以下文章来源于深蓝AI ，作者深蓝学院 深蓝AI . 专注于人工智能、机器人与自动驾驶的学习平台。 作者 | 深蓝学院 来源 | 深蓝AI 点击下方 卡片 ，关注" 自动驾驶之心 "公众号 戳我-> 领取 自动驾驶近15个 方向 学习 路线 >>自动驾驶前沿信息获取 → 自动驾驶之心知识星球 本文只做学术分享，如有侵权，联系删文 FIRI: Fast Iterative Region Inflation for Computing Large 2-D/3-D Convex Regions of Obstacle-Free Space 作为机器人与无人机领域的杰出学者，高飞老师始终站在研究的最前沿。 高飞：国家优青。浙江大学控制学院长聘副教授，研究员，博士生导师。主要研究方向：空中机器人、轨迹规划、自主导航、集群协同、定位感知。近年来，以第一作者/通讯作者身份 在知名机器人期刊、会议发表论文70余篇；获IEEE TRO 2020年最佳论文荣誉提名奖、国际基础科学大会ICBS 2024前沿科学奖、IEEE ICRA 2024年最佳论文提名等学术荣誉；入选爱思 唯尔数据库2023/24年度全球前2％顶尖科学家。 实 ...

Core Viewpoint - The interview with Wang Puzhong, a key figure at Meituan, highlights the competitive landscape in the food delivery market, emphasizing the challenges posed by rivals like JD.com and Ele.me, and the aggressive strategies being employed by major players in the industry [1][3][21]. Group 1: Competitive Landscape - Wang Puzhong asserts that the food delivery market is difficult to penetrate, indicating that the growth in orders from competitors is merely a bubble and does not guarantee future success [3][4]. - Meituan's business structure, based on a diversified strategy, faces challenges from multiple competitors across various sectors, including Ele.me, JD.com, and others [4][5]. - The competition is characterized as a war where each giant is willing to invest heavily, with JD.com and Alibaba leading the charge in terms of financial backing and strategic moves [7][10]. Group 2: Strategic Responses - Meituan is committed to defending its core business in food delivery, even at the cost of abandoning community group buying initiatives, and is focusing on expanding its instant retail infrastructure [12][20]. - The company has increased its investment in technology, particularly in robotics and AI, to enhance operational efficiency and maintain its competitive edge [17][19]. - Meituan's delivery network is currently the most reliable in the industry, but the short-term advantages may be overshadowed by competitors' aggressive subsidy strategies [11][14]. Group 3: Future Outlook - The potential outcomes of the ongoing competition could range from a successful defense by Meituan to a stalemate or intervention by external forces, with various scenarios predicted by analysts [21][23]. - The entry of other players like Pinduoduo and Douyin into the instant retail space indicates that the competition will remain fierce and multifaceted [23].

Core Viewpoint - Hangzhou Yushu Technology Co., Ltd. has officially initiated its IPO counseling process, with CITIC Securities as the counseling institution, marking a significant step towards its public listing [1][3]. Company Overview - Hangzhou Yushu Technology was established on August 26, 2016, with a registered capital of 364 million yuan, and is classified under the C39 category of computer, communication, and other electronic equipment manufacturing [2][3]. - The controlling shareholder and actual controller of the company is Wang Xingxing, who directly holds 23.8216% of the shares and controls an additional 10.9414% through Shanghai Yuyi Enterprise Management Consulting Partnership, totaling 34.7630% [2][3]. IPO Counseling Process - The counseling agreement was signed on July 7, 2025, and the counseling will include comprehensive training on legal knowledge related to IPOs, ensuring that participants understand their responsibilities as a public company [2][3]. - The counseling institution will conduct a comprehensive assessment in October 2025 to determine if Yushu Technology meets the conditions for issuing and listing its shares [3]. Recent Developments - On June 18, 2025, Yushu Technology completed a capital increase from 2.889 million yuan to 364 million yuan, representing a 125-fold increase [4][5]. - The company has undergone a name change from Hangzhou Yushu Technology Co., Ltd. to Hangzhou Yushu Technology Co., Ltd. to align with its IPO preparations [3][4]. Financial Performance - Yushu Technology focuses on the research, development, production, and sales of high-performance quadruped and humanoid robots, achieving annual revenues exceeding 1 billion yuan [6]. - The company holds a 69.75% market share in global quadruped robot sales and ranks first in humanoid robot shipments as of 2023 [6]. - Yushu Technology has completed 10 rounds of financing, with notable investors including Meituan, Sequoia China, and Tencent, and its pre-IPO valuation exceeds 10 billion yuan [6][7].

