On Monday, at the CES exhibition in Las Vegas, Qualcomm introduced a new-generation "full-stack" robotics architecture that integrates hardware, software, and composite AI, aiming to power various forms ranging from small home robots to full-size humanoid robots. Simultaneously, Qualcomm also launched its latest high-performance robotics processor—the Qualcomm Dragonwing™ IQ10 series—targeting industrial Autonomous Mobile Robots (AMRs) and advanced full-size humanoids. This represents the company's newest generation of processors specifically designed for robotics, further expanding its existing robotics product roadmap and delivering high-performance, low-power "robotic brain" capabilities. Qualcomm stated that leveraging its mature expertise in edge AI and high-performance, low-power systems, this innovation can transform prototype products into deployable intelligent machines. Regarding this, Nakul Duggal, Executive Vice President and Group General Manager of Qualcomm's Automotive, Industrial & Embedded IoT and Robotics business, commented: "As a pioneer in energy-efficient, high-performance physical AI systems, we deeply understand what is required to achieve reliable and safe operation of the most complex robotic systems at scale. By building upon our low-latency, safety-certified high-performance technology—spanning the complete chain from perception and cognition to planning and execution—we are redefining the possibilities of physical AI, moving intelligent machines from the laboratory into real-world application environments." Qualcomm indicated that it is collaborating with multiple companies within its robotics platform ecosystem, including Advantech, APLUX, AutoCore, Booster, Figure, Kuka Robotics, and Robotec.ai, aiming to achieve large-scale deployment-ready robotics applications. Among these, as Figure expands its humanoid robot platform, Qualcomm Technologies is partnering with them to jointly define the next-generation computing architecture. Brett Adcock, Founder and CEO of Figure, stated: "Figure's mission is to develop general-purpose humanoid robots driven by advanced AI to eliminate dangerous and undesirable jobs, enhance productivity across industries, and create economic abundance, thereby enabling humans to lead happier, more purposeful lives. Qualcomm's platform achieves an excellent balance between powerful computing capability and superior energy efficiency, serving as a critical foundation to help Figure turn this vision into reality."

Built on a mature foundation: From concept to deployment Qualcomm stated that this universal robotics architecture leverages the company's expertise in energy efficiency, scalability, and edge AI performance, ushering in a new era of autonomous robots and connected intelligence. Currently, the Dragonwing industrial processor roadmap already supports multiple general-purpose robot forms, including industry-leading humanoid robots from global manufacturers such as Booster and VinMotion. The architecture supports advanced perception capabilities and motion planning combined with end-to-end AI models (like VLA and VLM), enabling generalized operational capabilities and human-robot interaction. Among these collaborations, Qualcomm partnered with Vietnamese robotics company Vinmotion to introduce the Motion 2 robot—a general-purpose humanoid robot. In a video released by Vinmotion on YouTube last week, the robot demonstrated some of its capabilities: it can punch through a wooden board, squat down to pick up a small teddy bear from the ground, and bend its back in a way that most people over 30 can only "dream about." The company claimed that the launch of the Dragonwing IQ10 helps Qualcomm take a significant step toward real-world large-scale deployment in industrial applications. Meanwhile, Qualcomm is also engaged in discussions with Kuka Robotics regarding its next-generation robotics solutions.

Full-stack architecture According to Qualcomm, the general-purpose robotics architecture powered by the Dragonwing IQ10 redefines the possibilities of robotics by combining powerful heterogeneous edge computing, edge AI, mixed-criticality systems, software, MLOps, and an AI data flywheel. The company stated that this system is supported by a growing partner ecosystem and complemented by a comprehensive developer tool suite. The end-to-end holistic solution enables robots to perform intelligent reasoning and adaptation within spatial and temporal contexts, achieving industrial-grade reliability and scalable expansion for multiple forms. This collaborative network accelerates the development of deployable robotics solutions, addressing the "last mile" challenge and driving faster, more scalable innovation adoption across various industries.