Since the 5G cycle, innovation in the smartphone industry has been dead, and it has been running like a headless fly for at least five years. AI is expected to be the life-saving straw of the game changer.

According to the latest news, Apple plans to hold a fall new product launch on September 10, 2024. From the previous preheating, it is not difficult to know that the focus of Apple's launch conference will be AI, and how to define AI mobile phones will be the highlight. It is understood that IOS will usher in the biggest update in 10 years.

As the biggest change point in the past five years, the promotion of AI mobile phones is bound to be overwhelming. The overwhelming traffic will urge the industry chain to imagine more from an optimistic perspective: AI will start a new round of replacement cycle.

But we seem to have missed out on thinking about the overall situation, especially the fate of domestic mobile phone brands: will the industry prosper, or will it be a drama where newcomers laugh while old ones cry?

01 The difficulties we face are greater than we imagined

Compared with mobile phones three or four years ago, the performance of mobile phones from major brands has not improved significantly. Most people change their mobile phones not because they are looking for better products, but because their old phones are no longer usable.

No matter what manufacturers brag, there are no secrets in the industry without technological innovation, so everyone is at the same level.

Except for Apple, user loyalty of all other mobile phone brands is very low. Ultimately, the industry is always trapped in a state of infinite involution. Choosing a mobile phone has become choosing an air conditioner. Among a bunch of parameters that I don’t care about, I finally choose to look at the price.

2021 is the last peak of mobile phone manufacturers. Take Xiaomi Group as an example. Its total revenue has fluctuated between 200 billion yuan and 300 billion yuan in the past five years. The main growth still depends on non-mobile phone IOT. The shipment volume of pure mobile phones has remained in the range of 150 million to 200 million units for many years.

OPPO, VIVO, and Honor, although they are not listed, their performance in the past three years is only slightly worse than that of Xiaomi, which is still quite diligent.

Source: Company financial reports

1) The dream of total growth is shattered

Before AI brought about expected innovation, the actual difficulties faced by domestic mobile phone manufacturers were actually greater than what was reflected in the financial data.

As early as 2016, global smartphone shipments hit a historic peak of 1.47 billion units, and then began a downward trend. During the seven-year decline, only Apple, relying on the sanctions against Huawei, partially proved its brand value and moat.

All other domestic brands are just a market share game of one rising while the other falling. Maybe a certain model sells well, and the market share of this brand can rise a little in the short term, but it will be copied by competitors and stab you in the back with a lower price.

There are actually not many technical secrets between brands. What may not be widely known is that among the different models of Android phones, more than half rely on the same OEM factory to complete all research and development and design, which is a bit like you eat the same pre-prepared dishes in different restaurants.

Source: IDC

2) The end of structural innovation

Why didn’t the mobile phone industry feel so miserable after the total volume stopped growing? This is because after realizing that the total volume was not growing, a term began to become popular in the industry: micro-innovation. For example:

The manufacturing process of mobile phone chips has been upgraded from 14nm to the latest 3nm, which indirectly drove the prosperity of TSMC; mobile phone storage has been upgraded from 64G to 1T, boosting the performance of Hynix.

The improvement in mobile phone communication quality and the upgrade of signals from 4G to 5G have increased the demand for PCBs, RF devices, and inductors.

The inward-rolling screen and camera, the screen has been upgraded to a folding screen, and is further divided into up and down folding, left and right folding, and three folding. The camera is upgraded from a single camera to a dual camera, and finally to a drum washing machine.

In the past seven years of total volume decline, mobile phone brands relied on over-the-top materials, piled up functions that consumers did not need, and then forcibly raised the prices of mid- and high-end mobile phones, so consumers could only buy them with a pinch of their noses. It is undeniable that this process also brought structural opportunities to the industrial chain, such as the prosperity of the camera optical industry chain, but the model of mobile phone brands maintaining the growth of mobile phone sales revenue is a typical way of eating before the grain is earned.

Micro-innovation will be counterproductive if it is used to deceive consumers. For example, after the camera upgrade to a "drum washing machine", it seems to have lost its power. Xiaomi 14Ultra has caused quite a bit of controversy. It can be said that it is just a digital camera with a call function. The mobile phone industry has even seen the spectacle of patching up for another three years. The refurbishment of second-hand mobile phones was once very popular. The mobile phone industry has become no different from home appliances.

If there is no more innovation, we can confidently draw a seemingly scary but inevitable conclusion: the decline in the total number of mobile phones is a collective chronic suicide of the Android system.

02 Good news: AI is a revolution that truly touches the soul of mobile phones

1) Does AI need a mobile phone as a carrier?

There is endless debate about how AI should be commercialized. Even if it is combined with hardware, there are many carriers to choose from, such as cars, computers, glasses, mobile phones, speakers, and watches. Why must it stick to mobile phones?