Group 1 - The core viewpoint of the article highlights the increasing interest in the robotics industry, with approximately 10 companies seeking to list on the Hong Kong Stock Exchange this year, including Megatech [1] - Megatech has submitted its IPO application to the Hong Kong Stock Exchange, with a focus on smart laboratories and intelligent manufacturing, achieving a compound annual growth rate (CAGR) of 43% over the past three years [2][12] - The company has incurred significant losses, totaling 2.28 billion yuan over the past three years, primarily due to high research and development expenditures [12][14] Group 2 - Megatech's revenue for 2022, 2023, and 2024 was 455 million yuan, 663 million yuan, and 930 million yuan, respectively, with gross margins of 28.1%, 23.9%, and 29% [13][15] - The company has a diverse client base, serving over 880 customers across various sectors, with a significant portion of revenue coming from its top five clients [18] - The robotics market, particularly in the field of autonomous intelligent agents, is projected to grow significantly, with the global market expected to reach approximately 383.7 billion yuan by 2030 [23][24] Group 3 - Megatech's products are categorized into two main areas: smart laboratories and intelligent manufacturing, with the latter accounting for a larger share of revenue [16][18] - The company faces competition from both international and domestic players, ranking sixth in the smart laboratory autonomous intelligent agent market in China [26][28] - The robotics industry is undergoing a transformation driven by advancements in AI and machine learning, creating substantial opportunities for growth [22][23]

Core Viewpoint - The article highlights the urgent need for Europe to revitalize its robotics industry, as it risks falling behind China and the US in the global robotics market, which is projected to reach $100 billion by 2030 [1][2]. Summary by Sections Current State of Robotics in Europe - Europe accounted for only 17% of global industrial robot installations in 2023, significantly lower than the 72% in the Asia-Pacific region [3]. - From 2019 to 2024, global investment in robotics surged by 230%, yet European startups received only $3.7 billion in funding in 2024, less than a quarter of the US amount [3]. Structural Issues Facing Europe - The fragmentation of the European market hinders collaboration and resource allocation, contrasting with the unified markets of China and the US [4]. - Europe struggles with technology transfer from research to market, with only 7 out of the top 20 AI patent holders being European, while the US and China hold nearly 80% of the patents [4][6]. - Regulatory rigidity, particularly the EU's AI Act, stifles innovation and complicates compliance for robotics companies [6]. - A lack of cohesive national strategies has led to missed opportunities in emerging technologies like humanoid robots and autonomous vehicles [6]. - Demographic challenges, such as an aging population and labor shortages, have not been effectively leveraged to promote robotics adoption [7]. Strategic Recommendations for Europe - The report suggests focusing on industries with low automation levels and significant labor shortages, such as logistics, healthcare, and agriculture, rather than competing in high-cost areas like humanoid robots [10]. - It advocates for increased investment in traditional strengths like collaborative robots and medical assistance robots to build a robust technological foundation [10]. - To address funding gaps, the report recommends establishing a European robotics "mother fund" to attract private capital and leveraging the recent surge in AI investment [12]. - A proposed increase in robotics-related budgets by at least 5% from 2028 to 2034 could facilitate faster technology transfer from research to industry [14]. - A unified skills framework for robotics training across the EU is essential to address the skills gap in the workforce [15]. - The establishment of regulatory sandboxes could allow for innovation while ensuring safety, enabling companies to test new technologies in controlled environments [17]. Future Outlook - The report emphasizes that the integration of AI with robotics could significantly enhance productivity, with predictions indicating a 31% contribution to manufacturing value added by 2035 [20]. - If Europe can effectively harness its engineering education and manufacturing expertise, it has the potential to redefine its role in the global robotics landscape [20].

