Nvidia(US:NVDA) Zheng Quan Shi Bao·2025-09-28 18:28Core Insights - The demand for AI computing power has surged this year, leading to significant stock price increases for AI chip and related industry companies, such as Cambricon, Shenghong Technology, and Industrial Fulian [1][2] - Concerns about potential market bubbles have emerged, including issues related to valuation, technological bottlenecks, and capital competition [1] - NVIDIA and OpenAI announced a groundbreaking investment plan, committing $100 billion to build a 10 GW super AI data center, which is expected to further enhance market expectations for AI computing power [1][2] AI Infrastructure Competition - The global AI competition has entered a new phase characterized by "super large scale, super high energy consumption, and super high investment" [2] - NVIDIA's partnership with OpenAI marks a significant milestone in AI computing, with NVIDIA's stock price reaching an all-time high following the announcement [2] - This collaboration is expected to boost the entire AI computing supply chain, from chips to data center operations, while intensifying the competition between nations over computing power sovereignty [2][3] Shift from Model to Computing Power - The competition in AI is shifting from model development to foundational computing power, with the U.S. controlling 75% of global computing power [3][4] - China's efforts to catch up in computing hardware are likely to intensify, with a focus on semiconductor collaboration to overcome hardware bottlenecks [3][4] - The emergence of domestic AI models like DeepSeek and Kimi K2 indicates progress in narrowing the gap with U.S. models, although challenges in computing power remain [4] Opportunities and Challenges in Data Center Development - China is advancing in the construction of super AI data centers, supported by a vast market and abundant resources [5][6] - The "East Data West Computing" initiative aims to create a network of super data centers, enhancing the country's AI infrastructure [6][7] - However, challenges persist, particularly in the semiconductor supply chain and GPU availability, which are critical for AI data centers [8] Investment Logic Restructuring - The core of AI competition is shifting from "model chasing" to "building an autonomous computing foundation," impacting industry development paths and investment strategies [9] - There is a growing trend towards investing in hardware first, followed by software, with significant opportunities expected in AI edge devices in the near future [9][10] - The market is likely to see a transition from cloud-based AI to edge-based applications, creating broader opportunities in sectors like edge devices and communication modules [9][10]