Workflow
人形机器人本体软硬件系统
icon
Search documents
半年完成5亿元融资,阿里、蔚来、联想等投了人形机器人|早起看早期
36氪· 2025-03-13 00:01
Core Viewpoint - The article discusses the recent A+ round financing of "Zhuji Power," a company specializing in embodied intelligent robots, highlighting its strategic support from leading institutions and its focus on humanoid robot technology [3][4]. Financing and Support - "Zhuji Power" has completed a total of 500 million yuan in A round financing over the past six months, with strategic investments from major players like Alibaba Group, China Merchants Venture Capital, and NIO Capital [3]. - The company has received continuous follow-up investments from several existing shareholders and has attracted notable financial investors such as Gaojie Capital and Oasis Capital [3]. Technology and Product Development - The company focuses on three core embodied intelligent technologies: design and manufacturing of hardware, full-body motion control based on reinforcement learning, and training strategies for embodied brain models [3][4]. - "Zhuji Power" offers humanoid robot hardware systems and embodied agent development toolchains to various sectors, including research, manufacturing, and commercial applications [4]. Data Utilization and Algorithm Development - The company has made significant progress in AI exploration, algorithm development, and hardware design, addressing challenges in training data collection for humanoid robots [5]. - "Zhuji Power" has introduced the VideoGenMotion (LimX VGM) algorithm, which utilizes human operation video data to train existing video generation models, enabling robots to understand and execute tasks effectively [5][6]. Product Features and Market Presence - The latest iteration of the humanoid robot features increased degrees of freedom and stability, enhancing posture control and balance [6]. - The TRON 1, the world's first multi-form bipedal robot, has been successfully delivered to multiple countries, achieving a commercial closed loop in design, development, mass production, and sales [7][8].
阿里投的具身智能公司,半年融了5个亿!
量子位· 2025-03-06 08:29
Core Viewpoint - The article highlights the recent financing success of LimX Dynamics in the field of embodied intelligence, emphasizing its strategic support from leading institutions and its focus on advancing core technologies in humanoid robotics [2][4][30]. Financing Overview - LimX Dynamics has completed an A+ round of financing, accumulating a total of 500 million yuan in A-round series financing within just six months [3][4]. - The financing has received strategic support from major investors, including Alibaba Group, China Merchants Venture Capital, and NIO Capital, among others [4]. Technology Focus - The funding will primarily be directed towards three core technologies in embodied intelligence: hardware design and manufacturing, full-body motion control based on reinforcement learning, and training strategies for embodied brain models [5][12]. - LimX Dynamics has introduced the TRON 1, the world's first multi-modal bipedal robot, which supports various locomotion forms and aims to lower the barriers for reinforcement learning research [7][10]. Product Development - The company has successfully delivered TRON 1 to multiple countries, achieving a commercial closed loop in design, development, mass production, and sales [11]. - LimX Dynamics has also showcased a full-size humanoid robot capable of performing complex movements, highlighting its high degrees of freedom, flexibility, and stability [13][18]. Algorithm Innovation - The company has developed the LimX VGM, an embodied intelligence operation algorithm that utilizes video generation technology to enhance data training and algorithm performance [19][20]. - This innovation allows for the direct application of human operation data to robot operations, marking a significant advancement in the field [21][24]. Key Features of LimX VGM - The workflow of LimX VGM includes three critical steps: training with human operation videos, generating operation videos with depth information, and executing robot operations based on the generated behaviors [22][30]. - The algorithm's decoupling from the robot body allows for cross-platform deployment, enhancing operational efficiency and reducing costs [28][29].