以下文章来源于芯片小韭菜 ，作者芯片小韭菜 芯片小韭菜 . 专注芯片行业及公司基本面研究，不提供操作建议。 长假就要结束了，对A股板块梳理一下，不做为投资建议。 目前最热的版块是 存储 。各种宏大叙事，AI导致存力的需求大增，sora2的视频生成，就是一个存储需求大增的典型案例。下图形象的说明了存储的需求 增量： 长假期间，海外从芯片的海力士，三星，美光，华邦电，南亚科，涨到了ssd的恺俠，涨到了硬盘的希捷。还有各种存储占比大的设备，拉姆研究，爱德 万等等。映射到国内就是各种存储芯片股，存储模组股，存储相关的设备股。 设备 板块 也很热，这和国内的扩产预期相关，正好国内的cx，cc也正在扩产，所以存储相关的设备也很热门。从前道设备到后道设备都有很多机会，尤 其后道设备的长川科技出了一个很好的业绩预增以后。日本新领导人上台，美国收紧对实体公司的子公司的销售等，也属于对设备板块的利好。假如在去 美化以后的下一步是去日化，那有些个股会有机会。gkj相关的零部件也值得关注。 AI算力 这边，hwj节前偏弱，现在定增落地，不知道是否有机会起来。海光有个订单。芯原也有点走弱了。翱捷在减持影响了一个月后，正开始走好。 代工 ...

德州仪器发表了以下声明： 此举正值德州仪器计划关闭其位于北德克萨斯州的剩余150mm半导体晶圆厂之际。 该员工表示，此前他们曾得到承诺，如果他们签约帮助关闭德州仪器老旧的150mm工厂，他们将在德州仪器新 工厂获得优先录用。 据该员工称， 德州仪器将有近400名员工失业。 据报道，德州仪器位于达拉斯北中央高速公路的150mm晶圆厂的150多名员工已收到通知，他们的最后工作日 为12月12日。 德州仪器（TI）已通知员工， 其现有的达拉斯和谢尔曼制造工厂预计将于12月进行一轮裁员。 一位不愿透露姓名的员工表示，该公司于9月25日向员工透露了这一消息。 "我们定期研究如何提高运营效率，并更好地支持公司的长期战略。因此，我们近期进行了一些组织架构调整， 影响了部分员工。这些调整包括关闭位于达拉斯和谢尔曼的剩余150mm晶圆厂，并已进入我们计划的多年过渡 期的最后阶段。这些都是艰难的决定，我们的首要任务是支持我们的长期员工度过这些变化。德州仪器对北德 克萨斯州的长期承诺，包括我们对谢尔曼新工厂的投资，保持不变。" 转自： 天天 IC 加入"中国IC独角兽联盟"，请点击进入 是说芯语原创，欢迎关注分享 ...

Core Insights - OpenAI and AMD have announced a groundbreaking strategic partnership to deploy a total of 6 gigawatts (60 billion watts) of AMD GPU computing power over the next few years, marking a new phase in global AI infrastructure development and significantly impacting the AI chip market competition landscape [1][6]. Partnership Details - The collaboration will utilize a multi-year, multi-generation product deployment model, with the expansion of OpenAI's computing power set to officially begin in the second half of 2026, starting with an initial deployment of 1 gigawatt (10 billion watts) based on the AMD Instinct MI450 series GPUs [3]. - This partnership aims to create a continuous iterative computing support system, leveraging AMD's next-generation AI acceleration products alongside rack-mounted AI solutions, enhancing the synergy between hardware and software since the MI300X and MI350X series [3]. Equity and Financial Implications - AMD has granted OpenAI warrants to purchase up to 160 million shares of common stock, potentially giving OpenAI approximately 10% ownership in AMD if fully exercised. The exercise of these warrants is tied to specific milestones, including AMD's stock price reaching $600, while the current closing price is $164.67, with an exercise price as low as $0.01 [5][6]. - This profit-sharing mechanism is expected to align the strategic goals of both companies, fostering a long-term stable cooperative ecosystem. AMD's CEO highlighted the collaboration as a true win-win, supporting ambitious global AI initiatives and advancing the entire AI ecosystem [5][6]. Market Impact - Following the announcement, AMD's stock price surged over 35% in pre-market trading, reaching a high of $207, indicating strong market confidence in the partnership's prospects [6]. - This collaboration not only provides AMD with a flagship customer endorsement for its AI chip technology but also helps OpenAI diversify its computing supply chain, reducing reliance on a single supplier. This follows OpenAI's recent agreement with NVIDIA for 10 gigawatts of computing power, further enhancing its diversified computing strategy [6]. Industry Implications - The strategic partnership, valued at hundreds of billions of dollars, represents AMD's strong challenge in the AI accelerator market and is a significant step for OpenAI in building a trillion-dollar computing infrastructure. It signals the emergence of a deep cyclical "closed-loop economy" model in AI, which is expected to profoundly influence the development trajectory of the global AI industry [6].

