| | 沪股通 | | | | | --- | --- | --- | --- | --- | | 名称 | 涨跌幅 | 净买入额(亿) | 成交额 | 名称 | | 阿里巴巴-W | 1.3% | 1.43 | 30.01亿 | 阿里巴巴-W | | 长飞光纤光缆 | 21.2% | 2.03 | 20.20 Z | 腾讯控股 | | 腾讯控股 | 1.4% | 2.10 | 14.40 Z | 小米集团-W | | 中国海洋石油 | -0.2% | -3.85 | 14.35 Z | 美团-W | | 小米集团-W | 0.8% | 1.74 | 12.89 Z | 长飞光纤光缆 | | 中芯国际 | 2.1% | -0.31 | 11.37 Z | 中本国际 | | 工商银行 | 0.8% | 0.24 | 8.67亿 | 南方恒生科技 | | 美团-W | 1.8% | 0.57 | 8.66亿 | 中国海洋石油 | | 紫尖矿业 | 1.8% | 3.58 | 8.29亿 | 中国移动 | | 中国人寿 | 4.3% | 3.13 | 8.20 Z | 快手-W | 12月17日，南下资金净买入港股 ...

另外，他还宣布，openvela开源生态持续拓展，全球合作伙伴突破100家。赋能1500+品类，搭载设备超 1.6亿台；生态覆盖升级，首次实现从IoT芯片到车用MCU的关键拓展。（新浪科技） 【#卢伟冰公布小米软硬件生态进展#：拥有120万软件开发者，硬件月活用户数7.42亿】在今日的2025 小米人车家全生态合作伙伴大会上，小米集团合伙人、集团总裁卢伟冰发表《一路同行，澎湃未来》的 主题演讲。#小米开源最新大模型##卢伟冰回应小米自研大模型开源上线# 他公布了小米软硬件生态最新进展。软件生态方面，拥有120万全球软件开发者，国内全端月分发11 亿，1800万游戏付费用户数，1300万内容与服务订阅用户。 硬件生态规模方面，全球月活用户数7.42亿，同比增长8.2%；loT设备全球连接量（不包括手机、平板 和电脑）10.4亿，同比增长20.2%；全球硬件合作伙伴15000+。 ...

他公布了小米软硬件生态最新进展。软件生态方面，拥有120万全球软件开发者，国内全端月分发11 亿，1800万游戏付费用户数，1300万内容与服务订阅用户。 他公布了小米软硬件生态最新进展。软件生态方面，拥有120万全球软件开发者，国内全端月分发11 亿，1800万游戏付费用户数，1300万内容与服务订阅用户。 硬件生态规模方面，全球月活用户数7.42亿，同比增长8.2%；loT设备全球连接量（不包括手机、平板 和电脑）10.4亿，同比增长20.2%；全球硬件合作伙伴15000+。 另外，他还宣布，openvela开源生态持续拓展，全球合作伙伴突破100家。赋能1500+品类，搭载设备超 1.6亿台；生态覆盖升级，首次实现从IoT芯片到车用MCU的关键拓展。 新浪科技讯 12月17日上午消息，在今日的2025小米人车家全生态合作伙伴大会上，小米集团合伙人、 集团总裁卢伟冰发表《一路同行，澎湃未来》的主题演讲。 责任编辑：江钰涵 新浪科技讯 12月17日上午消息，在今日的2025小米人车家全生态合作伙伴大会上，小米集团合伙人、 集团总裁卢伟冰发表《一路同行，澎湃未来》的主题演讲。 硬件生态规模方面，全球月活用户数 ...

12月9日，"昇腾CANN开源开放沙龙·上海站"成功举办。本次活动由华为技术有限公司主办，上海昇腾 人工智能生态创新中心（以下简称"创新中心"）、中国计算机学会上海活动中心承办，上海昇思生态创 新中心、上海市人工智能学会、上海市漕河泾新兴技术开发区发展总公司等单位共同协办。活动聚焦 CANN开源开放的技术路径与生态共建，吸引了众多华东地区的开发者、企业代表及专家学者积极参 与，现场交流氛围热烈，共同探讨算力技术的深度落地与产业融合。 共话开源生态，展望智能未来 中国计算机学会（CCF）上海副主席丁炎在开场致辞中表示，上海作为中国人工智能产业高地，展现出 强劲的创新活力与丰硕成果。他呼吁共同构建开源开放的AI基础软硬件体系，并建议以开发者为中 心，依托创新中心推动区域技术生态建设，为人工智能发展注入新动能。 中国计算机学会（CCF）上海副主席丁炎 华为上海战略与业务发展部部长董浩在致辞中指出，昇腾扎根上海以来，已联合300多家生态伙伴孵化 超1000个AI解决方案，培养5000余名开发者。他表示，昇腾将持续践行"硬件开放、软件开源、使能伙 伴、发展人才"战略，坚持全面开源开放，以开发者为中心加速创新。 华为上海 ...

