Investment Rating - The report maintains a "Recommendation" rating for the computer industry, expecting the industry index to rise more than 5% over the next 3-6 months compared to the benchmark index [15]. Core Insights - The release of DeepSeek V3.2-Exp model is expected to drive the development of the domestic AI ecosystem, showcasing the synergy between domestic chips and large models, which leads to significant improvements in computational efficiency and cost reduction [5]. - The introduction of the DeepSeek Sparse Attention mechanism allows for a substantial enhancement in long text processing efficiency, overcoming traditional computational complexity limitations [5]. - The cost of AI applications is anticipated to decrease significantly, with the DeepSeek API prices reduced by over 50%, making advanced AI models more accessible to developers and businesses [5]. - Investment focus is recommended on domestic chip manufacturers like Cambricon and Hygon, as well as companies within the Huawei supply chain and various AI application firms [5]. Industry Basic Data - The computer industry comprises 337 listed companies with a total market capitalization of approximately 61,619.92 billion and a circulating market capitalization of about 55,710.57 billion [2]. Relative Index Performance - The absolute performance of the computer industry over the past 1 month, 6 months, and 12 months is -1.6%, 24.3%, and 72.1% respectively, while the relative performance is -4.3%, 6.3%, and 47.3% [3].

Core Insights - The article discusses how venture capital firms can seize investment opportunities and create long-term value through empowerment in the rapidly changing technology landscape, particularly focusing on AI [2] Investment Focus on AI - Xunfei Venture Capital, established in 2016, specializes in investments in the technology sector, particularly artificial intelligence [4] - The investment strategy emphasizes core capabilities and future-oriented investments, focusing on the main AI track and various application layers such as AI + new hardware, AI + life sciences, and AI + energy revolution [4][7] Project Selection Methodology - The team employs a funnel model for project selection, which includes having excellent traffic entry points, clear algorithms to filter out unqualified projects, and providing genuine empowerment [5] - Each stage of the funnel model increases the success probability by 20%-30%, enhancing the likelihood of successful investments [5] Unique Challenges in Hard Technology - The article highlights that the hard technology era differs from the internet era, requiring distinct methodologies and valuation models [6] - Investments in hard technology projects often involve a lengthy process from laboratory to market, necessitating a tailored approach for each industry [6] Dual Empowerment and Ecosystem Building - Xunfei Venture Capital's second core investment logic is "dual empowerment and ecosystem building," aiming for mutual benefits between the firm and its portfolio companies [8] - The firm has achieved business synergy with over 70% of its invested projects, focusing on collaboration across technology, business, and branding to support project growth [8] Investment Philosophy - The investment philosophy combines industry, technology, and capital to aid in constructing an AI ecosystem [9] - The goal is to create a "carrier fleet" in the AI field, enabling collaborative exploration and advancement in the AI wave [9]

Core Insights - The article discusses how venture capital institutions can seize investment opportunities amidst rapid technological changes and a new investment landscape, with insights from Xu Jingming, co-founder of iFlytek and chairman of iFlytek Venture Capital [1] Investment Focus on AI - iFlytek Venture Capital, established in 2016, focuses on investments in the technology sector, particularly artificial intelligence (AI) [2] - The investment strategy emphasizes a "funnel model" for project selection, which includes having excellent traffic entry points, clear algorithms to filter out unqualified projects, and providing genuine empowerment [2][3] - The team aims to invest in a "1+3" strategy, focusing on the main AI track and three "AI+" areas: AI + new hardware, AI + life sciences, and AI + energy revolution [3] Industry Collaboration - iFlytek Venture Capital has achieved industry collaboration with over 70% of its invested projects, aiming for mutual empowerment through technology, business, and brand synergies [4][5] - The company emphasizes a dual empowerment approach, where both the venture capital firm and the invested companies benefit from collaboration [5] Ecosystem Building - The investment philosophy of iFlytek Venture Capital is centered on "industry + technology + capital," aiming to build a robust AI ecosystem [6] - The goal is to create a "carrier fleet" in the AI field, exploring and advancing together in the wave of artificial intelligence [6]

