.87%到496.02%；实现净利18.5亿元到21.5亿元，扭亏 为盈；扣除非经常性损益后实现净利16亿元到19亿元。 寒武纪亦在业绩预告中提示，本次业绩预告是公司财务部门的初步核算，尚未经注册会计师审计，具体准确的财务数据以公司正式披露的经审计后的 《2025 年年度报告》为准。 值得注意的是，近日，福布斯更新了截至2026年 ...

今日寒武纪调整点评20260203 ＃关于寒武纪公司的2026年盈利预测，国海计算机是少有的一直坚持相对清晰且冷静判断的卖方。 我们认为：2026年对于国产AI芯片行业整体来说，＃市场规模翻倍已经是比较中性偏乐观的判断了，且具体还要 看后续大厂的Capex指引，但我们对此是乐观的。 ＃关于寒武纪公司的2026年盈利预测，国海计算机是少有的一直坚持相对清晰且冷静判断的卖方。 我们认为：2026年对于国产AI芯片行业整体来说，＃市场规模翻倍已经是比较中性偏乐观的判断了，且具体还要 看后续大厂的Capex指引，但我们对此是乐观的。 ＃我们看2028年国内三大CSP需求乐观可支撑3-4万亿市值的AI芯片市场，这其中包括互联网自研，先看ASIC与 GPU比例55开。 我们看建议理性看待每一家AI芯片公司的市占率，理论上其中不会有一家超过30%，且具体要依靠对产品、生 态、供应链及产业趋势的比较分析。 ＃我们看2028年国内三大CSP需求乐观可支撑3-4万亿市值的AI芯片市场， 今日寒武纪调整点评20260203 ...

Core Viewpoint - Peter Thiel's investment firm has completely divested from Nvidia, following similar actions by other major institutions, indicating a potential shift in market sentiment towards the AI chip giant [1][2][5]. Group 1: Thiel's Macro Fund Actions - Thiel Macro Fund sold approximately 537,742 shares of Nvidia, representing nearly 40% of its portfolio, and as of September 30, it no longer holds any Nvidia shares [2][4]. - The estimated value of the shares sold during the third quarter is close to $100 million based on Nvidia's average stock price during that period [4]. - The fund also liquidated its position in Vistra and reduced its Tesla holdings by 207,613 shares, while initiating new positions in Microsoft and Apple [4]. Group 2: Other Institutional Actions - SoftBank has completely exited its Nvidia position, realizing $5.83 billion from the sale [5]. - Bridgewater Associates significantly reduced its Nvidia holdings from 7.23 million shares to 2.51 million shares, a decrease of 65.3% [5]. - Michael Burry disclosed a substantial short position against Nvidia, purchasing $186 million in put options, indicating a bearish outlook [6]. Group 3: Market Sentiment and Analyst Predictions - Nvidia is set to release its Q3 fiscal 2026 earnings report on November 19, with expected revenue of $55.28 billion, a year-over-year increase of over 55% [7]. - Analysts from Jefferies and Wedbush express optimism, predicting Nvidia will exceed expectations and raise future guidance [7]. - Oppenheimer and Citigroup have raised their target prices for Nvidia, with Citigroup initiating a "30-day bullish" outlook ahead of the earnings report, emphasizing the supply-demand imbalance in the AI chip market [8].

