证券日报网讯1月9日，江波龙（301308）在互动平台回答投资者提问时表示，公司提供从消费级、工规 车规级，直至企业级服务器存储等多样化存储产品，并服务于众多优秀头部客户。公司是国内少数具 备"eSSD+RDIMM"产品设计、组合以及规模供应能力的企业，适配数据中心场景对存力的高性能需 求，公司将持续深化与多个领域知名客户的合作。 ...

Core Viewpoint - The storage industry is experiencing a "super cycle" driven by AI demand, leading to significant expansion plans from major companies like Jiangbolong and Demingli, with expectations of continued price increases into the first half of next year [1][5]. Group 1: Expansion Plans - Jiangbolong has announced a fundraising plan of 3.7 billion yuan, while Demingli has proposed a 3.2 billion yuan expansion plan [1]. - Jiangbolong's fundraising will focus on three key areas: storage product application technology development, NAND Flash main control chip design, and storage chip packaging and testing [1]. - Demingli aims to enhance its full-chain technical capabilities from storage main control chip R&D to storage module delivery through its fundraising projects [1]. Group 2: AI Market Focus - Jiangbolong's fundraising will support high-end storage development and industrialization projects aimed at the AI sector, with a total investment of 930 million yuan for AI-related storage solutions [3][4]. - The demand for storage in AI servers is significantly higher than in traditional servers, with DRAM usage approximately eight times and NAND Flash usage three times that of conventional servers [3]. - Both companies are targeting high-capacity and high-performance solid-state drives (SSDs) and memory products to meet the growing needs of AI applications [4]. Group 3: Market Outlook - The global semiconductor storage market is projected to grow by 16.2% by 2026, reaching a size of 214.8 billion USD, presenting opportunities for domestic storage manufacturers [5]. - Jiangbolong and Demingli are both focusing on enhancing their production capabilities to meet the increasing demand from data centers and AI applications [5]. - The storage industry is characterized by a supply-demand imbalance, with expectations of limited supply from upstream manufacturers in the coming years [6]. Group 4: Financial Health and Strategy - Both companies are facing cash flow challenges, with Demingli reporting negative operating cash flow for several consecutive quarters and a high debt ratio of 73.28% [7]. - Jiangbolong has initiated a plan for H-share listing to address its urgent funding needs, indicating a strong demand for capital in the storage sector [7]. - Companies are adopting inventory management strategies to ensure supply chain stability amid market volatility, with a focus on aligning procurement with sales plans [7][8].

Core Viewpoint - Jiangbolong plans to raise 3.7 billion yuan through a private placement to actively invest in high-end AI storage, storage controller chips, and advanced packaging and testing [1] Group 1: Fundraising and Investment Plans - The company intends to use 880 million yuan of the raised funds to develop storage products targeting the AI sector, including enterprise-level SSDs and high-end consumer SSDs [1] - The company is expanding its enterprise storage business by leveraging self-developed technology and actively participating in major client technology and new product bidding [1] Group 2: Market Growth and Opportunities - The global AIGC wave and the "Digital China" top-level planning are providing new growth momentum for the storage market [1] - As a leading independent memory manufacturer in China, the company is advancing its domestic and international dual-wheel layout in the storage industry, indicating significant growth potential [1] Group 3: Financial Projections - The company's projected net profit attributable to shareholders for 2025, 2026, and 2027 is expected to be 1.16 billion, 1.85 billion, and 2.32 billion yuan, respectively [1] - Corresponding price-to-earnings ratios for 2025, 2026, and 2027 are forecasted to be 99, 62, and 49 times, respectively, maintaining a "buy" rating [1]

