Investment Rating - The report maintains a "Recommended" investment rating for the humanoid robot industry, indicating that the industry index is expected to outperform the benchmark index by more than 10% [6]. Core Insights - The humanoid robot concept index increased by 2.06% from February 22 to February 27, 2026, outperforming the CSI 300 index by 0.98 percentage points. Year-to-date, the humanoid robot index has risen by 7.52%, surpassing the CSI 300 index by 5.77 percentage points [11][15]. - Key developments include the release of China's first national humanoid robot and embodied intelligence standard system, which marks a new phase of standardized development in the industry [18]. - Significant financing activities were reported, including several companies securing substantial investments, indicating strong market interest and growth potential in the humanoid robotics sector [20][21][22]. Weekly Market Review - The humanoid robot index saw a weekly increase of 2.06%, with notable individual stock performances such as Andar Intelligent rising by 37.86% and Tianqi Co., Ltd. declining by 11.84% [11][15]. - The report highlights the performance of the humanoid robot index against the broader market, showcasing its resilience and growth trajectory [11][12]. Weekly Hotspot Review Policy Developments - The first national standard for humanoid robots and embodied intelligence was officially released, covering the entire industry chain and lifecycle, which is expected to facilitate standardized development [18]. - Shanghai is focusing on creating a comprehensive fund matrix to support emerging industries, including humanoid robotics, and aims to establish benchmark application scenarios for new technologies [19]. Product and Technology Iteration - New product launches include the Unitree As2 quadruped robot and the Honor Robot Phone, showcasing advancements in humanoid robotics technology [20]. - BMW has initiated pilot projects integrating humanoid robots into its production lines, demonstrating practical applications and value creation in real-world manufacturing environments [4]. Investment and Financing - Significant financing rounds were reported, including a multi-billion yuan investment in Self-Variable Robotics, which has become the highest-valued company in the embodied intelligence sector for 2026 [21]. - NIO's chip subsidiary secured over 2.2 billion yuan in financing, enhancing its capabilities in high-end chip development for autonomous driving and humanoid robotics [21]. - Qianxun Intelligent completed nearly 2 billion yuan in financing, with a valuation exceeding 10 billion yuan, indicating strong investor confidence in the sector [22]. Key Company Announcements - Haitai Technology has received small batch orders from humanoid robot clients, indicating initial market traction [23]. - Green's Harmonics reported a 122.40% year-on-year increase in net profit for 2025, attributed to the recovery of the industrial robot market and significant growth in its embodied intelligence robotics business [24].

Core Viewpoint - The article outlines the key economic goals and tasks for 2026, emphasizing sustainable growth, employment generation, and advancements in various sectors, including technology and agriculture [2][7][10]. Economic Growth and Employment - The GDP growth target for 2026 is set between 4.5% and 5% [7]. - More than 12 million new urban jobs are expected to be created [8]. - The average annual growth rate of disposable income for residents is projected at 5.4% [5]. Agricultural and Industrial Production - The grain production target for 2026 is approximately 1.4 trillion jin [8]. - The manufacturing value-added scale is expected to maintain its global leadership [5]. - The production capacity for grain is aimed to reach 1.45 trillion jin [7]. Research and Development - The total R&D expenditure is expected to grow by over 7% annually [6]. - The average years of education for the working-age population is targeted to increase to 11.7 years [7]. Environmental Goals - A cumulative reduction of 17% in carbon dioxide emissions per unit of GDP is aimed for [6]. - The target for the reduction of carbon emissions in 2026 is approximately 3.8% [8]. Fiscal Policy and Budget - The budget deficit is projected to increase by 230 billion yuan, reaching a total deficit scale of 1.3 trillion yuan [8]. - The general public budget expenditure is expected to reach 30 trillion yuan for the first time [8]. - Special long-term bonds worth 1.3 trillion yuan are planned to be issued [8]. Sectoral Development Initiatives - The government plans to support emerging industries such as integrated circuits, aerospace, and biomedicine [10]. - Initiatives to promote future industries like quantum technology and 6G are also highlighted [10]. - Measures to enhance the agricultural sector through a national agricultural census and urbanization strategies are outlined [13].

