Core Viewpoint - The A-share Sci-Tech Innovation Board is experiencing a strong performance, particularly in the semiconductor sector, driven by policy support, AI catalysis, and the trend towards domestic production [1][6]. Industry Summary - The Sci-Tech Innovation Board is seeing a resurgence in activity, with the Sci-Tech Chip 50 ETF (588750) rising over 2% and trading volume increasing significantly [1]. - Key stocks in the semiconductor sector, such as Stmicroelectronics and Cambricon, are leading the gains, with Stmicroelectronics up over 7% and Cambricon up 4% at one point [5]. - The government is emphasizing the importance of technological innovation, aiming to strengthen the position of enterprises in this area and accelerate breakthroughs in key technologies [3]. - Recent policies, including the "6+1" measures proposed at the Lujiazui Forum, aim to deepen reforms in the Sci-Tech Innovation Board [4]. Company Summary - Stmicroelectronics announced a cash dividend of 60.2 million yuan, distributing 1.5 yuan for every 10 shares [2]. - Cambricon has stated that its general-purpose AI chips are highly adaptable, supporting mainstream open-source AI models, which enhances its competitive edge in AI computing power [2]. - The semiconductor sector is expected to see an increase in IPOs for domestic computing chip companies, with companies like Suiruan Technology and Biran Technology likely to go public in the second half of the year [6]. Market Trends - The semiconductor sector is benefiting from a combination of policy support, AI-driven demand, and a strong push for domestic production capabilities [6]. - The demand for AI chips is expected to remain robust, driven by the need for pre-training and inference capabilities in AI models [6]. - The domestic semiconductor industry is poised for growth, with significant potential for expansion in computing power and advanced manufacturing processes [6].

Core Insights - OpenAI has begun utilizing Google-made AI chips, specifically Tensor Processing Units (TPUs), for its products including ChatGPT, marking a significant shift from its previous reliance on NVIDIA chips [1][2] - This collaboration indicates OpenAI's strategy to diversify its suppliers, as it has historically depended on NVIDIA for both training AI models and executing inference calculations [1] - OpenAI anticipates that leasing TPUs from Google Cloud will help reduce inference-related costs, potentially positioning TPUs as a cheaper alternative to NVIDIA GPUs [1] Group 1 - The partnership between OpenAI and Google represents a surprising collaboration between two major competitors in the AI field, aimed at addressing OpenAI's growing computational needs [1] - Morgan Stanley has released a report supporting Google, suggesting that if the agreement is confirmed, it would reflect Google's confidence in its long-term search business and accelerate the growth of Google Cloud, with a valuation exceeding 18 times [1] Group 2 - For Google, this collaboration coincides with its efforts to expand the external availability of its self-developed TPUs, which were previously used mainly for internal projects [2] - The partnership has attracted interest from other tech giants like Apple and competitors of ChatGPT, indicating a broader market interest in Google's TPU technology [2] - However, Google has reportedly not leased its most powerful TPU models to OpenAI, indicating a strategy to reserve its advanced versions for internal projects, including its own Gemini large language model [2]

Core Insights - The report highlights the recent developments in the financial market, including the performance of major indices and sector trends [4][12]. - It emphasizes the impact of government policies on various industries, particularly in the context of digital assets and environmental initiatives [6][8]. Financial Market Overview - The Shanghai Composite Index closed at 3424.23, down 0.7%, while the Shenzhen Component Index rose by 0.34% to 10378.55 [4]. - The report notes a total trading volume of 15,411 billion, with 3,227 stocks rising and 1,662 falling, indicating a slight contraction in market activity [12]. Industry Developments - Hong Kong has introduced a new digital asset policy, accelerating the ecosystem for stablecoin applications, with related stocks including Jingbeifang (002987) and Sifang Jingchuang (300468) [8]. - The offshore engineering equipment industry is being catalyzed by new policies, with companies like Hailanxin (300065) and Yaxing Maolian (601890) positioned to benefit [8]. - Platinum prices have surged to a 10-year high, significantly outpacing gold, with related stocks such as Gebijia (835438) and Guiyan Platinum (600459) highlighted [8]. Company Focus - Wenkang New Energy (688779) is focusing on the development of solid-state battery materials, achieving cumulative shipments of over 100 kilograms for some products [10]. - Zhongji Renjian (301508) has established a testing facility for new energy batteries, capable of meeting the demands for solid-state battery testing [10]. - Changliang Technology (300348) is actively researching digital solutions related to stablecoins [10]. Market Review and Outlook - The report indicates that the market is experiencing a slight contraction, with a notable performance in industrial metals and technology sectors, while banking and oil sectors are facing declines [12][13]. - The report suggests that the market is showing resilience, with strong support levels for the Shanghai Composite Index, indicating a lower probability of significant downturns in the short term [12]. - There is a strong interest in technology sectors such as chips and lithium batteries, suggesting a continued aggressive investment approach in these areas [13].

