2025 年 12 月 29 日 行业研究 评级：推荐(维持) 研究所： 证券分析师： 刘熹 S0350523040001 liux10@ghzq.com.cn [Table_Title] H200"有条件"对华开放，长期依然看好国产 算力 ——计算机事件点评 最近一年走势 | 行业相对表现 | | 2025/12/26 | | | --- | --- | --- | --- | | 表现 | 1M | 3M | 12M | | 计算机 | -1.4% | -6.5% | 11.6% | | 沪深 300 | 3.1% | 2.4% | 16.8% | 相关报告 《计算机行业专题报告：GPU+ASIC 渗透加速，液 冷市场规模再添增量（推荐）*计算机*刘熹》—— 2025-12-25 《计算机行业专题报告：互联总线——Scale-Up 能力跃升的关键通路（推荐）*计算机*刘熹》—— 2025-12-25 《计算机专题报告：超节点渐成共识，产业链成长 动能明确 ——AI 算力与模型应用月 报（202510）（推荐）*计算机*刘熹》——2025-11-17 《计算机事件点评：百度昆仑芯新品发布，助力中 国 AI ...

Core Viewpoint - Baidu's AI chip business, Kunlun, has completed a new funding round with a valuation nearing $3 billion and is preparing for an IPO in Hong Kong by early 2027 [1] Group 1: Company Developments - Baidu's stock rose over 6% to $126.42 following the news of Kunlun's funding and IPO plans [1] - Kunlun aims to submit its IPO application to the Hong Kong Stock Exchange in the first quarter of 2026 [1] Group 2: Financial Highlights - The recent funding round has valued Kunlun at approximately $3 billion [1]

Group 1 - Kunlunxin, a subsidiary of Baidu focused on AI chips, is preparing for an IPO in Hong Kong and has completed a new financing round with a post-money valuation of approximately 21 billion RMB (about 3 billion USD) [1][2] - Following the news, Baidu's stock in Hong Kong surged by 7.51%, reaching a peak increase of 8.70 [2] - Kunlunxin is recognized as a key player in China's high-performance computing sector, with the latest financing involving multiple institutional investors and the IPO preparations in preliminary stages [2][3] Group 2 - The push for Kunlunxin's market entry is driven by the tightening of U.S. export restrictions on advanced semiconductors and the growing domestic demand for high-end AI chips, indicating a significant shift towards local supply chain alternatives [3] - In the past year, several Chinese AI chip companies, including Cambricon and Suirian Technology, have also sought to go public, reflecting an accelerated move of the domestic computing ecosystem towards capital markets to support larger R&D investments [3] - As of now, Baidu has not responded to inquiries regarding the IPO plans [3]

Core Viewpoint - Kunlun Core plans to submit its IPO application to the Hong Kong Stock Exchange as early as the first quarter of next year, aiming to complete the IPO by early 2027 [1] Group 1 - Kunlun Core is targeting a listing on the Hong Kong Stock Exchange [1] - The company aims to submit its application no later than the first quarter of next year [1] - The completion of the IPO is expected by early 2027 [1]

Core Viewpoint - The recent announcements from major Chinese tech companies like Huawei and Baidu regarding their AI chip development signify a shift in the domestic semiconductor landscape, aiming to fill the gap left by Nvidia and enhance the competitiveness of Chinese AI chips [3][4][5]. Group 1: AI Chip Development - Major Chinese companies are increasingly revealing their AI chip roadmaps, with Huawei planning to release four Ascend AI chips over the next three years, while Baidu is set to launch two Kunlun AI chips in the next two years [2][3]. - Huawei's detailed disclosure of technical parameters for its chips, including bandwidth, computing power, and memory, marks a significant change in the traditionally low-profile approach of Chinese chipmakers [2][7]. - The introduction of supernodes and clusters is seen as a critical strategy for overcoming the limitations of China's semiconductor manufacturing processes, which are currently capped at the 7nm node [10][12]. Group 2: Competitive Landscape - The competition in the global AI chip market is characterized as asymmetric, with Chinese chips lagging behind North American counterparts like Nvidia in various technical specifications, yet capable of leveraging networking capabilities to surpass them in performance [5][6]. - Huawei's Ascend series has been recognized as a formidable competitor, with its first chip released in 2018 and subsequent iterations showing significant performance improvements despite challenges posed by U.S. sanctions [6][8]. - Baidu's Kunlun chip, while still behind in performance compared to Nvidia's offerings, is focusing on cost-effectiveness and specific use cases, indicating a strategic approach to market entry [8][9]. Group 3: Market Dynamics - The domestic AI chip market is witnessing a shift towards inference tasks, with inference scenarios accounting for 42% of the GenAI IaaS service market, while training scenarios have decreased to 58% [14][15]. - The challenges of using domestic AI chips for large model training are acknowledged, with companies like Huawei and Baidu working to adapt their technologies to meet these demands [14][15]. - The push for self-developed chips by major cloud providers is seen as a way to reduce costs and improve performance, with companies like Kunlun seeking to penetrate external markets [16][17].

