The need for artificial intelligence (AI) on edge devices has been realized and the race to design edge-optimized chipsets has begun. According to new research from Omdia, AI processing on the edge device circumvents privacy concerns while avoiding the bandwidth, latency, and cost concerns of cloud computing. Edge inference has emerged as a key workload in 2019–20, and many companies have introduced chipset solutions to accelerate AI workloads. Omdia has identified 10 device categories that will drive AI edge chipset demand including mobile phones, drones, HMDs, robots, smart speakers, PCs/Tablets, security cameras, automotive, edge servers and machine vision.
Shipments of devices needing AI are increasing rapidly as the need for AI acceleration has been realized on the edge. New AI driven applications that demand higher compute are continuing to increase, thus fueling the need for innovation in chipset architectures. According to Omdia principal analyst, Anand Joshi, “Different architectures provide different benefits for AI edge acceleration, and it’s becoming common to mix architectures to get the best of breed in AI edge processing. This trend will continue, with multiple architectures merging into AI edge chipsets. The integrated chipsets will become more widespread at the edge in comparison with the discrete chipsets.” Omdia forecasts that global AI edge chipset revenue will grow from $7.7bn in 2019 to $51.9bn by 2025.
Omdia’s report, “Artificial Intelligence for Edge Devices,” provides a quantitative and qualitative assessment of the opportunity for AI edge processing across several consumer and enterprise device markets. Global revenue and shipment forecasts, segmented by chipset architecture, power consumption, compute capacity, training versus inference, and application attach rate for ten device categories, extend through 2025. An Executive Summary of the report is available for free download on the firm’s website.