公众号记得加星标⭐️，第一时间看推送不会错过。 来源：内容 编译自techxplore 。 纽约大学坦登工程学院的研究人员创建了 VeriGen，这是第一个成功训练生成 Verilog 代码的专用人 工智能模型，Verilog 代码是一种描述芯片电路如何运作的编程语言。 该研究刚刚获得了 ACM 电子系统设计自动化学报 2024 年度最佳论文奖，肯定了它在自动化创建传 统上需要深厚技术专业知识的硬件描述语言方面的重大进步。 "通用人工智能模型不太擅长生成 Verilog 代码，因为互联网上可供训练的 Verilog 代码非常少，"该 研究的主要作者、纽约大学坦顿分校电气与计算机工程系 (ECE) 教授 Siddharth Garg 表示。Garg 教授同时还在纽约大学无线项目和纽约大学网络安全中心 (CCS) 任教。"这些模型在 GitHub 上表现 良好的编程语言（例如 C 和 Python）上往往表现良好，但在 Verilog 等表现较差的语言上表现往往 要差得多。" 与Garg一起，由纽约大学坦顿分校博士生、博士后研究员以及教员Ramesh Karri和Brendan Dolan- Gavitt组成的团 ...

极米科技 20250627 摘要 极米科技国内市场毛利率约为 30%，海外市场更高，近期已接近 45%。毛利率提升主要得益于 TI DLP 芯片价格下降 40%及光机结构方 案设计迭代和零部件国产化替代。 国内市场需求承压，但利润端持续好转，规模保持稳定。海外市场持续 增长，欧洲市场去年收入 6 亿人民币，增速超 40%，美国市场收入近 2 亿人民币，增速相近。日本市场因渠道调整拖累海外整体增长，海外市 场增速约为 20%以上。 车载业务聚焦智能座舱和智能大灯，已与华为智选车合作落地，预计今 年贡献 3-4 亿人民币收入。目前车载业务处于投入期，预计收入达 10 亿人民币左右时实现稳定盈利，目标后年毛利率达 20%，净利率达 10%左右。 商用投影业务 6 月底量产，7 月开始出货，目标占据国内 100 亿人民币 商用市场 10%-20%份额。前期规划两款商用 SKU，亮度分别为 5,000 流明和 8,000-10,000 流明，预计今年实现数千万人民币收入，毛利率 预计 40%-50%，规模化后稳态净利率目标 15%-20%。 Q&A 极米科技家用产品线的毛利率结构如何，以及国内外市场的毛利率表现是否存 ...

摘要 光模块市场景气度持续上修，受益于互联网云厂商规划新发展及乐观的 投关交流，预计 2026 年 800G 和 1.6T 光模块需求量将分别增至 4,000 万只和 600-800 万只，主要驱动力来自 CSB 厂商架构变化和市 场需求增加。 旭创保持市场领先地位，新易盛份额快速提升，与 Meta、亚马逊等合 作紧密，并积极拓展英伟达等客户。预计 2025 年龙头公司收入可达 360-370 亿人民币，2026 年达 440-450 亿人民币，净利润率提升潜 力巨大。 二线光模块厂商如索尔思、联特科技和剑桥科技有望分享市场份额，行 业贝塔不断上修。国内互联网音响加大投入，华工科技和光迅科技等国 内龙头企业受益，预计 2025 年下半年业绩强劲。 光模块跳线市场潜力巨大，800G 光模块跳线需求量约为 2000 万支， 市场空间达数百亿人民币。泰胜光博创、亨通光电、爱康得等厂商有望 受益，业绩弹性大。 Q&A 近期光模块行业的市场表现如何？ 近期光模块行业表现突出，尤其是在 6 月份通信行业中排名第一。头部公司如 旭创、新易盛、太辰光和博创的涨幅显著，达到百分之四五十。整个互联网云 场对 ASEC 的重视程度 ...

以下文章来源于半导体行业观察 ，作者杜芹DQ 半导体行业观察 . 半导体深度原创媒体，百万读者共同关注。搜索公众号：半导体芯闻、半导体产业洞察，阅读更多原创 内容 本文来自微信公众号： 半导体行业观察 （ID：icbank） ， 作者：杜芹DQ，原文标题：《这类芯 片，寒冬已过？》，题图来自：AI生成 去年摩根士丹利发布的《寒冬将至》报告曾为半导体行业浇下一盆冷水。然而，2025年上半年以 来，全球芯片市场正悄然释放出复苏信号。尤其是内存芯片领域，从通用DRAM到高带宽内存 （HBM） ， 市场价格回升、库存去化、订单恢复，似乎正在书写"春天将至"的新篇章。 信号一：韩国通用DRAM出口激 增 韩国作为全球DRAM产业的核心基地，其出口数据是观察行业周期的风向标。自今年2月起，韩国 DRAM出口额结束此前的同比下滑态势，连续四个月实现两位数增长：3月增长27.8%，4月增长 38%，5月增长36%，6月前20天增长25.5%。 （来源：Daeun Lee） 这些数据与"半导体寒冬"的主流叙事背道而驰。相反，它们表明，通用DRAM市场 （主要应用于PC 和基础服务器） 正快速形成供需紧张的格局。据韩国KED A ...

