Core Viewpoint - The article discusses the transformation of the AI industry, emphasizing that the competition has shifted from hardware capabilities to a battle for software, developers, and ecosystem building, with Huawei's Ascend and its heterogeneous computing architecture CANN at the forefront of this change [1][4]. Summary by Sections CANN Open Source Announcement - Huawei's rotating chairman Xu Zhijun announced that the CANN hardware enabling will be fully open-sourced by December 30, 2025 [2]. Significance of CANN Open Source - The open-sourcing of CANN represents a profound self-revolution in the domestic AI infrastructure, aiming to break the closed model traditionally dominated by hardware manufacturers and embrace a more open and community-driven future [4][19]. - The success of the ecosystem relies on attracting academic innovation and creating a stable, universal, and efficient foundational tool for developers [5][18]. Developer Perspectives on CANN - Developers describe CANN's evolution as a challenging journey, with early versions requiring low-level programming skills, which hindered productivity [10][11]. - The introduction of the Ascend C programming language marked a significant improvement, aligning more closely with mainstream programming practices [15]. Challenges Faced by Developers - Early developers faced high technical barriers and a lack of stable architecture, leading to a difficult development environment [11][13]. - Systemic issues persisted, such as the inability to reproduce model accuracy across different frameworks due to a lack of transparency in the underlying systems [17]. The Role of Open Source - Open sourcing CANN is seen as a means to break down technical barriers and empower developers by providing transparency and control over the platform [21][23]. - The open-source model aims to foster a vibrant community where developers can contribute and innovate, moving away from reliance on a few official experts [29]. Ecosystem Empowerment - Open source provides unprecedented opportunities for deep integration between academia and industry, allowing researchers to address real-world problems and convert solutions into academic contributions [26]. - The shift from users to contributors is expected to cultivate a new generation of developers who can engage in high-quality projects [28]. Future Outlook for CANN - The current focus is on matching CUDA's capabilities while fostering original innovations within the CANN ecosystem [44]. - Huawei has committed to investing significant resources, including 1,500 petaflops of computing power and 30,000 development boards annually, to support the open-source community [45].

Group 1: Market Performance - As of August 27, 2025, the China Securities Financial Technology Theme Index (930986) increased by 0.40%, with notable gains in constituent stocks such as Lingzhi Software (688588) up 13.36% and Wealth Trend (688318) up 11.49% [3] - The Financial Technology ETF (516860) rose by 0.58%, with a latest price of 1.74 yuan, and has accumulated a 4.67% increase over the past week [3] - The Financial Technology ETF's trading volume was active, with a turnover rate of 17.42% and a transaction value of 369 million yuan [3] Group 2: Policy and Industry Outlook - The State Council recently released the "Artificial Intelligence+" action plan, aiming for over 70% application penetration of new intelligent terminals and agents by 2027, and over 90% by 2030, focusing on six key areas for AI integration [4] - The policy emphasizes the collaborative development of models, data, and computing power, and plans to establish several national AI application pilot bases [4] - Open-source securities believe that the domestic AI ecosystem is expected to accelerate its prosperity under the resonance of policy and technology [4] Group 3: Financial Technology ETF Metrics - The latest scale of the Financial Technology ETF reached 2.118 billion yuan, marking a one-year high, with the latest share count at 1.226 billion shares [4] - The ETF saw a net inflow of 19.05 million yuan, with a total of 287 million yuan net inflow over the past five trading days [4] - The leveraged funds continue to invest, with a net purchase amount of 1.8509 million yuan this month and a latest financing balance of 119 million yuan [4] Group 4: Performance and Returns - As of August 26, 2025, the Financial Technology ETF's net value increased by 181.20% over the past year, ranking 3rd out of 2977 index equity funds [5] - The ETF has recorded a maximum monthly return of 55.92% since inception, with an average monthly return of 10.10% and a historical three-year profit probability of 97.77% [5] - The Sharpe ratio for the ETF over the past year is 2.01, indicating strong risk-adjusted returns [5] Group 5: Risk and Fee Structure - The Financial Technology ETF has a management fee rate of 0.50% and a custody fee rate of 0.10%, which are among the lowest in comparable funds [6] - The tracking error for the ETF over the past year is 0.044%, demonstrating the highest tracking precision among comparable funds [6] - The index closely tracks the performance of companies involved in financial technology, with the top ten weighted stocks accounting for 51.26% of the index [6]

Group 1 - The AI industry is experiencing a trend driven by the resonance of policy and technology, with Huawei's HarmonyOS terminal count surpassing 10 million [1] - Major applications like Taobao are expected to complete HarmonyOS adaptation by the end of September, and Tencent has migrated over 50 products to HarmonyOS, indicating strong native intelligence performance [1] - The release of DeepSeek-V3.1, which utilizes a hybrid inference architecture and UE8M0 FP8 parameter precision, is expected to enhance the domestic AI ecosystem by improving performance and compatibility with domestic chip designs [1] Group 2 - The computer industry is showing strong performance, with promising prospects for AI computing power and applications, driven by improvements in leading large model performance and domestic computing power adaptation [1] - The implementation of stablecoin regulations is fostering innovation in the industry, contributing to a high level of activity in the capital markets [1] - The Software ETF (code: 515230) tracks the software index (code: H30202), which reflects the overall performance of listed companies in the software development and IT services sectors in the A-share market [1]