Core Viewpoint - The AI training market is expected to grow rapidly, driven by the increasing adoption of AI GPUs from companies like NVIDIA and AMD, as well as the expansion of the ASIC market propelled by major American CSPs [2][3]. Group 1: ASIC Market Growth - The ASIC market is projected to experience a compound annual growth rate (CAGR) of up to 70% from 2024 to 2026, with an expected shipment of over 5 million ASIC units this year, reflecting a year-on-year growth of over 20% [2]. - The shipment ratio of AI GPUs to ASICs is expected to shift from 62:38 this year to 60:40 by 2026, indicating a growing importance of ASICs in the AI server category [2]. Group 2: Major CSP Developments - AWS plans to launch the Teton 2 cabinet using Trainium 2/2.5 chips in the second half of the year, which is expected to boost its ASIC chip shipments by over 40% [3]. - Meta is set to begin mass production of its Minerva cabinet using its MTIA chips, benefiting its main assembly partners, including Celestica and Quanta [3]. Group 3: ASIC vs. GPU - ASIC chips are custom-designed for specific applications, while NVIDIA's GPUs are general-purpose processors suitable for a wider range of functions [4]. - The sales of ASIC chips, led by Broadcom, are expected to reach between $60 billion and $90 billion by 2027, highlighting their growing significance in the AI chip market [4]. Group 4: Market Dynamics - Experts believe that the AI chip market is not a zero-sum game; both ASICs and GPUs can coexist and share the growth of the AI industry [5][6]. - Morgan Stanley forecasts that the ASIC chip market for AI will grow from $12 billion in 2024 to $30 billion by 2027, with a CAGR of 34% [6]. Group 5: Cost Efficiency - Amazon's Trainium chips reportedly reduce inference task costs by 30% to 40% compared to NVIDIA's H100 GPUs, while Google's latest TPU v6 shows a 67% improvement in energy efficiency over its predecessor [7]. Group 6: NVIDIA's Perspective - NVIDIA's CEO Jensen Huang acknowledges the value of ASICs but emphasizes their limitations in flexibility, arguing that GPUs are better suited for the rapidly changing AI landscape [8][9]. - Huang believes that NVIDIA's strategy focuses on leveraging the versatility of GPUs and maintaining a comprehensive software ecosystem, which is crucial for its competitive edge [10].

Core Viewpoint - TSMC is significantly expanding its CoWoS capacity, with projections indicating a rise from 70k wpm at the end of 2025 to 100-105k wpm by the end of 2026, and further exceeding 130k wpm by 2027, showcasing a growth rate that outpaces the industry average [1][2]. Capacity Expansion - TSMC's CoWoS capacity will reach 675k wafers in 2025, 1.08 million wafers in 2026 (a 60% year-on-year increase), and 1.43 million wafers in 2027 (a 31% year-on-year increase) [1]. - The expansion is concentrated in specific factories, with the Tainan AP8 factory expected to contribute approximately 30k wpm by the end of 2026, primarily serving high-end chips for NVIDIA and AMD [2]. Utilization Rates - Due to order matching issues with NVIDIA, CoWoS utilization is expected to drop to around 90% from Q4 2025 to Q1 2026, with some capacity expansion plans delayed from Q2 to Q3 2026. However, utilization is projected to return to full capacity in the second half of 2026 with the mass production of new projects [4]. Customer Allocation - In 2026, NVIDIA is projected to occupy 50.1% of CoWoS capacity, down from 51.4% in 2025, with an allocation of approximately 541k wafers [5][6]. - AMD's CoWoS capacity is expected to grow from 52k wafers in 2025 to 99k wafers in 2026, while Broadcom's capacity is projected to reach 187k wafers, benefiting from the production of Google TPU and Meta V3 ASIC [5][6]. Technology Developments - TSMC is focusing on advanced packaging technologies such as CoPoS and WMCM, with CoPoS expected to be commercially available by the end of 2028, while WMCM is set for mass production in Q2 2026 [11][14]. - CoPoS technology offers higher yield efficiency and lower costs compared to CoWoS, while WMCM is positioned as a cost-effective solution for mid-range markets [12][14]. Supply Chain and Global Strategy - TSMC plans to outsource CoWoS backend processes to ASE/SPIL, which is expected to generate significant revenue growth for these companies [15]. - TSMC's aggressive investment strategy in the U.S. aims to establish advanced packaging facilities, enhancing local supply chain capabilities and addressing global supply chain restructuring [15]. AI Business Contribution - AI-related revenue for TSMC is projected to increase from 6% in 2023 to 35% in 2026, with front-end wafer revenue at $45.162 billion and CoWoS backend revenue at $6.273 billion, becoming a core growth driver [16].

