【文/观察者网 熊超然】"华为发布全新人工智能（AI）芯片技术，挑战英伟达主导地位。"9月18日，随 着华为全联接大会2025（HC 2025）发布多个振奋人心的消息，外媒再度对这场中美科技巨变予以关 注，其中就出现了这样醒目的报道标题。 当天，华为副董事长、轮值董事长徐直军宣布了多颗昇腾系列芯片和演进路线，包括昇腾950系列、昇 腾960系列和昇腾970系列。他表示，预计华为2026年第一季度推出昇腾950PR芯片，第四季度推出昇腾 950DT，2027年第四季度推出昇腾960芯片，2028年四季度推出昇腾970芯片。 徐直军还公布了以昇腾950芯片为基础的新型超节点（SuperPoD），将成为全球最强超节点（算力规模 8192卡），甚至比英伟达将在2027年推出的NVL576系统更强。不仅如此，以昇腾960芯片为基础的超 节点（算力规模15488卡），也将在2027年第四季度上市，持续提供充沛算力。 美国彭博社提到，近期传出中国买家已不愿为英伟达对华"减配特供版"芯片买单，而华为的最新研发进 展，标志着中国为摆脱对所谓"AI行业黄金标准"英伟达硬件的依赖，并推动国产替代产品而作出的最新 尝试。 英国路透社 ...

炒股就看金麒麟分析师研报，权威，专业，及时，全面，助您挖掘潜力主题机会！ 【文/观察者网 熊超然】"华为发布全新人工智能（AI）芯片技术，挑战英伟达主导地位。"9月18日，随 着华为全联接大会2025（HC 2025）发布多个振奋人心的消息，外媒再度对这场中美科技巨变予以关 注，其中就出现了这样醒目的报道标题。 当天，华为副董事长、轮值董事长徐直军宣布了多颗昇腾系列芯片和演进路线，包括昇腾950系列、昇 腾960系列和昇腾970系列。他表示，预计华为2026年第一季度推出昇腾950PR芯片，第四季度推出昇腾 950DT，2027年第四季度推出昇腾960芯片，2028年四季度推出昇腾970芯片。 徐直军还公布了以昇腾950芯片为基础的新型超节点（SuperPoD），将成为全球最强超节点（算力规模 8192卡），甚至比英伟达将在2027年推出的NVL576系统更强。不仅如此，以昇腾960芯片为基础的超 节点（算力规模15488卡），也将在2027年第四季度上市，持续提供充沛算力。 美国彭博社提到，近期传出中国买家已不愿为英伟达对华"减配特供版"芯片买单，而华为的最新研发进 展，标志着中国为摆脱对所谓"AI行业黄金 ...

图源：观察者网 他还公布了以昇腾950芯片为基础的新型超节点，将成为全球最强超节点，甚至比英伟达将在2027年推 出的NVL576系统更强。 （文/观察者网 吕栋 编辑/张广凯） 9月18日，在华为全连接大会2025上，华为轮值董事长徐直军宣布了多颗昇腾系列芯片和演进路线，包 括昇腾950系列、昇腾960系列和昇腾970系列。 图源：观察者网 他提到，华为自研了低成本HBM，将以一年一次算力翻倍的进度推进，支持FP8等更多精度格式，更大 的互联带宽。 徐直军表示，有了昇腾芯片为基础，就能满足客户的算力需求，超节点将成为AI基础设施建设的新常 态。 不仅如此，以昇腾960为基础的超节点，也将在2027年四季度上市，持续提供充沛算力。 徐直军还提到，华为还将更新通用计算鲲鹏处理器，包括鲲鹏950系列和鲲鹏960系列，以及以这些芯片 为基础的超节点。 图源：观察者网 他还宣布了面向超节点的互联协议"灵衢"，把更多计算资源连接在一起，以昇腾950为基础可以组成超 50万卡集群，以昇腾960为基础甚至可以组成超过99万卡的集群。 图源：观察者网 徐直军提到的这些芯片和技术，将在2026-2027年相继上市。 "我们单颗 ...