Group 1 - The Hong Kong stock market shows a significant recovery, with strong gains in semiconductor stocks, supported by breakthroughs in domestic AI models that enhance the application of domestic chips [1] - Dongfang Securities highlights the release of DeepSeek's V3.1 version, designed for next-generation domestic chips, which is expected to promote the large-scale application of domestic AI chips in the inference market [1] - The domestic AI chip sector is experiencing continuous improvements in manufacturing processes and technical designs, providing a solid foundation for the localization of the AI computing power chain [1] Group 2 - CITIC Construction Investment believes that the domestic computing power sector is entering a window of innovation, with recent catalysts for domestic computing power chips [2] - NVIDIA has announced the launch of its Spectrum-XGS Ethernet technology aimed at connecting independent data centers to create an AI super factory with a computing capacity in the hundreds of billions [2] - As of August 25, 2023, the A-share NVIDIA industry chain index saw a rise of over 2%, with notable gains in stocks such as Zhongke Shuguang and Aobi Zhongguang [2] Group 3 - The Hong Kong Stock Connect Technology ETF has seen significant inflows, with a total net inflow of 57.96 million yuan over four days, reaching a new high of 3.33 billion yuan in total assets [3] - Major AI companies like Xiaomi, Alibaba, and Tencent account for nearly 30% of the ETF's weight, alongside core "hard tech" stocks like SMIC and Huahong Semiconductor [3] - The market liquidity is improving, with expectations of a Fed rate cut boosting global risk appetite, which may benefit the Hong Kong stock market [3] Group 4 - The Hong Kong Stock Connect Technology ETF focuses on wafer fabs and AI applications, aiming to capture opportunities in the "pure technology revolution" of the AI era [4]

Core Viewpoint - The surge in A-share semiconductor and computing power stocks, led by companies like Cambricon, Haiguang Information, and Zhongke Shuguang, was triggered by a comment from DeepSeek on August 22, resulting in significant price increases and trading halts for these stocks [1][2]. Company Performance - Cambricon's stock price broke through the 1100 yuan and 1200 yuan marks, closing at 1243.20 yuan with a total market capitalization of 520 billion yuan [2]. - Year-to-date, Cambricon's stock has increased by nearly 90%, with significant gains attributed to major shareholder Zhang Jianping, who has seen substantial profits from his holdings [4]. - In Q1, Cambricon reported a revenue of 1.11 billion yuan, a year-on-year increase of over 420%, and a net profit of 355 million yuan [4]. Shareholder Activity - Zhang Jianping increased his holdings in Cambricon by 14% in Q1, bringing his total to over 6 million shares, with a reference market value of approximately 3.79 billion yuan [4]. - In addition to Cambricon, Zhang Jianping also invested heavily in four other "AI+" stocks, totaling nearly 6 billion yuan across five stocks [5]. - Zhang Jianping's recent activities include a significant reduction in his holdings in Hangang Co., where he decreased his stake by over 82% in Q2 [11]. Market Trends - The AI and computing power sectors are experiencing rapid growth, with companies like Hainan Huatie and Zhongjian Technology also seeing substantial stock price increases [7][15]. - Hainan Huatie signed a major computing power service agreement worth 3.69 billion yuan, contributing to its stock price surge of nearly 85% in Q1 [7]. - The Chinese computing power market is expected to grow significantly, with projections indicating a growth rate exceeding 40% by 2025 [18].

在人工智能的浪潮中，算力一直是推动技术发展的核心动力。如今，国产算力正以前所未有的速度崛起，成为全球AI领域的一匹黑马。近日，潞晨科技 与华为昇腾的强强联合，推出了一系列基于国产910B算力的DeepSeek R1推理API及云镜像服务，不仅性能追平高端GPU，还免费开放体验，这无疑是 国产AI生态建设的一大里程碑，也标志着智能新时代的到来。 一、国产算力的华丽转身：从追赶者到引领者 曾经，高端GPU在AI算力领域占据着绝对的主导地位，国产算力似乎总是难以望其项背。然而，随着技术的不断进步和创新，国产算力正 在悄然改变这一格 局。 潞晨科技与华为昇腾的合作，正是这一转变的生动体现。 （二）灵活选择：从大模型到小模型 潞晨云平台为DeepSeek R1系列模型提供了从满血671B大模型到高效蒸馏小模型的多种选择。 开发者可以根据具体的业务场景，精准选择最适合的模型，实 现精度与效率的最佳平衡。 （一）技术突破：深度适配与优化 潞晨科技凭借其自研的推理引擎，成功实现了与华为昇腾910B算力的深度适配与优化。这一技术突破不仅让国产算力在性能上追平了高端GPU，还为开发者 提供了更加高效、灵活和稳定的AI推理服务。 ...