Investment Rating - The report maintains an "Outperform" rating for the company, projecting significant revenue and profit growth over the next few years [3][7]. Core Insights - The company is expected to achieve revenues of RMB 95.2 billion, RMB 114.7 billion, and RMB 136.2 billion for the years 2025, 2026, and 2027 respectively, with net profits attributable to shareholders of RMB 2.04 billion, RMB 2.71 billion, and RMB 3.36 billion during the same period [3][7]. - The quarterly revenue has shown acceleration, with a 24.96% year-on-year increase in 2025H1, reaching RMB 47.43 billion [3][7]. - The growth in the internet business is significant, with a 37.75% increase in revenue for New H3C in 2025H1, driven by strong demand in the internet sector [3][7]. Financial Summary - The total revenue for 2023 is projected at RMB 77.31 billion, with a growth rate of 4.4%. By 2027, revenue is expected to reach RMB 136.23 billion, reflecting an 18.8% growth rate [2][4]. - Net profit attributable to shareholders is forecasted to decline to RMB 2.10 billion in 2023, but is expected to recover to RMB 3.36 billion by 2027, with corresponding EPS increasing from RMB 0.74 to RMB 1.17 [2][4]. - The gross profit margin for 2025H1 was reported at 15.24%, a decrease of 3.80 percentage points year-on-year [3][7]. Business Segments - The domestic government and enterprise revenue for New H3C reached RMB 31.04 billion in 2025H1, a 53.55% increase, while international revenue grew by 60.25% to RMB 1.92 billion [3][7]. - The ICT infrastructure and services revenue was RMB 36.04 billion in 2025H1, with a gross profit margin of 18.06%, down 6.32 percentage points year-on-year [3][7]. Inventory and Liabilities - As of the end of 2025H1, inventory stood at RMB 43.84 billion, up 19.0% from the beginning of the year, while contractual liabilities increased by 9.7% to RMB 15.91 billion [3][7].

Core Viewpoint - The company Shenglan Co., Ltd. is actively expanding its data center business and has established connections with major clients in the industry, including Industrial Fulian, Inspur, and H3C [2] Group 1 - Shenglan Co., Ltd. has confirmed its client base for data communication connectors, which includes notable companies such as Industrial Fulian, Inspur, and H3C [2] - The company is committed to continuously exploring opportunities within the data center sector [2]

Core Insights - The Chinese accelerated server market is projected to reach $16 billion in the first half of 2025, more than doubling from the first half of 2024, with expectations to exceed $140 billion by 2029 [1][3] Market Overview - The top three vendors by sales revenue in the first half of 2025 are Inspur, H3C, and Lenovo, collectively holding nearly 50% of the market share [3] - In terms of server shipments, Inspur, H3C, and Ningchang lead the market with approximately 43% share [3] - The internet sector remains the largest purchaser, accounting for nearly 69% of the overall accelerated server market, with significant growth observed in finance, education, telecommunications, and healthcare sectors [3] Policy and Industry Dynamics - The computing power voucher policy has significantly lowered the entry barriers for domestic AI computing power through substantial funding and multi-level support systems [3] - This initiative has facilitated a strategic shift for domestic manufacturers from single product replacements to comprehensive ecosystem innovation [3] - The market is transitioning from a phase of demand driven by large-scale commercial deployment in 2024 to a focus on product competitiveness in the first half of 2025 [3] Application and Technology Trends - In the industrial sector, AI servers are enhancing real-time process optimization, predictive maintenance, and digital twin simulations, improving manufacturing precision [3] - In finance, applications include high-frequency trading risk control and intelligent investment research, capable of processing high-dimensional real-time data [3] - In smart healthcare, technologies such as genomic analysis and surgical robot control are emerging, emphasizing ultra-low latency and high reliability [3] - Traditional industries like retail, energy, and agriculture are leveraging edge computing for real-time analysis and localized decision-making [3] Accelerated Chip Market - The market for accelerated chips in China is expected to exceed 1.9 million units in the first half of 2025 [5] - Demand for non-GPU cards is growing rapidly, supported by policies, capturing 30% of the market share, while domestic chip brands account for approximately 35% of the overall market [5] Competitive Landscape - The Chinese AI accelerated server market is currently in a phase of scale expansion and domestic substitution, with a need for breakthroughs in high-end computing efficiency and ecosystem maturity [7] - Future competition will shift focus from single-chip performance to system efficiency, open ecosystem collaboration, and green computing cost control [7] - The industry is encouraged to avoid low-level redundant construction and enhance international competitiveness through technological collaboration and standard optimization [7]