Regardless of the various technical analyses, the debate has been ended as Apple and OpenAI embraced each other. At this year's Apple WWDC24 Developer Conference, Cook announced that he would cooperate with OpenAI, and it was a deep cooperation model.

The integration of ChatGPT, powered by GPT-4o, will be integrated into all Apple operating systems this year, including iOS, iPadOS, and macOS; ChatGPT subscribers can also connect their accounts and use paid features directly in the Apple system. The cooperation between Apple and OpenAI is a win-win situation for both sides: OpenAI lacks application scenarios, and Apple lacks high-quality large models.

The hardware system, operating system and ecosystem of mobile phones are already very mature. Implanting AI in mobile phones will directly face billions of consumers and will most easily form a closed loop of the business model.

In 2023, OpenAI's revenue is $2 billion, and in 2024 it is optimistically expected to be $5 billion, which is obviously not enough to support the company's huge expenses. Although it is okay to rely on Microsoft for blood transfusions, it is not a long-term solution after all. If AI is installed on mobile phones, it may have a fission effect.

The most comparable case is the cooperation between Google and Apple. In order to obtain the default search engine status on Apple's entire product line, Google is estimated to pay Apple $15 billion per year. As Apple has more than 1 billion users, Google has also benefited greatly from the tie-in. The search-related revenue obtained through Apple devices may be $50 billion. After deducting the $15 billion in fees, it can still generate more than $30 billion in revenue.

If OpenAI and Apple cooperate to produce similar results, it may directly increase its revenue by more than $10 billion. This speculation is not unreasonable. After announcing its cooperation with OpenAI, Apple said that it might charge a monthly rent of about $20 for Apple Intelligence's advanced AI features.

Let's do a simple calculation. If 10% of the 1.5 billion Apple users choose to pay, the annual turnover they contribute will reach 36 billion US dollars. If half of that is returned to OpenAI, it will indeed be a huge sum of tens of billions of dollars.

2) What changes can AI bring to mobile phones?

The business model works, and what the public is most concerned about is, in terms of user experience, what kind of changes can AI bring to mobile phones?

A superficial analysis may start with specific functions, such as AI phones will provide conference transcription, real-time special effects rendering, simultaneous interpretation, smart photo editing, email sending and receiving, etc., to improve user work efficiency and provide interesting and novel experiences. However, this interpretation is prone to fall into the fallacy of partial correctness, such as the blind man touching the elephant.

What is the essence of a mobile phone? We believe that the essence of a mobile phone is the interaction method. When smartphones replaced feature phones, the interaction method changed from buttons to touch. Apple's Siri failed because its voice interaction function was too weak.

The reason why AI touches the soul of mobile phones is that it will subvert the way mobile phones interact. From passive input interaction to active interactive interaction; the establishment of a multimodal large model makes this method all-round, including text, voice, and video. Your mobile phone may become your personal AI assistant in the future.

03 The bad news: AI is more like poison for domestic mobile phone manufacturers

From the second half of this year to the first half of next year, the penetration rate of AI mobile phones will try to challenge the key threshold of 10%. Once the judgment of the innovation of interaction mode is verified, it is estimated that within three years, the penetration rate of AI mobile phones will quickly reach 50%.

On paper, it seems that all mainstream mobile phone manufacturers will actively embrace this change, and everyone is on the same starting line. Each brand has launched its own terminal language model with comparable parameters and its own operating system.

But the reality is that the pressure faced by Android brands may be greater than the opportunities. If the difficulty of traditional competition is 100 points, Apple can get 100 points. In the past few years when the total volume of Android has declined, its product strength has risen from 60 points to 90 points; but AI will directly increase the difficulty of games to 200 points, and the gap between Android and Apple will be ruthlessly widened again.

Figure: Comparison of terminal model parameters of different AI mobile phone manufacturers Source: Everbright Securities

1) The real reason why Apple hit a new high

Historically, although Apple has invested a lot in AI, it made a serious mistake in terms of direction, that is, it did not care about the big model, thinking that the big model was only used for search and had no value to mobile phone manufacturers. The emergence of the intelligence of the big model made Apple realize that it had made a complete mistake.

In mid-2024, Apple will completely step out of the quagmire of car manufacturing and MR and choose to go all in on AI. By connecting to external large models such as OpenAI, Apple has completed the last link in the AI ​​puzzle and returned to the right direction of the industry.

Unlike other manufacturers who use AI as a gimmick and selling point, Apple's thinking on AI is to return mobile phones to the most natural form of interaction, rather than being limited to passive structured commands.

