Investment Rating - The report maintains a "Positive" investment rating for the communication industry [10]. Core Insights - The communication sector saw a 1.99% increase in the 28th week of 2025, ranking 14th among major industries in the Yangtze River region. Year-to-date, the sector has risen by 10.00%, placing it 13th [2][4]. - OpenAI plans to launch an AI browser and open-source a new model, while xAI has released Grok 4 and is initiating a new round of financing to accelerate AI data center construction. The acquisition of Fokexima by Shijia Photon aims to enhance the MPO supply chain [2][5][6][7]. - The current valuation of the communication sector is at a historical low, indicating a high cost-performance ratio for investment [2]. Summary by Sections Market Performance - In the 28th week of 2025, the communication sector increased by 1.99%, with a year-to-date rise of 10.00%. The top three gainers among companies with a market cap over 8 billion yuan were Hengbao Co. (+21.1%), Changxin Bochuang (+19.8%), and Shijia Photon (+14.8%). The top three decliners were Runze Technology (-4.1%), Canqin Technology (-3.4%), and Guomai Technology (-3.4%) [4]. AI and Computing Developments - OpenAI is set to release a browser integrated with a generative AI assistant, which could challenge Chrome's market dominance. Additionally, OpenAI will soon unveil an open-source language model, marking its first release since 2019 without Microsoft's involvement [5]. - xAI's Grok 4, touted as the "world's strongest AI," has surpassed doctoral-level performance in standardized tests. The company is preparing for a new financing round valued between 170 billion to 200 billion USD, with participation from Saudi Aramco's PIF fund [6]. Strategic Acquisitions - Shijia Photon plans to acquire 82.4% of Fokexima, a leading supplier of MT connectors, to strengthen its MPO product line. Fokexima's production capacity has exceeded 10 million units per month, and its revenue for 2024 is projected at 270 million yuan, reflecting a 225% year-on-year increase [7]. Investment Recommendations - The report recommends focusing on key players in various segments: - Telecom operators: China Mobile, China Telecom, China Unicom - Optical modules: Zhongji Xuchuang, Xinyi Sheng, Tianfu Communication, Shijia Photon - Domestic computing: Fenghuo Communication, Huafeng Technology, Yingweike, Runze Technology, Guanghuan New Network, Aofei Data, Guangxun Technology, ZTE, Unisplendour - AI applications: Heertai, Tuobang Technology, Yiyuan Communication, Meige Intelligent, Guanghetong, Boshijie, Aojie Technology - Satellite applications: Haige Communication, Huace Navigation [8].

Core Insights - The third China International Supply Chain Promotion Expo (Chain Expo) opened, highlighting China's remarkable supply chain achievements and innovation support [2] - NVIDIA's CEO emphasized the company's journey from a small startup in 1993 to a leader in AI and GPU technology, showcasing its significant contributions to the gaming and AI industries [3][4] Group 1: NVIDIA's Contributions to AI - NVIDIA invented the first programmable graphics processor (GPU) in 1999, which opened doors for accelerated computing [3] - The introduction of CUDA in 2006 transformed GPUs into general-purpose computing engines, marking the beginning of the AI era [3] - Since the launch of AlexNet on NVIDIA GPUs, the company has significantly enhanced AI computing performance by 100,000 times, surpassing Moore's Law by 1,000 times [4] Group 2: AI's Impact on Various Industries - AI is revolutionizing multiple sectors, including scientific research, healthcare, energy development, transportation, and logistics [4] - Major Chinese platforms like Tencent, Alibaba, and ByteDance leverage AI to enhance their services, demonstrating the technology's widespread application [5] - Over 1.5 million developers in China are utilizing NVIDIA's platform to bring their innovations to life, contributing to the global AI landscape [5] Group 3: Future of AI and Supply Chain - AI is becoming a fundamental infrastructure, akin to electricity and the internet, reshaping supply chains and production logistics [6] - Hundreds of projects in China are using NVIDIA Omniverse for digital twin simulations to optimize factory and warehouse designs [6] - The next wave of AI will focus on robotic systems capable of reasoning and executing tasks, leading to AI-driven factories and products [6]