Core Insights - China has achieved a world-class breakthrough in the integration of marine technology and digital economy with the successful deployment of the world's first commercial underwater data center core equipment in Hainan Ling Shui, marking a significant advancement in digital infrastructure [1] Project Overview - The project was initiated in 2022 and is divided into three phases, with plans to deploy 100 data pods [3] - The first underwater data center is expected to be completed by November 2023, with the Hainan Underwater Intelligent Computing Center cluster set to be operational by February 2024 [3] Technological Advantages - The underwater data center utilizes seawater for natural cooling, achieving low energy consumption (PUE value below 1.1) and high reliability (server failure rate only one-eighth that of land-based data centers) [5] - The modular design allows for rapid deployment within 90 days, enhancing operational efficiency [3][5] Strategic Significance - This breakthrough provides robust green computing support for China's "East Data West Computing" initiative, addressing the urgent demand for efficient, low-carbon digital infrastructure in coastal regions [7] - It exemplifies the deep integration of the "Marine Power" and "Digital China" strategies, showcasing China's global leadership in cross-disciplinary technological innovation [7]

Core Viewpoint - The global DRAM super cycle has officially arrived, leading to a prolonged supply tightness in memory products [1] Group 1: Price Trends - Since the second quarter, DDR4 prices have been rising, which has also led to an increase in DDR5 prices, putting cost pressure on consumers with installation needs [1] - Micron Technology and SanDisk have announced price hikes, with SanDisk increasing prices by over 10% and Micron notifying channel partners of a 20% to 30% increase [1] Group 2: Drivers of Demand - The explosive growth of AI is driving DRAM products into a new super cycle [2] - The core driver is the widespread application of HBM, particularly as tech giants actively develop customized ASIC chips to enhance AI system performance [4] Group 3: Supply Constraints - The rapid development of AI is placing unprecedented pressure on DRAM capacity, with each AI computing cluster requiring a significant amount of HBM, further increasing the demand for basic DRAM wafers [5] - Global DRAM suppliers currently have only about three weeks of inventory, a near seven-year low, significantly below the industry average of ten weeks, indicating a rapidly increasing supply-demand imbalance [5] Group 4: Future Projections - Major tech companies, including AMD and NVIDIA, have a substantial demand for HBM, and the market's demand expectations for DRAM are further elevated as these companies develop AI ASIC chips [6] - UBS analysts predict that OpenAI's upcoming ASIC chips will utilize 12-layer stacked HBM3E technology, potentially consuming 500,000 to 600,000 DRAM wafers per month globally between 2026 and 2029 [6] - OpenAI's Stargate supercomputing cluster is expected to consume 900,000 DRAM wafers monthly, accounting for about 40% of global supply, elevating the strategic importance of DRAM to that of advanced process chips [6] - In the short term, suppliers will struggle to significantly increase capacity, as major DRAM manufacturers like Samsung, SK Hynix, and Micron have shifted some production lines to HBM products and upgraded to 1c nm processes to enhance capacity [6] - The concentration of global DRAM capacity in South Korea poses a key challenge for the industry in meeting the massive demand from the AI sector in the coming years [6] Group 5: Future Technology - With the upcoming introduction of next-generation technologies like HBM4, demand is expected to further expand, accelerating the growth of the DRAM industry [7] - Expanding DRAM supply remains the only way to meet the strong demand for HBM from tech giants [8]