Core Viewpoint - Huawei demonstrates strong technological capabilities in the semiconductor industry, particularly with its Ascend series chips and the recent launch of CM384, positioning itself as a leader in domestic AI chips [2][3]. Group 1: Financial Performance - In 2024, Huawei achieved a total revenue of RMB 862.072 billion, representing a year-on-year growth of 22.4% [5]. - The smart automotive solutions segment saw a remarkable revenue increase of 474.4%, while terminal business and digital energy businesses grew by 38.3% and 24.4%, respectively [5]. - Revenue from the Chinese market reached RMB 615.264 billion, driven by digitalization, intelligence, and low-carbon transformation [5]. Group 2: Huawei Cloud - The overall public cloud market in China is projected to reach USD 24.11 billion in the second half of 2024, with IaaS accounting for USD 13.21 billion, representing a year-on-year growth of 14.4% [6]. - Huawei Cloud holds a 13.2% market share in the Chinese IaaS market, making it the second-largest cloud provider after Alibaba Cloud [6]. - Huawei Cloud's revenue growth rate reached 24.4%, the highest among major cloud vendors in China [6]. Group 3: Ascend Chips - The CloudMatrix 384 super node integrates 384 Ascend 910 chips, achieving a cluster performance of 300 PFLOPS, which is 1.7 times that of Nvidia's GB200 NVL72 [10]. - The single-chip performance of Huawei's Ascend 910C is approximately 780 TFLOPS, which is one-third of Nvidia's GB200 [10][11]. - The Ascend computing system encompasses a comprehensive ecosystem from hardware to software, aiming to meet various AI computing needs [15][20]. Group 4: HarmonyOS - HarmonyOS features a self-developed microkernel, AI-native capabilities, distributed collaboration, and privacy protection, distinguishing it from Android and iOS [12]. - The microkernel architecture enhances performance and fluidity, while the distributed soft bus technology allows seamless connectivity among devices [12][13]. Group 5: Kirin Chips - The Kirin 9020 chip has reached high-end processor standards, comparable to a downclocked Snapdragon 8 Gen 2 [23]. - The Kirin X90 chip, based on the ARMv9 instruction set, features a 16-core design with a frequency exceeding 4.2GHz, achieving a 40% improvement in energy efficiency [25][26]. Group 6: Kunpeng Chips - Kunpeng processors are designed for servers and data centers, focusing on high performance, low power consumption, and scalability [27]. - The Kunpeng ecosystem strategy emphasizes hardware openness, software open-source, enabling partners, and talent development [29].

Core Viewpoint - The article discusses the security risks associated with NVIDIA's H20 chip, particularly its potential to compromise AI security in China, and highlights the challenges faced by domestic chip manufacturers in meeting the growing demand for AI computing power [1][3][8]. Group 1: Security Risks and Concerns - The Chinese authorities have raised concerns about the H20 chip's vulnerabilities, which could allow data to be uploaded to Singapore and accessed by the U.S. Department of Commerce [1]. - There is a fear that the H20 chip may contain backdoors that could threaten China's AI security [1][8]. - The urgency of addressing these risks is underscored by the fact that domestic orders for H20 chips surged immediately after the U.S. lifted restrictions on their sale to China [4][8]. Group 2: Domestic Chip Manufacturing Challenges - Domestic AI chip manufacturers, such as Huawei's Ascend series and Cambricon's series, are reportedly capable of outperforming the H20 chip in key metrics, yet the lack of production capacity remains a significant issue [3][4]. - The demand for AI computing power in China is projected to reach 211.50 EFlops in 2024, while domestic supply can only meet 98.24 EFlops, resulting in a substantial shortfall of 113.26 EFlops, or 53.5% reliance on imports [4][3]. - The current domestic production capacity is insufficient to meet the rapidly increasing demand for AI chips, leading to a market gap of 61.3% [4][3]. Group 3: Competitive Landscape in AI - The global AI market is highly competitive, with companies that can secure advanced computing chips gaining a significant advantage in deploying their AI models [4][5]. - Despite the technological capabilities of domestic chips, the inability to meet market demand forces companies to rely on imported chips, such as the H20, to remain competitive [4][5]. - The article emphasizes that while the U.S. holds a technological edge, China's rapid application and commercialization of AI technologies pose a significant threat to U.S. dominance in the field [5][6]. Group 4: U.S. Strategy and Implications - The U.S. has adopted a strategy of allowing the sale of modified versions of its advanced AI chips to China, which may include monitoring capabilities to track the development of China's AI ecosystem [8][6]. - This approach aims to slow down China's AI advancements while still engaging with the lucrative Chinese market [8][6]. - The potential for the U.S. to manipulate the AI development environment in China through backdoor access to chips raises significant concerns about the future of China's AI innovation [8][6].

在全球科技产业格局中，芯片领域的一举一动都备受瞩目。近期，英伟达 CEO 黄仁勋的北京之行以及与之紧密 相关的美国芯片出口管制新规，犹如投入湖面的巨石，激起层层涟漪。 4 月 17 日下午，中国贸促会会长任鸿斌与黄仁勋举行会谈，这一消息迅速引发各界关注。值得注意的是，这是 黄仁勋时隔 3 个月再次到访北京。在会谈中，黄仁勋明确表示，中国是英伟达非常重要的市场，希望将继续与 中国合作。这一表态背后，有着诸多深层次的原因。 从市场层面来看，中国市场对于英伟达而言举足轻重。据相关数据显示，中国区年营收达 171 亿美元，创下历 史新高，占其全球营收的 13%。如此庞大的市场份额，使得英伟达难以割舍中国市场。而且，中国不仅是全球 最具规模的消费市场之一，其蓬勃发展的产业生态与领先的软件实力，也成为英伟达持续创新的重要动力。在 中国市场的成功经验，推动着英伟达不断加大研发投入，与中国企业的深度合作，也助力其成长为更具竞争力 的国际化企业。 然而，就在黄仁勋访华的背景下，美国政府的一系列动作却给英伟达与中国的合作蒙上了阴影。2025 年 4 月 9 日，美国政府通知要求英伟达向中国（包括香港和澳门），或总部位于上述地区或 ...