以下文章来源于Omdia ，作者Omdia Omdia . Omdia隶属于Informa TechTarget，是一家专注于技术研究与咨询的机构。通过深刻的科技市场洞察力和 可操作的建议，Omdia帮助组织做出明智的增长决策。 如果您希望可以时常见面，欢迎标星收藏哦~ 根据Omdia发布的半导体产业观察，AI应用模型正在中国各垂直行业广泛落地，标志着边缘AI时 代的到来。众多大语言模型(LLM)正在各行业积极部署垂直应用模型。具备边缘推理能力的数字终 端将快速增长，成为中国半导体产业扩张的重要驱动力——尤其是成熟工艺技术领域。2025年四 季度最新数据显示，2026年中国半导体市场预计增长31.26%，市场规模将达到5465亿美元。 《半导体应用领域市场预测工具(AMFT)- 中国地区(2Q25)》预计中国半导体市场在2025年增长 16.17%，2026年增长13.63%。在《半导体应用领域市场预测工具(AMFT)- 中国地区(4Q25)》报 告中，2025年四季度的更新将预测上调为2025年增长21.63%、2026年增长31.26%。 2025年第四季度预测中，存储市场规模显著上调。与二季度版本相比， ...

根据Omdia发布的半导体产业观察，AI应用模型正在中国各垂直行业广泛落地，标志着边缘AI时代的到 来。众多大语言模型(LLM)正在各行业积极部署垂直应用模型。具备边缘推理能力的数字终端将快速增 长，成为中国半导体产业扩张的重要驱动力——尤其是成熟工艺技术领域。2025年四季度最新数据显 示，2026年中国半导体市场预计增长31.26%，市场规模将达到5465亿美元。 2025年第四季度预测中，存储市场规模显著上调。与二季度版本相比，对2026年存储市场的预测在新版 本中增长了62.8%、对2027年的预测增长了53%、2028年的增长了36%、2029年的增长了25.8%. 随着2025年第四季度开始的全球AI大基建的狂潮，中美双方AI对于内存的消耗量是指数级增长，数据 中心的大规模部署带来的高性能存储芯片(HBM)需求暴涨。 供应受限而需求增加，在供需失衡的大背景下，内存价格经历着巨大的波动，并且根据Omdia的预测， 这样的紧缺会延续到2027-2028年。 随着AI发展势头日益明确，存储芯片供需进一步失衡，全球短缺状况持续蔓延。中国高端存储芯片 (HBM、高端DRAM、NAND)自给率仍然较低，约90 ...

截至今日（1月26日）收盘，上证指数报收4132.61点，下跌0.09%；深证成指收于14316.64点，下跌 0.85%。创业板指下跌0.91%；科创50指数下跌1.35%。 不含未开板新股，今日可交易A股中，上涨个股有1604只，占比29.36%，下跌个股有3771只。其中，收 盘封死涨停的有77只，跌停股有42只。另外，53股封板未遂，整体封板率为59.23%。 德邦证券指出，贵金属与有色金属表现强势。具体来看，国际商品中"贵金属领涨，银价飙升，锡、 金、镍价大涨"。国内市场同样呈现"贵金属、有色、能化强势，白银领涨"的格局。 从封单力度（封单量占流通A股比例）来看，中农联合、湖南黄金、天奇股份等力度较大，分别为 14.40%、11.25%、7.92%。 | | | | 1月26日涨停封单资金居前的个股 | | | --- | --- | --- | --- | --- | | 代码 | 简称 | 涨停封单额 (亿元) | 涨停板情况 | 涨停原因 | | 002155 | 湖南黄金 | 44.43 | 首板 | 重组复牌+黄金预增+湖南国资 | | 601615 | 明阳智能 | 8.71 | 3连板 ...

