节后四季度电子行业投资机会梳理 20251007 摘要 算力需求持续增长，除英伟达外，AMD、ATC 等厂商或将获得更多市场 份额，海外 ASIC 和国产算力存在补涨机会，A 股通富微电及 PCB、液 冷产业链值得关注。 华为升腾系列芯片 2025 年表现低于预期，但最新发布的 950、960、970 系列在生态兼容性和技术支持上取得进展，预计 2026 年华为升腾产业链将迎来增长。 PCB 领域正焦倍版和预埋电容等新技术值得关注，液冷技术因散热密度 增加和节能需求将持续发展，英维克、飞荣达等公司存在潜在机会。 存储行业三季度价格环比二季度增长显著，业绩表现亮眼，DRAM 和 SSD 价格持续提升，产业链机会包括 SSD、DRAM 以及 HDD 领域。 美股 HPD 表现推动股价上涨，HDD 短缺导致价格上涨，企业加速转向 SSD，尽管 HDD 单单位容量成本较低，但 SSD 在企业级存储速度上具 有优势，国内厂商 ESSD 解决方案预计明年量产。 存储行业本轮周期由供需共振驱动，超越上一轮周期，近期 DRAM 和 SSD 价格均有所提升，10 月第一周和 9 月下旬期间 SSD 涨幅较明显。 端侧设备方面， ...

2025 年我们看到了人形机器人的强劲发展机会，主要催化因素来自于特斯拉 等头部厂商的产业进展，例如 Gen 3 发布及其对零部件需求的推动。预计到 2026 年底，人形机器人月度产销可达 1 万台左右，全年总产量可能在小几万 到 5 万台之间。考虑到头部份额约为 20%-30%，行业总量应在 10 多万台水 平。单台价格合理区间为 2 万至 25,000 美元，其中 BOM 占售价比例约六七 成，即每套 BOM 约 10 万美元。因此，人形机器人零部件全球市场容量预计可 达 100 至 200 亿美元。这一增量市场相当于千元级别汽车零部件市场规模， 其中核心零部件公司如德昌电机控股、拓普集团和肇民科技值得重视。 汽车行业四季度展望 20251008 摘要 欧洲新能源车市场增长确定性强，受益于欧盟碳排放政策，中国车企如 上汽、比亚迪等表现突出，预计将持续受益并成为全球增长最强劲的市 场之一。 人形机器人领域 2025 年迎来发展机遇，特斯拉等头部厂商的产业进展 是主要催化剂，预计 2026 年底月度产销可达 1 万台，零部件市场容量 预计达 100-200 亿美元。 汽车零部件企业在算力投资方面涌现机会，如潍 ...

摘要： 2025年Q3的A股，科技成长主线彻底"霸屏"——AI算力、半导体、机器人、消费电子轮番冲高，我们持续带领会员精准踩中机构资金节奏，重仓出海链、 国产半导体、机器人等核心标的，超额收益拉满。 展望Q4，科技成长仍将是绝对主角，半导体、AI算力、机器人、消电端侧四大方向值得超配，还能兼顾双创弹性与龙头红利。今天就把Q3的赚钱逻辑拆 透，再给大家划Q4的黄金布局点。 01回顾Q3的3个轻松"超额"策略，吃透科技牛 策略1：认准"机构主导的科技牛"，抓准资金流向 继去年11月到今年2、3月持续强调科技牛大机遇后，7月16日我们又在公众号文章《》提出需要重视基本面投研并及时转变为机构手法，同时重点抓住包 括AI算力、机器人等为主的科技成长主线方向核心标的趋势机会。 逻辑很简单：国家队（社保、大基金）+两融资金带头冲，险资、公私募跟着加仓。Q3两融余额从1.8万亿飙到2.4万亿，科技股融资占比超60%，相当 于"聪明钱"集体用脚投票科技主线。而且中国经济向科技转型，资本市场"含科量"越来越高，国家队配置自然向科技倾斜。 行情验证超给力：寒武纪涨120%、海光信息涨78%，工业富联更猛，直接涨218%创历史新高； ...