Group 1 - The core viewpoint is that the domestic AI computing power ecosystem is evolving rapidly, with significant server procurement results announced by major companies like Industrial and Commercial Bank of China and China Unicom, where domestic suppliers won over 90% of the bids [1][2][8] - The server procurement results include a total of approximately 10 billion yuan, with Industrial and Commercial Bank of China awarding 3 billion yuan for Haiguang chip servers and China Unicom awarding a total of 7.96 billion yuan across various server categories [2] - The release of Alibaba's large model Qwen3-VL-30B-A3B and Huawei's support for it indicates advancements in domestic AI models, with the model being competitive in various tasks with only 30 billion activation parameters [3][4] Group 2 - Tencent's latest visual model Hunyuan-Vision-1.5-Thinking achieved third place globally and first domestically in the LMArena rankings, showcasing the rapid iteration of domestic large models [4] - The trend towards self-controlled domestic solutions is becoming prominent, with significant investments expected in domestic computing power, especially as overseas AI chip supply is constrained due to geopolitical factors [5] - The overall sentiment is optimistic regarding the acceleration of domestic computing power and AI development, suggesting that investors should pay attention to leading companies in this sector [5][8]

Core Insights - The rapid advancement of domestic computing chips is complemented by positive developments in general-purpose processors in China [1] - Major procurement announcements from Industrial and Commercial Bank of China (ICBC) and China Unicom indicate a significant shift towards domestic CPU usage in server procurement [1][6] - The overall market for intelligent computing is expected to grow substantially, with projections indicating that by 2027, intelligent computing power in China will reach 1117 EFLOPS [1][9] Group 1: Procurement Announcements - ICBC announced a procurement of Haiguang chips with a total value of approximately 3 billion yuan, with candidates including Inspur, ZTE, and Lenovo [1] - China Unicom's procurement project has a budget exceeding 7.9 billion yuan, with a significant portion allocated for domestic general-purpose processors [1] - The procurement data shows that over 90% of the servers will use domestic CPUs, with Haiguang and Kunpeng CPUs being the most utilized [3][4] Group 2: Market Trends and Projections - The share of domestic chip servers has increased to 90.1%, a significant rise from around 40% in previous years [4] - The demand for server CPUs is expected to grow, with projections indicating that by 2028, the demand for data center servers will reach 13.99 million units, with a compound annual growth rate of 6% from 2024 to 2028 [7] - The domestic server CPU market is rapidly evolving, with X86 architecture still dominating but Arm architecture gaining market share [8][9] Group 3: Domestic CPU Development - Domestic CPUs such as Haiguang, Kunpeng, and Alibaba's Xuantie are rapidly developing, with a notable increase in market presence [8] - The domestic CPU localization rate has risen from 0% in 2018 to 36% in 2023, with expectations to reach 85% by 2028 due to policy support and technological advancements [9] - The current landscape shows a strong push towards domestic solutions in the server CPU market, driven by both demand and government initiatives [9]

