WAIC 2025 世界人工智能大会 论坛：2025年7月26日-28日 上海世博中心 展览：2025年7月26日-29日 上海世博展览馆 当人工智能深度渗透文娱产业，一场关于技术赋能与人文坚守的行业变革正在上演。AIGC提升内容生产效率，却致创意同质化与版权困境；虚拟人实现 情感交互突破，却因伦理模糊引发数字身份焦虑；文旅视听借AI升级体验，却因文化差异与数据壁垒遇出海瓶颈；音乐疗愈等创新展现融合潜力，却受 限于临床标准缺失与产业转化断层。文娱行业亟需重构创意价值体系，破解标准化与个性化、效率与伦理、本土与全球的矛盾。 在此背景下，2025世界人工智能大会暨人工智能全球治理高级别会议聚焦AI艺术与文娱多元化赋能发展，将从听、看两方面呈现。 「智响未来」AI艺术与文娱论坛矩阵 WAIC 2025艺术与文娱板块将以多场重要论坛的构造，探索AI在艺术、文旅、游戏、传媒等领域的行业解决方案及前景展望。 重点议题 AIGC内容版权确权的技术方案与法律框架 虚拟人情感交互的伦理边界与商业化模式 邀请来自艺术、学术与产业的国内外重要嘉宾，围绕艺术实践的新范式、未来教育的重构、创造力的哲学基础以及政策支持与行业演进等多个维 ...

通信 优于大市（维持） 证券分析师 李宏涛 资格编号：S0120524070003 邮箱：liht＠tebon.com.cn 研究助理 市场表现 -20% -10% 0% 10% 20% 29% 39% 49% 2024-07 2024-11 2025-03 通信 沪深300 资料来源：德邦研究所，聚源数据 相关研究 1.《海内外算力共振，AIDC 景气将 持续》，2025.7.7 2.《算力软硬件投资和用户仍快速增 长 ， 业 绩 和 市 值 预 期 向 好 》， 2025.6.30 3.《中国新型基础设施：昇腾芯片+低 轨卫星+数字货币》，2025.6.23 4.《星链卫星坠落，低轨卫星新型基 础设施价值凸显》，2025.6.16 5.《中国电信（601728.SH）：高股息 带来稳健收益，数据入表+前沿布局 打开成长空间》，2025.6.12 [Table_Main] 证券研究报告 | 行业周报 通信 2025 年 07 月 15 日 芯片与操作系统筹备国产替代，生态 和出货值得关注 [Table_Summary] 投资要点： 1. 投资策略 3.本周回顾及重点关注组合 2 / 17 请务必阅读正文之后 ...

Group 1: Market Overview - On July 14, the main funds in the Shanghai and Shenzhen markets experienced a net outflow of 26.576 billion yuan, with the ChiNext board seeing a net outflow of 12.112 billion yuan and the CSI 300 index stocks a net outflow of 6.366 billion yuan [1] - Among the 21 industries tracked, mechanical equipment, public utilities, and home appliances saw the largest gains, each rising over 1%, while real estate, media, and non-bank financial sectors fell more than 1% [1] - Four industries received net inflows from main funds, with mechanical equipment leading at 394 million yuan, followed by home appliances at 117 million yuan, and coal and petrochemical sectors each exceeding 37 million yuan [1] Group 2: Individual Stock Performance - Among individual stocks, 37 saw net inflows exceeding 100 million yuan, with six stocks surpassing 200 million yuan in inflows [2] - Zhongji Xuchuang, a leader in optical modules, topped the inflow list with 497 million yuan, benefiting from increased demand for ASIC servers driven by growth in ARR for AWS's Anthropic and Google's Gemini [2] - Notable stocks such as Xiangyang Bearing, Brother Technology, and Kelu Electronics also saw significant inflows, with Kelu Electronics reaching a net inflow of 314 million yuan, the highest since June 23, 2016 [2] Group 3: Notable Outflows - Over 90 stocks experienced net outflows exceeding 100 million yuan, with 21 stocks seeing outflows over 300 million yuan [3] - Leading the outflows were BYD and Northern Rare Earth, each with net outflows exceeding 1.1 billion yuan, with Northern Rare Earth's outflow reaching 1.105 billion yuan, marking a new high since November 4, 2024 [3][4] Group 4: Tail-End Market Activity - At the market close, there was a net outflow of 1.483 billion yuan, with the ChiNext board seeing an outflow of 806 million yuan [6] - Notably, the digital currency concept stock Chutianlong saw a significant reversal, with a net inflow of 204 million yuan, despite a projected loss of 35 to 40 million yuan for the upcoming half-year [7] - Other stocks with notable tail-end inflows included Tianyang Technology and Hanyu Pharmaceutical, each exceeding 40 million yuan [8]

