从乌兰察布市大数据局了解到，近日，华为云基于CloudMatrix384超节点的新一代昇腾AI云服务在乌兰 察布正式上线，这标志着当地在AI算力基础设施领域取得重大突破，将为千行百业智能化转型提供高 效稳定的算力支撑。华为云CloudMatrix384超节点首创将384颗昇腾NPU通过全新高速网络对等互联， 打破传统AI服务器边界，形成一台超级"AI服务器"，可提供高达300PFLOPs的算力，从根本上解决传统 架构的瓶颈。与和林格尔、贵安、芜湖等地相比，此次乌兰察布上线的CloudMatrix384超节点，具有多 项显著优势，其单卡推理吞吐量达2300Tokens/s，较传统架构提升近4倍，甚至超越英伟达H100，实现 了性能倍增、以存强算、MoE亲和、长稳可靠、朝推夜训、即开即用六大特性。同时依托百TB级带宽 的光纤骨干网，乌兰察布覆盖京津冀10ms时延圈，全国主要流量高地城市都能在30毫秒之内访问超节 点资源，为用户提供低延迟、高效率的算力服务。(新浪财经) ...

Core Viewpoint - The article emphasizes that the AI industry is transitioning into a new phase where system architecture and efficiency in communication are becoming more critical than just chip performance. This shift is highlighted by the introduction of Huawei's CloudMatrix384 super node, which aims to address the communication bottlenecks in AI data centers [1][4][80]. Group 1: AI Industry Trends - The AI competition has evolved from focusing solely on chip performance to a broader dimension of system architecture [2][80]. - The current bottleneck in AI data centers is the communication overhead during distributed training, leading to a significant drop in computing efficiency [4][80]. - A fundamental question arises: how to eliminate barriers between chips and create a seamless "computing highway" for AI workloads [5][80]. Group 2: Huawei's CloudMatrix384 - Huawei's CloudMatrix384 super node features 384 Ascend NPUs and 192 Kunpeng CPUs, designed to create a high-performance AI infrastructure [5][11]. - The architecture employs a fully peer-to-peer high-bandwidth interconnectivity and fine-grained resource disaggregation, aiming for a vision of "everything poolable, everything equal, everything combinable" [8][80]. - The introduction of a revolutionary internal network called "Unified Bus" allows for direct and high-speed communication between processors, significantly enhancing efficiency [13][15]. Group 3: Technical Innovations - CloudMatrix-Infer, a comprehensive LLM inference solution, is introduced alongside CloudMatrix384, showcasing best practices for deploying large-scale MoE models [21][80]. - The new peer-to-peer inference architecture decomposes the LLM inference system into three independent subsystems: prefill, decode, and caching, enhancing resource allocation and efficiency [23][27]. - A large-scale expert parallel (LEP) strategy is developed to optimize MoE models, allowing for high expert parallelism and minimizing execution delays [28][33]. Group 4: Cost and Utilization Benefits - Directly purchasing and operating CloudMatrix384 poses significant risks and challenges for most enterprises, including high initial costs and ongoing operational expenses [44][46]. - Huawei Cloud offers a rental model for CloudMatrix384, allowing businesses to access top-tier AI computing power without the burden of ownership [45][60]. - The cloud model maximizes resource utilization through intelligent scheduling, enabling a "daytime inference, nighttime training" approach to optimize computing resources [47][60]. Group 5: Performance Metrics - Huawei Cloud deployed a large-scale MoE model, DeepSeek-R1, on CloudMatrix384, achieving impressive throughput metrics during both the prefill and decode stages [62][70]. - The system demonstrated a throughput of 6,688 tokens per second during the prefill phase and maintained a decoding throughput of 1,943 tokens per second, showcasing its efficiency [66][69]. - The architecture allows for dynamic adjustments to balance throughput and latency, adapting to different service requirements effectively [73][80].

Market Overview - The A-share market showed mixed performance with the Shanghai Composite Index rising by 0.26% while the Shenzhen Component Index and the ChiNext Index fell by 0.85% and 1.21% respectively. The total trading volume in the Shanghai and Shenzhen markets reached 11,117.15 billion yuan [2][7] - In the industry sectors, transportation, banking, and coal led the gains, while the comprehensive, machinery equipment, and automotive sectors experienced declines. Concept sectors such as Tianjin Free Trade Zone, China-South Korea Free Trade Zone, and dairy industry saw gains, while animal vaccines, controllable nuclear fusion, and digital watermarking faced declines [2][7] Economic Indicators - In the first quarter, China's GDP grew by 5.4% year-on-year, with a total GDP of 318,758 billion yuan. The industrial added value increased by 6.5% year-on-year, and the service sector's added value grew by 5.3%. Retail sales of consumer goods reached 124,671 billion yuan, up by 4.6% [3][8] - Fixed asset investment (excluding rural households) was 103,174 billion yuan, growing by 4.2%. The urban unemployment rate averaged 5.3%, and the per capita disposable income was 12,179 yuan, reflecting a nominal increase of 5.5% [3][8] Industry Insights Battery Industry - The new national standard for electric vehicle batteries, effective from July 1, 2026, includes significant upgrades in safety requirements, particularly in thermal diffusion, bottom impact, and fast charging cycle safety [9][11] - Solid-state batteries are expected to see accelerated industrialization due to their superior safety performance compared to traditional lithium batteries. The new standards will likely increase demand for solid-state batteries, facilitating their transition from research to mass production [11][12][13] Semiconductor Industry - The U.S. government's licensing requirements for exporting H20 chips to China indicate increased trade restrictions, which may lead to a loss of market share for Nvidia in China. This situation presents opportunities for domestic AI chip manufacturers to capture more market share [14][15] - The ongoing trade tensions are expected to accelerate the domestic semiconductor industry's development, enhancing the importance of self-sufficiency in the supply chain. Domestic chip companies may benefit from cost advantages in design, manufacturing, and packaging [15][18] Company-Specific Analysis Beijing Bank - Beijing Bank reported a 12.6% year-on-year growth in total assets as of the end of 2024, with loans growing by 9.8% and financial investments by 12.8%. The average net interest margin for 2024 was 1.47%, reflecting a 7 basis points decline year-on-year [20][21] - The bank's non-performing loan ratio stood at 1.31%, a slight decrease from the previous year, while the coverage ratio was 209%. The bank is actively addressing problem assets and investing in digital transformation [21][22]