截至今日(1月26日)收盘，上证指数报收4132.61点，下跌0.09%；深证成指收于14316.64点，下跌 0.85%。创业板指下跌0.91%；科创50指数下跌1.35%。 德邦证券指出，贵金属与有色金属表现强势。具体来看，国际商品中"贵金属领涨，银价飙升，锡、 金、镍价大涨"。国内市场同样呈现"贵金属、有色、能化强势，白银领涨"的格局。 从封单力度(封单量占流通A股比例)来看，中农联合、湖南黄金、天奇股份等力度较大，分别为 14.40%、11.25%、7.92%。 | | | | 1月26日涨停封单资金居前的个股 | | | --- | --- | --- | --- | --- | | 代码 | 简称 | 涨停封单额 (亿元) | 涨停板情况 | 涨停原因 | | 002155 | 湖南黄金 | 44.43 | 首板 | 重组复牌+黄金预增+湖南国资 | | 601615 | 明阳智能 | 8.71 | 3连板 | 收购芯片+太空光伏+风电龙头 | | 002009 | 天奇股份 | 7.32 | 首板 | 人形机器人+锂电循环 | | 300017 | 网宿科技 | 4.65 | 首板 | 边缘AI ...

Core Insights - TSMC is actively expanding its advanced packaging capacity to secure orders for Apple's next-generation smartphone chips, particularly the A20 series for the upcoming iPhone 18, which will utilize 2nm process technology [1][2] - The transition from InFO to WMCM packaging technology is expected to enhance interconnect density, yield, and thermal management capabilities, addressing future demands for edge AI computing [2] - TSMC's aggressive capacity expansion includes upgrading existing InFO equipment and establishing a new WMCM production line at its AP7 facility in Chiayi, Taiwan, with plans to double monthly capacity from 60,000 wafers by the end of 2026 to over 120,000 by 2027 [3] Capacity Expansion and Production Line Adjustments - TSMC is implementing a dual-track capacity expansion strategy, upgrading existing InFO equipment while building a new WMCM production line [3] - The company is also reconfiguring mature process capacity to support advanced packaging, with potential transformations of its Fab 18 P9 into an advanced packaging facility and expanding 40nm and 65nm capacities at Fab 14 for critical components [3] - Testing requirements for 2nm chips are increasing, with TSMC actively procuring hundreds of final test and system-level test machines to meet the growing demand [3] Demand Drivers from Apple's Product Line - The expansion of TSMC's capacity is driven not only by the iPhone but also by Apple's plans to implement 2nm technology across its product lines, including MacBook M series chips and R2 chips for head-mounted devices [4] - The collaboration between Apple and Google to enter the AI space is expected to further increase demand for high-performance chips and advanced packaging solutions [4] - Despite concerns over potential price increases for 2nm chips and memory, the competition for AI computing power from cloud to edge has established a growth logic for TSMC and the semiconductor ecosystem [4]