Core Insights - Domestic AI chip companies have been relatively low-profile in recent years, but recent announcements from major players like Huawei and Baidu have broken this silence, revealing their AI chip development roadmaps [1][2][4] - The competitive landscape is shifting as domestic companies aim to capture market share left by Nvidia, with a focus on clear product roadmaps and advanced capabilities [2][10] Domestic AI Chip Development - Huawei plans to release four Ascend AI chips over the next three years, while Baidu has announced two Kunlun AI chips in the next two years [1][4] - The Ascend 950 series will include two models, 950PR and 950DT, designed for different stages of AI inference and training, with specific memory and bandwidth capabilities [7][8] Performance and Technology - Despite advancements, domestic AI chips still lag behind international competitors in terms of performance metrics such as process technology and memory bandwidth [3][10] - The "super node + cluster" strategy is being adopted by major companies to enhance AI computing capabilities, compensating for limitations in individual chip performance [14][17] Market Dynamics - The AI chip market is becoming increasingly competitive, with a focus on both training and inference capabilities, as companies like Huawei and Baidu seek to establish their products in the market [19][20] - The demand for AI inference is rising, with predictions that it will become a core segment of AI infrastructure services [20][21] Future Outlook - Companies are exploring IPO opportunities and external market engagements, as seen with Baidu's Kunlun chip seeking to expand its market presence [12][13] - The development of super nodes and clusters is seen as crucial for overcoming the limitations of current chip manufacturing processes, particularly in the context of U.S. sanctions [16][18]

Core Insights - Domestic AI chip companies are becoming more vocal about their product roadmaps, with major players like Huawei and Baidu announcing upcoming AI chip releases, breaking a period of silence in the industry [1][2][4] - The shift towards showcasing product capabilities is seen as a response to market opportunities left by Nvidia, with analysts suggesting that a clear product roadmap is essential for capturing market share [2][3] - Despite advancements, domestic AI chips still lag behind their international counterparts in performance, necessitating innovative solutions like "super nodes" and clusters to meet AI computing demands [3][4][11] Huawei's Developments - Huawei plans to release three new Ascend AI chip series (950, 960, and 970) between 2026 and 2028, marking a significant shift from its previous strategy of limited product announcements [4][5][8] - The Ascend 950 series will include two models, focusing on different stages of AI inference, with specifications indicating substantial improvements in memory bandwidth and processing power [6][7] - Huawei's super node strategy aims to enhance computing power by interconnecting multiple chips, allowing for greater efficiency in large-scale AI model training [12][14] Baidu's Strategy - Baidu has announced its Kunlun chip roadmap, with the M100 and M300 chips set for release in 2026 and 2027, respectively, targeting large-scale inference and multimodal model training [9][10] - The Kunlun chip is designed to support high-performance computing tasks, with plans for super nodes that can accommodate multiple chips for enhanced processing capabilities [10][22] - Baidu's recent disclosures may be influenced by competitive pressures and potential IPO considerations for its chip division [10][22] Industry Trends - The domestic semiconductor supply chain has made significant strides, filling gaps left by U.S. sanctions and enhancing the predictability of future product iterations [2][3] - The focus on "super nodes" and clusters is seen as a critical strategy for overcoming limitations in individual chip performance, particularly in the context of large AI model training [11][12] - The competition in the AI chip market is intensifying, with various companies exploring specialized designs to meet specific application needs, particularly in inference tasks [20][21] Market Dynamics - The demand for AI inference capabilities is rising, with a notable shift in the market towards optimizing chips for specific tasks rather than general-purpose applications [18][20] - Companies are leveraging their cloud services to validate and promote their self-developed chips, creating a direct internal demand that supports their market positioning [22] - The landscape is characterized by fragmentation, with both integrated and specialized chip manufacturers vying for market share in the rapidly evolving AI sector [20][21]

Core Viewpoint - Baidu Group's stock price increased by 3.05% following the announcement that Kunlun Chip Technology won a significant procurement project with China Mobile, indicating a positive market reaction to the company's advancements in AI chip technology [1] Group 1: Company Developments - Kunlun Chip Technology secured a multi-billion yuan order from China Mobile for AI general-purpose computing devices, marking a significant expansion of its application scenarios beyond internal use by tech companies to telecommunications [1] - The company achieved a 70% market share in two of the three bidding packages and a 100% share in the third package for the AI server products based on Kunlun chips [1] - Kunlun Chip Technology, previously part of Baidu's smart chip and architecture department, has successfully completed the mass production and application of two generations of general-purpose AI chip series [1] Group 2: Technological Advancements - Baidu's AI training efficiency and energy efficiency have reached near-international advanced levels, as demonstrated by the large-scale cluster based on the "Kunlun P800" accelerator card being used in the training of the new version of Wenxin Yiyan [1]

Group 1 - Baidu Group-SW (09888) opened over 3% higher, currently up 3.05% at 135 HKD with a trading volume of 134 million HKD [1] - Kunlun Core Technology won a large-scale order in the China Mobile centralized procurement project, indicating a shift in application scenarios from internal use by tech companies to telecommunications [1] - Kunlun Core's AI server products achieved a market share of 70%, 70%, and 100% in the respective bidding packages for the AI general computing devices project from 2025 to 2026 [1] Group 2 - Kunlun Core Technology, previously part of Baidu's smart chip and architecture department, completed independent financing in April 2021 and has since launched two generations of general AI chip series products [1] - Baidu's large-scale cluster based on the "Kunlun P800" accelerator card has entered the training trial of the new version of Wenxin Yiyan, achieving training efficiency and energy efficiency close to international advanced levels [1]