Market Overview - On September 12, the storage sector experienced a significant opening surge, with companies like Deminor hitting the daily limit, and others such as Dongxin Co., Jiangbolong, and others seeing substantial increases [1] Company Insights - **Xingsen Technology (002436.SZ)**: A pioneer in the domestic IC packaging substrate industry, with applications in mobile memory modules [3] - **Deminor (001309.SZ)**: Transitioning from losses and planning to reduce holdings; has established a comprehensive storage product matrix including mobile storage and solid-state drives [3] - **Dongxin Co. (688110.SS)**: One of the few companies in mainland China providing a full range of storage chips including NAND Flash, NOR Flash, and DRAM [3] - **Jiangbolong (301308.SZ)**: A leading domestic storage module company, with memory products covering mainstream types like LDDR memory modules [3] - **Beiyi Innovation (603986.SS)**: A leader in the storage field, focusing on the development and sales of memory chips and microcontrollers, ranking first globally in non-foundry Flash supply [3] Industry Trends - Major US storage leaders like Micron and SanDisk saw collective gains, with SanDisk announcing a 10% price increase across all channels for consumer products, indicating a potential new pricing cycle due to changing supply-demand dynamics [5][6] - The global DRAM market size increased approximately 17% quarter-on-quarter to $30.9 billion in Q2 2025, driven by the rise of generative AI and increasing demand for DRAM contracts and HBM shipments [6] - Nvidia's launch of the RubinCPX GPU aims to enhance speed in handling massive data, utilizing GDDR7 memory, which could lead to increased demand for DRAM [5][6] - The establishment of Changchun's third phase company marks a significant step in China's semiconductor industry, potentially breaking the monopoly of Samsung and SK Hynix in the NAND market [6]

Core Viewpoint - Citigroup expects Micron Technology (MU.US) to provide better-than-expected guidance for Q4 FY2025 earnings, driven by growth in DRAM and NAND product sales and price increases [1] Group 1: Earnings Guidance and Projections - Citigroup maintains a "Buy" rating on Micron Technology and raises the target price from $140 to $175 [1] - The revenue forecast for FY2026 is increased from $54.5 billion to $56 billion, with EPS expectations raised from $14.62 to $15.02, exceeding market expectations by 26% [1] - For Q1 FY2026, Citigroup projects revenue of $13 billion and EPS of $3.23 [1] Group 2: Market Demand and Trends - The storage market is recovering due to limited capacity and unexpected demand, particularly from the data center end market, which accounts for 55% of Micron's revenue [1] - There is a significant increase in demand from the AI sector, with capital expenditures for C25 chip packaging rising by $18 billion during the earnings season [1] Group 3: Stock Performance - As of Thursday's market close, Micron Technology shares rose by 7.55% to $150.57, with a year-to-date increase of 79% [2]

需求端，AI应用推动及数据中心、客户端、移动领域存储需求强劲；供应端，行业面临供应紧缩，一 方面NAND生产商将产量转向下一代节点致低密度芯片供应紧缺，另一方面部分供应商因财务困境短期 难扩产。 此外，近日，长存三期（武汉）集成电路有限责任公司宣布成立，据此前媒体消息称，长江存储正积极 扩产，其NAND今年有望在国内市场占据约三成的份额，而到2026年底长江存储产能有望达到全球 NAND闪存的15%。我们认为，国产存储产业链成长有望加速。 隔夜美股存储公司集体大涨，闪迪涨超14%，美光科技涨7.55%。 近日闪迪宣布将面向所有渠道和消费者客户的产品价格上调10%以上，且考虑未来几个季度进一步调 整。东方证券认为此举或将标志着新一轮定价周期的拉开帷幕，预计竞争对手也有望效仿。此次提价源 于供需格局转变，供需缺口或持续扩大。 *免责声明：文章内容仅供参考，不构成投资建议 *风险提示：股市有风险，入市需谨慎 公司方面，据东方证券、互动平台等表示， 普冉股份、万润科技、朗科科技等为国内存储产业公司。 精测电子：长江存储以及长鑫存储均为公司客户。 ...

