Advances in AI model performance and the computing capabilities of chips embedded in end-user devices are driving the adoption of AI beyond cloud-based infrastructure. AI applications are increasingly being deployed across a wide range of devices, including AI Phone, AI PC, robots, industrial intelligent IoT, smart wearables, smart vehicles and smart home devices. On-device AI enables devices to perform all or part of AI model inference locally, supporting perception, understanding, content generation and decision-making. By coordinating processing tasks across devices, edge computing nodes and cloud platforms, on-device AI helps reduce response latency, maintain functionality under limited network connectivity, and support personalized services and local data management.
Frost & Sullivan, a leading global industry research and growth consulting firm, has released the China On-device AI Industry Blue Book on the Development and the Application of the industry, drawing on its global industry databases, expert interviews and industry case studies. The report provides a comprehensive analysis of the definition, technical architecture, key growth drivers and development trends of on-device AI, with a focus on four core layers: sensing and input, edge computing, on-device AI models and intelligent applications. Looking ahead, the evaluation of on-device AI technologies is expected to place greater emphasis on task execution performance, response latency and energy efficiency. Advances in AI agent technologies will further drive the transition from standalone AI functions to coordinated execution of multiple tasks, while expanding opportunities for cross-device collaboration and industry-specific applications. The report aims to provide a structured industry reference to support technology selection, product development and the deployment of AI solutions across different application scenarios.
