chinese manufacturers race with apple, ai phones enter the era of "autonomous driving"
2024-09-11
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
the super-wave of smartphone replacement has not yet arrived, but the tension between technology giants is growing.
in the past two days, when people were hotly discussing the differences between the iphone 16 and the previous generation, the topic of huawei's "tri-fold screen mobile phone" came up again. not only did the launch time of the two "collide", but the product release time was also set on the same day.
the innovation track does not need any humble players. when apple was late in the field of foldable screens and ai, a new round of competition for technological innovation in global smartphones has begun.
according to the reporter, in the past year, not only in the folding screen track, chinese mobile phone manufacturers have also joined hands with more large model companies to promote the implementation of ai strategies at home and abroad, and even invested heavily to deploy ai to the operating system layer and reconstruct the underlying software and hardware systems of mobile phones. among them, there are many domestic and foreign ai giants in the cooperation, including bytedance, alibaba, as well as google, meta and microsoft.
according to analysts, ai native applications built on generative ai will be the key to stimulating the growth of ai mobile phones in the future. if ai native applications can reshape the existing habits of mobile internet content consumption, there will be an opportunity to shorten the replacement cycle of mobile phones.
chinese mobile phone manufacturers "outflank" apple
when ai can be used on iphones is one of the focuses of chinese apple users.
in the early morning of september 10th, beijing time, at the autumn new product launch conference, apple successively updated three major hardware products: apple watch, airpods and iphone. many ai functions were integrated into its various product series, including the affordable iphone 16 series.
however, the launch time of ai functions is different in different regions. apple said that the beta version of apple intelligence will be launched in the us version in october. it will cover other localized english versions by the end of the year, and chinese, french, japanese and spanish versions will be launched next year. in this regard, an industry insider joked: "iphone 16 is here, but ai is still on the way."
apple has previously announced a partnership with openai, but it has yet to announce an ai partner in china.
at the same time as apple's press conference was held, some management teams from chinese mobile phone manufacturers were also watching the live broadcast. a person from the honor management team told reporters after the press conference that the ai mentioned by apple is more about reconstructing mobile phone applications from the system level, using the "intelligent body" method to enable ai to achieve "autonomous driving" on the mobile phone side. "but in fact, chinese mobile phone manufacturers' deployment of ai is not slower than apple, and may even be faster," the person said.
three days before apple's press conference, honor announced its ai agent strategy at the consumer electronics show in berlin, germany. the official said that this is the industry's first cross-application open ecological intelligent body released globally, and will be commercially available on the next generation of honor flagship phones. the above-mentioned honor person said, "in fact, users do not need to wait for apple to learn chinese, as chinese manufacturers have already started similar arrangements."
in their view, ai agent mobile phones need at least four capabilities, including "natural semantic understanding and computer vision", "user behavior habit learning and scene environment perception", "intention recognition and decision-making", and "in-application and cross-application operation". in other words, only based on the understanding of user habits and the environmental perception of the current usage scenario can it understand user needs and respond quickly, execute and call various mobile phone resources to meet user needs.
in addition to honor, oppo and vivo will also update their ai progress in the second half of the year. previously, the two companies have successively released their own ai models.
vivo insiders told reporters that chinese mobile phone manufacturers have been exploring the application scenarios of ai on mobile phones, covering language big models, voice big models, vision big models, multimodal big models, and end-side big models. huawei's "pure-blooded hongmeng" is planned to be launched in the fourth quarter, and the application of pangu ai on mobile phones is also one of the focuses of this operating system.
at this point, domestic mobile phones will be updated to varying degrees before the arrival of apple's ai intelligent body. analysts believe that this will also boost the overall sales of ai phones.
according to data provided by counterpoint, the penetration rate of generative ai mobile phones will reach about 43% in 2027. at the same time, the stock size of generative ai mobile phones will increase from only millions in 2023 to 1.23 billion in 2027. idc believes that the shipment of ai mobile phones in the chinese market this year will reach 37 million units. the global shipment of ai mobile phones will reach 170 million units, accounting for about 15% of the total shipment of smartphones.
ai mobile phone reaches a critical point
according to the analysis data of major institutions, ai mobile phones released this year are mainly concentrated in high-end models, and mid-range smartphones will gradually adopt them later. it is expected that by 2028, the proportion of ai mobile phones will surge to 54%. this growth is mainly driven by consumers' demand for enhanced functions such as ai assistants and end-side processing.