Core Viewpoint - The article discusses the IPO of Shenzhen-based storage chip company, ChipX, which aims to capitalize on the growing demand for storage driven by AI and the recovery of the storage industry. The company faces challenges from market competition and supply chain fluctuations while leveraging its technological strengths in the code-type flash memory segment [1][4][7]. Company Overview - ChipX was established in April 2014 and focuses on code-type flash memory, which is critical for system operations. The company operates under a Fabless model, offering a product line that spans from 1Mbit to 8Gbit, and has diversified into analog chips and MCUs [3]. - According to data from Zhaoshang Consulting, ChipX ranks sixth globally among fabless code-type flash manufacturers by revenue in 2024, with fourth place in SLC NAND Flash and fifth in NOR Flash [3]. Market Performance - The global flash memory market size decreased from $58.5 billion in 2020 to $40.9 billion in 2023, impacting ChipX, which reported a 33.3% revenue drop in 2024 and incurred losses exceeding 51 million yuan over two years. However, a recovery began in 2025, with the market size rebounding to $68.4 billion, a 67% year-on-year increase [4][5]. - In the first nine months of 2025, ChipX achieved a profit of 380 million yuan, with a revenue increase of approximately 10% and a gross margin of 18.8% [4]. Industry Trends - The code-type flash segment's market share increased from 6.1% in 2020 to 7.2% in 2024, with projections to reach 9.3% by 2030, driven by the rapid development of edge AI [5]. - Despite the industry's recovery, ChipX faces intense competition, with pricing strategies from peers affecting its pricing power. The company's revenue growth has been relatively slow, with a 10% increase in the first nine months of 2025 [6]. Challenges and Risks - ChipX's R&D expenditures have fluctuated, decreasing from 85.2 million yuan in 2023 to 33.3 million yuan in the first nine months of 2025, which raises concerns about maintaining competitiveness in a rapidly evolving industry [6]. - The company's reliance on external suppliers for wafer manufacturing and testing has increased, with procurement from five major suppliers rising from 75.4% to 83.2% between 2023 and 2025. This dependency, coupled with high wafer-related costs, poses risks to profit margins [6]. - The anticipated surge in storage chip prices due to rising demand for AI CPUs and memory could lead to increased costs for ChipX, further squeezing profit margins [6]. Conclusion - ChipX's decision to pursue an IPO aligns with the recovery of the global storage industry and the deepening of domestic alternatives. Its technological expertise and market position in the code-type flash segment are significant assets. However, challenges such as slow growth, reduced R&D investment, and supply chain concentration must be addressed for the company to solidify its market presence [7].

Core Insights - The rise of AI PCs, capable of independently processing AI tasks, is becoming increasingly significant as the cost of computing power for large models continues to rise [2][3] - Intel's CEO, Pat Gelsinger, announced the launch of the third-generation Core Ultra processor using Intel 18A technology at CES 2026, which is expected to enhance performance and efficiency in AI PCs [2][4] - The integration of CPU, GPU, and NPU in Intel's AI PC processors allows for more local processing of AI tasks, attracting interest from consumers and cloud computing companies [3][4] Group 1: AI PC Development - As of January 2026, over 100 million AI PCs equipped with Intel processors have been shipped globally, with their combined computing power equivalent to that of 40 data centers [4] - Intel has partnered with over 200 independent software vendors (ISVs) to develop more than 500 AI applications based on AI PCs, indicating a growing ecosystem [4][10] - The AI PC is seen as a resource that is currently undervalued, with potential to alleviate financial pressures on software companies by allowing local processing of AI tasks [3][4] Group 2: Technological Advancements - The Intel 18A process technology improves energy efficiency by 15% and increases chip density by 30%, allowing for more features to be integrated into smaller chip areas [7][10] - The third-generation Core Ultra processor is designed to achieve 40% better performance at the same power consumption compared to previous generations [7] - AI PCs are expected to enhance user experiences in various applications, including gaming, by providing smoother performance through AI-generated frames [9][10] Group 3: Collaboration with Cloud Services - Intel is exploring the collaboration between AI PCs and cloud services, with the aim of providing users with greater control, lower costs, and improved privacy [20][21] - The current strategy involves a static allocation of tasks between local and cloud resources, with ongoing efforts to create a more dynamic system [18][20] - Applications that benefit from the collaboration include content creation tools, where local processing can enhance performance and reduce reliance on cloud resources [19][20] Group 4: Future Directions - The company aims to drive the large-scale deployment of AI PCs in 2026, focusing on optimizing applications to enhance user experience [9][20] - Intel is committed to supporting developers by providing tools and frameworks that facilitate the integration of AI capabilities into existing software [12][22] - The company recognizes the importance of local AI capabilities and is actively working to strengthen its partnerships with OEMs and ODMs, particularly in the Chinese market [23]

