电子行业算力周跟踪：英伟达Blackwell推迟持续发酵，持续重视国产算力需求释放

Investment Rating - The industry is rated as "Outperform the Market" with a focus on maintaining this rating [21]. Core Insights - The delay in the launch of Nvidia's Blackwell architecture GPU due to design defects has caused significant market reactions, impacting companies like Intel and domestic players in the computing power sector [3][4]. - Despite the negative sentiment, the demand for computing power remains strong, suggesting potential growth opportunities in the computing power chain [2][3]. - The procurement results from China Mobile indicate a significant focus on domestic AI server production, with a total of 7,994 AI servers procured, amounting to 19.1 billion yuan [4]. - The successful operation of large models on general-purpose servers highlights opportunities in the CPU server market, which can reduce deployment costs compared to AI acceleration chips [4][6]. - Lightweight open-source models are expected to lower computational requirements, potentially impacting mobile AI and edge computing significantly [6]. Summary by Sections Nvidia Blackwell Architecture - The launch of Nvidia's Blackwell AI GPU has been delayed by design defects, with potential impacts on major clients like Meta, Google, and Microsoft [3]. - Key estimates suggest that Blackwell could generate revenue for Nvidia's data center segment, increasing from $47.5 billion in 2024 to over $200 billion in 2025 [3]. Domestic Computing Power Demand - China Mobile's recent procurement of AI servers emphasizes the ongoing trend of domestic substitution in computing power [4]. - The procurement includes a significant share for Kunlun Technology (21.05%) and other domestic firms, indicating a shift towards local suppliers [4]. AI Model Tracking - The download rates for various AI applications, including those from iFlytek and Tencent, show a growing interest in AI technologies, with iFlytek's app downloads exceeding 3.6 million in a week [7][8]. - The performance of lightweight models and their accessibility on platforms like Google Colab is expected to democratize AI usage [6].