证券研究报告|行业简评报告 2026 年 02 月 26 日 本季营收与指引均高增，战略备货以满足未来市场需求 英伟达（NVDA.O）FY26Q4 跟踪报告 TMT 及中小盘/电子 事件： 英伟达发布 FY26Q4 季报，本季营收 681 亿美元，同比+73%/环比+20%，创 历史新高；non-GAAP 毛利率为 75.2%，同比+1.7pcts/环比+1.6pcts。综合财 报及交流会议信息，总结要点如下： 评论： 1、FY26Q4 营收高增再超预期，公司开启战略备货以应对未来市场需求。 FY26Q4 营收 681 亿美元，同比+73%/环比+20%，超出此前预期（650 亿美 元），本季营收、营业利润和自由现金流均创历史新高；non-GAAP 毛利率为 75.2%，符合此前指引（74.5%-75.5%），同比+1.7pcts/环比+1.6pcts，环比 增长主要得益于 Blackwell 架构产能持续爬坡；本季库存环比增长 8%，采购承 诺也大幅增加，公司已战略性储备库存并锁定产能，以满足未来数个季度的市 场需求。 2、数据中心营收再创新高网络业务表现亮眼，游戏受供应链影响环比下降。 1）数据中心：营收 ...

报道称，由200余人组成的团队正致力于开发一系列人工智能硬件产品。除智能音箱外，还包括智能眼 镜和智能灯等。其中，智能音箱预计最早于2027年2月面市，售价在200至300美元之间。该设备将配备 摄像头，可获取用户及其周围环境的多模态信息，以提供更智能的交互体验。智能眼镜则有望于2028年 投入量产。 为进入硬件赛道，OpenAI去年斥资65亿美元收购了由前苹果设计总监乔纳森·伊夫创立的io Products公 司。此举旨在顺应物理人工智能和增强现实技术的市场趋势，拓展其生态边界。 当前，人工智能硬件市场竞争日趋激烈。Meta与雷朋联合推出的智能眼镜凭借视频录制与直播功能取 得不俗市场表现。与此同时，苹果和谷歌也被曝正在研发各自的人工智能智能眼镜产品。（勃潺） 图为外媒报道截图 来源：环球网 【环球网科技综合报道】据海外媒体The Information周五援引知情人士消息报道，OpenAI正加速布局人 工智能硬件领域，计划于2027年前推出一款带摄像头的智能音箱，这将是该公司首款自研人工智能设 备。 ...

Group 1: Elon Musk's Perspectives on AI - The ultimate control of "digital superintelligence" cannot be achieved, similar to how chimpanzees cannot control humans, but the construction and values embedded in AI are crucial [4] - The key to AI safety is the "maximum pursuit of truth," avoiding the imposition of false beliefs on AI [5][6] - The danger of "thought viruses" being implanted in AI is not yet fully recognized by most [7] - AI will significantly transform the nature of work, with many current jobs being replaced by AI and robotics, leading to a future where work may become optional [10][11][12] - By 2026, true Artificial General Intelligence (AGI) is expected to emerge, with AI intelligence surpassing human intelligence by 2030 [17][21] - AI will revolutionize productivity, leading to unprecedented efficiency in the production of goods and services, potentially resulting in a significant drop in prices and an increase in purchasing power [24][32] Group 2: Jensen Huang's Insights on AI Infrastructure - The value of computer infrastructure accumulated over the past decade, approximately $10 trillion, is being modernized to adapt to the new computing methods of the AI era [26] - Open-source models are disrupting the AI industry, with explosive growth in downloads as various stakeholders seek to participate in the AI revolution [27] - The core of physical AI is to enable AI to understand the "rules of the real world," with a focus on creating data to train AI through simulation and synthetic data [29][30] - The AI revolution is not just about artificial intelligence but also about generative AI leading a new era, with Nvidia's AI generator producing tokens that can transform various industries [32][33] - The next generation of AI will be rooted in an understanding of physical laws, requiring the development of physics-based AI that can integrate into daily life [37][41]