繁花满树，必深植其根；星火已燃，亟待乘风而上。笔者认为，想让这片"数字雨林"筋骨强健、气血通 畅，产业界各方需精准发力、协同共振，在以下三个关键方向凝聚智慧、倾注力量。 首先，以标准筑基，让创新有序生长。开源不等于无序，繁荣更需规则。建议由开放原子开源基金会牵 头，联合头部企业、科研院所加快制定跨行业兼容性标准，明确核心接口与安全规范。唯有标准统一， 才能让开发者"一次开发，多端复用"，让用户实现"无缝切换、体验一致"。 其次，以场景破局，让技术创造价值。生态整合并非消除差异，而是让差异在协同中创造价值。下一 步，可借鉴"智慧交通"模式，在电力、航天、水利等重点领域打造跨场景示范工程：推动电网调度系统 与卫星通信数据融合，实现"地天一体"的能源管理；依托开源鸿蒙的分布式能力，让矿山设备与水利监 测系统共享环境数据，构建"山海联动"的灾害预警网络。场景的深度融合，将倒逼技术迭代与生态优 化。 最后，以开放聚智，让生态生生不息。开源鸿蒙的生命力在于"全球开发者共建"。产业界需进一步完善 贡献者激励机制，让中小企业、个人开发者都能在生态中获得价值回报；同时，积极推动开源鸿蒙与国 际标准组织对接，吸引全球力量参与技术 ...

Core Concepts - The article emphasizes the importance of performance per total cost of ownership (Perf per TCO) and operational flexibility in the design and deployment of AWS Trainium3 [4][8] - AWS adopts a multi-source component supplier strategy and custom chip partnerships to optimize TCO and accelerate time to market [4][8] AWS Software Strategy - AWS is transitioning from internal optimization to an open-source ecosystem, aiming to leverage contributions from external developers to enhance its software offerings [5][10] - The strategy includes releasing and open-sourcing new native PyTorch backends and developing an open software stack to expand AWS's ecosystem [5][10] Market Competition Landscape - The competitive landscape for Trainium3 includes major players like NVIDIA, AMD, and Google, with AWS needing to accelerate development to maintain its market position [7][10] - Trainium3's market strategy focuses on delivering strong performance per TCO and supporting a wide range of machine learning workloads [7][10] Hardware Specifications and Generational Comparison - Trainium3 features significant upgrades over its predecessor, Trainium2, including a doubling of performance metrics and increased memory capacity [12][11] - The article highlights the confusion caused by inconsistent naming conventions in AWS's product lineup and calls for clearer naming similar to NVIDIA and AMD [12][11] Architectural Evolution - The architecture of Trainium3 has evolved to include switched scale-up rack types, which provide better performance and flexibility compared to previous toroidal designs [25][26] - The article details the physical layout and key features of Trainium3's rack architecture, emphasizing its design philosophy focused on maintainability and reliability [27][28] Packaging and Manufacturing Technology - Trainium3 utilizes advanced packaging technologies such as CoWoS-R, which offers cost advantages and improved mechanical flexibility compared to traditional silicon interposers [18][19] - The manufacturing challenges associated with the N3P process node are discussed, highlighting the need for careful management of leakage and yield issues [15][20] Commercialization Acceleration Strategies - AWS is implementing strategies to enhance assembly efficiency, including a cableless design and the use of retimers to optimize supply chain management [43][44] - The company aims to adapt to data center readiness and accelerate commercialization through flexible deployment options [43][44] Network Architecture and Scalability - The article outlines the network architecture of Trainium3, focusing on its horizontal and vertical scaling capabilities, which are designed to optimize performance for machine learning tasks [48][49] - AWS's strategy includes minimizing total cost of ownership while maximizing flexibility in network switch options [48][49]

Core Viewpoint - The article emphasizes the successful development of the HarmonyOS ecosystem over five years, highlighting its role as a self-developed and controllable infrastructure for various industries in China [1][20]. Group 1: Development and Impact - A short film titled "Together" showcases the inspiring stories of the HarmonyOS ecosystem, demonstrating its integration into critical sectors such as energy, education, and healthcare [2][4]. - The codebase of the open-source HarmonyOS community has surged from 7 million lines to 130 million lines, with over 9,700 contributors, reflecting a collective effort towards innovation and sustainable development [5][6]. - The HarmonyOS has become a foundational element in China's software landscape, facilitating significant advancements in various industries, including transportation and energy [8][24]. Group 2: Industry Applications - In the transportation sector, the Hebei Expressway Group has implemented the "Jihong" smart tunnel solution using HarmonyOS, enhancing operational efficiency and safety while reducing maintenance costs by automating 80% of manual tasks [10][11]. - The energy sector has seen the development of the "Mine Hong" smart mining solution, allowing remote control of underground equipment, significantly reducing risks associated with high-risk operations [12][14]. - In healthcare, HarmonyOS is transforming patient monitoring systems, enabling real-time health data tracking in hospitals, thereby improving response times and patient care [16][17]. Group 3: Ecosystem and Security - The HarmonyOS ecosystem is characterized by its open-source nature, promoting collaboration among various industry partners and developers, which is crucial for its rapid adoption across multiple sectors [24][25]. - The system's full-stack self-research ensures technological autonomy, mitigating risks associated with external dependencies, while its open code allows for community-driven security enhancements [26]. - The ecosystem's growth is symbolized by the emergence of 84 industry-specific versions and over 1,400 hardware and software products, expanding its reach and fostering a truly interconnected environment [25].