Core Viewpoint - Qualcomm is re-entering the data center market by acquiring Alphawave, a leading semiconductor IP company, to enhance its competitive edge in this sector, particularly in AI chip development [1][2]. Group 1: Acquisition Details - Qualcomm announced the acquisition of Alphawave IP Group plc for an implied enterprise value of approximately $2.4 billion, expected to be completed by Q1 2026 [1][2]. - The acquisition aims to accelerate Qualcomm's expansion into the data center market and provide critical assets, particularly in high-performance, low-power computing [2][4]. Group 2: Financial Performance of Alphawave - Alphawave reported revenues of $308 million for FY2024, a 4% decline year-over-year, while FY2023 revenues were $322 million, showing a significant growth of 74% [2]. - The company achieved a booking amount of $520 million in FY2024, a 34% increase year-over-year, with a gross margin of 59%, up by 8 percentage points [2]. Group 3: Market Position and Technology - Alphawave ranks fourth globally in the semiconductor design IP market with a market share of 3.2% [3]. - The company specializes in high-end interface IP products, which are crucial for building high-performance computing applications, particularly in AI systems [3]. Group 4: Strategic Implications for Qualcomm - The acquisition is expected to enhance the data transmission efficiency of Qualcomm's Oryon CPU and Hexagon NPU in AI training and inference scenarios [4]. - Analysts suggest Qualcomm may pursue two development paths: launching Arm-based data center processors to compete directly with Nvidia and AMD, or adopting a chip design service model similar to Broadcom [4][5]. Group 5: Historical Context and Future Outlook - Qualcomm previously attempted to enter the data center market with the Centriq 2400 processor in 2017 but faced challenges due to the dominance of Intel's x86 architecture [5]. - The company has been gradually building its capabilities, including the acquisition of Nuvia in 2021, which focused on data center CPUs [5][6]. - Qualcomm's CEO indicated that the company is well-positioned to capitalize on the growing demand for AI data centers and plans to re-enter the server market with new products by 2025 [7][8].

Core Viewpoint - AMD's stock price has been on a downward trend for the past nine months, falling below significant long-term support levels, despite a strong rebound from its April low of approximately $75 per share [1][3]. Financial Performance - In Q1 2025, AMD reported revenue of $7.44 billion, a year-over-year increase of 36% but a quarter-over-quarter decrease of 3%, exceeding market expectations of $7.12 billion, driven by strong sales of EPYC CPUs, Instinct Mi300 AI GPUs, and Ryzen processors [4][5]. - The gross profit for Q1 2025 was $3.74 billion, up 46% year-over-year, with a gross margin of 50%, which is a 3 percentage point increase from the previous year [5]. - Operating income reached $806 million, a significant increase of 2,139% year-over-year, with an operating margin of 11% [5]. - Net income for Q1 2025 was $709 million, up 476% from the previous year, resulting in diluted earnings per share of $0.44, a 529% increase [5]. Segment Performance - The Data Center segment generated revenue of $3.67 billion, up 57% year-over-year, primarily due to growth in EPYC CPU and Instinct GPU sales [6]. - The Client and Gaming segment revenue was $2.94 billion, a 28% increase year-over-year, with Client revenue specifically rising by 68% [6]. - The Embedded segment saw a slight decline in revenue, down 3% year-over-year to $823 million [6]. Future Outlook - AMD's management expects Q2 2025 revenue to be around $7.4 billion, indicating a slowdown in growth to approximately 26%, although this guidance exceeds the general market expectation of $7.24 billion [8]. - The company anticipates strong performance in the Data Center market in the second half of 2025, driven by the Mi350x AI GPU, with a projected market share of 5-10% in the AI chip market by 2027, which is expected to reach $400 billion [7][9]. Shareholder Returns - AMD generated $727 million in free cash flow in Q1 2025, with a free cash flow margin of about 10%, and returned $749 million to shareholders through stock buybacks [10]. - The company has a buyback authorization of approximately $4 billion and a net cash balance exceeding $3 billion, indicating potential for continued shareholder returns [10]. Valuation and Investment Thesis - Following the strong Q1 2025 results, AMD's fair value estimate has increased from $151 to $162 per share, suggesting a potential upside of about 64% from current levels [14]. - Analysts project a five-year target price of approximately $291 per share, implying a compound annual growth rate (CAGR) of about 24.1% [14]. - Despite challenges in the semiconductor industry, AMD's expected CAGR exceeds the analyst's investment hurdle rate of 15%, making it an attractive accumulation opportunity at current levels [15].