Core Viewpoint - The establishment of the "Hefei Jinghui Chuangxin Investment Fund" with a scale of 300 million yuan focuses on investing in core areas of the semiconductor industry chain, including chip manufacturing, packaging testing, and key materials [1]. Group 1: Fund Details - The fund was officially established in July, with contributions from three listed companies: Jinghe Integrated, Huicheng Co., and Guanggang Gas, with Jinghe Integrated contributing 200 million yuan, accounting for 66.7% of the fund [1]. - The fund management is handled by "Hefei Jinghe Huixin," a joint venture established by Jinghe Integrated and Huicheng Co. last year [1]. Group 2: Contributor Background - Jinghe Integrated is the third-largest wafer foundry in China, with revenue growth from 230 million yuan in 2018 to 10 billion yuan in 2022, and it went public on the Sci-Tech Innovation Board in 2023 with a valuation of 40 billion yuan [3]. - Huicheng Co. is a leading enterprise in display driver chip packaging and has achieved mass production of 12-inch wafer gold bumps, also listed on the Sci-Tech Innovation Board [3]. - Guanggang Gas is the largest supplier of high-purity electronic gases in China, with a market share of 15.3% in the electronic bulk gas market [3]. Group 3: Investment Landscape - Hefei has cultivated semiconductor companies like Changxin Storage and Jinghe Integrated through a model of "state-owned capital leading + precise investment," which has transformed these companies into "industrial investors" that support the industry chain ecosystem [4]. - The local government has invested in over 80 new generation information technology companies through a "leading CVC + government fund" model, with plans to promote projects like high-end photomasks by 2025 [5]. - Recent statistics indicate that over 40% of early-stage investments in the semiconductor sector in China will come from industrial capital by 2025, an increase of 22 percentage points compared to three years ago [6].

Core Viewpoint - The article highlights the significant growth and potential of Haiguang Information, particularly in the context of the recent 30 billion procurement project by Industrial and Commercial Bank of China, with major players like Inspur, ZTE, and Lenovo as candidates. Haiguang is positioned as a major beneficiary due to its advancements in domestic chip technology and its strategic focus on AI and domestic substitution [1][3]. Group 1: Market Position and Growth - Haiguang Information's market capitalization has surged from around 300 billion to nearly 600 billion, driven by domestic chip stock growth and the dual forces of domestic substitution and AI [3]. - The company has successfully capitalized on the domestic market's need for high-end computing solutions, filling the gap left by international firms due to export restrictions [8][9]. Group 2: Technological Advancements - Haiguang has made significant investments in R&D, with over 90% of its workforce dedicated to this area, and its R&D expenditure reached 3.446 billion, accounting for 37.61% of its revenue [6]. - The C86 architecture of Haiguang is fully compatible with the Wintel ecosystem, allowing it to compete directly with Intel's mainstream products, while its DCU architecture aligns closely with NVIDIA's offerings [7]. Group 3: Security and Performance - The company has integrated advanced security features into its CPUs, achieving high security standards and reducing the risk of vulnerabilities that have plagued international competitors [12]. - Haiguang's C86 chips demonstrate stable performance and compatibility with mainstream ecosystems, minimizing system migration issues and enhancing reliability in high-demand sectors like finance [13].

Core Viewpoint - The article discusses the generational differences in humor appreciation, highlighting how younger audiences find joy in absurdity and spontaneity, while older generations seek structured narratives and deeper meanings [5][11]. Group 1: Humor and Entertainment - Younger generations are immersed in fragmented entertainment, leading to a preference for humor that is immediate and unexpected, contrasting with older generations' expectations for a coherent storyline [5][6]. - The humor in contemporary performances, such as "Skills Gobang," relies on absurdity and non-logical elements, which serve as a form of emotional release for younger audiences [7][8]. Group 2: Emotional Needs and Expectations - Older generations have a "high demand" for entertainment that offers a complete story and meaningful themes, while younger generations exhibit a "low demand" for entertainment, seeking light-heartedness and minimal emotional burden [11]. - The shift in emotional needs reflects broader societal changes, where younger individuals face high-pressure environments and prefer entertainment that allows for carefree enjoyment without the need for deep analysis [11].

