On the first day of construction in 2024, the mobile phone circle is particularly lively.
First of all, Meizu Technology stated that it was all in AI and stopped new projects of traditional smartphones, and then Chen Mingyong, founder of OPPO, issued an internal letter saying that 2024 is the first year of AI mobile phones, and AI mobile phones will become the third stage after feature phones and smart phones; Xiaomi, which has been warming up for a long time, also announced that its own image flagship Xiaomi 14Ultra is set for February 22, and will join forces with SU7 to create the first people, cars, and home full ecological press conference, counting the previous Samsung S24 series and Honor MagicOS 80, in 2024, AI mobile phones have become the main theme of the industry. So the question is, what exactly is an AI phone?
From concept hype to functional perfection, AI has gradually become the main front
Strictly speaking, there is no specific quantitative standard for AI mobile phones in the current market. Although as early as 2015, AI-based APP preloading and smart album classification functions appeared on mobile phone products; In 2017, AI chips, AI dual cameras, and NPU with integrated neural processing units have also been implemented, but the chips, algorithms, systems, and cloud service modules required by AI mobile phones are lacking.
In addition to the basic AI functions, manufacturers have expanded AI functions around performance, games, images, and fast charging tracks (such as the V1 image chip of the blue factory, Xiaomi's surging G1 battery management chip, etc.). These auxiliary tools developed based on AI computing power can smoothly perform complex tasks, continuously learn and optimize user habits, and launch intimate services in a targeted manner, covering speech recognition, natural language processing, autonomous driving, and intelligent recommendation.
From the initial software underlying application, APP function beautification to the optimization of the underlying system architecture, AI integrates fragmented demand services to make smartphone interaction more humane and more flexible in application.
The Huawei Mate60 series, which will be launched at the end of August 2023, has a built-in Pangu model, which can directly use the AI model in the HUAWEI cloud to add color to the local**, and even significantly improve** the clarity to 8K visual effects; On September 1, at the Guangdong-Hong Kong-Macao Greater Bay Area Artificial Intelligence Industry Development Technology Forum of "Gathering the Bay Area and Discussing the Future", Xiaomi stated that its self-developed 6 billion parameter large model can not only achieve offline Wensheng diagrams and AI to remove passers-by, but also be built into the voice assistant in the future (Xiao Ai passed the large model filing on January 31) and promote related development tools and ecological construction. The recently launched OPPO Find X7 series, Honor Magic6, and Vivo X100 series have all supported 7 billion parameter large models.
Although in mid-2023, Samsung stated that it would build the world's first AI large model mobile phone, and announced its self-developed Gauss large model and three AI applications of Language, Code, and Image at the end of 2023. But the pace and determination of domestic manufacturers are stronger than Samsung. Interestingly, Apple, which has always lagged behind, also stated at the annual earnings conference that it is developing generative artificial intelligence technology, which will change Siri, message and other application functions, and even add large language models to Siri to achieve the deep integration of Siri and shortcut applications, and vigorously promote AI function mining.
Obviously, on February 18, OPPO Chen Mingyong and Meizu Technology's statement is not empty words, and it is an industry consensus to embrace AI at this stage, and the upcoming Xiaomi 14Ultra and Huawei P70 series will bring you new answers in terms of AI functions, AI processing capabilities, and AI integration optimization.
So, what are the trade-offs of the manufacturers behind betting on AI, and can Dongfeng manufacturers change the status quo of the industry with the help of AI?
Seizing the AI highland is expected to give rise to a new market pattern
In my opinion, there are three main considerations for leading manufacturers to vigorously enter the AI mobile phone market. First of all, at this stage, the scale of smart phone screens, images, fast charging, and storage hardware has reached a bottleneck, and the transparent first-chain system makes everyone basically unable to open the strength gap, and the fast-paced new frequency is easy to make consumers aesthetically fatigued and reduce their desire to change to a new one. As a new outlet in the industry, manufacturers need to use AI to expand the track, stimulate user consumption and increase the market.
Secondly, AI mobile phones have great market prospects. On January 23, market research agency Canalys issued a document saying that 5% of smartphone shipments in 2024 will be AI mobile phones, and this proportion will rise to 45% in 2027, referring to the global smartphone shipment volume in 2023, and the global market AI mobile phone sales are expected to reach 60 million units in 2024. In other words, there are 60 million potential consumer groups waiting for manufacturers to compete under the stimulation of AI, and once the pace of their own brand products lags behind or there are technical shortcomings, it is easy to cause the loss of brand users.
Finally, there is the issue of the market structure that we have always talked about. In the era of smart phones, many niche brands have withdrawn from the market due to market pressure, and it is the norm to stay with these manufacturers to fight closely, in 2024, Apple will not only win the first place in the global market and the domestic market, but also dominate the high-end market with more than 70% market share, coupled with the disruption brought about by Huawei's strong return, manufacturers including Samsung need sales to stimulate the market, to some extent, the sales data is linked to the brand characteristics, and when you sell well, it is easy to stimulate the herd effect. Of course, consumers who uphold the principle of "I can do without you can't do without" have played the role of account managers in recent years, and once the software and hardware functions of the manufacturer's products cannot meet the needs of users, it is easy for everyone to switch to the camp.
As for whether AI mobile phones can change the industry pattern, I think the international market has not changed much, and Huawei is expected to sit in the top five in the domestic market. The reason is easy to understand,Domestic manufacturers generally lack a SoC that can break through the upper limit of performance,At this stage3nm process mobile phone processor does not have a specific AI-specific unit,NPU and TPU This kind of AI task performance also depends on a unified running frequency cache memory bandwidth to improve。 Although HiSilicon Kirin has returned to 5G, the process technology and public architecture are limited, and the extent to which domestic manufacturers can develop depends on Qualcomm and MediaTek's face.
At the same time, at present, AI mobile phones still lack mature application scenarios, even the Samsung S24 series only stays at real-time translation, recording and transcription, AI frame insertion and other basic functions, it is difficult for consumers to feel the differentiated functions brought by AI to mobile phones, and it is difficult to impress consumers with AI in the short term. In other words, Meizu technology, which is all in AI, may be the first to go out under the new outlet, after all, BYD's market effect is not something that anyone can copy.