Gasgoo Munich- GIGAAI and D-Robotics officially entered a strategic partnership on September 14.Under the agreement, the two sides will systematically integrate key links in the deployment of embodied intelligence. They aim to build a comprehensive solution combining a "world model brain, an on-device embodied brain, a high-compute on-device base, and a physical execution loop."Specifically, within this architecture, GIGAAI handles scene generation, action outcome prediction, strategy evaluation, and failure sample retraining — providing an interactive, assessable, and iterative virtual training environment. GigaBrain serves as the on-device embodied brain, managing natural language task understanding, visual-spatial perception, task decomposition, skill routing, and on-device decision-making. Meanwhile, GIGAAI-Policy oversees the generation of action candidates, strategy output, and failure recovery.D-Robotics' RDK S600, for its part, takes on the computational and engineering deployment tasks, including on-device multimodal inference, task scheduling, trajectory planning, status monitoring, and data collection.The RDK S600 is D-Robotics' flagship high-compute platform designed for embodied intelligence, with 560 TOPS of computing power. Built on a mature multi-core heterogeneous architecture, it supports efficient and stable on-device inference for VLA, VLM, and embodied foundation models across hardware architecture, compute scheduling, and model deployment toolchains.Image Credit: GIGAAIFrom there, physical robot bodies such as GIGAAI’s Maker and SeeLight S1 execute tasks in the real world, continuously feeding data back into the model iteration pipeline. This creates a full closed loop spanning "visual perception, language understanding, task decomposition, action generation, physical execution, data feedback, and world model iteration."Leveraging the S600's on-device computing power and AI-native toolchain, the partners will also work on model conversion, quantized deployment, operator adaptation, inference acceleration, memory optimization, and power tuning for GigaBrain-0.7. The goal is to continuously refine a software-hardware integrated solution that delivers higher performance, lower latency, and greater stability.Furthermore, the two companies will accelerate on-device testing and engineering applications of GigaBrain across various form factors, including Maker, Shiguang, robotic arms, dual-arm systems, and mobile chassis. They aim to deliver mature, deployable solutions for industrial manufacturing, home services, commercial services, and the developer ecosystem.Currently, GIGAAI's GigaBrain series is already advancing toward commercial deployment in multiple home and industrial scenarios.