To achieve this goal, Apple has planned a 10-year vision, divided into three time periods:

Phase 1: The Siri stage of voice interaction. The iOS18 update at the end of this year is the largest update in the past 10 years. Developed using AI architecture, Siri will be built into an intelligent control center, providing mobile phones with intelligent human-computer interaction methods.

The second stage: using external models and Apple's own multimodal large models, relying on its own cloud hybrid computing power, using AI technology to solve problems that are difficult to handle with traditional computer languages, and achieving breakthroughs in the two modalities of image and text. It is expected that by the iPhone 18 series, AI technology will be fully introduced.

The third stage: In the next ten years, AI will become a real part of the operating system, with real intelligence to understand the world and user needs.

In addition to the completely different strategic planning, Apple also has more cards to play in terms of technical feasibility.

While enjoying the low-power ARM ecosystem, Apple can also use its own IP to the greatest extent possible, combined with its own operating system and the most advanced manufacturing process. In theory, Apple's hardware advantage is still far ahead of its competitors. Therefore, in theory, among the terminal models, Apple can support the largest model, which is estimated to be 50 billion parameters.

Since AI requires more payment, it is bound to start penetrating from high-end phones. In addition, Apple accounts for 20% of the mobile phone market share and is the most high-quality customer willing to pay a premium for high functions.

Apple defines AI as an innovation in interaction methods, and it has the greatest probability of success, as well as a bright prospect of opening up business models as long as it succeeds. These are the real reasons that support Apple's market value to hit an all-time high.

2) Does Android have the strength to compete? The gap is widening

Faced with the surging tide of AI mobile phones, although every mobile phone manufacturer has declared that they attach great importance to AI, objectively speaking, due to backward performance and strategic contempt, Xiaomi and other companies are not on the same starting line as Apple in the field of AI.

Different from Apple's full-family bucket strategy, Android relies on the power of the masses, with Qualcomm/MediaTek responsible for chips, Google responsible for operating systems, and brand manufacturer OVM responsible for products. Each of these links has a visible gap with Apple, and once multiplied, it becomes a product of two dimensions.

If the performance gap can ultimately be made up with time, then the differences in strategic choices will easily lead to a situation where the harder you try, the more you fail.

Among them, Xiaomi has the biggest strategic deviation. While Apple has already anchored the innovation of interactive models, Xiaomi is still emphasizing the "full ecosystem of people, cars and homes" strategy in its latest financial report, as if it is stuck in the "1+N" strategy of the previous era.

Source: Xiaomi Group 2024 Interim Report

The discussion on AI also reflects the company's understanding, which currently remains at the "empowerment" level. Xiaomi still regards AI as a bonus, seeing it as an additional function of mobile phones, cars, and IOT, and does not mention innovation in interaction methods at all.

Source: Xiaomi Group 2024 Interim Report

It has to be admitted that Xiaomi's supply chain capabilities are absolutely world-class. The first car it produced in only three years not only sold well, but also had a gross profit margin of more than 10%, which slapped the faces of a number of new car-making forces. Even looking across the country, only BYD's supply chain capabilities can compete with it.

At that time, Lei Jun chose to build cars because he believed that cars were the most advanced technology and would have a huge siphon effect on mobile phones. If Xiaomi did not build cars, it would have no future. However, with the establishment of AI multimodal big models, the most advanced technology has now become AI. Apple's decision to completely stop its car-building project and embrace AI is an example.

So in 2024, Apple will present Apple Intelligence, while Xiaomi will still focus on people and cars. Apple wants to grasp the revolution of interaction methods, while Xiaomi wants to expand application scenarios and define AI as empowerment.

Due to the great success of Xiaomi SU7, we are also optimistic about the competitiveness of Xiaomi Automobile. For Xiaomi, investing in automobiles will definitely involve too much energy. It will not be so easy to turn around in the future.

If Xiaomi is not able to cope with the challenge, then OPPO, VIVO and Honor may not fare much better. In addition to not having as close a relationship with Qualcomm as Xiaomi, these brands are not listed and have cross-holdings with the supply chain, which leads them to pay more attention to cash flow and profits, which will only constrain their R&D investment.

Someone will definitely mention Transsion, which is completely misaligned in competition. It makes low-end mobile phones with an average price of less than $150. Can it bring unexpected surprises to everyone with "inclusive AI"? We suggest not to have too high expectations. For Transsion, let's not talk about the feasibility of the technology. Since AI is bound to increase the daily use cost of mobile phones, it will take at least five years for it to penetrate into low-end mobile phones.

If we say that in the past seven years, Xiaomi and others were able to maintain their market share through cost-effectiveness, channels, marketing and other means, as the growth rate of mobile phone sales declined, they wrote the song of domestic mobile phones selling well around the world. But in the next 10 years, AI has awakened the sleeping lion of Apple.