The market for AI devices will exceed the smartphone market because it encompasses not just phones but a new generation of wearables (glasses, rings, watches) that will serve as constant companions connected to AI agents.
AI devices must be close to human senses to be effective. Glasses are the most natural form factor as they capture sight, sound, and are close to the mouth for speech. This sensory proximity gives them an advantage over other wearables like earbuds or pins.
Unlike the vertically integrated smartphone market, AI wearables will be dominated by a horizontal model where diverse fashion brands integrate technology. Consumers will prioritize personal style and choice, preventing a single tech giant from winning with one design.
Qualcomm's CEO argues the immediate value of AI PCs is economic, not experiential. SaaS providers, facing massive cloud AI costs, will drive adoption by requiring on-device processing to offload inference, which fundamentally improves their business model.
The gap between the promise and reality of personal AI assistants stems from two bottlenecks: immature AI models that lack "physical AI" context, and the latency of cloud computing. Real-time usefulness requires powerful, on-device processing to eliminate delays.
Qualcomm's CEO argues that real-world context gathered from personal devices ("the Edge") is more valuable for training useful AI than generic internet data. Therefore, companies with a strong device ecosystem have a fundamental advantage in the long-term AI race.
The adoption of humanoid robots will mirror that of autonomous vehicles: focus on achievable, single-task applications first. Instead of a complex, general-purpose home robot, the market will first embrace robots trained for specific, repeatable industrial tasks like warehouse logistics or shelf stocking.
The intense power demands of AI inference will push data centers to adopt the "heterogeneous compute" model from mobile phones. Instead of a single GPU architecture, data centers will use disaggregated, specialized chips for different tasks to maximize power efficiency, creating a post-GPU era.