证券研究报告 2025年06月25日 AI算力系列之交换机：算力网络稳定"核心"， 助力AI高效发展 更多一手海外资讯和海外投行报告加V：shuinu9870 更多一手海外资讯和海外投行报告加V：shuinu9870 更多一手海外资讯和海外投行报告加V：shuinu9870 更多一手海外资讯和海外投行报告加V：shuinu9870 更多一手海外资讯和海外投行报告加V：shuinu9870 更多一手海外资讯和海外投行报告加V：shuinu9870 作者： 1 行业评级： 上次评级： 强于大市 强于大市 维持 （ 评级） 行业报告| 行业深度研究 更多资料加入知识星球：水木调研纪要 关注公众号：水木纪要 摘要 分析师 王奕红 SAC执业证书编号：S1110517090004 分析师 张建宇 SAC执业证书编号：S1110525040002 分析师 唐海清 SAC执业证书编号：S1110517030002 分析师 康志毅 SAC执业证书编号：S1110522120002 请务必阅读正文之后的信息披露和免责申明 交换机是实现网络互联的关键设备，数据中心建设推动交换机需求增长。以太网交换机是用于网络信息交换的网络设备， ...

但是现在，珠峰大本营的网络信号已经十分通畅。在大本营，拿出手机进行视频连线的游客随处可见， 珠峰美景通过网络与更多人共享。 中国移动西藏公司日喀则分公司网络部网络运维班长拉平是这里网络发展的见证者。 2020年，珠峰高程测量登山队在5月重测珠峰高程，中国移动全程负责通信保障任务，这是5G信号第一 次"登上"珠峰。作为保障团队工作人员，拉平的任务是在珠峰大本营到峰顶沿途的5300米、5800米和 6500米3处营地新建8个基站，确保登山队的信号稳定。 拉平至今都对那一段路记忆犹新——走到最后，海拔超过5800米，8吨重的设备是46头牦牛驮上去的， 而重达5.8吨的光缆，则全部要靠人工扛上去。"运输一盘3公里长的光缆需要40个人同时肩扛，我们每 走几步就得休息一下，既是因为累，更是因为缺氧。"拉平回忆。 "我们珠峰大本营足足住了200多名游客，大家晚上直播、刷视频，网络信号都流畅得很。"刷着直播， 阿旺久美告诉记者。 这是西藏日喀则定日县珠峰景区，阿旺久美在珠峰大本营经营着一家超市。"前几年，这里别说直播看 视频，连发个信息都费劲，有些帐篷旅馆把'有无线网络'作为卖点。很多游客都是回到镇上，才把在这 里拍摄的照片 ...