Investment Rating - The industry investment rating is "Accumulate (Maintain)" [2] Core Viewpoints - The report indicates that the high-quality development of the Beijing Stock Exchange (北交所) is expected to yield significant results in 2025, with a positive outlook for the overall performance of the exchange in the coming year [21] - The report highlights the performance of the North Exchange 50 Index, which experienced a decline of 3.11% to close at 1528.98 points as of September 28, 2025, while other indices such as the Shanghai and Shenzhen 300, ChiNext, and Sci-Tech 50 showed increases of 1.07%, 1.96%, and 6.47% respectively [3][6] - The report emphasizes the average market capitalization of the 277 constituent stocks in the North Exchange, which stands at 3.138 billion yuan, and notes a significant valuation premium compared to other A-share markets [8][11] Summary by Sections 1. North Exchange Market Overview - As of September 28, 2025, the North Exchange has 277 constituent stocks with an average market capitalization of 3.138 billion yuan [6] - The North Exchange A-shares had a median PE ratio of 47 times, significantly higher than the median PE ratios of 29 times for the entire A-share market, 39 times for ChiNext, and 41 times for Sci-Tech Board, indicating a valuation premium of 164%, 121%, and 116% respectively [8] 2. Industry Performance - The top five performing industries in the A-share market for the week were electrical equipment, non-ferrous metals, electronics, media, and public utilities, with respective increases of 3.86%, 3.52%, 3.51%, 0.63%, and 0.28% [11] 3. North Exchange Individual Stock Performance - Among the 277 stocks listed on the North Exchange, 19 stocks increased in value, while 256 stocks decreased, resulting in an increase ratio of 7% [13] 4. New Stocks on the North Exchange - As of September 28, 2025, 19 companies, including Dongsheng Jin Material and Baiying Biological, have reached the inquiry stage, while three companies, including Jingchuang Electric and Dapeng Industrial, are in the registration stage [17][18] 5. Investment Strategy for the North Exchange - The report suggests focusing on several sectors for investment in 2025, including: 1. Data Centers: KLT and Shuguang Shuchuang 2. Robotics: Suzhou Axis, Audiwei, Jun Chuang Technology, and Fuheng New Materials 3. Semiconductors: Hualing Co. and Kaide Quartz 4. Consumer Goods: Taihu Snow, Boshenglong, Lusi Co., Kangbiter, and Thunder God Technology 5. Military Information Technology: Chengdian Guangxin and Xingtuzhihui [21]

Group 1 - The commercialization of AI applications is approaching a critical turning point, driven by the maturation of underlying technology and the validation of business models [8][23][24] - The efficiency and cost-effectiveness of AI computing power have significantly improved, with the cost of inference for AI models dropping dramatically, providing economic support for commercialization [9][20] - The capabilities of large models have reached commercial standards, with advancements in multi-turn dialogue, long-term memory, and multi-modal perception, exemplified by models like GPT-5 [19][11][12] Group 2 - The demand for AI applications is rapidly expanding, as indicated by the significant increase in token consumption, which reflects the growing commercial value of AI across various sectors [24][26][28] - AI programming has emerged as a leading area for commercialization, with notable advancements and user growth in tools like GitHub Copilot and Cursor, indicating a clear path to revenue generation [28][29][49] - AI multi-modal applications, particularly in visual content generation, are becoming commercially viable, with companies like Kuaishou and Meitu demonstrating significant revenue growth [33][35] Group 3 - AI is reshaping the advertising value chain, enhancing return on investment (ROI) through improved user insights and targeted advertising strategies [39][40] - AI-driven enterprise services are catalyzing performance across various sectors, with companies like Salesforce and Palantir reporting substantial revenue growth attributed to AI integration [44][45] - The education sector is experiencing a transformation through AI, with personalized learning solutions driving user engagement and revenue growth for platforms like Duolingo [46][49] Group 4 - The healthcare industry is leveraging AI for precise diagnostics and accelerated drug development, with companies like Tempus achieving significant revenue increases through AI applications [51][52] - The current AI application market is influenced by overseas business models, with a focus on sectors that have demonstrated clear revenue validation, such as AI programming and AI healthcare [53][54]

