Workflow
具身智能布局“交卷”,腾讯、京东、商汤猛掐机器人生态位
Nan Fang Du Shi Bao·2025-07-31 05:28

Group 1: Core Insights - Major tech companies are accelerating their investments in the robotics sector, with a total of 23 funding rounds in embodied intelligence enterprises this year [1][6] - Companies like Tencent, Alibaba, Meituan, Baidu, JD.com, and Xiaomi are building their own robotics teams to enhance their capabilities in embodied intelligence [1][6] - The World Artificial Intelligence Conference (WAIC) showcased various products and collaborations from these companies, highlighting their focus on developing industry-specific models and partnerships [1][2][4] Group 2: Company Strategies - Tencent launched the Tairos open platform, which provides modular software solutions for the robotics industry, aiming to empower developers with essential software capabilities [2][8] - JD.com has established its JoyInside brand, collaborating with numerous robotics companies to integrate dialogue-driven intelligent agents into various smart hardware [4][8] - NetEase introduced the "Lingjue" model for outdoor mining excavators, focusing on specific applications and open-sourcing its training dataset [5][10] Group 3: Investment Trends - Among the seven major companies, Alibaba has been the most aggressive with six investments, followed by Meituan with five, and JD.com with four [6][7] - The investments are primarily directed towards startups in the embodied intelligence space, with a focus on enhancing their own technological capabilities [6][11] Group 4: Future Directions - Companies are increasingly focusing on data, platforms, and models rather than hardware, as they seek to establish a strong position in the embodied intelligence ecosystem [11][12] - There is a recognition that while major companies have advantages in data and algorithms, they face challenges in hardware development, creating opportunities for startups [12][14] - The industry is expected to see further consolidation as major players look to acquire companies that specialize in perception and control layers to complete their embodied intelligence ecosystems [14]