Group 1 - China National Offshore Oil Corporation (CNOOC) has developed the country's first deep-water oil and gas ROV (remotely operated vehicle) with seven functions, which has successfully completed its first application in the Pearl River Mouth Basin project [1] - The ROV features high operational sensitivity, a wide range of motion, and a strong load capacity, capable of operating at depths of up to 7,000 meters and withstanding pressures of approximately 700 atmospheres [1] - The project team utilized a digital testing system to ensure smooth operation, conducting continuous automated testing for 10 hours and multiple rounds of verification including land functional tests and load tests [1] Group 2 - The world's first vacuum suction cup high-speed rail wall-climbing robot was showcased at the 12th World High-Speed Rail Conference in Beijing, developed by the Guangzhou Railway Group in collaboration with a robotics company [5][6] - This robot can navigate complex steel structures and perform tasks such as rust removal, inspection, and cleaning, with a modular design that enhances operational efficiency by 40% [6] Group 3 - The University of Southern California has developed the MOTIF robotic hand, which integrates thermal, inertial, and force sensors to enhance robotic dexterity [10] - Inspired by the human hand, the MOTIF hand can safely grasp high-temperature objects and accurately identify objects of the same shape but different weights, showing promising results in laboratory tests [10] - Future plans include integrating high-resolution fingertip sensors and optimizing algorithms for applications in home and industrial settings [10] Group 4 - A team from KAIST and Chungnam National University has developed a bionic sensory neural system for robotic hands, capable of simulating human-like responses to touch [13] - This system uses a new type of memristor to mimic biological functions, allowing the robotic hand to ignore repeated harmless stimuli while responding quickly to dangerous signals [13] - The technology has potential applications in micro-robots and prosthetics, aiming to enhance energy efficiency in these fields [13] Group 5 - Nagoya University has released Japan's first real-time AI dialogue system, J-Moshi, which can mimic natural Japanese conversation patterns [16] - J-Moshi captures short responses typical in Japanese dialogue, addressing limitations of traditional AI in simultaneous listening and speaking [16] - The system has practical applications for non-native speakers learning Japanese and is being explored for use in call centers, healthcare, and customer service [16]

Group 1 - Nvidia's CEO Jensen Huang emphasized the importance of the Chinese market, stating it has cultivated 50% of the world's AI talent [3][9][10] - The company has made significant advancements in AI technology, with the introduction of the DeepSeek open-source inference model, which is considered revolutionary [3][4][10] - Nvidia's stock reached an all-time high, with a market capitalization exceeding $4.1 trillion, reflecting strong investor confidence [2][5] Group 2 - Nvidia has a long history in the AI infrastructure sector, starting from gaming chips in 1993 and evolving into a leader in AI computing [4][5] - The company has developed technologies that have increased AI computing power by 100 times, significantly outpacing Moore's Law [5][10] - Nvidia's partnerships with major Chinese tech companies like Tencent, Alibaba, and Xiaomi have been crucial for its growth and innovation in the region [4][6] Group 3 - Huang highlighted the rapid development of AI applications in China, driven by intense market competition and a strong educational system [3][6][10] - The company is focused on creating a global AI ecosystem, with applications ranging from healthcare to autonomous vehicles [6][8] - Nvidia's H20 chip is particularly suited for training large models, while the RTX Pro is designed for digital twins and robotic simulations [7][10] Group 4 - Huang praised Huawei for its strong technological foundation in chips, systems, and networks, asserting that underestimating Chinese manufacturing capabilities is naive [9][10] - The future of AI is expected to be driven by robotics, with a focus on automation to address global labor shortages [11][12] - Huang expressed optimism about the unique advantages China possesses in AI technology, mechanical engineering, and manufacturing, which will foster the growth of the robotics industry [11][12]

Core Insights - NVIDIA's CEO Jensen Huang emphasized the company's commitment to the Chinese market during his recent visit, highlighting the importance of AI development and collaboration with local tech firms [2][3][4] - The introduction of the H20 chip to the Chinese market and the upcoming RTXpro graphics card were significant announcements, indicating NVIDIA's ambition to expand its product offerings in China [2][7][8] Group 1: Company Strategy and Market Position - Huang expressed that NVIDIA has been operating in China for 30 years and has established partnerships with major companies like Baidu, Alibaba, Tencent, and Xiaomi [4] - The company aims to maintain its growth momentum in China, with Huang stating that NVIDIA's growth is just beginning, not ending [8] - Huang acknowledged the strong capabilities of competitors like Huawei, particularly in chip design and cloud services, indicating a respect for the competitive landscape [4][5] Group 2: AI and Technological Advancements - Huang highlighted that AI has become a fundamental infrastructure, comparable to electricity and the internet, and that China is at the forefront of this transformation [6] - He noted the impressive advancements in AI models and applications in China, citing companies like DeepSeek and the global impact of TikTok as examples [6] - Huang predicted significant developments in the robotics sector in China, driven by the country's robust manufacturing base and supply chain capabilities [6] Group 3: Product Launch and Supply Chain - The H20 chip, designed specifically for the Chinese market, was recently approved for sale, following previous export restrictions that cost NVIDIA an estimated $8 billion in losses [7] - Huang mentioned that there are numerous orders for the H20 chip, but it will take time to ramp up production capacity to meet demand [7] - He expressed optimism about future advancements in NVIDIA's technology, indicating that better products will be available in the coming years [8]