Investment Rating - The report assigns an "Accumulate" rating for the industry, indicating a potential increase of over 15% relative to the CSI 300 index [4][9]. Core Insights - The release of DeepSeek-V3.2 and its enhanced version V3.2-Speciale marks a significant breakthrough in reasoning capabilities, tool invocation, and the open-source ecosystem, promoting the prosperity of large model open-source and developer ecosystems [2][4]. - DeepSeek-V3.2 achieves top-tier performance comparable to closed-source models, particularly in reasoning capabilities, and integrates innovative thinking modes into tool invocation, providing more efficient and cost-effective solutions for AI application development [4]. - The Speciale version has excelled in international competitions, achieving second place in the ICPC and demonstrating the potential of open-source models to reach human-level intelligence in complex reasoning tasks [4]. Summary by Sections Industry Overview - The report highlights the computer industry, focusing on advancements in AI and large models, particularly the DeepSeek series [4]. Investment Recommendations - The investment suggestion emphasizes that the release of DeepSeek-V3.2 signifies a new phase in the performance and practicality of open-source large models, with a balanced focus on performance and efficiency [4]. Technological Advancements - DeepSeek-V3.2 incorporates a systematic approach to chain thinking within tool invocation processes, significantly enhancing the model's generalization and execution capabilities in complex scenarios [4]. - The model has undergone reinforcement learning across over 85,000 complex instructions in more than 1,800 environments, achieving the highest level among open-source models in tool invocation assessments [4]. Ecosystem Development - The comprehensive upgrade of DeepSeek-V3.2's open-source and API services is expected to accelerate technological penetration and drive transformative changes in industry applications [4]. - The open strategy is anticipated to attract numerous developers to build vertical applications based on DeepSeek, further solidifying its leading position in the open-source domain [4].

华泰证券表示，展望2026年，推荐三条主线：1）航空：供给进一步放缓+需求边际改善，有望从25年 的客座率提升，切换到26年的票价提升，叠加油/汇走势利好，盈利具备高弹性。2）油运：受益于 OPEC+增产、长航距国家原油出口量增加与地缘扰动，油运运价中枢有望显著抬升。3）α个股：估值 具备吸引力的行业龙头、细分赛道空间广阔的个股、受益于配置盘增加的高股息个股。 中信建投：阿里云持续加速，开源生态+模型性能构建B端壁垒 中信建投表示，阿里依托Qwen大模型底座实现业务全面重塑，并凭借开源策略与强劲性能加速构建B 端生态壁垒。同时，公司坚定上修资本开支以应对旺盛的算力需求，云收入持续高增，验证了"基础设 施投入-技术迭代-商业变现"的闭环逻辑。建议关注1）阿里生态玩家；2）Pre-AI 的收入率先落地，推 荐OA+ERP环节；Pre-AI。3）部分细分垂直场景 AI 收入更快。降本关注AI-coding和多模态。4）本地 推理逐步起量。5）端侧AI。 华泰证券：关注交运三条投资主线 航空、油运、α个股 中信证券：12月基本不存在流动性缺口 资金面对债市的风险有限 中信证券指出，经测算，12月基本不存在流动性缺口， ...

Investment Rating - The report maintains a rating of "Outperform" for the computer sector [5] Core Insights - Alibaba is leveraging the Qwen large model foundation to comprehensively reshape its business and is accelerating the construction of B-end ecological barriers through an open-source strategy and strong performance [1][2] - The company is increasing capital expenditure to meet the surging demand for computing power, with cloud revenue continuing to grow significantly, validating the "infrastructure investment - technology iteration - commercial monetization" closed-loop logic [1][4] Summary by Sections Section 1: Alibaba's AI Layout - Alibaba has transitioned from a cloud-based service model to a comprehensive AI-driven approach, utilizing its self-developed "Flying" system to support internal and external business operations [10][11] - The company has established a solid technical foundation through its research institutions, including the Damo Academy and Tongyi Laboratory, to drive AI development [10][33] Section 2: Model Performance and Market Position - The Qwen series flagship model has rapidly iterated and is now among the top tier globally, with performance approaching that of leading overseas closed-source models [3][46] - Alibaba's open-source strategy has positioned it to capture market share in the domestic B-end market, building a strong user moat and ecological barrier [3][37] Section 3: Cloud Infrastructure and Revenue Growth - To address the explosive demand for AI computing power, Alibaba is increasing its infrastructure investments, achieving rapid revenue growth in AI-related products [4][76] - The company has developed self-research chips and a global data center network to ensure soft and hard resource coordination [4][76] Section 4: Investment Recommendations - The report suggests focusing on various players within the Alibaba ecosystem, particularly those involved in Pre-AI revenue generation and specific vertical AI applications [1] - Recommended companies include Fengwei Network, Kingdee International, and others in the AI coding and multimodal sectors [1]