Group 1 - The company, Fortior, plans to globally offer 16,299,500 H-shares, with 1,630,000 shares available in Hong Kong and 14,669,500 shares for international offering, with the pricing expected to be up to HKD 120.50 per share [1] - The company specializes in the design and development of BLDC motor drive control chips, establishing a strong market position in the industry [2] - As of December 31, 2023, the company holds a 4.8% market share in China's BLDC motor control and drive chip market, ranking sixth among competitors [2] Group 2 - The company has entered into cornerstone investment agreements, with cornerstone investors agreeing to subscribe for shares totaling approximately USD 112 million (or about HKD 879 million) [3] - Assuming the maximum offer price of HKD 120.50 per share, cornerstone investors will subscribe for a total of 7,295,700 shares [3] - The estimated net proceeds from the global offering are approximately HKD 1.846 billion, with planned allocations for R&D, product expansion, overseas market promotion, strategic investments, and working capital [4]

Group 1 - The company, Fortior, is planning to conduct an initial public offering (IPO) from June 30 to July 4, 2025, with a global offering of 16.2995 million H-shares, where 10% will be offered in Hong Kong and 90% internationally, with a maximum offer price of HKD 120.50 per share [1] - Fortior is a chip design company specializing in the design and development of BLDC (Brushless Direct Current) motor drive control chips, establishing a strong market position in the industry [1] - According to Frost & Sullivan, BLDC motors are widely used due to their high efficiency, low power consumption, high control precision, and low noise compared to traditional motors [1] Group 2 - As of December 31, 2023, Fortior holds a 4.8% market share in China's BLDC motor control and drive chip market, ranking sixth and being the only Chinese company among the top ten players [2] - The company's revenue increased from RMB 323 million in 2022 to RMB 411 million in 2023, representing a growth of 27.4%, and is projected to further increase to RMB 600 million in 2024, a growth of 45.9% [2] - Fortior's gross profit rose from RMB 185 million in 2022 to RMB 219 million in 2023, an increase of 18.1%, and is expected to reach RMB 316 million in 2024, a growth of 44.5% [2] Group 3 - Assuming the maximum offer price of HKD 120.50 per share, the estimated net proceeds from the global offering will be approximately HKD 1.846 billion [3] - The company plans to allocate the net proceeds as follows: approximately 34% for enhancing R&D and innovation capabilities, 10% for expanding the product portfolio, 16% for expanding overseas sales networks, 30% for strategic investments or acquisitions, and 10% for working capital and general corporate purposes [3] - Fortior has entered into cornerstone investment agreements with investors such as Taikang Life, Baoyin, and 3WFund, committing to subscribe for shares totaling USD 112 million [3]

Group 1 - TSMC is expected to produce Apple's A19 processor using 2nm technology in the second half of this year, while competitors Samsung and Intel are facing challenges in yield and progress, respectively [1] - Japan's 2nm technology has reportedly entered trial production and is expected to achieve mass production by next year, positioning Japan to compete directly with TSMC in chip manufacturing [3] - Japan established a chip company, Rapidus, in 2022 with support from major Japanese firms and the government, aiming to revitalize its semiconductor industry, which once held a significant global market share [3] Group 2 - The decline of Japan's semiconductor industry was accelerated by U.S. policies, leading to a loss of market dominance, with companies like Elpida falling behind [5] - The U.S. is seeking to diversify its chip supply sources, as TSMC currently produces 70% of advanced chips for American companies, and TSMC is raising prices for its upcoming 2nm technology by 50% [5] - Japan's Rapidus aims to revitalize its semiconductor sector, aligning with U.S. interests to create competition against TSMC [7] Group 3 - Japan's rapid advancement in 2nm technology is supported by IBM's technological expertise, which has historically played a significant role in chip manufacturing [7] - Japan has acquired advanced 2nm EUV lithography machines from ASML, becoming the third company after Intel and Samsung to do so, enhancing its production capabilities [7] - TSMC aims to maximize the value of its first-generation EUV lithography machines while Japan focuses on quickly achieving mass production of 2nm technology, leveraging its advantages in chip materials [9]