Group 1 - The core viewpoint is that AI and software will drive factories in the next decade, creating new opportunities for China's supply chain ecosystem [1][3] - NVIDIA's CEO Jensen Huang highlighted the significance of China's supply chain, calling it a miracle and emphasizing the role of AI in transforming manufacturing processes [3][4] - NVIDIA has evolved from a gaming chip company to a provider of foundational infrastructure for AI, indicating a major shift in the industry [4] Group 2 - Huang noted that AI has enhanced computational capabilities by 100 times compared to previous architectures, significantly outpacing Moore's Law [3] - The company is focused on building a global AI ecosystem, with applications ranging from healthcare to transportation, showcasing the versatility of AI technology [3][4] - Over 1.5 million developers in China are currently utilizing NVIDIA's platform for AI development, indicating a robust ecosystem of innovation [3]

2024财年日本26.7%的日本民众使用过生成式AI服务，远低于中国的81.2%与美国的68.8%。 日本使用生成式AI现状 本财年白皮书介绍了世界范围内AI技术研发的现状和动向，认为大语言模型（LLM）的发展主要由资 金实力雄厚的海外科技巨头和初创企业引领。日本本土的大语言模型研发日趋活跃，但日本在AI领域 的研发能力和应用方面仍落后于中美为代表的世界领先国家。 白皮书调查数据显示，2024财年日本26.7%的日本民众使用过生成式AI服务，远低于中国的81.2%与美 国的68.8%。其中，对比2023与2024财年，中国民众使用生成式AI的占比增幅显著，一年内上升了约35 个百分点，美国同期也上升了约22个百分点。 就年龄结构而言，在日本，20~29岁的年轻人是使用生成式AI服务最多的群体，占比约为44.7%。40~49 岁的群体紧随其后（29.6%）。"不知道如何使用"和"生活中不需要"成为日本民众不使用AI的最主要原 因。 在日本，生成式人工智能（AI）使用情况如何？ 近日，日本总务省发布2025年版《信息通信白皮书》（下称"白皮书"）。白皮书显示，日本在AI研究开 发能力和应用方面仍比较落后，2024 ...

作者｜Calvin French-Owen 译者｜核子可乐 策划｜冬梅、褚杏娟 近日，据《连线》援引多位知情人消息，OpenAI 研究员 Jason Wei 即将加盟 Meta 新成立的超级智 能实验室。 据 Jason Wei 个人网站信息，他曾参与 OpenAI 的 o3 模型及深度研究模型开发。2023 年加入 OpenAI 前，他曾在谷歌任职，期间专注于思维链研究 —— 这种研究的核心是逐步训练 AI 模型处理 复杂查询。在 OpenAI 工作期间，Wei 曾坦言自己是强化学习的 "忠实拥趸"。强化学习是通过正反馈 或负反馈来训练、优化 AI 模型的技术，如今已成为 AI 研究的热门领域，而 Meta 超级智能团队此前 聘请的多位研究员，恰好都深耕这一方向。 另有消息人士向《连线》杂志透露，OpenAI 的另一位研究员 Hyung Won Chung 也将一同加入 Meta。多位消息人士证实，两人在 OpenAI 内部的 Slack 账户现已停用。目前，OpenAI、Meta 以 及 Wei 和 Chung 本人都未回应《连线》杂志的置评请求。 这些核心研究员的流动，无形中让外界对 OpenAI 的团 ...

黄仁勋现场图。本文 来源：腾讯科技 脱掉皮衣，穿上唐装，中文开场！黄仁勋链博会演 讲实录全文 7月16日，黄仁勋身着唐装出席了第三届链博会。在此之前，他身着标志性皮衣出席了多个活动。 在开幕式演讲环节中，黄仁勋一度尝试用中文开场，他说，"我在美国长大，学到了很多汉语。"然 后他又自嘲中文不好，不想折磨在场听众，随即切换成英语演讲。 我现在要用英语讲完我的演讲的剩余部分，不会在剩下的时间里折磨你们。 这是我第一次来到这里。 这一切始于很久以前。英伟达当时还只是一个小型初创公司，我们为PC产品带来了3D显示技术， 因此催化了世界上最大的娱乐产业之一的PC游戏产业，这令人自豪。 我们的成就建立在先驱者的基础上，1999年，我们首次开发出了GPU硬件——全世界首个可编程 图形处理器，打开一扇新的大门，进入到一个加速计算的新时代。将GPU转换为通用计算引擎， 这为AI时代的发展奠定了基础。 黄仁勋表示，英伟达始于GPU，在发展的过程中打开了通用计算的大门，为人类的AI奠定了基 础，"从Kepler架构，到最新的Blackwell架构，英伟达在人工智能计算方面取得了进步。我们的系 统和算法历经了成百上千次的迭代，让英伟达 ...

