Core Viewpoint - China aims to increase the domestic self-sufficiency rate of semiconductors used in AI data centers to over 70% by 2027, with Beijing targeting 100% self-sufficiency, reducing reliance on US products like Nvidia [2][4]. Group 1: Government Plans and Policies - Shanghai plans to achieve over 70% self-sufficiency in semiconductors for AI data centers by 2027, while Beijing aims for 100% [4]. - Guizhou province's Gui'an New Area mandates that approximately 90% of semiconductors for AI-focused data centers be sourced from Chinese manufacturers [4]. - The Chinese Communist Party emphasizes the importance of self-reliance in AI semiconductor development, following a collective study session on AI technologies [4]. Group 2: Market Dynamics and Competitors - Nvidia currently holds an 80% market share in China's AI semiconductor sector as of early 2024 [4]. - Huawei's Ascend 910 series aims to replace Nvidia's H20 products, with the current 910B model achieving about 85% of H20's computing power [5]. - Market predictions suggest Nvidia's share may drop to 50-60% in the next five years, while Chinese companies could increase their share to 40-50% [5]. Group 3: Adoption of Domestic Products - The Chinese government encourages major tech companies to switch to domestic products, with ByteDance reportedly considering Huawei's offerings [7]. - There are strong recommendations against using Nvidia's H20 in government-related projects, promoting the use of domestic alternatives even if they are less performant [7]. Group 4: Challenges and Balancing Act - Despite efforts to reduce dependence on US technology, challenges remain, such as slow progress in AI model development due to reliance on Huawei products [7]. - Experts suggest that China may seek a balance between importing Nvidia products and developing its own technology to enhance AI capabilities [8].

Core Insights - The article discusses the fluctuations in capital expenditure among major tech companies following the release of Deepseek, highlighting a surge in domestic computing power and a subsequent cooling off after Tencent's capital expenditure announcement [1] - The ongoing US-China tensions are identified as a significant factor affecting the ability of major companies to invest, particularly in data center expansions due to difficulties in acquiring necessary hardware [1] - Despite the challenges, demand for computing power remains strong, with companies like Nvidia recovering from early-year declines and domestic firms like Huawei and others making strides in the computing power sector [1][2] Group 1 - Major tech companies have increased capital expenditures in response to the booming domestic computing power market, but this trend has cooled following Tencent's announcements and the impact of the "tariff war" [1] - The US-China friction is causing difficulties in hardware acquisition, which is expected to slow down the expansion of data centers across major companies [1] - Nvidia has rebounded from earlier declines, and overseas demand for computing power remains robust, with domestic companies also achieving historical highs in certain sectors [1] Group 2 - Nvidia's founder has acknowledged Huawei as a strong competitor, noting that a significant portion of global AI scholars are based in China, indicating the competitive landscape in the AI sector [2] - Major domestic companies are not reducing their investments in AI, with Alibaba emphasizing its strategic focus on AI for the next decade [2] - Progress in US-China negotiations may lead to improved access to overseas computing power cards, with expectations for more AI applications to be launched in the second half of the year [2] Group 3 - The article suggests that the domestic computing power sector is poised for a resurgence, indicating optimism about future developments [3]

Core Viewpoint - The article discusses the growing interest and procurement plans for domestic GPUs, particularly Huawei's Ascend series, in light of increasing restrictions on GPU imports from the U.S. [1] Domestic GPU Procurement Situation - ByteDance is the most proactive in adopting domestic GPUs, primarily using Cambrian and Ascend, with plans to increase procurement in 2025 [2] - Alibaba is more cautious, planning to purchase Haiguang and Cambrian cards in 2025 while continuing to stockpile NVIDIA products [3] - Tencent is the most conservative, preferring NVIDIA products, followed by H company's offerings, and only considering Cambrian and Haiguang as alternatives [3][4] Ascend 910 Series Latest Situation - As of early June, the Ascend 910B has shipped approximately 160,000 units, while the 910C has only shipped about 27,000 units due to production delays [5] - The 910C has received positive feedback for its cluster capabilities but still lags behind NVIDIA in performance [5][6] - The annual shipment target for the Ascend series has been adjusted from 850,000 to around 700,000 units due to demand fluctuations and market conditions [6] Application Scenarios and Challenges - Approximately 15% to 20% of Ascend cards are used for model training, with the majority for inference tasks [6] - High costs and technical challenges limit the widespread use of Ascend cards in high-end training applications [6][8] - The development of multi-modal models is driving higher hardware standards, with domestic manufacturers striving to catch up with international brands [7] Custom Development and Market Outlook - Collaborating with Ascend for customized development offers cost advantages but requires significant human resources for adaptation [8] - The domestic chip industry is expected to expand its market share as technology advances and market demands evolve [8]

