Another billion euros! Apple's German core, the signal of the next iPhone is expected to get better?

Apple, which is addicted to core manufacturing, has a new move - on March 2, local time in the United States, Apple announced that it will expand its silicon design center in the center of Munich, and is expected to invest an additional 1 billion euros in the next 6 years.

It is reported that the new investment will be used to create "state-of-the-art research facilities", which can create more open and convenient research and development conditions for Apple's research and development team. Johny Srouji, Apple's senior vice president of hardware technology, said that there are more than 2,000 Apple engineers working at the center, and the scale will be further expanded in the future.

At MWC 2023 held at the end of February, Qualcomm CEO An Meng revealed that Apple's 5G modem chip is about to be ready, which may be used in the iPhone 16 series at the earliest. Qualcomm is mentally prepared to lose all orders from Apple.

These two dynamics reveal Apple's strength in chip research and development. But how many people remember that more than ten years ago, Apple was still stuck in the neck by semiconductor manufacturers such as Samsung, Qualcomm, and Intel? In order to achieve chip freedom, how much effort and how much money has Apple put in?

These questions are actually not difficult to answer. What is really worth thinking about is how other mobile phone manufacturers should steal from Apple and create their own chip moats.

(Image via UNsplash)

1 billion plus code Munich laboratory, Apple insists on creating a closed-loop self-developed chip

Munich has in fact become one of Apple's most important "overseas bases", and this additional investment is not surprising.

As early as March 2021, Apple announced the construction of a new European semiconductor laboratory in Munich, with an initial investment of 1 billion euros. According to Apple's planning at the time, the Munich semiconductor laboratory covers an area of 30,000 square meters and is located in Karlstrasse in the middle of Munich, which is also a technology center for German and even European semiconductor companies.

According to Reuters' revelations that year, one of the reasons why Apple chose Munich was because of its rich talent resources. Google, NXP Semiconductors, and ARM had all established R&D centers in Munich at that time, and leading car companies such as Audi and BMW also had bases there. After Apple's arrival, it poached a lot of talent from these giants. In addition to this home base, Apple has seven other bases in Munich, and the total number of employees in Germany is estimated to exceed 4,000.

There's a reason why Apple attaches so much importance to the Munich semiconductor lab. From the beginning, the Munich semiconductor lab has focused on the development of 5G and wireless technology – that is, to build 5G modem chips for Apple.

As we all know, Apple's obsession with 5G baseband chips has never been broken, and getting rid of Qualcomm's constraints and creating a closed-loop fully self-developed chip is also Apple's long-term pursuit. The iPhone 14 released last year is equipped with a Qualcomm X65 modem chip and radio frequency IC, and the iPhone 15 released this year is expected to remain inseparable from Qualcomm. For this reason, the Munich Semiconductor Laboratory had high hopes from the very beginning.

According to public information, the specific work of the laboratory also includes long-term monitoring of the performance of iPhones, iPads, Apple Watch and Macs, and improving M1 series technology based on this. After the completion of the first phase of the center's project in 2022, Apple also reorganized the local chip supply chain in Europe, gradually transferring high-quality talents and R&D equipment scattered throughout Europe to Munich.

The Munich semiconductor laboratory also lived up to Cook's trust and commitment. Qualcomm CEO Anmeng has accepted the fact that Apple has lost this big customer, and once the 5G baseband chip comes out, Apple will also clear the last obstacle to achieving the "fully self-developed chip closed loop".

In 2021, Apple launched a MacBook with M1 Pro and M1 Max processors to get rid of its dependence on Intel. The iPhone has already enabled all A-series chips, and even power chips and security management chips are Apple's own products. Now the only thing that can jam Apple's neck in this link is Qualcomm's exclusive monopoly of 5G baseband chips.

As for the WiFi chip research and development plan that was postponed earlier, the impact on Apple is actually not so great. After all, from the perspective of cost, there is a certain gap between WiFi chips and 5G baseband chips, and Apple has been strictly controlling the cost of WiFi chips. On the 5G baseband chip, starting from the Qualcomm external baseband chip X55 equipped with the iPhone 12, the cost far exceeds Apple's self-developed A series chip.

In Cook's supply chain management philosophy, the highest-margin and core components are best in their own hands, or scattered among sufficiently high-quality suppliers, and try to avoid dependence on a single supplier. Under the trade-off, the strategic significance of 5G baseband chips far exceeds that of WiFi and RF chips, and naturally receives more attention and becomes the core of Apple's core manufacturing plan.

Of course, the Value Research Institute (ID: jiazhiyanjiusuo) believes that after Apple conquers 5G baseband chips, it is still possible to increase WiFi and RF chip research and development. Analyst Guo Mingxi broke the news that Apple only postponed the WiFi chip project and did not completely stop it.