Core Viewpoint - The acquisition of Haowei Technology by Degute represents a significant "elephant swallowing snake" transaction, with a substantial disparity in revenue and asset size between the two companies [1][2]. Group 1: Acquisition Details - Degute plans to acquire 100% of Haowei Technology through a combination of share issuance and cash payment, making Haowei a wholly-owned subsidiary [1]. - The specific transaction pricing has not been disclosed, but the scale of the acquisition is notable given the financial metrics of both companies [1]. - The transaction involves 14 shareholders of Haowei Technology, with the top three shareholders holding 27.83%, 27.62%, and 13.85% of the shares respectively [2]. Group 2: Financial Comparison - In 2024, Degute's revenue is projected to exceed 500 million yuan, while Haowei Technology's revenue for 2023 and 2024 is reported at 3.861 billion yuan and 3.654 billion yuan respectively, indicating a revenue scale several times larger than Degute's [2][3]. - As of March 31, 2024, Degute's total assets are 1.156 billion yuan with equity of 765 million yuan, whereas Haowei Technology's total assets are 5.617 billion yuan with equity of 2.954 billion yuan [1][2]. Group 3: Business Transition - Following the acquisition, Degute aims to expand its business from energy-saving equipment manufacturing to telecommunications software development, cloud and AI software services, and industry digital solutions, thereby creating a second growth curve [4]. - The company has been transitioning from a "product supplier" to a "system integration service provider," focusing on a full value chain system for clean combustion and efficient heat exchange [5]. Group 4: Historical Context of Haowei Technology - Haowei Technology, previously known as Zhongxing Ruanchuang, was listed on the New Third Board before delisting in 2017 and has undergone multiple rounds of listing guidance since then [1][5]. - The company has received high-level certifications from major cloud service providers, indicating its strong capabilities in cloud management services [2].

证券研究报告 | 2025年07月13日 通信行业周报 2025 年第 28 周 优于大市 xAI 推出 Grok4 模型，算力基础设施企业 Q2 业绩预告较好 行业要闻追踪：全球算力继续呈现高景气度。（1）xAI 正式推出 Grok4，支 持更复杂的交互方式。7 月 10 日，马斯克旗下的人工智能公司 xAI 正式发布 AI 聊天机器人最新版本 Grok 4。该模型在 Humanity's Last Exam 中，通 过文本输入达到了约 25%的正确率。Grok 4 模型订阅费为 30 美元/月，Heavy 版本的费用为 300 美元/月，价格超过了 OpenAI 最贵的 200 美元 Pro 会员月 费。Grok4 提供代码领域的专业支持，同时主打多模态功能，支持更复杂的 交互方式。（2）7 月 10 日，博通管理层会议指引 AI 推理需求被低估，市场 规模预期存上修空间。博通管理层指出过去两个月 AI 推理订单大幅激增且 超当前产能，这一需求未被纳入此前对三个现有 AI 客户 600-900 亿美元的 市场规模预测。此外，非 AI 业务复苏与结构优化，形成增长合力，相关算 力产业链持续收益。（3）英伟达 ...

行 业 研 通信 2025 年 07 月 13 日 投资评级：看好（维持） 行业走势图 数据来源：聚源 -29% -14% 0% 14% 29% 43% 58% 2024-07 2024-11 2025-03 通信 沪深300 相关研究报告 《AI 驱动光铜共进，AEC 等受益于 高速短距连接需求—行业深度报告》 -2025.7.8 《万国和润泽 REIT 均获超百倍认 购，世纪互联上调业绩指引，AIDC 产业链发展或提速 —行业周报》- 2025.7.6 《 世 纪 互 联 上 调 业 绩 指 引 ， 看 好 AIDC 产业链投资机会—行业点评报 告》-2025.6.30 算力公司业绩亮眼，Grok-4 发布，AWS 推出 GB200 UltraServers，看好全球 AIDC 产业链 ——行业周报 | 蒋颖（分析师） | 雷星宇（联系人） | 陈光毅（联系人） | | --- | --- | --- | | jiangying@kysec.cn | leixingyu@kysec.cn | chenguangyi@kysec.cn | | 证书编号：S0790523120003 | 证书编号：S0790 ...