Core Viewpoint - 欧陆通 has developed a diverse core product matrix, focusing on high-efficiency power solutions for data centers and server systems [1] Product Offerings - The company offers a range of power solutions including: - 3,200W Titanium M-CRPS server power supplies - 1,300W to 3,600W Titanium CRPS server power supplies - 3,300W to 5,500W Titanium and Super Titanium GPU server power supplies - Immersion liquid cooling server power supplies - Rack power shelf solutions [1] - A new rack power solution compliant with the Open Compute Project (OCP) third-generation open rack (ORV3) specifications has been introduced, featuring six 5,500W server power supplies capable of delivering up to 33KW of power output with a conversion efficiency of 97.5% [1] Support for Domestic Industry - The company is committed to supporting the development of the domestic industrial chain and is actively collaborating with clients to develop fully domesticated data center power products [1] - Multiple domesticated power supply models are currently in research and production stages [1] Partnerships and Collaborations - 欧陆通 has successfully supplied products to well-known domestic server system manufacturers such as Inspur, Foxconn, Huaqin, Lenovo, ZTE, and H3C [1] - The company maintains close cooperation with leading domestic internet enterprises and other end customers [1]

Core Insights - The company has developed a diverse core product matrix, including various high-capacity power supply solutions for servers, demonstrating its commitment to innovation and meeting market demands [1] Product Offerings - The product range includes 3,200W titanium M-CRPS server power supplies, 1,300W to 3,600W titanium CRPS server power supplies, and 3,300W to 5,500W titanium and ultra-titanium GPU server power supplies [1] - The company also offers immersion liquid cooling server power supplies and rack power shelf solutions, with a configuration of six 5,500W server power supplies capable of delivering up to 33KW of power output and achieving a conversion efficiency of 97.5% [1] - Expandable solutions are available, including options for 2OU 66KW, 3OU 66KW, and 4OU 132KW [1] Support for Domestic Industry - The company is actively supporting the development of the domestic industrial chain by collaborating with clients to develop fully localized data center power products, with several models currently in research and production [1] - The company has successfully supplied products to well-known domestic server system manufacturers such as Inspur, Foxconn, Huqian, Lenovo, ZTE, and H3C, while maintaining close cooperation with leading domestic internet companies and end customers [1]

Investment Rating - The report maintains an "Overweight" rating for the electronic industry [1] Core Insights - The report highlights the competitive landscape of Switch chip manufacturers, with NVIDIA dominating the market through proprietary protocols, while Broadcom is gaining traction with its open-source SUE architecture [6][13] - It emphasizes the potential for domestic manufacturers to replace imported Switch chips, with companies like Shengke Communication leading the way in Ethernet chip production [35] - The report suggests investment in key players such as Haiguang Information and Shengke Communication, while also recommending attention to companies like ZTE and Lanke Technology [6][35] Summary by Sections 1. Switch Chip Manufacturers - NVIDIA's NVSwitch is noted as the highest bandwidth and most mature private solution in the Scale-Up segment [11] - Broadcom holds a 90% market share in cloud data center switches and has introduced the SUE architecture for open Ethernet interconnects [13] - Astera Labs is recognized for its complete product chain, being the only company that integrates switch chips, extension lines, and software management platforms [20] 2. Achieving Domestic Replacement of Switch Chips - Shengke Communication is identified as a leading domestic Ethernet switch chip manufacturer, with 12.8T and 25.6T chips entering customer promotion stages [35] - Other domestic players like Shudao Technology and Lanke Technology are making strides in the PCIe segment, with Shudao expected to achieve breakthroughs in domestic replacement by the end of 2025 [41] - Major manufacturers such as Haiguang, Huawei, and ZTE are also developing self-researched chips to support domestic replacement efforts [43][45][50] 3. Investment Recommendations - The report recommends focusing on Haiguang Information and Shengke Communication for investment opportunities, while suggesting to keep an eye on ZTE, Wantong Development, and Lanke Technology [6][35]