编者按 人类对机器人的想象并不陌生。从小时候看的铁臂阿童木、变形金刚擎天柱，到《终结 者》、《超能陆战队》、《西部世界》，如今2026年中央广播电视总台春晚的机器人表演，更让 观众意识到人形机器人正从想象一步步走入现实。 优必选在十年前就登陆央视春晚，是全球少数具备人形机器人全栈式技术能力的公司，并已 投入量产。2023年12月在香港交易所主板挂牌上市，成为"人形机器人第一股"。这样的企业，为 我们解读这一产业提供了极佳的"硬核样本"。今天我们和它的创始人、董事会主席兼CEO周剑一 起探讨背后的几大关键问题。 机器人斗舞 是未来发展方向吗？ 问题一 十年前优必选机器人就已经登陆春晚，500台小机器人充当伴舞。但当2026年谈到机器人斗舞这个问 题，周剑的第一反应是表示怀疑，跳舞机器人不是他的理想。要做会思考的机器人，应该投入到全栈式的 技术布局，包括大脑、软硬件耦合，尤其关键的是：一定要去了解客户的场景需要，去解决人类的真实问 题，提供具身智能的人形解决方案。而不是去把精力、经费用于吸睛、声量。 机器人蓝领进工厂有多远？ 问题二 美国科技趋势未来学家凯文·凯利在谈到人形机器人时，指出面对的制约包括：电池能耗、 ...

Core Viewpoint - The article discusses the insights of Zhou Yunjie, Chairman and CEO of Haier Group, regarding the integration of user feedback into business management and the future of robotics in households, emphasizing safety and the potential of embodied intelligence in the context of aging populations [1][3]. Group 1: User Engagement and Feedback - Zhou Yunjie emphasizes the importance of learning from users and maintaining a close relationship with them, using social media feedback to improve management systems [1]. - He dedicates about one and a half hours each night to review user comments and suggestions, highlighting the dual nature of online traffic in amplifying both strengths and weaknesses [1]. Group 2: Robotics in Households - The focus of household robots should not solely be on humanoid designs but prioritize safety first, with applications tailored to specific household scenarios such as assisting with getting up and walking [3]. - The market for elderly care robots is seen as vast, with a gradual adoption process expected, positioning robots as a new category of home appliances with unlimited market potential [3]. Group 3: Recommendations for Embodied Intelligence - Zhou proposes four key recommendations to advance embodied intelligence: 1. Breakthrough in core technologies by establishing national R&D projects focused on multi-modal perception, adaptive learning, and biomimetic operations [3]. 2. Build foundational infrastructure by creating national-level open innovation platforms and datasets for embodied intelligence [4]. 3. Develop a standardized system for embodied intelligent robots to facilitate large-scale implementation in manufacturing [4]. 4. Support leading enterprises in creating benchmark applications for embodied intelligence in complex tasks, promoting replicable and scalable case studies [5].

Core Insights - The article discusses the emergence of a new paradigm in robotics, focusing on "human-first perspective data" as a key to achieving advanced robotic intelligence, outperforming major players like NVIDIA by over 20% in various benchmarks [1][4][6]. Group 1: Human-First Perspective Data - The concept of "human-first perspective data" is gaining traction in Silicon Valley, with companies like Tesla and Generalist AI investing heavily in this data type to enhance robotic capabilities [3][7]. - NVIDIA's recent EgoScale framework demonstrates that increasing human demonstration data can significantly improve robotic dexterity, emphasizing the importance of human data over machine-generated data [4][6]. - Deep Intelligence, founded in 2025, is recognized as a pioneer in leveraging human-first perspective data to decode physical common sense, which is crucial for the advancement of embodied intelligence [11][8]. Group 2: Understanding Physical Common Sense - The article highlights the critical role of physical common sense in achieving true robotic intelligence, with Generalist AI labeling it as the "dark matter" of robotics [8][14]. - Current domestic discussions in embodied intelligence often overlook the significance of physical common sense, focusing instead on trajectory fitting from real or simulated data [17][18]. - Deep Intelligence's approach prioritizes "understanding first, action next," aiming to equip robots with a deep understanding of physical world operations before executing tasks [20][21]. Group 3: Technological Innovations - Deep Intelligence has developed a comprehensive technology stack that includes data, architecture, and algorithms to enhance the efficiency of learning from human-first perspective data [24][22]. - The company has created a translation pipeline, Egocentric2Embodiment, to convert human perspective videos into structured learning signals for robots, ensuring they understand the underlying physical interactions [25][34]. - The PhysBrain model, trained on human-first perspective data, has achieved a 67.4% success rate in tasks, outperforming competitors that relied on extensive machine trajectory data [27][29]. Group 4: Advanced Model Architectures - The TwinBrainVLA architecture allows for the simultaneous training of a "left brain" for general understanding and a "right brain" for specific robotic actions, preventing knowledge loss during optimization [31][32]. - The integration of various innovations has led to the development of PhysBrain1.0, which achieved a remarkable 79.8% success rate in testing, surpassing industry benchmarks [37][38]. - The model's ability to generalize across different tasks and platforms indicates a significant advancement in robotic intelligence, showcasing the potential for real-world applications [39][40]. Group 5: Future Directions - Deep Intelligence plans to scale human-first perspective data collection to a million hours by mid-2026, aiming to fully reveal the scaling laws of physical common sense [43][46]. - The company’s approach is expected to create a competitive edge that is difficult for others to replicate, as it emphasizes data efficiency and systematic modeling of physical common sense [42][46].