Group 1: Economic Indicators - The core CPI growth in the U.S. for August met expectations, while initial jobless claims rose to a nearly four-year high, indicating a soft labor market [1] - Traders are increasingly betting on at least two interest rate cuts by the Federal Reserve within this year, with the prospect of a rate cut next week appearing solid [1] Group 2: AI Storage Stocks Performance - Major U.S. storage companies, Micron Technology (MU) and SanDisk (SNDK), saw significant stock price increases, with Micron rising over 7% and SanDisk soaring 14%, marking a rare seven-day consecutive rise [2][5] - This surge is supported by strong fundamentals, driven by a golden cycle in storage due to AI advancements [5] Group 3: Demand Drivers - AI is pushing a comprehensive upgrade in storage needs, with Citigroup predicting that Micron will provide guidance that exceeds market expectations in its upcoming earnings report [6] - The demand for DRAM and NAND is surging, with data center demand accounting for 55% of Micron's revenue, and cloud giants increasing capital expenditures by $18 billion, directly boosting AI storage demand [7] Group 4: Price Trends - The storage industry, which has faced price declines in recent years, is now entering a price increase phase, with Micron, Samsung, and Kioxia announcing price hikes [8] - Channel and consumer prices have already seen a 10% increase, and Morgan Stanley anticipates that NAND prices will exceed expectations in Q4 and continue to rise until 2026 [8] Group 5: Institutional Outlook - Citigroup has raised Micron's target price to $175, anticipating strong earnings, while Morgan Stanley has upgraded SanDisk to a "Top Pick," increasing its target price from $70 to $96 [9] - The rationale includes substantial orders from large cloud providers and a shift in the NAND market from consumer to enterprise, which is expected to enhance average selling prices [9] Group 6: Investment Logic - The current market uptrend is driven by both cyclical recovery and innovation, with supply constraints and surging demand leading to strong profitability in the industry [10] - Innovations such as QCL NAND, near-line SSDs, and next-generation high-bandwidth storage are expected to open up long-term growth opportunities, with AI and cloud demand being more certain and sustainable compared to previous cycles [10]

明日主题前瞻 1、固态硬盘 | 据IT之家报道，存储大厂铠侠（Kioxia）宣布与英伟达（NVIDIA）展开合作，打造读取速度较传统SSD快近百倍的固态硬盘，预计在2027年 之前推出。据悉，该产品将主要服务于生成式AI运算所需的服务器市场，预计将带来市场的新一代革新。有别于传统SSD需经由CPU连接GPU的方式，这种 新型SSD将达成GPU的直接连接与数据交换。这种直连模式能大幅提升数据传输效率，对于AI运算至关重要。 点评：分析认为，SSD直连GPU技术作为AI算力基础设施的重要创新，其市场前景呈现技术驱动与需求爆发的双重红利。2024年中国智能算力规模达725.3 EFLOPS，同比增速74.1%，预计2025年智能算力规模将突破1037.3 EFLOPS。SSD直连技术通过提升数据吞吐量，成为满足算力需求的关键路径。SSD直连 GPU技术凭借其在算力效率、成本控制和场景适配性上的优势，将成为AI算力市场增长的核心驱动力。预计未来5年，全球AI存储市场规模将突破千亿美 元，中国凭借政策支持、技术迭代和应用场景优势，有望占据全球超3成的市场份额。 5、我国在研新药数量占全球数量的比例超过20%，跃居全球新 ...

在人工智能火热的当下，关于算力的讨论已经充满各大报端。但其实作为这轮AI崛起的关键组 成，存储在其中扮演的角色不容忽视。熟悉人工智能原理的读者都知道，只有将大量准确的数据提 供给大模型训练，才能获得更好的AI应用体验，而这正是闪存发力的地方。 IDC其发布的《理解人工智能数据周期（AI Data Cycle）和闪存在各行业的应用》报告中也直 言，尽管近年来有关AI如何赋能企业的研究主要聚焦于大语言模型（LLMs）以及和图形处理单元 （GPUs）有关的算力方面，但同样重要的是，相关机构也需同步提升其数据存储与管理能力。这 一部分正是支撑AI系统高效处理并应用海量数据、实现先进分析与智能决策的关键所在。 那么，人工智能究竟需要怎样的闪存？ 如果您希望可以时常见面，欢迎标星收藏哦~ 人工智能中的存储 如前文所说，在人工智能中，存储的数据不仅存在于云端，端侧也是需要发力的地方。尤其是随着 AI应用不断扩展，对战略性数据管理的需求也将持续上升，进而使得存储解决方案的整合与优化 成为面向未来的企业所需优先考虑的核心任务。在闪迪看来，这时候就需要一个将一堆原始数据转 化为能为大模型使用的"知识"的人工智能数据周期（AI D ...