"historically, apple has experienced two key super cycles: the first occurred in 2016, triggered by the large-screen iphone 6, and the second occurred in 2021, jointly stimulated by the combination strategy of the iphone 12/13 series and the rebound in demand after the epidemic." counterpoint believes that apple intelligence will drive the step-by-step growth of iphone revenue rather than a single super cycle. the agency predicts that by 2025, apple's service revenue will exceed us$100 billion, accounting for a quarter of its total revenue.
the reason why this predicted volume has been reached is that ai mobile phones have gone through multiple rounds of development.
when apple spent $200 million to buy the company that founded siri in 2010, investment in ai phones had already begun. however, it was not until seven or eight years later that the "label" of ai phones was widely used by mobile phone manufacturers around the world.
in 2018, huawei announced that smartphones will move from the "app era to the ai era". honor, which carries huawei's ai strategy, has begun testing the waters on the "magic" phone model, and xu zhijun, a member of the executive team, is personally supervising the battle. at that time, samsung also told the story of "visual recognition" on the bixby platform on the galaxy s9. executives of other domestic mobile phone manufacturers are frequently visiting scientific research institutions in the united states while also planning their own ai strategic plans.
at this time, ai functions on mobile phones were mainly concentrated in the app stage. for example, various ai assistants, translation software, and image processing applications all provide ai functions through independent apps. these apps have different focuses, and users need to install and open different apps to use ai services. with the development of ai technology, ai mobile phones have entered the integration stage, and ai functions have begun to be integrated into other applications on mobile phones. for example, ai beauty in camera applications, natural language processing functions in chat tools, etc.
at present, the development of ai mobile phones has entered the system-native stage. ai technology is integrated into the native system of mobile phones, that is, the underlying hardware and software, and users obtain services through voice, text, etc. multimodal ai technology improves the ability of mobile phones to understand and process user needs, enhances the naturalness and accuracy of interaction, and is more intelligent than the previous two stages.
for example, apple demonstrated the combination of ai and apple's voice assistant siri at the press conference. apple said that siri can now extract information from users' text messages, and can also let siri send specified photos to specified users.
currently, this feature of apple can be realized on android phones. "apple's built-in siri can only mobilize the functions of apple's native applications, and cannot be linked with third-party applications such as wechat and alipay. we can already activate the intelligent body through voice commands, and realize functions such as one-click closing of automatic renewal, one-click ordering of drinks, and one-click travel planning and ticket booking." honor staff told reporters.
the head of another domestic mobile phone manufacturer also told reporters that with the application of ai in mobile phones, mobile phones can begin to actively observe and understand users, predict weather conditions for business trips based on authorized positioning, and provide reminders. "mobile phones used to be just a tool, but in the future they will become an external brain, no longer just a hardware state."
"in terms of functions, first of all, at the human-computer interaction level, with the support of llm, the new multimodal interaction will replace the traditional, single touch screen interaction, and gradually realize the leap from graphical user interface gui to voice user interface vui. secondly, the combination of multimodal input and output capabilities can greatly enhance the productivity tool attributes of smartphones. it can generate charts, texts, music, pictures and even videos required by users based on various forms of input information, and can also edit the input pictures and videos." archie, a smartphone industry analyst at counterpoint research, told reporters that these new technological changes can gradually stimulate new demands in the consumer market.
however, it can also be seen that the ai chip reasoning capabilities at the hardware level and the interaction methods at the software level have put forward higher requirements on the ai mobile phone itself, and the cost and technical solutions have reached a critical point.
honor ceo zhao ming told reporters that the era of massive ai applications is coming. just like the current mobile app ecosystem, ai applications will also show a fragmented trend. what path to use to call these capabilities on the terminal side is also the biggest challenge facing major mobile phone manufacturers in the future.
"for consumers, ai in the future will mean a variety of services. to some extent, it is no different in essence from today's apps. but for mobile phone manufacturers, if they just package it at this level and claim it as their own, it is not a true ai phone. i believe that the ai solution released by apple will make the entire industry more sober. is the third-party ai translation yours? in fact, mobile phone manufacturers have only integrated it, which cannot be said to be an ai phone." zhao ming told reporters that a true ai phone should have the underlying interaction logic of "building human-computer interaction for intent recognition", which is the core of edge ai.
as for the cost issue of ai mobile phones, chen yiqiang, deputy general manager of the wireless communications business unit of chip maker mediatek, told reporters that the cost of hardware will decrease with scale, but in the short term, due to the insufficient scale of applications, there is actually no way to form an effective business model based on ai. only when the platform can scale it up, will the price of hardware come down.
"we optimize computing power by at least 30%-50% at different costs every year, but the cost will not be reduced much in the short term," said chen yiqiang.
neil shah, partner and vice president of research at market research firm counterpoint research, told reporters that integrating intelligence into mobile devices is an expensive investment, requiring a lot of r&d resources to develop and train models with billions of parameters. neil shah predicts that the commercialization of ai will become an inevitable trend, and in the future, mobile phone manufacturers including apple may adopt a variety of business models to meet this challenge, such as charging developers or asking users to pay for using advanced ai features.
(this article comes from china business network)