海光信息 20250926 摘要 海光信息 2024 年收入达 91 亿元，复合增长率超 50%，主要受益于国 产化和服务器 CPU 爆发。公司于 2022 年登陆科创板，虽上市初期利润 较低，但业绩持续释放推动估值消化，并维持高速增长。 公司上市时 PE 超 100 倍，但因业绩增速高达 145%，PEG 小于 1，估 值合理。2019 年至 2024 年，公司复合增速达 89.09%，利润从亏损 1 亿元增长至 19.31 亿元，显示市场对其高增长预期。 2025 年，海光信息与中国曙光整合，形成芯片设计到方案全链条能力。 深算 3 号量产标志着公司向 AI 算力领域延伸，AI 收入占比约三成，未来 预计显著增加，因 AI 芯片市场容量远大于 CPU 市场。 合并后，海光信息与曙光总体利润超 30 亿元，预计未来几年随市场放 量利润将持续上升，有望突破 50 亿甚至向百亿迈进。海光信息掌握 x86 指令集，DPU 产品兼容 CUDA 架构，深算 3 号已进入销售阶段。 国内 GPU 市场容量预计超 500 亿美元，海光信息 DPU 占据优势。国内 日均 Tokens 消耗量快速增长，大模型需求旺盛，为海光 ...

证券研究报告 | 行业周报 gszqdatemark 2025 09 27 年 月 日 电子 周观点：AI 算力持续演进，把握存储大周期机遇 阿里云重磅升级全栈 AI 体系，涉及大模型、智能体、AI 基础设施等各个 方面。1）模型方面，通义大模型实现七连发，在模型智能水平、Agent 工具调用和 Coding 能力、深度推理、多模态等方面实现多项突破。其中， 阿里通义旗舰模型 Qwen3-Max 全新亮相，预训练数据量达 36T tokens， 总参数超过万亿，拥有强 Coding 编程能力和 Agent 工具调用能力，性能 超过 GPT5、Claude Opus 4 等，跻身全球前三；下一代基础模型架构 Qwen3-Next 及系列模型正式发布，模型计算效率实现突破。截至目前， 阿里通义开源 300 余个模型，在中国企业级大模型调用市场中，阿里通义 占比第一。2）Agent 方面，发布全新 Agent 开发框架 ModelStudio-ADK， 升级了低代码 Agent 开发平台 ModelStudio-ADP，随着模型能力的不断提 升以及 Agent 应用的爆发，过去一年，阿里云百炼平台的模型日均调用量 ...

Investment Rating - The semiconductor industry is rated as "Maintain Recommendation" [1] Core Viewpoints - The semiconductor sector is experiencing strong overall performance, with a long-term positive outlook driven by AI computing demand, domestic substitution logic, and policy support [3] - The semiconductor equipment segment has shown robust performance, driven by increased demand for storage due to AI data centers and ongoing domestic substitution [3] - The semiconductor materials and electronic chemicals segment is active, with domestic substitution extending into high-end fields, although top-tier products still rely on imports [3] - The integrated circuit packaging and testing sector is focusing on technological upgrades and application expansion, particularly in advanced packaging to meet high-end chip demands [3] - The analog chip design sector is slowly recovering, with demand returning in industrial and communication markets, while automotive electronics recovery remains lagging [3] - The digital chip design sector is benefiting from AI computing demand, with notable IPO activities boosting market sentiment [3] Summary by Sections Semiconductor Equipment - The semiconductor equipment sector is a key focus for capital allocation, driven by new project launches from leading storage manufacturers and advanced logic expansion [3] Semiconductor Materials & Electronic Chemicals - The semiconductor materials segment is benefiting from vertical upgrades in the semiconductor industry chain, with a focus on high-end products like advanced photoresists [3] Integrated Circuit Packaging and Testing - The packaging and testing sector is transitioning towards advanced packaging solutions to meet the demands of AI and high-performance computing [3] Analog Chip Design - The analog chip sector is in a slow recovery phase, with potential improvements in profit margins due to anti-dumping investigations [3] Digital Chip Design - The digital chip design sector is experiencing structural growth driven by AI computing needs, with significant developments in domestic GPU iterations [3] Investment Recommendations - The report suggests focusing on companies with increasing domestic substitution rates and technological breakthroughs, recommending specific companies such as Cambricon, Haiguang Information, and others [3]

