TENCENT(HK:00700) 机器人圈·2025-08-01 11:55Core Viewpoint - Tencent is positioning itself as a platform provider in the field of embodied intelligence rather than a competitor in hardware production, focusing on software capabilities and collaboration with other enterprises [3][5]. Group 1: Tencent's Strategy on Embodied Intelligence - Tencent's strategic view on embodied intelligence was articulated by Zhang Zhengyou at the WAIC, emphasizing the company's long-term vision and commitment to developing a robust platform [2]. - The launch of the Tairos platform signals Tencent's intent to provide tools and support for other companies in the embodied intelligence sector, rather than competing directly in hardware production [3][5]. Group 2: Role and Timing of Tairos - Tairos serves as a modular platform that allows companies to select specific capabilities according to their needs, addressing the challenges of data requirements and model training in the embodied intelligence field [7][8]. - The decision to open Tairos was influenced by insights gained from visiting numerous enterprises, revealing a significant demand for the tools and technologies that Tencent has developed over the past seven years [6][7]. Group 3: Hardware Development and Prototypes - Tencent is not pursuing the mass production of hardware like robotic dogs but is instead focused on developing research prototypes for internal validation and exploration [5][6]. - The small robot prototype, Xiaowu, is being tested in complex scenarios such as elderly care, where safety and robustness are critical, highlighting Tencent's commitment to refining technology through practical applications [5][6]. Group 4: Challenges in the Embodied Intelligence Industry - The embodied intelligence industry faces several core technical challenges, including dynamic 3D modeling, standardization of tactile sensors, and insufficient model generalization capabilities [6][7]. - There is a significant need for large datasets to train effective embodied intelligence models, and the current trend towards end-to-end model training is limited by data requirements [7][8]. Group 5: Future of Embodied Intelligence - The industry is currently experiencing a phase of exploration, akin to the early days of mobile communication, with the potential for significant advancements as the market matures [9][10]. - Zhang Zhengyou suggests that the ultimate form of embodied intelligence should not be limited to human-like robots, advocating for diverse forms that better suit modern living environments [10][11].