Core Viewpoint - The company, iFlytek, has revised the use of funds from its 4 billion yuan private placement, allocating 2.4 billion yuan for computing power platform construction, while the remaining 1.6 billion yuan will be evenly distributed for the development of the Spark Education large model and typical products, as well as for working capital [2] Group 1: Fund Allocation and Strategy - The initial plan on August 22 allocated 3.2 billion yuan for working capital and 800 million yuan for the development of the Spark Education large model and typical products [2] - iFlytek aims to expand its computing power scale through a computing power leasing model to support large model research and algorithm upgrades [2] - The company is currently researching a new version of the iFlytek Spark large model that requires a larger-scale domestic computing power cluster for support [2] Group 2: Domestic Computing Power Development - iFlytek claims to have independent and controllable computing power compared to other domestic large model manufacturers [2] - In October 2023, iFlytek launched the first fully domestic intelligent computing platform, "Feixing No. 1," based on Huawei's Ascend 910B chip, achieving over 90% performance of a similarly scaled NVIDIA A800 cluster [2] - The company plans to train China's first trillion-parameter large model based on domestic computing power by January 2024 [3] Group 3: Challenges and Collaborations - The industry still heavily relies on NVIDIA's ecosystem for larger parameter models and algorithm innovations, facing challenges in adapting and optimizing algorithms based on domestic computing power [3] - iFlytek collaborates with domestic computing power manufacturers like Huawei and Cambrian to address hidden issues in domestic computing hardware and software [3][4] - The company acknowledges that adapting to new algorithms and architectures on domestic computing platforms incurs additional costs and time, slowing down model release progress [4]

Group 1 - Company plans to invest 2.4 billion yuan in computing power leasing projects to enhance its autonomous and controllable computing foundation [1] - The adjustment in fundraising usage includes 800 million yuan for working capital and 2.4 billion yuan for computing power leasing [1] - The computing power leasing project is essential for the continuous growth and competitiveness of the company's research and various business lines [1][2] Group 2 - The company launched the "Feixing No. 1" platform, supporting trillion-parameter model training, and plans to train the first domestic billion-parameter model by January 30, 2024 [2] - The computing power leasing project is expected to meet the increasing demands for computing power as large models evolve, with each token input requiring 6 to 8 floating-point operations per parameter [2] - The company has served 806,000 enterprise users with its large model, covering over 300 application scenarios, and sufficient autonomous computing power will lower the customization threshold for industry models [2] Group 3 - The investment in computing power reflects the company's commitment to autonomous controllable technology and strategic investment in the AI industry [3] - The continuous strengthening of the computing power foundation is expected to lead to breakthroughs in more core areas for the company's large model [3] - The company emphasizes the importance of autonomous innovation and competitive advantages in key areas to seize strategic development opportunities [3]

人民财讯9月30日电，企查查APP显示，近日，星火智法(上海)科技有限公司成立，法定代表人为雍文 渊，注册资本1000万元，经营范围包括人工智能应用软件开发、计算机软硬件及辅助设备零售、电力电 子元器件销售等。企查查股权穿透显示，该公司由科大讯飞（002230）旗下讯飞智元信息科技有限公司 全资持股。 ...

9 月 29 日晚，科大讯飞发布《 2025 年度向特定对象发行 A 股股票预案（修订稿）》，拟 变更募集资金用途。 【导读】科大讯飞大幅调整定增方案 中国基金报记者 卢鸰 值得注意的是，在 8 月 22 日发布定增预案之后， 8 月 26 日国务院公开发布了《关于深入 实施 " 人工智能 +" 行动的意见》，明确 " 强化智能算力统筹；支持人工智能芯片攻坚创新 与使能软件生态培育，加快超大规模智算集群技术突破和工程落地 "，" 加强智能算力互联互 通和供需匹配，创新智能算力基础设施运营模式 " 。 科大讯飞称，国家高度重视自主可控大模型算力建设，租用算力模式在国家政策支持下更具 合理性。 基于国产算力进行训练和推理的效率与使用英伟达算力存在较大差距，科大讯飞通过与华 为、中科海光、寒武纪等国产算力厂商建立长期紧密联合攻关协作机制，深度使用国产算力 集群，发现并解决了多个国产算力底层软硬件的隐藏问题。 | 序号 | 募集资金投向 | 投资总额 | 拟使用募集资金 | | --- | --- | --- | --- | | | 星火教育大模型及典型产品 | 112,571.00 | 80,000.00 | | ...

