New knowledge of culture and technology.Almost overnight, the iPhone 15 in people's hands became the product of the previous era.
Just after the Spring Festival holiday, OPPO issued an internal letter saying that the first year of AI mobile phones has arrived, and the impact of AI on the mobile phone industry can be compared to the replacement of feature phones by smartphones that year; It also announced plans to invest 100 billion yuan in AI technology research and development in the next five years.
OPPO, which cut the chip team Zheku last year, now has an annual net profit of only 10 billion.
Meizu, which is more radical than OPPO, shouted the slogan "AI for new generations", and even cut itself off and stopped the follow-up products of traditional smartphone projects to show its determination.
In addition to these two public statements, other mobile phone manufacturers are more or less also deploying AI. Large models are on the machine, AI applications are prevalent, AI chip iteration and so on.
In the strategy of Texas Hold'em, it is generally when the chips are not enough or the hand is extremely firm, and the choice is all in, winning as the boss, and losing.
As the entire mobile phone market struggles to break through the ceiling, every manufacturer wants to be that successful revolutionary. But it is worth speculating that the current wave of AI in the mobile phone industry will be as earth-shaking as the replacement of feature phones by smartphones? What aspects should AI mobile phone manufacturers start with?
AI comes as standard
AI mobile phones have not yet come out with a real machine in a complete sense, and it has become the standard configuration of new products launched by manufacturers.
Looking back on the past, it can be found that domestic mobile phone manufacturers are the most active practitioners of the innovation and application of new technologies, and the same is true for AI mobile phones.
Huawei's mobile phones return to 5G's Mate 60 series, equipped with a Pangu model and a new generation of NPU, which can directly use Huawei's cloud AI model to enhance the image quality of the phone's local **; Xiaomi's self-developed 6 billion parameter-level large model can be used offline on mobile phones for the functions of "Wensheng Diagram", "AI Expanded Map" and "AI Removal of Passers-by".
At the end of last year, vivo and OPPO also successively launched self-developed blue heart models and Andean large models, seamlessly connecting with various ecosystems; MagicOS 80 is also based on platform-level AI, bringing many new highlights; Not long ago, Samsung's flagship model Galaxy S24 series was just released, and the first AI processing, AI chatbot, image processing, real-time call translation and other functions as the focus of publicity.
For manufacturers, it is difficult to tell the story of traditional smartphones, and in today's market, it is embarrassing to come up with new products without mentioning AI. The rush to put the large model into the mobile phone is also to inject a shot in the arm into the market that has been dormant for a long time.
Starting from the AIGC application, various small functions such as Wensheng diagram, beautification, and translation have become the top priority of various brands to integrate AI with mobile phones.
Among the reasons, in addition to the urgent need to break through the bottleneck in the mobile phone market, the rapid progress of AI technology in 2023 is also one of the conditions. With the performance of ChatGPT, Midjourney, Stable Diffusion and other Wensheng texts, Wensheng diagrams, and SORA Wensheng** various large-scale model products, the major manufacturers have further confirmed that AI technology can indeed provide productivity, and it will be greatly developed and applied in the next few years, and it will be too late to get on the car.
According to market tracker Counterpoint Research, the market share of generative AI smartphones will account for 4% in 2023 and is expected to double in 2024. IDC also predicts that the global shipment of a new generation of AI mobile phones will reach 1700 million units, accounting for 15% of overall smartphone shipments. The share of AI mobile phones in the Chinese market will also grow rapidly, accounting for more than 50% by 2027.
For major mobile phone manufacturers, this is a key battle for future survival, and it is also a tough battle that cannot afford to lose. In addition to the killer application winner-takes-all, AI technology will also change the experience logic of existing applications, which may be the dawn of technology to help smartphones complete leapfrog innovation. Mobile terminals, which have hundreds of millions of users, are not only a natural entrance, but also can accelerate the popularization of AI applications.
Smartphones have been rapidly iterating in the course of more than ten years of development, and various brands have rolled up the processor and then rolled up the screen, image, and memory, which has finally led to the difficulty of eye-catching disruptive innovation in recent years.
This wave of AI mobile phones may be a timely rain with rain and dew. However, it is foreseeable that the hardware foundation of each manufacturer's products is still determined by the upstream ** chain, and the construction of software ecology will be the main point of contention. What kind of capabilities should the AI mobile phone born along this line of thought provide?
A new round of technology competition
In line with the logic of replacing feature phones with smartphones, if AI phones want to replace smartphones, hardware and software need to go hand in hand.
From the perspective of user privacy, data service fees, response speed and other aspects, if the AI application on the mobile phone can only rely on cloud services, it is obviously not friendly to consumers. Therefore, at this stage, AI mobile phones must at least run locally.