Investment Rating - The report does not explicitly provide an investment rating for the industry or specific companies involved. Core Insights - The AI coding startup Windsurf, initially close to being acquired by OpenAI for $3 billion, opted to join Google DeepMind, focusing on agentic coding. Google executed a soft acquisition through non-exclusive technology licensing and talent absorption, with the deal valued at approximately $2.4 billion [1][2][8]. - Windsurf's core product, Agent IDE, is designed for multi-agent AI collaboration, highlighting the increasing importance of integrated development environments in AI programming [3][9]. - The competitive landscape has shifted, with platform risks escalating as independent AI tool providers face survival pressures. Windsurf's experience illustrates the dilemma of maintaining neutrality versus aligning with dominant platforms for resource support [4][10][11]. Summary by Sections Event - Windsurf was close to being acquired by OpenAI for $3 billion but chose to join Google DeepMind instead, focusing on agentic coding. Google did not acquire equity but engaged in a soft acquisition through technology licensing and talent absorption [1][2][8]. Commentary - The failed acquisition by OpenAI was primarily due to concerns over IP access rights granted to Microsoft, which raised fears within Windsurf's leadership about losing control over their core technology. This led to the collapse of the deal, allowing Google to seize the opportunity [2][8][10]. Product Overview - Windsurf's flagship product, Agent IDE, facilitates multi-agent AI collaboration, supporting task delegation, shared context, and persistent state management among AI agents [3][14]. Industry Implications - The situation faced by Windsurf reflects a broader trend in the AI industry where independent toolmakers must decide between maintaining platform neutrality or aligning with larger ecosystems for better resource access. This consolidation may accelerate standardization and innovation in AI development [11][12].

今年2月，孙正义宣布软银电信部门与OpenAI成立合资企业，双方各占50%股份，彰显出他支持OpenAI AI事业的承诺。该合资企业将向本地产业推广一款名为"Cristal Intelligence"的企业级AI产品，覆盖从汽 车制造商到零售商在内的多个行业。软银集团旗下各公司正采用OpenAI的工具，并计划每年在相关服 务上投入30亿美元。 孙正义与OpenAI的合作远不止于企业客户层面。软银已表示，将向OpenAI投资高达300亿美元，这项投 资将取决于OpenAI对其复杂运营结构的重组情况。此外，孙正义还与奥特曼携手参与了价值5000亿美 元的星际之门项目，在美国建设数据中心及其他AI基础设施。 "当我们考虑未来大规模扩大基础设施，远超10吉瓦时，我们将需要新的技术以及新的建设方案。"奥特 曼对孙正义说。两位高管并未提及为如此大规模项目提供能源的挑战，也未讨论如果他们对需求激增的 预期未能完全实现可能带来的潜在负面影响。 软银股价在今年6月上涨了38%，创下20年来的最佳表现，原因是股东们对孙正义的大胆计划和积极投 资表示欢迎。不过，受复杂融资计划存在风险的拖累，该股价格对应的市值依然低于公司总资产。( ...

本文第一作者密振兴，香港科技大学计算机科学与技术学院人工智能方向博士生，研究方向是多模态理解与生成，3D/4D 重建与生成，目前正在寻找工业 界全职职位或实习职位。 自 Stable Diffusion、Flux 等扩散模型 (Diffusion models) 席卷图像生成领域以来，文本到图像的生成技术取得了长足进步。但它们往往只能根据精确的 文字或图片提示作图，缺乏真正 读懂图像与文本、在多模 态上下文中推理并创作 的能力。 能否让模型像人类一样真正读懂图像与文本、完成多模态推理 与创作 ，一直是学术界和工业界关注的热门问题。 OpenAI 的 GPT-4o image generation 和 Google 的 Gemini Pro 用超大规模参数和海量数据，展示了强大的多模态推理与生成能力。但在学术与产业 环境中算力和数据并不充裕时，用 较少数据与计算资源 实现类似的功能，依然是一道难题。 在顶级学术会议 ICML2025 上，香港科技大学联合 Snap Research 提出了多模态理解与生成新方法： ThinkDiff 。该方法仅需较少的图文对和数小时训 练，就能让扩散模型具备思考能力，使其在复杂 ...

市场消息：OpenAI表示，已发现ChatGPT用户正在经历错误率升高的问题，公司团队正在努力实施缓解 措施。 ...