With shipments set to hit 1.6 billion by 2031, machine vision and IoT sensors are driving edge AI chip growth, reshaping intelligent devices globally.
Shipments of edge AI inference and training chipsets are set for a compound annual growth rate (CAGR) of 18% between 2024 and 2031, according to a recent report published by ABI Research. With this, the annual shipments are expected to reach 1.6 billion units by 2031.
Among the use cases, machine vision has emerged as the clear leader, accounting for the largest share of shipments. Units for machine vision applications are projected to climb sharply from around 150 million in 2024 to over 500 million by 2031. The surge highlights the expanding adoption of cameras and vision-based AI systems in robotics, autonomous vehicles, manufacturing automation, and retail analytics.
Sensor data analysis is also gaining momentum, with shipments expected to approach 370 million units by 2031. This growth reflects the proliferation of IoT (Internet of Things) sensors in industrial, smart city, and energy management applications, where on-device AI processing reduces latency and network dependency.
Other segments, including AI gateways and miscellaneous edge applications, are also rising. AI gateways, acting as intermediate processing nodes between devices and the cloud, are forecast to reach 250 million units, while the ‘Others’ category, encompassing predictive maintenance, anomaly detection, and embedded AI systems, could exceed 300 million shipments by 2031.
By contrast, audio and sound processing applications are expected to plateau around 100 million shipments, indicating market maturity in areas such as smart speakers and noise detection. On-premises servers remain negligible, underscoring that future growth in edge AI is largely device-driven rather than data-centre-focused.