Core Viewpoint - The article highlights the strategic move by Jiangbolong, a leading independent storage company in China, to raise up to 3.7 billion yuan for AI-related storage research and industrialization projects, amidst a surge in AI computing power demand [1][4]. Group 1: Fundraising and Investment Plans - Jiangbolong plans to raise funds through a private placement, targeting a total of 3.7 billion yuan, with a focus on AI-related storage development [1][4]. - The funds will be allocated to four key areas: high-end storage for AI, semiconductor storage control chip development, high-end packaging and testing projects, and working capital [3][4]. Group 2: Market Position and Growth - Jiangbolong is the second-largest independent storage company globally, with significant market presence through its brands FORESEE and Lexar [5]. - The company reported a 138.66% year-on-year increase in enterprise storage business revenue, reaching 693 million yuan in the first half of 2025 [5][7]. Group 3: Technological Advancements - The company aims to address the challenges of performance, capacity, and cost in storage technology, particularly for AI applications [4]. - Jiangbolong's self-developed control chips have surpassed 100 million units deployed, with ongoing validation by major manufacturers [5]. Group 4: Financial Performance - For the first three quarters of 2025, Jiangbolong achieved revenue of 16.734 billion yuan, a 26.12% increase year-on-year, and a net profit of 713 million yuan, up 27.95% [7].

Core Viewpoint - NVIDIA's recent decision to switch from DDR5 to LPDDR memory chips aims to reduce power costs for AI servers, which is expected to significantly boost industry demand and spill over into the consumer electronics market [1][3]. Group 1: Memory Chip Transition - NVIDIA is replacing traditional DDR5 with low-power LPDDR memory to lower AI server power costs, which is anticipated to increase demand in the industry [1]. - The shift to LPDDR is expected to enhance memory value in AI servers, which typically have a higher memory unit value compared to smartphones [1]. Group 2: SOCAMM Performance - SOCAMM, based on LPDDR, is designed for AI servers and offers a 2.5x increase in bandwidth while reducing power consumption by one-third [2]. - SOCAMM's compact design (14x90mm) allows for more efficient server configurations, and it supports a maximum capacity of 128GB per module, enhancing AI model training and inference performance [2]. Group 3: Memory Price Trends - Memory prices have increased by 50% this year, with expectations of a further 30% rise by Q4 2025 and an additional 20% increase early next year [3]. - The heightened demand for LPDDR from NVIDIA is likely to create broader and more long-term risk factors surrounding advanced chips, affecting the entire consumer electronics market [3].

Core Viewpoint - Nvidia has decided to switch from DDR5 to LPDDR memory chips for AI servers to reduce power costs, which may significantly impact the supply of consumer-grade memory [1] Group 1: Memory Chip Transition - Nvidia is replacing traditional DDR5 RDIMM with LPDDR to lower AI server power consumption costs [1] - The shift to LPDDR, typically used in mobile devices, is expected to increase demand in the industry significantly [1] Group 2: SOCAMM Performance - SOCAMM, based on LPDDR, offers 2.5x bandwidth improvement over traditional RDIMM and reduces power consumption by one-third [2] - SOCAMM is designed for AI servers, with a compact size of 14x90mm and a maximum capacity of 128GB per module [2] Group 3: Memory Price Trends - Memory prices have increased by 50% this year, with expectations of a further 30% rise by Q4 2025 and an additional 20% increase early next year [3] - The heightened demand for LPDDR from Nvidia is likely to extend price increases into the consumer electronics market [3]

Investment Rating - The industry investment rating is "Overweight" [6] Core Viewpoints - NVIDIA's recent decision to switch memory chip types aims to reduce AI server power costs by replacing traditional DDR5 with LPDDR, which is expected to significantly boost demand in the industry [6] - The shift to LPDDR is anticipated to have a ripple effect on the consumer electronics market, as LPDDR is commonly used in mobile devices [6] - Memory prices have increased by 50% this year and are projected to rise an additional 30% in Q4 2025, followed by a further increase of about 20% early next year [6] Summary by Sections - **Industry Overview**: The report highlights the growing demand for LPDDR memory due to NVIDIA's shift in AI server technology, which is expected to enhance memory pricing and availability in the consumer electronics sector [3][6] - **Technological Advancements**: SOCAMM, a memory module based on LPDDR, offers 2.5x bandwidth improvement over traditional RDIMM and reduces power consumption by one-third, making server designs more compact and efficient [6] - **Market Trends**: The report notes that the transition to LPDDR in enterprise solutions may lead to increased supply constraints in consumer-grade memory, further influencing market dynamics [6]

