According to market research firm Canalysup to datePrediction, in 2024, the worldSmartphoneOf the total shipments, it is estimated that about 5% will be equipped with end-side AI computing capabilities, which is equivalent to 60 million units.AI PhoneIndustry insiders analyzed that the huge user scale, portability and complete application ecosystem of smartphones make them the most ideal landing platform for edge AI.
Compared with PCs, it is not new for smartphones to run AI algorithms on chips such as image signal processors and neural network processing units. However, the emergence of large language models has prompted the industry to redefine what an "AI phone" is.
According to Canalys analysis, the most basic requirement for an AI mobile phone is that the SoC needs to contain a dedicated unit that can accelerate AI tasks and can smoothly run various generative AI models on the terminal side, such as language models, image generation, etc. This poses a new challenge to the computing power of mobile phones.
Source Note: The image is generated by AI, and the image is authorized by Midjourney
The report pointed out that unlike other electronic products, smartphones have the advantage of the largest number of users. By 2023, the number of smartphone users worldwide may reach 5 billion. This provides a broad space for the popularization of AI applications. In addition, the portability and rich application characteristics of smartphones are also conducive to the innovative application and personalized recommendation of end-side AI. In environments such as smart homes, mobile phones can also serve as control centers to output personalized device automation operations.
Industry insiders predict that edge AI will initially improve several application scenarios of smartphones. In terms of entertainment, AI can achieve better photo editing, videoSpecial EffectsIn terms of communication, it can automatically answer calls and generate reply messages based on voice intonation. In terms of information acquisition, it can realize content retrieval in videos and intelligent recommendation of search results. For mobile phone manufacturers, these early AI functions may increase the attractiveness of their products.