Chinese mobile phone manufacturers should stop focusing on chips

Recently, the news that Chinese mobile phone manufacturer OPPO is going to develop its own mobile phone chips (OPPO internal code name: Mariana) has excited many people in the industry. However, OPPO seems quite rational and official, and did not directly respond to the rumor, but only said that OPPO's core strategy is to make good products, and any R&D investment is to enhance product competitiveness and improve user experience.

Once upon a time, the success of Huawei's own chips not only made Huawei famous and wealthy in the smartphone market, but also made many domestic media look forward to whether manufacturers such as Xiaomi, OV and other manufacturers would be able to have independent mobile phone chips like Huawei in the future.

Faced with these expectations, Xiaomi and OV have always released news about building their own mobile phone chips on different occasions. Xiaomi even used the acquired Pinecone Technology in 2014 to release a mobile phone chip called Surge S1, which was applied to the Xiaomi 5c, a thousand-yuan phone. However, it ultimately failed due to poor results and subsequent capital investment.

Today in 2020, the industry has mentioned this topic again because of the exposure of OPPO's "Mariana" plan. So, compared with before, what are the chances of these Chinese mobile phone manufacturers making their own mobile phone chips?

Before we discuss, it is necessary to clarify the concept of mobile phone chips. Due to the complexity and integration of mobile phone chips, there are many types of so-called chips involved, but generally speaking, a mobile phone chip (SoC, System on Chip) is mainly composed of two core parts, namely the application processor chip AP (Application Processor) and the baseband chip BP (Baseband Processor). The former is mainly used to process various applications and data calculations in the mobile phone, while the latter is mainly used to realize the phone and data Internet functions of the mobile phone. The 4G and 5G networks mentioned daily are related to baseband chips. When the industry discusses mobile phone chips, it mainly refers to these two core parts, so our discussion is also mainly focused on this.

AP: Technological iteration is like rowing against the current. If you slack off, you will fall back a thousand miles.

As we all know, the improvement of chip performance, the number of transistors, and the reduction of power consumption/heat all rely on the improvement of process technology. These factors are directly related to the overall performance and user experience of mobile phones. Therefore, in recent years, mobile phone manufacturers have been competing to adopt newer chip process technology, which has accelerated the evolution cycle of process technology.

Take TSMC, which is now leading in process technology, for example. In 2011, it was still mainly based on the 28nm process, and it took 4 years to enter the 16nm process. But since then, the process technology upgrade cycle has been shortened, especially with the intensification of competition in the smartphone market. Due to competition and the need to continuously improve user experience, although the difficulty of process technology has increased, the upgrade cycle is getting shorter. For example, from 10nm to 7nm, then to 7nm+, and even this year's 5nm, the process is upgraded by one generation almost every year. As a result, the performance and cost of mobile phone chip manufacturers have increased.

Let's first look at the performance of mobile phone chips. Tech Centurion, a well-known foreign evaluation website, recently released the latest 2020 mobile processor performance ladder list. It should be noted that in order to test and rank processors more objectively, Tech Centurion has specially developed a new set of standards to evaluate processor performance. In addition to the original parameter information, the new standard will also take into account the actual performance of the processor, which improves data accuracy while helping consumers better understand the performance of the phone and is more objective.

From this ranking, it is not difficult to find that the top-ranked mobile phone chips, whether they are general mobile phone chip manufacturers (such as Qualcomm, MediaTek) or mobile phone manufacturers with independent mobile phone chips (such as Apple, Samsung, and Huawei), all use the most advanced 7nm+ process technology.

At the same time, we also see that many mobile phone chips that were previously launched were among the best in terms of performance last year, but in just one year, they are already ranked outside the top five, reflecting the fact that the current competition in mobile phone chip performance is like sailing against the current, and a slack in the oars will lead to a thousand miles of retreat. It should be noted here that because Apple has the support of its own iOS system, it can better design chips and maximize the performance of processors through the combination of software and hardware. This makes the performance of Apple mobile phone chips always in a leading position under the same process, which is an exception among mobile phone manufacturers.

The fierce competition in performance is accompanied by a substantial increase in costs. According to statistics, after the chip manufacturing process enters 28nm, the cost of chip tape-out increases exponentially, that is, the cost of a 28nm chip is only 2-3 million US dollars, and the cost of a 16nm chip rises to about 10 million US dollars. Based on the current investment, a 10nm chip may cost up to 130 million US dollars. When it comes to the 7nm process, the tape-out cost is more likely to exceed 300 million US dollars.

To more intuitively illustrate the increase in R&D costs for mobile phone manufacturers due to mobile phone chips, let's take Huawei as an example. Ten years ago, HiSilicon gave an average overall R&D cost of only 40 million yuan for each commercial chip. However, the Kirin 990 flagship chip released last year cost as much as 30 million US dollars for tape-out alone, which is about five times the cost of the entire chip R&D ten years ago.

Let's look at Apple. Historically, it has not been keen on developing the latest technology. Apple would observe other companies' progress in new technology and product defects, and then seize the best opportunity to launch products with first-class user experience and occupy the market. However, since 2016, Apple's R&D expenses have increased significantly, allegedly to expand chip and sensor research and development. It can be seen that even Apple dare not neglect its investment in the field of mobile phone chips, especially in the face of Qualcomm and Huawei's accelerated iteration of high-end mobile phone chips.

In summary, the current mobile phone chip market is no longer what it used to be. It faces the threshold of high technology and high investment. In addition, the acceleration of future iterations will inevitably further intensify the competition in technology and investment.

