On July 29, during the Caixin Roundtable-AI series event titled "The Next Stop for the Terminal Industry in the Agent Era," a white paper commissioned by Qualcomm and independently researched by IDC was officially released. Titled "From AI Devices to Personal AI: Agent-Driven Terminal Evolution and Industry Restructuring," the document systematically outlines and analyzes the current industry landscape based on technological and industrial trends. It explores the core characteristics, value potential, and implementation pathways of "Personal AI."

IDC China Vice President Wang Jiping noted that the AI super cycle is pushing the terminal market toward a "tipping point" of transformation. Terminal manufacturers currently face short-term pressures such as rising device prices and declining consumer willingness to upgrade. As the traditional logic of "parameter competition driving sales growth" gradually loses its effectiveness, AI terminals are emerging as the biggest growth highlight. Signals from the WAIC 2026 period indicate that intelligent agent terminals have begun to accelerate commercial deployment, with an increasing number of AI-native devices appearing. Wang Jiping emphasized that "Personal AI," which is user-centric and operates with intelligent agents as the core, represents a new form of personal intelligence in the AI era. This will offer manufacturers a breakthrough point for development and define the next wave of terminal growth. Regarding the practical needs for implementing Personal AI, he pointed out that a distributed computing architecture with end-edge-cloud collaboration, along with four key technological pillars—always-on sensing and natural interaction, on-device computing power and personal memory, security and privacy protection, and continuous computing and cross-device collaboration—will form the essential foundation for Personal AI's development. Open collaboration and ecosystem co-construction will become a necessary path for industry advancement, while the integration of hardware, software, and AI services will be a crucial source of terminal value enhancement.

At the event, Qualcomm product technology expert Zhu Yuankun shared Qualcomm's observations and practices regarding terminal innovation in the intelligent agent era. Citing the IDC white paper, Zhu Yuankun stated that Personal AI must possess three core capabilities: personalization, always-on sensing, and privacy security. Traditional computing architectures are facing a systematic overhaul. He explained that four key capability modules—perception, memory, reasoning, and tool execution—will together form the foundation of future intelligent agent experiences and become core issues for future chip hardware and software design, as well as operating system development. Furthermore, cloud-device collaboration will become an inevitable direction for intelligent agents to fully leverage the distinct computing and scenario advantages of the cloud and terminal devices, achieving truly personalized and efficient agent-based AI experiences. Zhu Yuankun emphasized that in the future, all devices will become endpoints of intelligent agents. Creating a user-centric Personal AI experience is not only a direction for technological development but is also becoming a shared goal for the industry.

Zhu Yuankun introduced Qualcomm's AI full-stack solution for the intelligent agent era. This includes the Qualcomm AI Engine, composed of the Oryon CPU, Adreno GPU, Hexagon NPU, sensor hub, and high-speed memory. Through heterogeneous computing, it creates an AI-native chip platform that delivers optimal performance and energy efficiency for different workloads, supporting the continuous operation of intelligent agents and cross-device collaboration. Additionally, Qualcomm's unified AI software stack helps large model manufacturers and application developers more conveniently deploy and optimize AI capabilities on the Qualcomm platform, accelerating AI innovation from concept validation to large-scale implementation.