Core Insights - Morgan Stanley's report compares the operational costs and profit margins of various AI solutions in inference workloads, highlighting that most multi-chip AI inference "factories" have profit margins exceeding 50%, with NVIDIA leading the pack [1][3]. Profit Margins - Among selected 100 MW AI "factories," NVIDIA's GB200 NVL72 "Blackwell" GPU platform achieved the highest profit margin of 77.6%, translating to an estimated profit of approximately $3.5 billion [3]. - Google's self-developed TPU v6e pod ranked second with a profit margin of 74.9%, while AWS's Trn2 UltraServer and Huawei's Ascend CloudMatrix 384 platform reported profit margins of 62.5% and 47.9%, respectively [3]. Performance of AMD - AMD's performance in AI inference is notably poor, with its latest MI355X platform showing a profit margin of -28.2%, and the older MI300X platform at a significantly lower -64.0% [4]. Revenue Generation - NVIDIA's GB200 NVL72 chip generates $7.5 per hour, while the HGX H200 chip produces $3.7 per hour. Huawei's Ascend CloudMatrix 384 platform generates $1.9 per hour, and AMD's MI355X platform only generates $1.7 per hour [4]. - Most other chips generate revenue between $0.5 and $2.0 per hour [4].

Core Viewpoint - The article discusses Huawei's advancements in AI training systems, particularly focusing on the MoE (Mixture of Experts) architecture and its optimization through the MoGE (Mixture of Generalized Experts) framework, which enhances efficiency and reduces costs in AI model training [1][2]. Summary by Sections Introduction to MoE and Huawei's Innovations - The MoE model, initially proposed by Canadian scholars, has evolved significantly, with Huawei now optimizing this architecture to address inefficiencies and cost issues [1]. - Huawei's MoGE architecture aims to create a more balanced and efficient training environment for AI models, contributing to the ongoing AI competition [1]. Performance Metrics and Achievements - Huawei's training system, utilizing the "昇腾+Pangu Ultra MoE" combination, has achieved significant performance metrics, including a 41% MFU (Model Floating Utilization) during pre-training and a throughput of 35K Tokens/s during post-training on the CloudMatrix 384 super node [2][26][27]. Challenges in MoE Training - Six main challenges in MoE training processes are identified: difficulty in parallel strategy configuration, All-to-All communication bottlenecks, uneven system load distribution, excessive operator scheduling overhead, complex training process management, and limitations in large-scale expansion [3][4]. Solutions and Innovations - **First Strategy: Enhancing Training Cluster Utilization** - Huawei implemented intelligent parallel strategy selection and global dynamic load balancing to improve overall training efficiency [6][11]. - A modeling simulation framework was developed to automate the selection of optimal parallel configurations for the Pangu Ultra MoE model [7]. - **Second Strategy: Releasing Computing Power of Single Nodes** - The focus shifted to optimizing operator computation efficiency, achieving a twofold increase in micro-batch size (MBS) and reducing host-bound issues to below 2% [15][16][17]. - **Third Strategy: High-Performance Scalable RL Post-Training Technologies** - The introduction of RL Fusion technology allows for flexible deployment modes and significantly improves resource utilization during post-training [19][21]. - The system's design enables a 50% increase in overall training throughput while maintaining model accuracy [21]. Technical Specifications of Pangu Ultra MoE - The Pangu Ultra MoE model features 718 billion parameters, with a structure that includes 61 layers of Transformer architecture, achieving high performance and scalability [26]. - The training utilized a large-scale cluster of 6K - 10K cards, demonstrating strong generalization capabilities and efficient scaling potential [26][27].

Market Overview - The Hang Seng Index showed upward movement at the end of April, providing guidance for May's market trends [1] - Historically, overseas markets tend to rise during long holidays, with notable increases in AI giants' stock prices, such as Microsoft and Mate [2] Economic Indicators - The U.S. non-farm payroll data for April exceeded expectations, with an increase of 177,000 jobs compared to the forecast of 138,000 [3] - The upcoming Federal Reserve meeting on May 7 is expected to maintain current interest rates, with a strong focus on the strengthening of the RMB, which recently surpassed the 7.20 mark for the first time since November [4] - The fiscal deficit for the year is projected at 5.66 trillion yuan, an increase of 1.6 trillion yuan from the previous year, indicating a higher deficit rate of 4% [4] Industry Trends - The release of the Harmony OS PC version is anticipated soon, which aims to reduce reliance on Windows and create a self-sufficient ecosystem, potentially reshaping the PC market [5] - Xiaomi's MiMo model, with a parameter scale of 7 billion, is designed for lightweight deployment, particularly in mobile and automotive applications [5] - The humanoid robot market is expected to see significant growth, with projections indicating that humanoid robots will enter the general product category by 2026, reaching a production or sales threshold of 100,000 units [5] Company Performance - UBTECH Robotics (09880) is projected to achieve a revenue of 1.305 billion yuan in 2024, a year-on-year increase of 23.7%, driven by growth in educational and customized intelligent robot products [6] - The company’s revenue from educational robots and solutions is expected to reach 363 million yuan, while revenue from customized robots is projected to grow by 126.1% to 141 million yuan [6][7] - The company has made advancements in navigation, machine vision, and AI integration, which are expected to accelerate the industrialization of humanoid robots [8] AI Industry Developments - The AI sector remains a focal point in the U.S.-China tech rivalry, with recent calls for restrictions on AI technology flow to China [9] - Major Chinese tech companies are ramping up AI development, with Alibaba and Tencent making significant advancements in their AI models [9] - Upcoming releases and investments in AI infrastructure are expected to catalyze growth in the technology sector, with Huawei set to launch new products in May [9] Market Sentiment - The Hang Seng Index is showing bullish sentiment, supported by a concentration of market interest in technology and consumer sectors, particularly in robotics and AI applications [10][11]