Core Viewpoint - The development of artificial intelligence and robotics in Guangdong requires a coordinated effort in talent, computing resources, and ecosystem support, with a focus on overcoming existing shortcomings in open-source data and reliance on imported components [1][2]. Group 1: Artificial Intelligence Development - Guangdong should enhance support for companies that actively contribute to open-source AI models, addressing the current shortfall in open-source initiatives [2]. - The establishment of a collaborative innovation center for AI and robotics is essential to lower the barriers for technology transfer and foster a more integrated development model [1][2]. Group 2: Robotics Industry - There is a pressing need for a national-level robotics training facility in Guangdong to facilitate the development and testing of high-end robots in realistic environments [1][2]. - Over 80% of core components for robotics, such as harmonic reducers and high-precision sensors, are still imported, highlighting the need for local innovation and production capabilities [1]. Group 3: Computing Infrastructure - The digital infrastructure, particularly in computing power, is critical for the growth of AI-related industries, with a goal for Guangdong to achieve 65% self-sufficient high-performance computing power by around 2027 [2]. - The establishment of a dedicated "computing power development fund" is recommended to support projects utilizing various domestic chip and server technologies [2]. Group 4: Data Utilization - To address the challenge of insufficient data for AI training, Guangdong should create a large-scale industry information database in collaboration with cities like Guangzhou, Shenzhen, and Dongguan, drawing on successful experiences from Shanghai [2][3]. - The proposal for a "Guangdong-Hong Kong-Macao Industrial Big Data Sharing Platform" aims to unlock the potential of data as a new resource for industry [3].

Core Insights - The AI training and inference integrated machine market in China is projected to reach 16.8 billion yuan in 2024, with expectations to exceed 20 billion yuan by 2025 due to the surge in demand for AI large models, accelerated domestic substitution, and policy support [1][3] - The industry is characterized by a focus on local deployment, particularly in sectors such as government, finance, and healthcare, which is driving the growth of the market [3][16] Industry Definition and Classification - The AI training and inference integrated machine is a specialized computing device or system that integrates AI model training and inference deployment functions, aiming for efficient collaboration from data training to model implementation [1] - It can be classified based on deployment location into cloud-based, edge-based, and hybrid machines, and by vertical specialization into general-purpose and industry-specific types [1] Current Industry Development - Following the release of ChatGPT and the GPT series by OpenAI, there has been a global surge in AI large model development, with a focus on commercializing these models [3] - The market for AI training and inference integrated machines is expected to grow significantly, driven by increased penetration in various sectors [3][16] Industry Value Chain - The upstream of the industry involves components such as CPUs, GPUs, storage devices, and foundational software like operating systems and databases [5] - The midstream focuses on the research, production, and system integration of training and inference integrated machines, while the downstream applications span government, finance, healthcare, manufacturing, energy, and education [5] Competitive Landscape - The industry is currently in its early stages, characterized by a "domestic-led, concentrated head, and accelerated vertical segmentation" competitive landscape [10] - Major players include Alibaba, Inspur, Sugon, iFlytek, Huawei, ZTE, and others, with a focus on enhancing hardware products related to computing power [10][11] Key Companies - **Inspur**: A leading cloud computing and big data service provider in China, with a projected revenue of 114.77 billion yuan in 2024, marking a 74.24% increase from 2023 [12] - **Sugon**: A core information infrastructure leader in China, with a revenue of 13.15 billion yuan in 2024, showing a 4.10% increase in net profit [14] - **iFlytek**: Known for its intelligent voice and AI solutions, it has launched new integrated machines supporting rapid deployment and data security [11] Industry Development Trends - The AI training and inference integrated machine is a key infrastructure for AI industrialization, with increasing demand for localized deployment in sectors like finance, healthcare, and government [16] - The industry is undergoing transformations such as technological integration, localized substitution, and green computing, with future developments expected in high-performance domestic solutions and multi-modal integration [16]