Guotai Junan: Mobile phones are a gold mine for AI data. AI-related phone companies have investment value.

Guotai Junan released a research report stating that mobile phones are a data goldmine waiting to be tapped, possessing the data and hardware interface that AI most needs. In the next five years, with the current development speed of TSMC's processes, other wearable hardware forms have not deviated from the computational and ecological foundation of mobile phone development. Glasses, earphones, PINs, etc. are still accessories to the phone. By 2025, "native AI smartphones" and "AI empowered smartphones" will confront each other head-on, and the superiority or inferiority of the two will soon be revealed, making AI smartphone-related companies worth investing in. Main points from Guotai Junan: - Mobile phones are a data goldmine waiting to be tapped, possessing the data and hardware interface that AI most needs. - Among all current hardware forms, mobile phones have the vast majority of user data, and are the only ones with cross-app permissions, making it possible for mobile phones to develop an "Jarvis" -like intelligent assistant that meets user needs. - Mobile phone companies and AI companies are both aware of the value of mobile phones as a data mine in the AI era. Mobile phone manufacturers have control over data permissions, while AI companies lead in large model technologies, leading to inevitable competition between the two until a new balance is reached. Mobile phone companies' AI has not yet impressed consumers. From the second half of 2023, mobile phone companies' AI has gone through three stages of definition: 1) APP-style AI: AI editing, AI Q&A, etc. 2) Intent understanding: leveraging the massive user data accumulated by mobile phones to deeply analyze user scenarios and hidden intentions. 3) Voice control - APP execution: users can easily perform operations like ordering food or calling a car with voice commands. Some phone brands are still in the first stage of cognition by late 2024, who will likely give up independent AI development and seek external cooperation opportunities. Apple and Honor are relatively leading in intent understanding and voice execution, with Honor being the first to launch a consumer product, and Apple brewing upgrades for Siri in iOS 18.3/18.4. However, the success of mobile phone companies' AI lies in natural semantic understanding, not automated execution programs. AI companies aim for "AI intelligent agents." Only by holding data permissions can the convenient experience of "asking AI for solutions to problems" be achieved. Developing APP-style AI requires data to be generated when users use the APP, while building AI for the phone system generates data continuously as long as the phone is in use. The target of AI companies is the "AI intelligent agent," with the current mainstream interaction mode of AI company APP users still centered on the main dialog box, mainly for one-time inquiries when encountering problems, similar to the use of AI search. It is worth noting that AI company APPs have opened up an "AI intelligent agent" community, where users can independently create agents, covering diverse parameters such as settings, voice, language, and even cloning the user's voice. Although this feature is still in the early stages of exploration, its ultimate goal is to tailor exclusive agents for users. Envisioning the future, agents can work together to complete tasks autonomously and be confirmed and accepted by users, truly shaping an "Jarvis" -like intelligent assistant that meets user needs. AI companies have gained a significant advantage in the consumer market and occupy nearly half of the market. By 2025, "native AI smartphones" and "AI empowered smartphones" will confront each other head-on, and the superiority or inferiority of the two will soon be revealed, making AI smartphone-related companies worth investing in. Risk Warning: AI smartphone penetration rate may not meet expectations, and technological development may fall short of expectations.