Arm China Unveils AI Ecosystem at WAIC 2026

ME AI News, Shanghai, China, July 19, 2026 – Arm China, a leading domestic provider of chip IP design and services, today announced its debut at the World Artificial Intelligence Conference (WAIC), partnering with the Arm ecosystem to comprehensively showcase its full AI deployment landscape through three business zones: Edge AI, Physical AI, and Cloud AI. The exhibition illustrated how ubiquitous Arm technology is driving the scalable adoption of AI across industries. The Future of AI is Built on Arm! At the concurrent AI Forward-looking Forum, multiple technical experts from Arm China delivered keynote speeches, presenting a comprehensive overview of Arm China’s forward-looking insights and innovative technologies for the AI era. In alignment with AI development trends, Arm China has built the “Star300” AIoT prototype platform, the “Zhouyi” X3-Pro NPU as a full-scenario high-efficiency computing foundation for Agentic AI, the “Linglong” AI VPU end-to-end technology, and initiated the “Open-Source AIOS Alliance,” firmly laying the technological groundwork for innovation in the AI era. [Highlights from the AI Forward-looking Forum: CPU, NPU, VPU, and AIOS Join Forces to Build a Solid Technical Foundation for Diverse AI Applications] Building the “Star300” AIoT Prototype Platform: Expanding AI Compute from Traditional MCUs The IoT ecosystem is rooted in Arm architecture. Over the years, Arm has kept pace with market demands and technological advancements, delivering a rich portfolio of IP solutions for embedded computing, including the Arm® Cortex®-M CPU series, Arm Helium™ vector extension technology, and the Arm Ethos™ NPU family of AI accelerators. Leveraging Arm’s mature ecosystem, Arm China has developed the “Star” AIoT prototype platform combining Cortex-M CPUs, Arm Helium vector extensions, and Arm Ethos NPUs for embedded AI scenarios. Zhu Xiaoming, Director of CPU Products at Arm China, stated that the first-generation platform, designated “Star300,” selects STAR-MC2 (i.e., Cortex-M52) and Ethos-U55 to fully support mainstream small models—extending AI compute capabilities from traditional MCUs and equipping embedded devices with an AI engine. Next-Gen “Zhouyi” NPU Preview: A Unified, Hierarchical, Scenario-Oriented Computing Architecture to Bring AI Agents into the Real World The Agentic AI era has arrived. AI models—particularly medium and small ones—are continuously improving in capability, inference paradigms are being restructured, inference workloads are becoming highly diversified, and edge-side AI inference requirements vary significantly across applications. These shifts demand a new generation of AI computing foundations. Arm China has developed the “Zhouyi” X3-Pro NPU as a full-scenario high-efficiency computing foundation tailored for Agentic AI. Dr. Shu Hao, Head of the NPU Product Line and Chief Architect at Arm China, said: “The ‘Zhouyi’ X3-Pro NPU adopts a unified, hierarchical, and scenario-oriented computing architecture. Its design philosophy—‘one-time adaptation, multi-chip reuse; scenario-driven, flexible configuration’—enables reuse of model adaptation, optimization, and deployment experiences across multiple industries and products while efficiently meeting diverse application requirements. This will significantly boost R&D efficiency and accelerate innovation deployment, bringing AI agents into every edge device.” Introducing “Wudang”: AI VPU Technology Enables Humans to See the World and Machines to “Understand” Visual Input In the AI era, AI itself has become the largest consumer of video content. Video processing demands are evolving toward higher resolution, superior image quality, and lower latency—requiring powerful VPUs to efficiently encode and decode massive volumes of video. The next-generation “Linglong” VPU “Wudang” was unveiled for the first time. This product fully inherits the previous-generation CAE technology while improving PPA by over 40% and offering a more comprehensive set of encoder specifications. Dr. Huang Xin, Director of VPU R&D at Arm China, said: “The ‘Linglong’ VPU’s AI encoding technology continues to evolve. Arm China has developed end-to-end AI video encoding technology based on a new AI VPU training framework to achieve ‘end-to-end global optimization,’ transitioning from enhancing ‘subjective visual quality’ to empowering ‘diverse AI applications.’ The ‘Linglong’ VPU is committed to becoming the engine of the AI visual world—enabling humans to see the world clearly and empowering AI to understand it.” Initiating the “Open-Source AIOS Alliance” to Explore the Next-Generation Infrastructure for AI Products Current breakthroughs in next-generation AI technologies—represented by AI agents and embodied intelligence—are driving the emergence of new computing platforms. Operating systems are rapidly evolving from traditional tools into agent-based forms. An AIOS intelligent system designed natively for these new AI products and embodied intelligence will become a critical enabler for unlocking edge-side AI potential and accelerating widespread AI adoption. Arm China officially announced the launch of the “Open-Source AIOS Alliance,” which has already attracted over 20 ecosystem partners to collaboratively explore next-generation AI product infrastructure and jointly