For a while, Apple really can't complete the closed-loop of fully self-developed chips. But there's no denying that Apple's ambitions have been swelling.

Looking back at Apple's core making road, banknote ability is not a winning weapon?

Back to the question posed at the beginning of the article: from being constrained to about to achieve a closed chip loop, what does Apple rely on to rise?

The most popular answer is to burn money.

The two phases of investment in Munich semiconductor laboratory are just the tip of the iceberg in Apple's core manufacturing plan. Banknote capacity has indeed played an important role in Apple's core-making plan.

In March 2010, Apple released the original A-series processor APL0398, which is known as the A4 processor. The processor uses a 45nm production process, ARMv7 architecture, and is only available on a few products such as iPhone 4 and iPod touch. A4 and the two subsequent generations of products A5 and A5 did not actually achieve a dimensionality reduction blow to Samsung and Qualcomm. Compared with A4, A5 only has a certain improvement in computing power and power consumption.

It wasn't until the A6 released in September 2012 that Apple's A-series chips really came to the altar. It was also from this year that Apple's research and development grew significantly, and its proportion of revenue directly doubled in 10 years.

Data show that in fiscal 2012, Apple's R&D expenditure and R&D accounted for 3.381 billion US dollars and 2.2% respectively, the former increased by 51.6% year-on-year, and the latter was basically flat. By fiscal 2022, Apple's R&D expenses have risen to $26.251 billion, accounting for 6.7% of revenue.

The Value Institute believes that large investment is of course an important guarantee for Apple's core manufacturing industry. But the ability to have banknotes alone is not enough. What's more, in recent years, Apple's business line has continued to expand, and automobiles, VR/AR, etc. have also consumed a lot of R&D expenses, and the group's R&D investment has repeatedly reached new highs is not separately linked to chip R&D.

Looking back at Apple's core manufacturing road, there are actually several keywords worth noting: distinguishing priorities, weakening opponents and breaking the rules and innovation.

First of all, Apple recognized the importance of architecture early on, and it was one of the first companies to develop its own processor architecture. The so-called clear distinction between priorities is to first resolve the main contradictions and lay a good foundation.

The processor architecture of the A4 and A5 generations is basically no different from Samsung, and the performance improvement is mainly achieved by adding cores, which is no different from the mainstream idea at that time. By the A6 launched in 2012, Apple adopted its own core architecture and launched a 64-bit mobile phone CPU the following year, a full two years earlier than Qualcomm.

Apple was able to find this direction, and two souls on the P.A. Semi team, Daniel W. Dobberpuhl and Jim Keller, contributed greatly. The latter directly participated in the research and development of A4 and A5 generation chips, and it was he who determined the strategies of the A series chips to pursue energy consumption ratio and power chip internal connection architecture. Even after leaving Apple, its influence continues to this day.

Second, Apple's core manufacturing process is indispensable for eclecticism, especially plundering the opponent's prey to weaken the opponent's strength. Apple does not exclude cooperation with semiconductor manufacturers, but often makes different partners hinder each other, and then divide the latter or even complete the merger. Samsung and Intel, which have worked closely with Apple, have a deep understanding of this.

Samsung reached a cooperation with chip manufacturer Intrinsity in 2009, and the S5PC110 chip developed with the latter's assistance and the original Galaxy S equipped with the chip became the strongest challenger to the iPhone. But soon after, Apple took the lead in acquiring Intrinsity, which directly led to Samsung losing the right to use Fast14 technology, and also greatly limited the acceleration ability of the ARM general architecture used by Samsung.

There are many similar examples. The aforementioned P.A.Semi was also acquired, and the acquisition of Anobit, an Israeli flash controller design company, is also one of the masterpieces. What is rare is that Apple has integrated these external teams and technologies into its own ecosystem.

Third, Apple's A-series chips have repeatedly acted as "white mice" and experimented with many new technologies.

If the early A4 and A5 series are still imitating Samsung, the Apple behind is gone on the road to originality. The A11 Bonic, launched in 2017, is the first generation of bionic chips, the first to realize the application of AI technology in smartphones; The T1 self-developed chip on the 2016 Mac gave hardware devices chip-level security for the first time.

Apple dares to experiment with its own chips and continue to innovate, and there is also an important reason: shipments and market share are stable, and there are enough business scenarios and users to test the technology.

In addition to compressing costs and increasing profits, Apple's self-developed chips are also a very important reason to improve product experience. User feedback allows Apple to understand which links need to be improved in the first time, and even predict the next needs of users, which in turn provides direction for chip research and development, forming a virtuous circle.

You can't do without money, but money isn't everything. Although Apple's core making road is not full of twists and turns, it has taken many dangerous moves. Today's results are not all paid for by burning money.

Apple's answer cannot be copied by friends

After Apple, mobile phone manufacturers have become a trend to develop their own chips, but the real threat to Apple is only Huawei a few years ago.