Core Viewpoint - Huang Renxun's visit to China highlights the significant advancements in AI technology, particularly by companies like Huawei, and emphasizes the competitive landscape in the AI sector, where both local and global players are rapidly innovating and collaborating [3][4][5]. Group 1: AI Innovation and Competition - Huang Renxun praised Huawei as an incredible technology company with deep traditions and strong capabilities in chip manufacturing and systems [3][4]. - He noted that underestimating Huawei and China's manufacturing capabilities is naive, highlighting the advanced nature of Huawei's mobile technology and communication systems [3][4]. - Huang expressed admiration for China's rapid progress in AI, stating that 50% of global AI researchers are in China, and the country is advancing quickly in AI infrastructure, models, and applications [5][6]. Group 2: DeepSeek and Open Source AI - Huang highlighted DeepSeek as a revolutionary model that is being applied across various fields, including healthcare and robotics, and emphasized the importance of open-source AI in democratizing access to technology [5][6]. - He mentioned that open-source models allow every country and industry to participate in the AI ecosystem, enhancing security through transparency [6]. Group 3: Supply Chain and Manufacturing - Huang acknowledged the significance of China's supply chain, noting that while most of NVIDIA's products are not manufactured in mainland China, many products are closely tied to the Chinese supply chain, especially in consumer electronics [6][7]. - He emphasized China's advanced manufacturing capabilities, particularly in consumer-oriented and precision manufacturing sectors [6][7]. Group 4: Future Trends in AI and Robotics - Huang predicted that the next wave of AI development will focus on robotics, citing NVIDIA's participation in the Chain Expo and the introduction of the new RTX PRO GPU aimed at smart factories and logistics applications [6][7]. - He expressed optimism about China's potential in developing robotics technology, given its unique advantages in AI and manufacturing [7]. Group 5: NVIDIA's Market Position and Strategy - Huang discussed NVIDIA's recent achievement of surpassing a market valuation of $4 trillion, marking it as the highest-valued company in history [13]. - He highlighted NVIDIA's strategy of investing heavily in building advanced AI systems and ensuring its architecture is widely applicable across various cloud platforms [14]. - Huang noted that competition in the AI space is intense, but he views many competitors as friends, fostering a spirit of collaboration that benefits the industry as a whole [14].

Group 1 - The world's first robot autonomously delivers goods via subway in Shenzhen, utilizing AI scheduling algorithms and robotic technology for unmanned delivery [1] - The robot, developed by Vanke's subsidiary, integrates multiple technologies including AI scheduling, panoramic laser radar, and a mechanical chassis [1] - The robot is designed to deliver popular convenience items to 7-Eleven stores within the subway stations [1] Group 2 - Elon Musk has launched a new "AI girlfriend" service, available for $30 per month, featuring characters with playful and inappropriate interactions [2][4] - The service includes two characters, a gothic girl named Ani and a cartoon panda named Bad Rudy, both designed to engage users in a unique manner [4] Group 3 - The 2025 World Humanoid Robot Games has commenced with an event called "Energy Cube Relay," showcasing eight humanoid robots collaborating to light up the energy cube [5][7] - This event aims to connect technology with culture and tradition, highlighting Beijing's development and global technological advancements [7] Group 4 - Zimmer Biomet announced the acquisition of Monogram Technologies for $177 million (approximately 1.2 billion RMB), marking a significant move into the orthopedic surgical robot sector [8][10] - The acquisition aims to enhance Zimmer Biomet's capabilities in total knee arthroplasty through Monogram's innovative robotic systems [10] Group 5 - New Zealand's Canterbury University has introduced an AI underwater robot named Poseidon, designed for marine farming, capable of autonomous underwater operations [11][13] - The robot employs advanced technologies such as Doppler speed recorders, inertial navigation systems, and AI path prediction algorithms to improve efficiency in monitoring marine farms [13]