Tencent Outlines Embodied Intelligence Cloud Provider Position at Embodied Perception Fusion and Multimodal Large Model Innovation Symposium

On September 15, at the Seminar on Embodied Perception Fusion and Multimodal Large Model Innovation, Liao Shaoyao, Director of Tencent Smart Mobility Solutions, outlined Tencent's positioning as a cloud provider in the field of embodied intelligence. Liao stated that Tencent aims to accelerate industry iteration. At the data collection stage, agents establish a connection with the cloud via an SDK and transmit data back through nationwide nodes. At the training stage, Tencent provides tidal computing resources—leveraging off-peak nighttime hours—and builds a thousand-GPU cluster by combining large and small GPUs, as well as memory-intensive and compute-intensive configurations. At the simulation stage, Tencent packages NVIDIA’s representative simulation scenarios into default images, enabling one-click deployment and browser-based access to cloud simulations, integrated with its machine learning platform for training. At the deployment stage, Tencent adopts a cloud-edge-device integrated architecture to dispatch lightweight computing tasks to nearby edge nodes. Additionally, Tencent’s Robotics Lab platform has opened free toolchains and software for agent manufacturers to validate vertical-specific scenarios.