Core Insights - Tesla's CEO Elon Musk has indicated a strategic shift towards humanoid robots, with a focus on leveraging China's manufacturing capabilities for the "Optimus" robot production [1][5] - The production line for Tesla's Model S and Model X will be repurposed for the manufacturing of humanoid robots, with a long-term goal of producing 1 million units annually [2] - The humanoid robot project is still in its early stages, with Musk expressing caution about its current performance and projecting a potential revenue scale of up to $10 trillion by 2027 [4] Manufacturing and Supply Chain - Tesla has been engaging with hundreds of Chinese component suppliers for over three years, indicating a strong reliance on China's rapidly developing robotics supply chain [5] - Approximately 50% to 70% of the manufacturing capabilities and core component production technologies in the humanoid robot sector are held by Chinese companies, with at least 55% of key components in the global supply chain sourced from China [5] - The U.S. leads in AI technology ("brain") while China dominates in manufacturing capabilities ("body"), creating a clear division in the global humanoid robot market [6][7] Technological Development - The U.S. is advancing the intelligent upgrade of robots through AI platforms, focusing on "physical AI" to enable autonomous actions in the real world [6] - Chinese companies have established a complete supply chain for producing core mechanical components necessary for humanoid robots, although they still have room for technological improvement [7] - Current assessments suggest that both the U.S. and China need to continue investing in R&D to enhance the technical capabilities required for humanoid robots, making it premature to predict the future competitive landscape [7]

Group 1 - The core point of the article is that German robotics startup RobCo has completed a $100 million Series C funding round to advance its physical AI roadmap, expand enterprise deployment, and deepen its business in the U.S. market [1] - The funding round was led by Lightspeed Venture Partners and Lingotto Innovation, with participation from Sequoia Capital, Greenfield Partners, Kindred Capital, Leitmotif, and The Friedkin Group [1]

Core Insights - The CES 2026 showcased physical artificial intelligence as a major trend, featuring applications across various sectors such as robotics, transportation, manufacturing, home appliances, and healthcare [1] - South Korean companies like Hyundai Motor Group, LG Electronics, and Doosan presented products utilizing physical AI, while companies like NVIDIA, Lenovo, and Qualcomm focused on platforms and chips [1] - China announced a strategy aimed at popularizing and promoting physical AI, with a particular emphasis on developing humanoid robots [1] Industry Strategies - The director of the AI Policy Research Laboratory at the Korea Software Policy Research Institute emphasized that physical AI is not just a technological innovation but a key strategy for maintaining national industrial competitiveness [1] - Proposed strategies include establishing data and computing infrastructure, cultivating interdisciplinary talent, creating a public-private collaborative R&D ecosystem, and developing a verification and safety certification system based on digital twins [1] - The need for a data center to collect and share vast amounts of cross-industry data was highlighted, ensuring high-quality standard datasets and expanding computing infrastructure to meet on-site demands [1]

Core Viewpoint - The article discusses the rising global recognition of China's renewable energy sector as a vital contributor to energy transition, countering criticisms from the U.S. regarding unfair competition and overcapacity [1][2]. Group 1: China's Renewable Energy Position - Zhang Lei, Chairman of Envision Group, asserts that Chinese renewable energy represents a "civilizational output," akin to the steam engine during the Industrial Revolution, establishing a new energy infrastructure globally [1]. - Al Gore supports this view, highlighting that by 2025, the total value of green technology exported by China will exceed that of all fossil fuels exported by the U.S. by 50% [1]. - Countries like Nigeria, Kenya, and Zambia are experiencing a surge in solar installations, indicating a global shift towards renewable energy solutions [1]. Group 2: Canada-China Energy Cooperation - Canadian Prime Minister's recent visit to China marks a significant thaw in relations, focusing on cooperation in key areas such as energy and agriculture, which is crucial for both nations [2]. - Meetings between Canadian officials and major Chinese energy companies, including China National Petroleum Corporation, Envision Group, and CATL, are pivotal for reshaping the global energy landscape [2][5]. - The collaboration is seen as a strategic response to Canada's energy transition challenges, showcasing China's renewable energy capabilities [2][11]. Group 3: Challenges in Canada's Energy Sector - Canada faces a dilemma with over 95% of its oil and most natural gas reliant on the U.S. market, posing strategic security risks amid politicized U.S. energy policies [9]. - The aging electrical grid in Canada struggles to integrate renewable energy, falling short of the 2030 and 2035 renewable energy targets [10]. - A weak green industrial base and incomplete supply chains hinder Canada's ability to leverage its resource advantages, necessitating external technological support [11]. Group 4: Technological and Systemic Advantages of China - China's renewable energy industry has developed a comprehensive "scale-technology-system" model, providing a robust foundation for international cooperation [12]. - Envision Group's AI-powered energy systems exemplify this systemic advantage, addressing the complexities of renewable energy management and enhancing operational efficiency [13][15]. - The AI energy system's capabilities align with Canada's long-term energy system upgrade plans, indicating a deepening of technological collaboration [14][19]. Group 5: Future Implications of Cooperation - The partnership between Canada and China transcends mere product trade, evolving into a collaborative model for technology standards and industrial practices [14][20]. - This cooperation is expected to reshape global energy collaboration, allowing Chinese renewable technology standards to influence international markets and contribute to global carbon neutrality efforts [20].