In addition, we also found that even if the mobile phone chips have the highest performance, their actual performance in the smartphone market, especially in the high-end flagship market, is often very different. This undoubtedly adds uncertainties in other technologies and markets besides the performance of the mobile phone chips themselves, thereby increasing additional risks.

In this regard, Xiaomi and OV should be extremely cautious compared with Apple, Samsung and Huawei.

BP: It takes time to achieve success. Technological accumulation and innovation are the key

As mentioned above, in addition to AP, BP (baseband) is equally important in mobile phone SoC, and even exceeds AP in determining the user experience of smartphones. Therefore, BP has been the top priority of competition from the beginning. However, with the iteration of communication technology, the required technical accumulation, threshold and investment are getting higher and higher. Before the advent of 5G, companies such as Texas Instruments and NVIDIA, which are also current chip giants, have withdrawn from the mobile phone chip market due to BP. In the 5G era, this trend has only increased.

The most typical example is Intel, the chip industry giant that entered the baseband market for nearly 10 years through the acquisition of Infineon. After spending tens of billions of dollars, it finally sold its baseband business to Apple for US$1 billion last year, leaving the market in disgrace just as the 5G curtain opened.

Speaking of Apple, the world's most financially wealthy tech giant, it also settled its patent licensing lawsuit with Qualcomm through a settlement due to the baseband problem. Importantly, the latest disclosed settlement agreement shows that Apple will have to purchase Qualcomm's 5G baseband for at least the next four years. Among them, the X55 baseband will be used from June 1, 2020 to May 31, 2021, the X60 will be used from June 1, 2021 to May 31, 2022, and the X65 or X70 will be used from June 1, 2022 to May 31, 2024.

It can be seen from this that even if Apple acquires Intel's baseband business and it is rumored that Apple has long intended to develop its own baseband, it will be difficult for Apple to break away from Qualcomm's baseband strength in the next three years at least. The high technical threshold is far from being solved by huge capital investment.

In fact, looking at the current 5G baseband market, as mobile phone manufacturers, only Huawei and Samsung have the strength to produce their own 5G baseband, which is closely related to their accumulation and continuous innovation in the communication field over the years. In this regard, some people in the industry will say that OV has caught up in 5G, proving that OV has the technical confidence to make 5G basebands and can even overtake others. Is this really the case?

According to the industry's search and analysis of data merging using INPADOC patent families, OV has a huge gap with Huawei and Samsung in terms of 5G patents. For example, in terms of the number of authorized patents, which measures the value of patents, Huawei and Samsung have 1,337 and 1,746 respectively. In comparison, OPPO has only 36 and vivo has 0. In terms of approved 5G standard patents, Huawei has 5,855, Samsung has 1,239, while OPPO has only 190 and vivo has only 159. In fact, not to mention compared with Huawei and Samsung, even compared with Intel, which has already withdrawn from the 5G baseband market, OV is at an absolute disadvantage.

Overall, OV is only at the middle and lower levels in terms of 5G-related technologies and patents. With this technical support and considerable investment, it is extremely risky to rashly enter this market.

Market: Economic costs and positioning difficulties need to be carefully considered

In addition to the above technical factors, there is a misunderstanding in many media reports that mobile phone manufacturers can reduce costs by making mobile phone chips themselves. But in fact, this view is one-sided. Admittedly, if the shipment volume of a mobile phone chip is large enough (such as Samsung, Huawei, and Apple), vertical integration can indeed reduce costs. But if the shipment volume is relatively small, the cost of using third-party Qualcomm or MTK mobile phone chips will be lower - because the cost of mobile phone chips is mainly spread by production, and the greater the production, the lower the cost.

According to Canalys' statistics on global smartphone shipments last year, Samsung, Huawei and Apple all shipped close to and exceeded 200 million units, with Samsung shipping close to 300 million units.

In addition to Apple using its own mobile phone chips, in the third quarter of last year, according to IHS Markit statistics, Samsung used its own Exynos chips in about 80.4% of its mid-range smartphones, up from 64.2% in 2018. Among the entire smartphone product line, 75.4% of smartphones are equipped with its own Exynos chips, up from 61.4% in the same period of 2018. Huawei used its own Kirin series chips in 74.6%, which was also an increase from 68.7% a year ago.

In comparison, the shipments of Xiaomi and OV are still at the level of about 100 million units. Even if Xiaomi and OV release their own chips now, their costs will definitely not be as high as Samsung, Huawei and Apple under the premise of using the same proportion of their own chips for shipments. More importantly, among the existing shipments, more than 50% of Xiaomi and OV are low-end phones priced at US$100-200. In addition, judging from the future market positioning and strategies of Xiaomi and OV, the Indian and Southeast Asian markets dominated by low-end phones are still the focus of their efforts to compete.

Given the current and future market factors mentioned above, it is meaningless for Xiaomi and OV to release their own high-end mobile phone chips. However, if they start from the mid- and low-end, they will lose the strategic value of increasing brand premium and thereby obtaining higher revenue and profits. Therefore, even if the above-mentioned technology and investment issues are solved, it is still quite challenging to fit the chips into their own smartphone market.

The independence of core chips in the field of science and technology has always been the pursuit and goal of our country and enterprises. Specifically in the smartphone industry, perhaps in view of the precedent of Huawei's successful mobile phones, it has been mentioned repeatedly. But we cannot simply see the success and ignore the development laws of the industry and our own strength and strategy in the industry.

In addition, as a smartphone, proprietary chips are only one of the core competitive advantages. It also requires the cooperation of innovation and application of many other components. If you blindly believe that having proprietary chips is everything, then you will inevitably fall into the misunderstanding of the "chip" theory. This is not only true for China's mobile phone industry, but also for other technology industries.