Why can't other mobile phone manufacturers keep up with Apple's research and development rhythm and fail to break through key technical barriers? Of course, the financial strength is not as good as Apple's is an important reason, but objectively speaking, the background of the times is also unfavorable to Huami OV - Apple entered the game early and ate the reform dividends of the semiconductor industry.

From the perspective of time, the development path of Apple's A-series chips and wafer foundry technology basically coincides. Samsung and TSMC are crazy involution, and the innovation cycle of Moore's Law is compressed, which objectively creates the performance myth of Apple's A series chips.

A6, A7, A8, A12 Bionic, A14 Bionic and other well-received products all correspond to the breakthrough of wafer foundry at key nodes. For example, A6 uses Samsung's 32nm process technology, and the previous two generations of products are still 45nm process. Later, after long-term cooperation with TSMC, the number of transistors in A-series chips soared from 3.3 billion to 16 billion, and the performance could naturally be greatly improved.

But the problem is that the room for progress and the speed of breakthrough in wafer foundry technology are not as fast as before. TSMC and Samsung have just finished 3nm technology, and the rumored 2nm or even 1nm may not come so soon. In addition to wafer foundry, other core technologies of chips are also basically stable, and pioneering innovations are becoming more and more difficult to achieve.

At present, the raw materials of chips, packaging and testing, and even lithography machines and other equipment are almost touching the ceiling, and doubts about Moore's Law are also rising one after another. In the semiconductor industry, the longer the technology develops and the deeper the research and development, the more difficult it is to make breakthroughs. Mobile phone manufacturers who joined the battle later can only adapt to this slow pace.

In the view of the Value Research Institute (ID: jiazhiyanjiusuo), it is very difficult for mobile phone manufacturers to copy Apple's core myth, and it is not necessarily in line with economic benefits. Because Apple and Samsung have now eaten most of the market share of the high-end market, chips are the trump card to increase the added value of mobile phones and hit the high-end, but if there are not enough shipments as support, it is difficult to cover research and development costs.

Of course, this is not to crack down on Huami OV's self-developed chip plan, but they do need to find a research and development route that is different from Apple and more in line with their own interests and brand positioning. Core processors and architectures are inseparable from giants such as Qualcomm, MediaTek, and ARM, and they must make efforts in other fields.

According to current observations, AI and image chips are the direction of OPPO and vivo, while Xiaomi's surging series chips take a diversified route, covering SoC, imaging, fast charging, power management and other links.

vivo X90 Pro is equipped with its own V2 processor, with AI function, and adopts TSMC's 6nm advanced process. Compared with the V1 chip launched in 2021, V2 has new breakthroughs such as FIT dual-core interconnection, high-speed and low-power buffer unit, and realizes the collaboration between AI chip and SoC. At least in the AI computing power acceleration link, V2 has made great progress compared to V1.

OPPO's self-developed chip plan is also focused on image chips, and its representative work is the Mariana image NPU released in December 2021. This chip based on AI algorithms can significantly improve the intelligence of imaging functions. And video is precisely the main selling point of OPPO and vivo all along.

After Huawei's HiSilicon development stagnated for well-known reasons, the core manufacturing business of domestic mobile phone manufacturers was also greatly hit. But the more difficult these times, the more potential can be unleashed. Although Apple's full score answer sheet cannot be copied, if it can go out of its own route, it is not a victory.

Write at the end

In the eyes of many people, the A series chip is Apple's masterpiece, and it is even equated with Apple's self-developed chip. But Apple's core-making plan did not begin with A-series chips, but goes back much further.

As early as 1994, Apple cooperated with IBM to launch the first generation of Power PC chips and achieved great success, but due to insufficient subsequent production capacity, it could only be abandoned. But by the time the IMBA and IMBA ended their switch to Intel in 2006, IBM couldn't keep up with Apple's pace and couldn't meet Apple's needs.

When Jobs launched the new processor chip in 2005, he said that the new product would maintain "the lead for the next 10 years." As for the old product that is 15 years old, even if there are no major mistakes, it is time to replace it.

This philosophy of always pursuing progress, and even forcing partners and suppliers to progress, runs through Apple's life, and also ensures that Apple has an inexhaustible sense of innovation and ambition. It's just that Jobs' judgment is still conservative, Apple's self-developed chip update speed far exceeds expectations, and the results achieved today are believed to be enough to satisfy Jobs in heaven.

Of course, other mobile phone manufacturers are now trying to catch up, hoping to copy Apple's successful route. However, the background of the times is completely different, copying Apple's answer is not advisable, competitors have to spend more time, resources, and explore a new core route.

After all, it took Apple more than ten years to have the strength it has today. Core making is a daunting task that cannot be rushed, but there is never a shortage of hope.