专题：2026年度国际消费电子展(CES) 新浪科技讯 2026 年美国CES展会期间，边缘 AI 芯片与解决方案厂商 Kneron（耐能）集中展示其覆盖 消费电子、智能家居、智能交通、工业物联网及企业级计算的全栈边缘AI 产品矩阵，系统性推动 AI 从 云端向边缘侧迁移。 耐能表示，其技术核心在于本地计算、超低时延、低功耗与高可靠性，通过"数据不出本地、智能运行 在边缘、集中统一管理"的架构，为全球用户提供更安全、更私密的 AI 体验。 随着生成式 AI 加速进入企业应用阶段，Kneron 在 CES 上推出全栈边缘AI 系统，覆盖从试点部署到企 业级规模化落地的完整路径。 专题：2026年度国际消费电子展(CES) 新浪科技讯 2026 年美国CES展会期间，边缘 AI 芯片与解决方案厂商 Kneron（耐能）集中展示其覆盖 消费电子、智能家居、智能交通、工业物联网及企业级计算的全栈边缘AI 产品矩阵，系统性推动 AI 从 云端向边缘侧迁移。 该系统基于 Kneron 自研 NPU 架构，整合 AI 芯片、安全操作系统、推理引擎与统一管理平台，构成端 到端的边缘智能基础设施。 耐能表示，其技术核心在于本地 ...

Core Insights - The artificial intelligence industry is currently engaged in a "GPU race," with a focus on cloud-based AI applications, but there is a call to shift attention towards physical AI and its potential breakthrough moments [1][5][16] - Dr. Allen Yang emphasizes that while AI has made significant strides with models like AlphaGo, physical AI is still awaiting its own "watershed moment" due to unique challenges rooted in the complexities of the physical world [2][6][12] Group 1: Challenges in Physical AI - Physical AI lacks comprehensive training data for extreme scenarios, unlike language models that can leverage vast internet data [2][13] - Real-time decision-making with millisecond-level latency is critical for applications like autonomous driving, where delays can lead to failures [2][14] - Many cutting-edge scenarios lack reliable cloud connectivity, necessitating the use of edge AI deployed locally on devices [2][15] Group 2: Innovations and Competitions - The Berkeley AI Racing Team has achieved significant milestones, including a top speed of 163 miles per hour in autonomous driving competitions, showcasing the need for complex real-time perception and planning [4][18] - The upcoming Tianmen Mountain humanoid robot challenge aims to test robots' mobility and decision-making in unstructured terrains, further pushing the boundaries of physical AI [4][29] - The collaboration with nine universities in China for the Tianmen Mountain challenge highlights the importance of interdisciplinary cooperation and real-world experience in advancing physical AI [4][26] Group 3: Future Directions - The focus on physical AI is expected to grow, with the potential for new breakthroughs that could redefine the field, similar to past milestones in AI history [2][18][29] - The upcoming competitions and challenges are designed to foster innovation and collaboration among institutions, aiming to discover the next "AlphaGo moment" in physical AI [25][29]

Core Viewpoint - The China Securities Regulatory Commission (CSRC) has issued supplementary material requirements for 10 companies, including Dongshan Precision (002384.SZ), which is preparing for a listing on the Hong Kong Stock Exchange [1][2]. Group 1: Regulatory Requirements - Dongshan Precision is required to provide detailed explanations regarding its medical device production, commission agency, IT and socio-economic consulting services, solar product system installation and sales, and solar industrial technology research consulting services, including necessary qualifications and licenses [1]. - The company must clarify whether its business scope involves areas listed in the "Negative List of Special Management Measures for Foreign Investment Access (2024 Edition)" [1]. - Dongshan Precision is asked to disclose the pledge status of shares held by its controlling shareholders and actual controllers before and after the issuance, and whether this could lead to significant ownership disputes or changes in control [1]. - The company needs to provide details on the regulatory procedures for overseas investments and foreign exchange registration related to its overseas subsidiaries, along with a conclusive opinion on compliance [1]. Group 2: Company Profile - Dongshan Precision is recognized as the world's largest supplier of edge AI device PCBs, with products primarily used in smartphones, personal computers, automobiles, industrial applications, and IoT devices, and is rapidly expanding its capabilities into the data center terminal market [2].