允中 发自 凹非寺 量子位 | 公众号 QbitAI 融资规模背后，是极佳视界在具身智能领域的全栈深耕： 具身基模、世界模型、原生本体、泛化场景 ，极佳视界正通过这"四位一体"的打法，叩开物理世界AGI的大门，加速通用机器人走进千家万 户。 产业巨头和顶尖资本集体押注，要做"物理世界的OpenAI" 资本正在加速押注具身智能的下一阶段。 近日，具身基模和通用机器人企业 极佳视界 ，最新完成近 10亿元Pre-B轮融资 。 本轮投资方包括： 其中， 中金资本、华强资本、财鑫资本、张科垚坤等作为老 股东持续超额重磅加持，庚辛资本中国担任财务顾问。 以大语言模型为代表的数字AGI已深度重塑了数字世界。 然而，全球GDP中仍有约50%根植于物理实体。极佳视界认为， 物理世界的基础模型突破 将是下一波变革的核心。 随着端到端架构、VLA （视觉-语言-动作模型） 及世界模型等技术的飞速演进，物理AGI正处于从技术积累转向产业爆发的临界点。 中芯聚源、上海半导体产投基金、临芯资本、星源资本、万林国际等 顶尖芯片和汽车产业资本 ； 中金资本、苏创投、华强资本、长江资本、光谷产投、锡山国投、金雨茂物、新鼎资本、领阳投资、财鑫 ...

Core Viewpoint - The humanoid robot industry is transitioning from a "technology demonstration" phase to a "mass delivery" phase, with companies needing to prove their ability to sell and deliver robots consistently to succeed in the competitive landscape [1][3]. Group 1: Delivery Data and Industry Insights - The significance of achieving a monthly delivery of over 100 units is highlighted, especially in the context of the global humanoid robot industry in 2025, where Chinese companies dominate the top six positions in annual shipment volume [3][5]. - High annual shipment numbers do not necessarily indicate sustainable monthly delivery capabilities, as many companies rely on end-of-year surges or single large project deliveries [5][6]. - The ability to achieve consistent monthly deliveries is a critical differentiator in the humanoid robot industry, reflecting supply chain maturity and engineering stability [7][9]. Group 2: Delivery Essentials - Key performance indicators such as delivery cycle, defect rate, return rate, and after-sales response speed are crucial for determining a company's viability in the humanoid robot sector [9][10]. - Many startups depend heavily on outsourced components, leading to longer delivery cycles and higher return rates, while companies like UniX AI have developed in-house core components, significantly reducing delivery times and improving quality [10][11]. Group 3: Manufacturing Strategy - UniX AI has opted for a heavier, more stable approach by building its own manufacturing base, investing 500 million yuan in a facility capable of producing thousands of robots annually [15][17]. - The local supply chain in Mianyang, where UniX AI is based, boasts a high localization rate of 80%, providing a solid foundation for robot manufacturing [19]. Group 4: Market Positioning - UniX AI has strategically chosen to focus on the "general B-end" market for bulk deliveries, avoiding the crowded C-end market dominated by competitors like Unitree [20][21]. - The company’s robots, such as the Wanda and Panther series, are designed to meet flexible demands across various sectors, ensuring scalability and adaptability [22]. Group 5: Capital Backing and Growth - UniX AI has successfully completed multiple rounds of financing, raising a total of 6 billion yuan within a short period, indicating strong market confidence in its scalable delivery capabilities [23][24]. - The company currently holds over 1,000 orders across high-value sectors such as hospitality, property management, security, and retail, showcasing its effective business model [24][26].

