衡宇 发自 上海 量子位 | 公众号 QbitAI 智元机器人实现对A股科创板公司上纬新材63.62%的控股权后，终于来了次公开、正式的亮相。 今年3月宣布"智元机器人董事长兼CEO"身份的 邓泰华，也在这场智元机器人首届合作伙伴大会上首次公开亮相。 倒也没讲虚的，一股脑吐出许多外界一直想知道又不知道的东西： 慷慨而谈的不只有邓泰华。 会前交流时， 智元机器人通用业务部总裁王闯 、 智元机器人合伙人&具身业务部总裁姚卯青 也出面回答了许多疑问。 我们整理其中精华，分成"吃瓜""技术""合作"三个板块，与大家分享。 融资这事儿智元想融就能融，年底将启动C轮，引入更多国际产业方； 无营收前提下，现金流能撑三年； 今年计划出货几千台，明年出货数万台，未来每年出货几十万台； 商业落地节奏，先ToB再ToC； 落地场景选择路径，简单环境+简单任务→简单环境+复杂任务→复杂环境； 当前主要在打基础 （能力、产品、团队） ，今年起产品成熟度提高，市场与团队具备基础，开始"逐步踩油门"； 智元团队超1000人，平均年龄31岁； 团队成员75%是研发，其中2/3投入AI，1/3投入本体。 未来三年，投入数十亿资金，孵化50个早 ...

Core Viewpoint - The article highlights the recognition of ZhiYuan Robotics as the only robotics company among the six winners of the SAIL Award, showcasing its innovative "QiYuan General Embodied Model" which bridges the gap between image-text input and robotic action execution, enhancing autonomous operational capabilities [2][3][4]. Group 1: SAIL Award Recognition - ZhiYuan Robotics won the SAIL Star Award at the 2025 World Artificial Intelligence Conference for its "QiYuan General Embodied Model," which excelled in multiple dimensions including technological innovation and application [2]. - The SAIL Award, established in 2018, aims to recognize projects that achieve significant technological breakthroughs and innovations in the AI field [2]. Group 2: Technological Innovations - The "QiYuan General Embodied Model" utilizes the ViLLA architecture, which introduces a method for predicting Latent Action Tokens, effectively bridging the semantic gap between image-text inputs and robotic actions [3]. - This technological advancement significantly enhances the robot's ability to understand commands and execute tasks in complex environments, marking a qualitative leap in robotic intelligence [4]. Group 3: Commercial Applications - The commercialization of the QiYuan model signifies a critical shift from specialized task tools to general intelligent agents, unlocking vast potential across various sectors including industrial production, commercial services, and home environments [4]. - In industrial settings, robots can adapt to different production line needs, while in commercial services, they can provide personalized and intelligent user experiences through natural language interaction [4]. Group 4: Data and Evaluation Innovations - ZhiYuan Robotics has established a comprehensive data closed-loop system to continuously collect, analyze, and optimize data, enhancing model performance and adaptability [4]. - The company has also introduced a novel evaluation method that combines simulation assessment with world models, improving testing efficiency and accuracy [5]. Group 5: Development Platforms - To lower the development threshold for embodied models, ZhiYuan Robotics has created Genie Studio, an all-in-one development and application platform that integrates data collection, model training, simulation evaluation, and model inference [5]. - This platform aims to provide a convenient and efficient development environment, fostering technological advancement and innovation in the industry [5].

Core Insights - The AgiBot World Challenge for the World Model track will officially commence in 2025 with a total prize pool of $560,000 [1] - Participants are required to train video generation models based on the AgiBot World dataset to simulate robot actions in real-world scenarios [1] - The competition server is open until September 1, with results to be announced on September 15, and winners will gain direct access to IROS 2025 and potential collaboration opportunities with the organizing company [1] Summary by Categories - **Event Details** - The AgiBot World Challenge will take place in 2025 with a prize pool of $560,000 [1] - The competition server is open until September 1, and results will be revealed on September 15 [1] - **Participation Requirements** - Contestants must develop video generation models using the AgiBot World dataset to mimic robot movements in realistic environments [1] - The organizing company provides the EVAC baseline model and evaluation tools to assist participants [1] - **Opportunities for Winners** - Winners will have direct access to IROS 2025 [1] - There is a potential for collaboration with the organizing company for successful participants [1]

Core Insights - The development of embodied intelligence is currently hindered by a shortage of data, with the available data for this industry being only a fraction of that used by large language models [1][2] - The establishment of data collection factories across various regions in China, such as the Super EID Factory in Tianjin and others in Shanghai and Beijing, marks the beginning of efforts to address this data scarcity [1][2] Group 1: Data Collection Facilities - The Super EID Factory in Tianjin spans 12,000 square meters and focuses on producing data rather than physical products, with a daily capacity to collect up to 550,000 data points and an annual output of 200 million high-quality data points [2][4] - The Zhiyuan data collection center in Shanghai, covering 3,000 square meters, has already collected over one million high-quality data points across various real-world scenarios [3][5] Group 2: Data Collection Methods - Different companies employ various methods for data collection; for instance,帕西尼 uses human data collection where operators wear devices to capture actions, while Zhiyuan employs robotic teleoperation to gather data through repetitive tasks [4][5] - The data collection process involves capturing both simulated and real data, with real data being the most valuable yet challenging to obtain [2][3] Group 3: Industry Challenges and Standards - The lack of standardized data collection practices poses a significant challenge for the industry, making it difficult to reuse and transfer training data across different applications [6] -帕西尼 and Zhiyuan are working towards building an open ecosystem to enhance data availability and quality, with帕西尼 actively participating in establishing data collection standards [6]