Core Viewpoint - The company Keda Xunfei has significantly revised its fundraising plan for the issuance of A-shares, changing the allocation of raised funds while maintaining the total amount at 4 billion yuan [1]. Fundraising Allocation - The revised fundraising plan includes a total investment of 432.571 million yuan, with 400 million yuan allocated for specific uses. The allocation is as follows: - 80 million yuan for the Spark Education large model and typical products - 240 million yuan for the computing power platform - 80 million yuan for supplementing working capital - Compared to the previous plan released on August 22, the amount for working capital has been reduced from 320 million yuan to 80 million yuan, while an additional 240 million yuan has been allocated to the computing power platform project, which now accounts for approximately 60% of the total fundraising [2][3]. Computing Power Platform Project - The computing power platform project aims to expand computing resources to support the iterative upgrades of the Spark cognitive large model. It will also provide computational support for the continuous optimization of various large models, including the Spark X1 model and multi-modal large models [3]. - The project aligns with national policies that emphasize the importance of autonomous and controllable large model computing power construction, making the rental of computing power more reasonable under government support [4]. National Policy Support - The recent government policy encourages the strengthening of intelligent computing power coordination and supports innovation in artificial intelligence chips. This includes accelerating breakthroughs in large-scale intelligent computing clusters and enhancing the operational models of intelligent computing infrastructure [4][5]. Global Market Strategy - Keda Xunfei plans to enhance the multi-language support capabilities of its large models, targeting overseas market demands. The company aims to optimize its technology to fit local language habits, cultural contexts, and industry applications, particularly for partners in countries along the "Belt and Road" initiative [5]. Stock Performance - As of September 29, Keda Xunfei's stock closed at 53.95 yuan per share, with a market capitalization of 124.7 billion yuan [5].

相对于8月22日发布的定增预案，科大讯飞本次修改后的方案中，40亿元的定增募资总额不变，星火教育大模型及典型产品项目8亿元的募资金额也不变； 但用于补充流动资金的金额由32亿元下调为8亿元，另外新增募集资金24亿元用于算力平台项目，该项目募资金额约占本次融资规模的60%。 【导读】科大讯飞大幅调整定增方案 中国基金报记者 卢鸰 9月29日晚，科大讯飞发布《2025年度向特定对象发行A股股票预案（修订稿）》，拟变更募集资金用途。 基于国产算力进行训练和推理的效率与使用英伟达算力存在较大差距，科大讯飞通过与华为、中科海光、寒武纪等国产算力厂商建立长期紧密联合攻关协 作机制，深度使用国产算力集群，发现并解决了多个国产算力底层软硬件的隐藏问题。 此外，《关于深入实施"人工智能+"行动的意见》还提出，要推动人工智能普惠共享。把人工智能作为造福人类的国际公共产品，打造平权、互信、多 元、共赢的人工智能能力建设开放生态。 对此，科大讯飞表示，公司算力平台项目的实施，将持续增强星火大模型的多语言支持能力，在为世界提供第二种选择的同时，将形成公司业务收入的新 增长点。 讯飞多语种大模型下一步将锚定海外市场落地需求，针对不同区域 ...

Group 1 - The core viewpoint of the article highlights the active performance of the robotics sector in the A-share market, with a notable increase in the robotics ETF (159770) and significant capital inflow [1] - As of September 30, the robotics ETF (159770) rose by 0.7% with a trading volume of 187 million yuan and a net subscription of 37.5 million units [1] - The component stocks of the robotics ETF saw substantial gains, with Yuntian Lifei-U increasing over 6%, and other stocks like Yijiahe and Keda Xunfei also showing strong performance [1] Group 2 - Data from Wind Financial Terminal indicates that the robotics ETF (159770) has experienced three consecutive days of net capital inflow, totaling nearly 470 million yuan [1] - The robotics ETF (159770) tracks the CSI Robotics Index, selecting securities from listed companies related to robotics as index samples [1] - The Ministry of Industry and Information Technology, along with five other departments, released the "Mechanical Industry Stabilization and Growth Work Plan (2025-2026)" [1] Group 3 - CITIC Securities anticipates a comprehensive bullish trend in the humanoid robotics sector for Q4, driven by optimistic business communication volumes for Tesla's humanoid robots and active shipments of domestic humanoid robots [1] - Recommendations include focusing on the T-chain and suggesting attention to sectors with better industrial trends, faster growth in sensors, dexterous hands, vertical applications, and domestic supply chains [1]