Core Viewpoint - Recent market fluctuations have led to significant rebounds in sectors such as computing power, PCB, and optical communication, driven by various factors including Oracle's substantial revenue growth and Nvidia's new GPU announcement [5][6][10]. Group 1: Market Rebound Factors - The rebound in the market is attributed to previous declines prompting investors to buy back in [5]. - Oracle reported a staggering 359% increase in its Remaining Performance Obligations (RPO), reaching $455 billion, indicating strong future revenue prospects [5][21]. - Nvidia announced the launch of a new GPU designed for large-scale context processing, which is expected to enhance computing capabilities [6]. Group 2: Sector-Specific Insights - Analyst Sun Xiaoya highlighted that the development of AI is driving demand for high-end PCB copper foil, suggesting that domestic manufacturers could benefit from industry growth [7]. - Citigroup expressed optimism about the optical communication sector, raising target prices for companies like Zhongji Xuchuang and Xinyi Communication to 569 RMB and 472 RMB respectively [8]. - Citigroup's analysis indicates that the recent rise in optical communication stocks is primarily driven by upward revisions in earnings forecasts rather than irrational valuation adjustments [9]. Group 3: Industry Demand Forecasts - Citigroup maintained its forecast for industry shipment volumes in 2026 while raising its demand estimates for 2027, projecting a year-on-year growth of 42% for 800G and 1.6T modules [10]. - The anticipated increase in AI investments by U.S. cloud service providers, expected to reach $600 billion by 2028, is likely to sustain demand for optical modules [10]. - Optimistic projections from Broadcom regarding ASIC growth and new customer orders further support the upward revision of demand forecasts [10]. Group 4: Economic Indicators - The recent CPI data showed a year-on-year decline of 0.4% in August, while the core CPI rose by 0.9%, indicating mixed inflationary pressures [15]. - The PPI's year-on-year decline narrowed to 2.9%, with upstream industries driving a month-on-month increase from -0.1% to 0.1% [18]. - Citigroup predicts that CPI readings may remain low in September but could rebound towards the end of the year, influenced by food prices, particularly pork [20].

Group 1 - Jiangbolong (SZ 301308) announced it will accept a research meeting with CICC on September 8, 2025, from 10:30 to 11:30 AM, with investor relations manager Huang Qi and senior supervisor Su Yangchun participating in the reception [1] - For the year 2024, Jiangbolong's revenue composition is 100.0% from the storage industry [1] - As of the report, Jiangbolong's market capitalization is 38.6 billion yuan [1]

Core Insights - The 2025 Open Data Center Conference (ODCC) held in Beijing focused on AI transformation and featured experts from data center, computing power, and storage sectors [1] - Jiangbolong launched the innovative enterprise-level memory product SOCAMM2, designed to address challenges posed by various AI workloads [1][7] - The company emphasized the integration of AI technology in storage systems to enhance performance and reliability [12][16] Group 1: Product Launch and Features - SOCAMM2 is based on LPDDR5/5x and features a modular design with a compact size of 14×90 mm, achieving a maximum capacity of 256GB and a data transfer rate of 8533 Mbps [7][10] - The product offers a significant energy efficiency improvement, consuming only one-third of the power of standard DDR5 RDIMM while increasing bandwidth by 2.5 times [7][10] - SOCAMM2 is positioned to complement HBM applications, balancing capacity, power consumption, and flexibility in high-performance computing environments [10] Group 2: AI Storage Solutions - Jiangbolong's enterprise-level SSDs are critical for AI clusters, facing performance bottlenecks and reliability challenges as data scales from GB to PB [12][14] - The company introduced intelligent storage solutions that utilize smart load sensing technology to optimize caching strategies and data layout, significantly enhancing system performance [14][16] - Machine learning models are employed to analyze real-time parameters for predictive maintenance, improving reliability and reducing operational complexity [16][17] Group 3: Future Developments - Jiangbolong plans to launch enterprise-level SSDs supporting QLC and PCIe Gen5, incorporating more intelligent innovations and advanced algorithms for AI data centers [25] - The company is establishing a comprehensive storage service model that integrates R&D, testing, and manufacturing to meet diverse AI application needs [21]

Group 1 - The storage industry maintained high prosperity in Q2, with DRAM industry revenue reaching $31.63 billion, a quarter-on-quarter increase of 17.1%, and NAND Flash top five brands' revenue increasing by 22.0% to $14.67 billion [1] - Demand from AI servers and data centers is expected to keep the storage industry prosperous in Q3, with DRAM's prosperity expected to be higher than NAND Flash due to its closer ties to the AI industry chain and higher demand for DDR5 and HBM [1] - The integrated circuit ETF (159546) tracks the integrated circuit index (932087), which selects listed companies involved in design, manufacturing, packaging testing, and related materials and equipment to reflect the overall performance of the integrated circuit industry [1] Group 2 - Investors without stock accounts can consider the Guotai Zhongzheng All Index Integrated Circuit ETF Initiated Link C (020227) and Guotai Zhongzheng All Index Integrated Circuit ETF Initiated Link A (020226) [1]