Core Viewpoint - Envision Energy is set to achieve a milestone of over 100 million kilowatts of installed wind power capacity globally by 2025, becoming the fastest wind turbine manufacturer to reach this mark, showcasing the potential of China's wind power industry and its commitment to long-termism and green technology innovation [1][3]. Group 1: Company Achievements - Envision Energy has maintained its position as the world's leading wind turbine order recipient for four consecutive years, with over 20 operational headquarters and R&D centers globally, and more than 60 manufacturing bases [5]. - The company has developed a comprehensive green energy ecosystem, integrating the entire value chain and industry chain of green energy, and has introduced groundbreaking products such as the Galileo AI wind turbine and the world's largest green hydrogen project [3][8]. - Envision's projects have become significant diplomatic symbols for China, including the first 100 MW wind project in Kazakhstan and the largest wind power project in Africa, demonstrating the company's global impact [5]. Group 2: Technological Innovations - Envision Energy has pioneered the concept of "intelligent wind turbines" and has made substantial advancements in physical artificial intelligence, which enhances renewable energy generation, storage, and grid management [3][8]. - The "Yuanjing Tianji" meteorological model and "Yuanjing Tianshu" energy model are central to Envision's AI systems, enabling real-time analysis and coordination of energy data to optimize renewable energy systems [8]. - The company plans to launch the world's largest AI power system by 2025, integrating all cutting-edge product technologies to drive global prosperity in the renewable energy sector [8].

Core Insights - The strategic partnership between Envision Group and Masdar aims to advance the large-scale application of AI energy systems and facilitate global energy transition [1][2] - Envision Group launched the world's first energy large model, "Envision Tian Shu," designed to create the largest physical AI system for energy [3][5] Group 1: Strategic Partnership - Envision Group signed a strategic cooperation agreement with Masdar during the 2026 Abu Dhabi Sustainability Week to promote AI energy systems [1] - The agreement was witnessed by key figures including Envision Group's Chairman Zhang Lei and Masdar's Chairman H.E. Dr. Sultan [1] Group 2: AI Energy System - "Envision Tian Shu" serves as the core of Envision's physical AI system, capable of analyzing vast amounts of real energy data to coordinate renewable energy generation, storage, grid dispatch, and energy demand in real-time [3][5] - The system aims to drive energy marginal costs close to zero and support the simultaneous scaling of energy systems and AI to meet unprecedented energy demands from AI infrastructure [3][5] Group 3: Weather Model Integration - "Envision Tian Shu" operates in conjunction with Envision's self-developed large-scale weather model, "Envision Tian Ji," which helps predict and manage energy fluctuations caused by weather conditions [5] - This integration ensures that the energy system can operate at scale, high efficiency, and reliability [5]

Core Insights - Physical AI is a key area of global technological competition, with Chinese companies emerging as leaders in the field of humanoid robots, automotive applications, and other physical AI patents [1][3] - According to a recent analysis, China ranks first globally in terms of comprehensive strength in patent applications, followed closely by the United States [1][3] - Major Chinese tech firms such as Baidu, Huawei, and Tencent lead in patent scores, while China Ping An Insurance ranks sixth [1][4] Patent Rankings - The analysis ranks Baidu, Huawei, and Tencent as the top three companies in the field of Physical AI, with scores of 4126, 3645, and 3043 respectively [4] - Samsung Electronics from South Korea ranks fourth with a score of 2734, followed by NVIDIA (2154) and China Ping An Insurance (1881) [4] - Other notable companies include Intel (1543), LG Electronics (1393), Alphabet (1325), and the Chinese Academy of Sciences (835) [4] Industry Context - The analysis indicates that while Chinese companies have a strong patent quantity, they still lag behind U.S. competitors like Intel, NVIDIA, and Alphabet in terms of patent quality [1][3] - The shift towards AI technologies is emphasized in China's 14th Five-Year Plan, which highlights the importance of high-quality development and technological advancement [4] - The CES 2024 showcased various Physical AI applications, indicating a competitive landscape among tech companies from China, the U.S., and South Korea [5] Future Developments - The Chinese government is actively supporting Physical AI as a national strategy, with plans to enhance AI integration across various industrial sectors by 2025 [5][6] - The application of AI in industrial enterprises is projected to rise significantly, with a forecasted increase from 9.6% in 2024 to 47.5% in 2025 [7] - China has established over 7000 advanced smart factories, demonstrating significant progress in the integration of AI and manufacturing [7]