Supply Chain and Market Trends - The company maintains strong relationships with major wafer suppliers through long-term agreements, ensuring supply chain resilience and diversity [2] - Recent trends show a price increase in storage products due to high demand from North American cloud service providers investing in AI, leading to a tightening supply for consumer and embedded storage products [3] Business Performance and Product Development - The company ranks third in total capacity for enterprise SATA SSDs in China and first among domestic brands, with ongoing expansion in enterprise storage product offerings [3] - The company has launched four series of self-developed controller chips, with over 100 million units deployed, indicating rapid growth in deployment scale [5] New Product Innovations - The company is actively developing high-performance storage products for data centers, including the newly released SOCAMM2, designed specifically for AI data centers, which shows significant improvements in bandwidth and power consumption [4] - The UFS4.1 product, developed in collaboration with SanDisk, has demonstrated superior performance in process, read/write speed, and stability compared to market alternatives, with accelerated adoption among Tier 1 clients [5]

Group 1: Market Trends and Price Increases - The recent increase in storage prices is attributed to the growing demand from North American cloud service providers investing heavily in AI infrastructure, leading to a significant supply shortage of HDDs [2][3] - DRAM and NAND product prices have seen a minimum increase of 20%, with some prices rising over 40% [3] Group 2: Impact on Profitability - The production cycle from wafer procurement to storage sales will positively impact the company's gross margin during periods of rising storage wafer prices, although raw material price fluctuations are only one factor affecting performance [4] - The company has made significant breakthroughs in enterprise storage, high-end consumer storage, overseas business, and self-developed controller chips, which will drive profitability growth more directly and sustainably [4] Group 3: Resource Supply Stability - As a leading independent storage manufacturer, the company maintains a strong inventory turnover rate and has established long-term partnerships with major wafer suppliers, ensuring resource supply stability [4] - The company has signed long-term agreements (LTA) or memorandums of understanding (MOU) with wafer suppliers, enhancing supply chain resilience and diversity [4] Group 4: Enterprise Storage Business Progress - The company ranks third in total capacity for enterprise SATA SSDs in China for the first half of 2025, leading among domestic brands [4] - The company is actively expanding its high-performance storage products for data center applications, including new memory types like CXL2.0 and MRDIMM, and has officially launched SOCAMM2 [5][6] Group 5: UFS4.1 and Controller Chip Developments - The company has successfully developed UFS4.1 products, which outperform comparable market products in terms of process, read/write speed, and stability, gaining recognition from major clients [6] - The deployment of the company's self-developed controller chips has surpassed 100 million units, with rapid growth expected in the coming year [6]

Supply Chain Management - The company maintains a robust supply chain system through long-term partnerships with major storage manufacturers like Yangtze Memory Technologies and ChangXin Memory Technologies, ensuring stability amid global supply shortages [1] - A diversified raw material supply system is being established to enhance risk management and align upstream supply capabilities with downstream business needs [1] Inventory Strategy - The company is enhancing inventory management to adapt to the tight supply situation in the storage market, focusing on stable procurement of core materials like storage chips [1] - Collaboration with upstream suppliers is deepened to better match business development needs [1] Market Trends and Pricing - The AI wave is driving explosive growth in data storage demand, with major companies like Samsung and SK Hynix recently raising prices for storage products, indicating a positive market outlook [2] - The storage price is expected to maintain an upward trend in Q4 2025 due to increased capital expenditure from leading tech firms [2] Profitability Indicators - The company is closely monitoring core profitability metrics, with a notable improvement in gross margin in Q3 2025, driven by favorable industry trends [3] - Continued price increases in storage products are anticipated to further enhance performance in Q4 [3] Product Development and Customer Engagement - The company is focusing on core business areas of storage control chips and solutions, increasing investment in R&D and product innovation to meet market demands [5] - There is an emphasis on expanding high-value enterprise storage and embedded storage businesses while optimizing customer structure and market share [5]