In October 2023, Qualcomm released the Snapdragon 8 Gen 3, the first mobile platform processor built specifically for generative AI, which is equivalent to giving all Android phone manufacturers a ticket to AI phones; MediaTek's latest Dimensity 9300 processor has 2 times higher integer and floating-point performance than the previous generation, and the operator acceleration performance has reached 8 times that of the previous generation.
In a certain sense, mobile phone products equipped with Qualcomm or MediaTek's latest flagship chips are equivalent to stepping into the threshold of AI mobile phones with one foot, which is not difficult for mobile phone manufacturers.
Of course, to make AI mobile phones, we must not only rely on stacking materials, but also do collaborative innovation of software and hardware, and optimize devices in combination with the new trend of small-scale parameter level models on the device side.
However, from past experience, only staying in the optimization of performance and UI cannot build competitive barriers for mobile phone brands, and users may not pay for it. The key is to see whether the landing function of AI technology is enough to attract consumers, whether it is supported by strong enough large-scale model products, and whether it can bring users practical scenario applications and stand up to verification.
Therefore, the only room left for mobile phone manufacturers to innovate is the OS system and software and hardware ecology.
Huawei's Hongmeng, Xiaomi's surging OS, and vivo's Blue River operating system are all laying the foundation for AI OS. More directly, Meizu announced its AI strategic plan, including building AI device products, reconstructing the Flyme system, and building an AI ecosystem.
In fact, for mobile phone manufacturers that have survived so far, it is not difficult to make an AI function or an OS system alone. However, to make AI OS, it is necessary to seriously consider the chain of problems such as self-developed large models, AI native frameworks, and new ecological applications.
The most important thing is the construction of hardware terminals. The development of AI technology may seem to be in full swing, but to effectively integrate it into operating systems and hardware devices, and create products that meet both current needs and have future development potential, requires not only technological innovation, but also a deep understanding of market needs and the ability to respond quickly. For example, which devices need to be AI, how to integrate with AI, and so on.
Judging from the current situation, Xiaomi's surging OS focuses on "people, vehicles, and homes" collaboration, and the Mi ecosystem also has enough terminal IoT devices; Huawei's HarmonyOS launched the "1+8+N" strategy in 2019. What needs to be urgent is vivo and OPPO, although the number of mobile phone shipments is considerable, but it has been in a backward position in ecological construction; Meizu relied on Geely to start in the corner, but it is difficult to overtake overnight in the field of competition technology accumulation rather than financial investment.
More realistically, whether AI mobile phones or AI OS can be successfully commercialized and bring a qualitative leap to the industry depends on whether users are willing to pay for their so-called disruptive features, just like the iPhone 4 that came out in 2011, completely changing the habits of using mobile phones.
AI Phone is still uncertain
The subversion of smart phones to feature phones is devastating, and the resulting mobile Internet revolution has had a significant impact on human production and life.
The difference between a feature phone and a smartphone can be seen at a glance from the size of the screen. AI already has the meaning of intelligence, and if it is still called a smartphone, it will not be disobedient.
In contrast, the foreseeable AI mobile phones only make the functional experience of smartphones more convenient and efficient, and the rapid replacement of feature phones by smartphones has fundamentally redefined products such as mobile phones.
If a smartphone goes from 0 to 1, then an AI phone is just from 1 to 10.
Qualcomm CEO Cristiano Amon mentioned in an interview that AI phones will be more responsive and enable deeper interactions. For example, 0An image can be generated in 6 seconds, and generative AI mobilizes all applications to work together.
In other words, in the future, people can do without AI phones, but not without smartphones. This means that the replacement of smart phones by AI mobile phones is obviously not as urgent and rigid as the replacement of feature phones by smart phones.
It is worth mentioning that whether the public on the consumer side is willing to pay for AI is also a question.
According to public information on Samsung's official website, Samsung plans to provide the Galaxy AI feature for free on supported devices by the end of 2025, and will charge for the AI feature in the future. This means that users have to pay for the functions of AI phones after buying huge sums of money, which is obviously not in line with the consumption habits of Chinese users. Even if the model works, the continuous investment in computing power and program upgrades of AI functions requires manufacturers to balance the trade-off between cost and profit.
Every technology or product outlet will inevitably have elements that are hyped and exaggerated.
At present, domestic manufacturers are fully pursuing AI mobile phones, led by OPPO and Meizu. What the two companies have in common is that they do not have the same mature IoT ecosystem as Xiaomi and Huawei, and the deployment of AI is not limited to mobile phones, but needs to be extended to a wider range of terminals such as PCs, wearables, homes, and even cars.
For manufacturers chasing the wind, AI mobile phones do a good job, which can radiate to ecological construction and even turn the tables against the wind, but it is also necessary to jump out of the logic of making mobile phones, look further, and make a fuss about mobile phone functions alone, which is obviously not a future-oriented approach.
It is foreseeable that the AI wind blowing into the mobile phone industry will be the beginning of a market revolution. In this process, the competition for technology iteration will be more intense, and it will be easier to see who is the "naked swimmer".