Core Viewpoint - TSMC has significantly reduced the peak power consumption of its advanced EUV lithography equipment by 44%, addressing a major challenge in the semiconductor industry's sustainability efforts [1][3][5]. Group 1: Technological Innovations - The energy efficiency improvement is attributed to deep innovations in equipment collaboration optimization, light source conversion efficiency, and system operation modes [3]. - TSMC's efforts in enhancing manufacturing energy efficiency will effectively reduce the carbon footprint of downstream products, strengthening their market competitiveness [3][5]. Group 2: Industry Context - The semiconductor manufacturing industry is one of the fastest-growing sectors in terms of energy consumption, with TSMC committing to 100% renewable energy usage and net-zero emissions by 2050 [5]. - The energy consumption of manufacturing equipment constitutes a significant portion of the total carbon footprint of chips, making efficiency improvements in core manufacturing equipment crucial for the industry's emission reduction goals [5]. Group 3: Impact of EUV Lithography - A single advanced EUV lithography machine consumes approximately 24,000 to 36,000 kWh per day, equivalent to the daily electricity usage of 1,200 to 1,800 average Chinese households [7][8]. - The reduction of 44% in power consumption translates to an estimated savings of about 13,200 kWh per day for TSMC, leading to annual savings in the range of hundreds of millions of kWh across its production lines [11]. Group 4: Future Outlook - TSMC plans to continue investing in green manufacturing technologies and integrate sustainability deeply into its technological blueprint, aiming to lead the global semiconductor industry towards a more efficient and cleaner future [5].

Core Insights - Huang Renxun views OpenAI not just as a customer but as a co-builder of the next-generation AI factory, predicting it could become a trillion-dollar tech company [6][8] - NVIDIA's investment in OpenAI could reach up to $100 billion, aimed at building a massive AI data center with 4-5 million GPUs, which is close to NVIDIA's entire shipment plan for 2025 [8] - AI is seen as a means to enhance human cognitive capacity, with 55%-65% of global GDP derived from human brainpower, suggesting significant economic growth potential if AI can double or triple productivity in various sectors [9][11] Investment and Infrastructure - Huang emphasizes the need for a robust infrastructure to support AI, which requires substantial energy and computational resources [11][12] - The demand for computational power is shifting from traditional software execution to real-time reasoning, necessitating a new approach to AI development [13][15] - NVIDIA's strategy focuses on maximizing output per watt of energy rather than competing on chip prices, establishing a long-term competitive advantage [20][22] Market Dynamics - NVIDIA's annual release of new architectures is essential to keep pace with the exponential growth in token generation, requiring continuous system upgrades [23][25] - Huang acknowledges the rise of self-developed AI chips by major companies but asserts that NVIDIA's general-purpose platform remains more adaptable in a rapidly changing AI landscape [28][29] - NVIDIA's ecosystem approach, including partnerships and open-source initiatives, positions it as a key player in the AI economy [31][32] Global AI Landscape - Huang discusses the concept of "sovereign AI," where nations recognize the importance of controlling their own AI systems and infrastructure [35][36] - He advocates for countries to build their own AI capabilities while leveraging existing models like OpenAI and Gemini [36] - Huang maintains a balanced view on competition with China, emphasizing the need for the U.S. to engage with the Chinese market while adhering to regulations [38] Future of Work - Huang addresses concerns about AI-induced job losses, suggesting that while some roles may be replaced, overall job opportunities will increase as AI enhances productivity [43] - He envisions a future where individuals have personal AI assistants that integrate into daily life, enhancing decision-making and productivity [46][48] - Huang encourages early participation in the AI revolution, suggesting that those who engage now will benefit the most as the industry evolves [49][51]