Core Viewpoint - The report from Mizuho Securities provides an analysis of companies including Broadcom, NVIDIA, AMD, Supermicro, and Huawei, highlighting the expected growth and challenges in the AI ASIC and GPU markets. Group 1: Broadcom and NVIDIA - Mizuho expects Broadcom's custom ASIC chips (TPUv7p/MTIA2) to accelerate in deployment by 2026, potentially being used in OpenAI's Strawberry and Apple's Baltra projects in the second half of 2026 [1] - In 2024, Broadcom's custom ASIC chips are projected to account for 70%-80% of usage, establishing it as a leader in AI ASICs, excluding self-manufactured AI ASICs like Google's TPU [1] - The UMAIN project in Saudi Arabia plans to deploy 4,000 GB200 NVL72 servers, corresponding to 280,000 NVIDIA GPUs and 350,000 AMD GPUs over the next five years [1] - The G42 project in the UAE has committed to importing 500,000 NVIDIA GB200 GPUs annually, valued at $15 billion, although the sustainability of this figure is questioned [1] Group 2: Huawei - The report anticipates that Huawei's Ascend 910 orders will exceed 700,000 units by 2025, with the next-generation Ascend 920 expected to launch in 2026 [2] - However, the current yield rate for the Ascend 910 is low at only 30%, a figure corroborated by previous reports [2][3] - Other estimates suggest that the shipment volume for the Ascend 910 series could be over 700,000 units this year [5]

Group 1: Macroeconomic and Industry Background - The return of the technology growth market in May is supported by a combination of loose liquidity and a recovery in risk appetite, with global liquidity remaining abundant and the offshore RMB surpassing 7.20 [2] - The US Federal Reserve's dovish signals and a shift in US fiscal policy towards balance have contributed to a weaker dollar, which further strengthens the rationale for allocating to growth stocks [2] - The end of the April earnings season has cleared risks for the technology sector, with the TMT sector's trading congestion dropping to an 11% historical low for 2023, providing space for capital rotation [3] Group 2: AI Revolution Deepening - Major overseas tech companies have significantly increased capital expenditures related to AI, with a year-on-year growth rate exceeding 30%, focusing on computing chips and cloud infrastructure [7] - The market anticipates that global AI computing investment will exceed $200 billion by 2025, indicating a robust demand for AI infrastructure [7] - Domestic companies are accelerating the deployment of multimodal AI agents, with new AI hardware products being launched, driving the transformation of consumer electronics towards "AI + terminals" [9][10] Group 3: Intelligent Robotics - The mass production of humanoid robots is accelerating, with predictions indicating that by 2050, there will be 100 million humanoid robots globally, generating annual revenues of $5 trillion [11] - The market for exoskeleton robots is expanding, with applications in medical and industrial fields, and the expected market space exceeding hundreds of billions [15] - The US's tariff exemptions on auto parts are reducing costs in the supply chain, benefiting the automotive and robotics sectors [17] Group 4: Autonomous Driving - Policies in both China and the US are accelerating the commercialization of autonomous driving, with new regulations and operational licenses being issued [18] - Key segments of the industry chain benefiting from this trend include high-precision sensors, computing chips, and high-precision mapping [19] Group 5: Investment Strategy for May - Focus on core sectors such as AI, robotics, and autonomous driving, with specific attention to AI applications and hardware, including AI PCs and AI glasses [20][21] - Identify key components in robotics, such as dexterous hands and lightweight materials, as well as critical elements in autonomous driving like lidar and vehicle communication [21] - Emphasize the importance of selecting companies with high R&D investment and strong order visibility while avoiding stocks with uncertain earnings [23]