build an open-source AIOS system foundation. Liang Quan, Vice President of Marketing & Ecosystem at Arm China, said: “Arm China will fully leverage Arm’s global ecosystem resources, end-to-end cloud-edge-device technical advantages, and model capabilities to provide a solid foundational technology base for innovation in next-generation AI products and drive the scalable deployment of AIOS across diverse emerging intelligent terminals.” [Three Business Zones in Action: Edge AI, Physical AI, Cloud AI – Arm Technology Empowering Full-Spectrum AI Applications] Arm technology is making AI accessible everywhere. The Arm China booth features three dedicated business zones: Edge AI, Physical AI, and Cloud AI—showcasing diverse AI applications powered by Arm across smart wearables, smart homes, AI PCs, smartphones, smart vehicles, robotics, data centers, and more. Edge AI Zone Arm delivers a full-stack solution for Edge AI—from IP and compute subsystems to software. For example: Arm Lumex compute subsystems and Arm Kleidi™ software are deployed in smartphones; Arm endpoint compute subsystems power AI PCs; Cortex-M CPUs + Helium + Ethos NPUs suit embedded devices; and Armv9 technology enables smart homes. Visitors can witness “how chips are made” and explore terminal applications of chip IP across smart wearables, consumer electronics, smart offices, smart homes, smart healthcare, and smart industry. “One grain of sand holds a world; one chip ignites countless industries”—clearly demonstrating how Arm technology empowers edge-side AI. Key demonstrations include three solutions: Qwen-Audio family TTS models based on Arm technology, an AI MCU product portfolio, and JishuShell—a one-click deployment tool for edge-side agents. Qwen-Audio Family TTS Models: Leveraging Arm SME2, Qwen-Audio family TTS models achieve efficient execution on a single Arm CPU—delivering low-latency, high-quality AI voice experiences across Arm-enabled devices.AI MCU Zone: Showcases a comprehensive solution matrix based on Arm Cortex-M CPUs with Helium and Ethos NPU, covering diverse applications such as automotive monitoring, traffic surveillance, AI companion toys, visual tracking, gesture recognition, and industrial predictive maintenance. The Arm China “Star 300” AIoT prototyping platform also demonstrated applications including object detection, speech recognition, keyword detection, and super-resolution on an FPGA platform. JishuShell Joint Exhibition: On-site demonstrations featured development boards equipped with JishuShell’s one-click deployment tool for edge agents, along with hands-on sessions. Leveraging its developer ecosystem, JishuShell is rapidly entering the Agentic AI domain, collaborating with Arm ecosystem chip and solution providers to explore next-generation AI applications and interactions on edge devices, with plans to further integrate AIOS capabilities. Physical AI Zone Arm provides full-stack support for Physical AI, including IP, compute subsystems, interconnect technologies, tools, and integrated software. Among these, the Arm Zena™ compute subsystem and AE (Automotive Enhanced) technology are applicable to both intelligent vehicles and robotics. The Physical AI Zone was one of the most popular areas at the event, highlighting Arm’s achievements in robotics and intelligent vehicles. A “robot all-star team” of over ten exhibits—including humanoid robots, emotionally interactive robots, coffee latte-art robotic arms, robot dogs, and robot pandas—made a collective debut. Arm technology underpins every aspect of these robots, from the central brain and interactive computing to battery management, local control, sensors, and actuators, forming a complete system architecture. Multiple intelligent vehicle chips and intelligent connected computing platforms were also unveiled simultaneously. Arm technology spans five key domains—powertrain, chassis, body, cockpit, and autonomous driving—enabling comprehensive intelligent upgrades across the automotive industry. Cloud AI Zone With its leading “performance per watt” advantage, the Arm Neoverse® compute platform is becoming a foundational pillar for cloud computing and AI infrastructure, delivering system deployments that balance performance, efficiency, and security for cloud and data center environments. Currently, nearly 50% of the compute power shipped to leading hyperscale cloud providers is based on Arm architecture. The Cloud AI Zone prominently featured Arm-based DPU chips, intelligent network cards, SuperNIC cards, and multi-type server racks (2U, 4U, 6U), covering the entire chain from network acceleration to full-system deployment. Beyond cutting-edge technology, the booth also offered engaging interactive experiences: a robotic arm brewed fresh coffee and performed precise latte art; a brain-computer interface “thought-powered popcorn” challenge delivered novel interactive fun. These playful interactions brought AI from the exhibition floor into everyday life, allowing the public to personally experience the innovative appeal of AI. Looking ahead, Arm China will follow its “AI Arm China” strategic direction, focusing on Edge AI, Physical AI, and Cloud AI as its three core pillars to advance its “All in AI” product strategy, build a robust computing foundation, and empower the AI era alongside ecosystem partners. (Source: Ifnar)