en.Wedoany.com Reported - Diji Robot held a media briefing in Shanghai, announcing that the Sunrise S600 computing platform has reached cooperation with over 20 leading industry clients and coordinated adaptation with more than 100 industry chain partners, covering fields such as humanoid robots, industrial embodied intelligence, and embodied large models. Multiple cooperative models have entered the stage of real-world testing and mass production verification.
On July 15, 2026, ahead of the World Artificial Intelligence Conference (WAIC) in Shanghai, Diji Robot held a media briefing in Shanghai to announce the client deployment, technical adaptation, and ecosystem co-building achievements of the Sunrise S600 computing platform. At the meeting, Hu Chunxu, Vice President of the Developer Ecosystem at Diji Robot, introduced the product capabilities and engineering deployment advantages of the Sunrise S600 for mass production of embodied intelligence. Key client representatives including Chen Tongqing, Co-founder of Tashi Zhihang; Nie Xiangru, Co-founder of PaxiNi Perception Technology; Lu Zongqing, Founder of Zhi Zai Wu Jie; and Fan Qingyuan, Co-founder of Octopus Dynamics, attended the meeting and shared practical experiences in real-machine R&D and mass production across various tracks.
Since its release in November 2025, the Sunrise S600 has gained recognition from industry clients for its single-chip full-stack computing power, high-reliability hardware design, and end-to-cloud integrated development toolchain. Within half a year, Diji Robot has reached cooperation with over 20 leading clients, including Tashi Zhihang, Qianxun Intelligent, Zibianliang Robot, Zhi Zai Wu Jie, Ubtech, PaxiNi Perception Technology, Beijing Humanoid Robot Innovation Center, Humanoid Robot (Shanghai) Company, and Xiaoyu Manufacturing, covering scenarios such as humanoid robots, industrial wheeled dual-arm, 3C flexible operations, embodied large models, and full-modal perception. Most projects have entered the stage of real-machine adaptation and scenario testing.
Currently, several benchmark projects have achieved clear deployment progress. The Tashi Zhihang A3 robot will be equipped with the Sunrise S600 to achieve large-scale deployment of industrial embodied robots. The Beijing Humanoid Robot Innovation Center's Embodied Tiangong 3.0 has completed computing power adaptation based on the Sunrise S600, with mass production and delivery planned for the second half of 2026. The Humanoid Robot (Shanghai) Company's Linglong 2.0 will complete multi-scenario application verification in the third quarter of 2026. The Sunrise S600 has moved from the technical verification stage into the deployment cycle of real hardware, real scenarios, and real mass production verification.
Chen Tongqing, Co-founder of Tashi Zhihang, stated that the S600 provides an edge computing foundation for the real-time stable operation of a general-purpose brain on real production lines. Edge computing power is the infrastructure for the industrialization of embodied intelligence, and victory belongs to complete systems that can integrate algorithms, hardware, chips, and scenarios.
Hu Chunxu, Vice President of the Developer Ecosystem at Diji Robot, stated that embodied intelligence is evolving towards batch deployment, and the core of mass production is stability, controllability, and reusability. The comprehensive capabilities formed by the Sunrise S600 in edge computing power, engineering toolchains, and industry chain collaboration can transform robot deployment from "trial and error" to "predictable."
Single SoC Achieves Full-Stack Intelligence: Sunrise S600 Becomes Industry Standard with High Reliability
The core challenge in deploying robots at the edge lies in the difficulty of efficiently balancing large model inference, multimodal perception, and millisecond-level motion control. Multi-chip architectures are costly and complex to debug, hindering mass production stability. The Sunrise S600 adopts a single-chip full-stack design, capable of simultaneously handling intelligent decision-making, environmental perception, and real-time motion control, simplifying the overall hardware architecture and adaptation difficulty. On the hardware level, the Sunrise S600 is equipped with a 4-core Nash BPU, providing 560 TOPS of edge computing power. Combined with an industrial-grade CPU and real-time MCU architecture, it can meet the high-load, low-latency operational requirements of various embodied devices.
In terms of model adaptation, the Sunrise S600 has completed edge-side optimization for mainstream VLA, VLM, LLM, and voice models. The inference frame rate for mainstream robot operation models is stable at 10 FPS, and models such as Pi0, Qwen3-VL-8B, and Whisper-medium all achieve efficient and stable edge-side inference.
The Sunrise S600 is paired with a self-developed full suite of scenario algorithms, covering capabilities such as spatial perception, omnidirectional positioning, and dynamic motion control. It can achieve spatial reconstruction and understanding using a pure vision solution, support stable traversal of complex terrain, and adapt to complex operational scenarios in industrial and humanoid applications.
In terms of mass production reliability, the RDK S600 module, built on the Sunrise S600, has passed 11 reliability tests, supports a wide operating temperature range from -40°C to 105°C, and can adapt long-term to mass production conditions such as industrial vibration, high-low temperature switching, and sustained high loads.
Diji Robot has launched the new Moss Agent Engine, the world's first open-source "Chat-to-Create" AI Agent framework. It runs through a three-layer collaborative system of "Cloud-PC-Edge," closing the loop of development, deployment, verification, and iteration. Paired with the RoboGo cloud platform, RDK Studio visual development tools, and edge scheduling capabilities, it can reduce cross-platform adaptation and repetitive development costs, compressing the robot development cycle from months to weeks.
100+ Industry Chain Partners Collaborate to Build a Mass Production-Level Software and Hardware Supporting System
The mass production deployment of embodied intelligence relies on a complete industry chain support system. Diji Robot has aggregated over 100 industry chain partners, covering the entire spectrum of hardware components, perception modules, controller solutions, software algorithms, and development ecosystems. It has completed deep adaptation based on the Sunrise S600 with hardware partners such as Bosch Semiconductor and GigaDevice, achieved solution integration with perception manufacturers like SmartSens and OV, and collaborated with robot teams from automotive industry Tier1 suppliers such as Luxshare Precision, Zhicheng Technology, Jinmai Electronics, and Huaqin Technology to launch mass production-level robot controller solutions based on the Sunrise S600.
In 2026, the embodied intelligence industry enters a critical phase of deployment verification and mass production pre-research. Relying on the edge computing power, full-stack toolchain, and mature industrial ecosystem of the Sunrise S600, Diji Robot has partnered with over 20 leading clients and coordinated with more than 100 industry chain ecosystem partners to steadily advance multi-scenario real-machine testing and pilot deployments, continuously streamlining the pre-mass production process.