十四届全国人大四次会议5日上午在人民大会堂开幕，习近平等党和国家领导人出席开幕会，国务院总理李强向大会作政府工作报告。 政府工作报告：培育壮大新兴产业和未来产业。建立未来产业投入增长和风险分担机制，培育发展未来能源、量子科技、具身智能、脑机接口、6G等未 来产业。 ...

Core Viewpoint - Xingdong Era has completed a strategic financing round of 1 billion RMB, with a valuation exceeding 10 billion RMB, attracting significant investment from both domestic and international industry leaders, indicating strong market confidence and growth potential in the embodied intelligence sector [3][4]. Group 1: Financing and Investment - The recent financing round occurred just two months after the previous one, with the intended investment amount significantly exceeding the fundraising target [4]. - A total of 16 domestic and international industrial investors have been attracted, including major players like Geely Capital, Alibaba, and Samsung, forming a robust ecosystem of "international giants + domestic leaders" [3]. Group 2: Technological Capabilities - Xingdong Era is the only company in China with full-stack self-research capabilities in embodied intelligence, having developed a comprehensive technical support system that includes perception, decision-making, and control [5][6]. - The company has achieved global recognition with its ERA-42 model, which integrates the entire process of perception, decision-making, and control, making it the first to accomplish full-body and dexterous hand operations [6][8]. Group 3: Market Performance and Applications - Xingdong Era has accumulated orders exceeding 500 million RMB, with international business accounting for 50%, indicating strong market penetration across North America, Europe, and Asia [12]. - The company has made significant advancements in logistics applications, including the first embodied intelligence logistics solution in China, achieving operational efficiencies of up to 70% in various cities [14][16]. Group 4: Future Outlook - The company plans to expand its logistics applications into e-commerce, industrial manufacturing, and pharmaceutical distribution, aiming to build a comprehensive embodied logistics service system [16]. - The recent financing is seen as a new starting point for accelerating technological and commercial upgrades, reinforcing the company's leadership position in the embodied intelligence industry [16].

Core Viewpoint - The article highlights the recent strategic financing of 1 billion yuan for the company Star Motion Era, which has achieved a valuation exceeding 10 billion yuan, driven by its leading technology in embodied intelligence and strong partnerships with major global institutions [1][3]. Group 1: Financing and Valuation - Star Motion Era completed a strategic round of financing amounting to 1 billion yuan, achieving a valuation over 10 billion yuan [1]. - The financing was backed by a consortium of investors including Samsung, Singtel, and several domestic industry leaders, reflecting strong market interest [1][7]. - The company has attracted 16 domestic and international investors, forming a robust ecosystem that integrates both international giants and domestic leaders across various industries [1][7]. Group 2: Technological Advancements - Star Motion Era is the only company in China with a fully self-developed stack that includes brain, control, ontology, joints, and dexterous hands, establishing a comprehensive technical support system [4]. - The company’s self-developed end-to-end embodied model, ERA-42, is the first globally capable of performing full-body and dexterous hand operations, enhancing the practical application of embodied intelligence [4][6]. - Collaborating with Stanford University's Chelsea Finn team, the company has developed the Ctrl-World model, achieving top rankings in global assessments and introducing innovative frameworks for embodied intelligence [6]. Group 3: Commercialization and Market Presence - Star Motion Era has secured over 500 million yuan in cumulative orders, with international business accounting for 50% of its operations, indicating strong market demand [7][8]. - The company has established logistics solutions that have significantly improved operational efficiency, with some applications achieving up to 70% efficiency in major cities [8]. - Major clients include nine of the top ten global technology companies, with high repeat purchase rates, demonstrating strong customer loyalty and market penetration [7][8]. Group 4: Future Prospects - The recent financing will bolster Star Motion Era's technological and commercial strategies, enabling further expansion into e-commerce, industrial manufacturing, and pharmaceutical distribution [8]. - The company aims to transition from technology development to industrial application, positioning itself as a leader in the global smart industry upgrade [8].