Investment Rating - The report maintains an "Overweight" rating for the computer industry, indicating an expectation that the industry index will outperform the market index by more than 10% over the next six months [1]. Core Insights - The report highlights that Huawei is expected to lead the domestic AI computing sector by developing "super node + cluster" computing solutions, which may allow China to achieve a significant leap in AI computing capabilities [1][28]. - The report notes that the SW computer sector has seen a cumulative increase of 3.42% over the past two weeks, outperforming the CSI 300 index by 2.42 percentage points, ranking it 5th among 31 first-level industries [2][10]. - The report emphasizes the rapid advancements in AI computing power, with Huawei's upcoming Ascend chips expected to double computing power annually, addressing the growing demand for AI capabilities [28]. Summary by Sections 1. Industry Performance Review - The SW computer sector has increased by 29.16% year-to-date, outperforming the CSI 300 index by 12.42 percentage points [10]. - In September, the sector experienced a decline of 1.09%, underperforming the CSI 300 index by 3.24 percentage points [10]. 2. Valuation Situation - As of September 25, 2025, the SW computer sector's PE TTM (excluding negative values) stands at 58.99 times, placing it in the 92.48th percentile over the past five years and the 87.60th percentile over the past ten years [20]. 3. Industry News - OpenAI and NVIDIA announced a partnership, with NVIDIA planning to invest up to $100 billion to support OpenAI's next-generation AI infrastructure [21]. - The Chinese AI industry is projected to exceed 900 billion yuan in 2024, with a year-on-year growth of 24% [21]. - Huawei announced a roadmap for its Ascend chips, with several new products expected to launch between 2026 and 2028, significantly enhancing AI computing capabilities [21][28]. 4. Company Announcements - Several companies in the sector have made significant announcements, including cash distributions and acquisitions, indicating ongoing strategic movements within the industry [24][25][27]. 5. Weekly Perspective - The report discusses Huawei's announcements at the 2025 Global Connectivity Conference, detailing the launch of new super node products that significantly enhance computing power compared to competitors [28]. 6. Recommended Stocks - The report suggests focusing on specific stocks such as GuoDianYunTong, ShenZhouShuMa, and LangXinXinXi, which are expected to benefit from the rising demand for domestic computing power [29].

长存长鑫新一轮扩产，供应链投资解读 20250924 摘要 美国禁令下，中国半导体设备国产替代加速，尤其在非光刻和检测设备 领域，国产化率逐步提高，缓解了产能扩张瓶颈，成熟逻辑、模拟、功 率等领域扩产或放缓，但先进制程和存储领域将加速。 中美贸易博弈导致芯片原产地认定，国内中芯国际、华虹等 Fab 厂受益 于订单回流，稼动率较高，表现优于其他地区，凸显国产替代在保障供 应链安全中的重要性。 国产替代和华为产业链是当前科技板块中相对低位且具投资潜力的领域， 尤其关注 3D 封装、先进制程、先进封装等，以及 AI 算力增长带来的市 场机会，如长江存储、长鑫存储、中芯国际等。 长存和长兴均在筹备 IPO，产能扩张温和但超年初预期，长存 FAB 3 将 在 2026 年进行层数升级和国产化率提高，Nan 产品层数将达 300 层以 上，国产化率预计达 60%-80%，利好供应设备厂商。 长存预计 2026 年二季度至三四月份分三阶段进行扩产招投标，总量约 5~6 万片，对应设备采购需求约 50~60 亿美金，刻蚀设备占比最大 （约 40%），薄膜沉积占 25%。 Q&A 长江存储和长鑫存储近期扩产的背景和原因是什么？ ...