Core Insights - The 22nd China-ASEAN Expo featured an AI-themed exhibition, showcasing advancements in artificial intelligence and its integration into various sectors, emphasizing the theme "Hand in Hand in the Intelligent Era, Building a Digital Silk Road" [18][19][26] Group 1: AI Innovations and Applications - Nearly 100 leading AI companies presented groundbreaking technologies, including humanoid robots performing various tasks, which attracted significant attention from attendees [19] - Local innovations were highlighted, such as the "Guizhi No. 1" data security and AI application platform, and the "Jian E" intelligent multi-dimensional road detection vehicle, showcasing the region's technological capabilities [20][22] - AI applications in agriculture and healthcare were demonstrated, with solutions enhancing efficiency in pig farming and providing intelligent diagnostic systems for medical services [22] Group 2: Interactive Experiences and Public Engagement - The exhibition featured interactive zones where attendees could experience AI technologies firsthand, such as AI coffee-making robots and smart home terminals, illustrating the technology's penetration into daily life [21] - Feedback from attendees indicated a strong sentiment that AI is becoming an integral part of everyday life, transforming both personal and professional environments [21][28] Group 3: Regional Cooperation and Market Expansion - The expo served as a platform for deepening cooperation between China and ASEAN countries in AI, facilitating discussions and partnerships aimed at mutual technological advancement [23][26] - Notable collaborations were established, such as the cross-border digital identity service platform between Malaysia and Guangxi, enhancing seamless verification processes for citizens [25] - The participation of ASEAN enterprises underscored the growing demand for AI solutions in various sectors, including education, healthcare, and agriculture [24][26] Group 4: Event Outcomes and Future Prospects - The AI exhibition attracted over 80,000 visitors, with around 28,000 professional attendees engaging in discussions, leading to an estimated cooperation intention amount exceeding 100 million yuan [28] - Companies expressed optimism about the future of AI in the region, viewing the expo as a critical opportunity for collaboration and market entry into ASEAN [26][27]

Core Viewpoint - The article provides a detailed comparison of Alibaba's super node with NVIDIA's NVL72 and Huawei's CM384, focusing on GPU count, interconnect technology, power consumption, and ecosystem compatibility. Group 1: GPU Count - Alibaba's super node, known as "Panjun," utilizes a configuration of 128 GPUs, with each of the 16 computing nodes containing 4 self-developed GPUs, totaling 16 x 4 x 2 = 128 GPUs [4] - In contrast, Huawei's CM384 includes 384 Ascend 910C chips, while NVIDIA's NVL72 consists of 72 GPUs [7] Group 2: Interconnect Technology - NVIDIA's NVL72 employs a cable tray interconnect method using NVLink proprietary protocol [8] - Huawei's CM384 also uses cable connections between multiple racks [10] - Alibaba's super node features an orthogonal interconnect without a backplane, allowing for direct connections between computing and switch nodes, reducing signal transmission loss [12][14] Group 3: Power and Optical Connections - NVIDIA's NVL72 uses copper for scale-up connections, while Huawei's CM384 employs optical interconnects, leading to higher costs and power consumption [15] - Alibaba's super node uses electrical interconnects for internal scale-up, with some connections made via PCB and copper cables, while optical interconnects are used between two ALink switches [18][19] Group 4: Parameter Comparison - Key performance metrics show that NVIDIA's GB200 NVL72 has a BF16 dense TFLOPS of 2,500, while Huawei's CM384 has 780, indicating a significant performance gap [21] - The HBM capacity for NVIDIA's GB200 is 192 GB compared to Huawei's 128 GB, and the scale-up bandwidth for NVIDIA is 7,200 Gb/s while Huawei's is 2,800 Gb/s [21] Group 5: Ecosystem Compatibility - Alibaba claims compatibility with multiple GPU/ASICs, provided they support the ALink protocol, which may pose challenges as major manufacturers are reluctant to adopt proprietary protocols [23] - Alibaba's GPUs are compatible with CUDA, providing a competitive advantage in the current market [24] Group 6: Supply Chain Insights - In the AI and general server integration market, Inspur holds a 33%-35% market share, while Huawei's share is 23% [33] - For liquid cooling, Haikang and Invec are key players, each holding 30%-40% of the market [35] - In the PCB sector, the number of layers has increased to 24-30, with low-loss materials making up over 60% of the composition, significantly increasing the value of single-card PCBs [36]