Rebirth of the internet boss

Chapter 211 Open and Closed (Two in One)

The reason why Li Feng chose the British Empire as his first stop was because of the British semiconductor company ARM, which later became the overlord of mobile chips.

Semiconductor is the core of electronic products, the cornerstone of the entire information industry, and a highly technology-intensive industry. So before talking about ARM, let us talk about several models of the semiconductor industry.

1. IDM mode (vertical integration mode).That is to say, a company takes care of the entire process from design, manufacturing to sales, and requires strong operating capital to support this business model, such as Intel and Samsung.

Samsung is very powerful. On the one hand, it is a vertically integrated model that can manufacture chips designed by itself; on the other hand, it also plays the role of a foundry, and at the same time provides foundry services for the A-series processors designed by Apple.

Of course, Intel is also awesome, and it is also the IP designer of CISC (Complex Instruction Set), which almost monopolizes X86 processors, mainly computer chips and server chips.

2. IP design mode. ARM is quite special. It belongs to the most upstream IP designer. It is the designer of RISC (Reduced Instruction Set). It does not manufacture or sell any chips, but only designs IP, including instruction set architecture, microprocessor, graphics core, and interconnection. Architecture, and then license its technical intellectual property (IP core) to other semiconductor factories.

3. Fabless mode (chip design + outsourcing production + sales).Such as Qualcomm, MediaTek, etc., get authorization from ARM, design their own chips, and be responsible for sales,
However, Qualcomm’s strength in mobile chips is not because of its advantages in chip design, but mainly because of its patent layout in the communication field. make money,
And Qualcomm's patent license fee is not calculated according to "single chip charge", but calculated according to the percentage of the cost price of the whole mobile phone. It is not an exaggeration to say that Qualcomm is the father of the entire smartphone.

4. Wafer foundry mode (pure chip manufacturing), such as TSMC, GlobalFoundries... TSMC alone has almost monopolized the entire foundry industry, with a market value of more than 2000 billion US dollars. This is not an ordinary foundry. Instead, it represents the top-level manufacturing industry. The back of the light involves 18 top-level suppliers, representing the highest level of the country's industry.

Take Huawei, which the Chinese people are most proud of, as an example, if Huawei goes public, its market value will be about the same as that of TSMC. Although Huawei is not listed now, its market value can be compared with similar ones. It is undeniable that Huawei is far from Samsung’s opponent.
The combined market value of the entire Samsung Group (Samsung Electronics + Samsung Semiconductor + Samsung Heavy Industries + Samsung C&T + Samsung Airlines...) is about US$3600 billion. If Huawei goes public as a whole, it should be about US$2000-3000 billion.
Of course, if Huawei is listed on the A-share market, it is possible to double the size of the Chinese people.

Let's go over the whole chip again, two IP designers: ARM (RISC Simplified Instruction Set), the development ecology, there are Qualcomm, MediaTek...
Intel (CISC complex instruction set), a closed ecology, only Intel and AMD (not ARM, but the agricultural enterprise AMD);
Three major chip manufacturers: TSMC (pure foundry), Intel, Samsung,
……

And Li Feng came here this time to obtain the RISC authorization of ARM's chip technology.

Now is not the era of mobile Internet in which more than 95% of smartphones and tablets in the world use the ARM architecture, and the era of ARM's mobile hegemony has not yet come.

Now the market value of ARM is only 10 billion US dollars, not even half of that of Fengrui. After the advent of the mobile era, the market value of ARM also began to skyrocket.
However, because of ARM's business model, the market value of ARM has not been high. Even in 2016, when Softbank acquired ARM for US$322 billion, it paid a premium of 43%.

As a company open to sell chip technology authorization platform, when it heard that a Chinese company wanted to obtain their chip architecture, ARM was very curious, but its attitude was very enthusiastic. CEO Warren East personally received it in the reception room Li Feng and his party.

This 40-year-old East is a typical British man, a gentle and polite gentleman, who makes people feel very comfortable when he speaks, and he didn't say anything because of Li Feng's youth.
After a while of courtesy, East briefly introduced some of the company's philosophy and situation, and asked his assistant to project the schematic diagram of the ARM-based chip directly on the screen behind him.

What does architecture mean?

Let's put it this way, making a CPU is actually similar to building a house. What ARM sells is the basic design blueprint of a house. After the IC design company (Qualcomm, etc.) buys the blueprint, it can modify and decorate it according to its own needs. For example, how many rooms do you want to design?How big is the dining room and bathroom?

Pointing to the schematic diagram, East said: "Mr. Li, these are our ARM chip architectures, with specific parameters on them. At present, we are relatively mature with the 6 series, which is now used by Nokia and Motorola. Of course , we also have the newly developed 7 series.”

As East said, the parameters and diagrams on the slides are very detailed.They don't have to play tricks at all. All these structures and details have been patented by ARM. You can produce them secretly, but once you mass-produce them, wait for a lawsuit.

After generally introducing the chip category, East then told Li Feng and the others about the charging situation: "The initial licensing fee for the 6 series is only 200 million US dollars, and we will charge 2 US dollars for each chip in the later period. enough!"

ARM's profit model is as follows: First, it depends on the initial licensing fee, ranging from US$100 million to US$1000 million, and pays off in one lump sum. The specific amount depends on the complexity of the licensed technology purchased.

The second is the royalties in the later period. The royalties are paid for every chip sold, usually 1 to 3 US dollars per chip.

The initial authorization fee is 200 million U.S. dollars, which doesn't matter.

但是每片芯片收2美金，那就得细算了：做个100万部的手机就得给他们200万美金，1000万部就是2000万美金，一亿部就是2亿美金了！

Besides, East didn’t have any good intentions. Other big manufacturers paid $1 in royalties, but Fengrui paid $2. Moreover, the 6 series is obviously a product that is about to become obsolete!

East even regarded himself as a bumpkin who didn't know anything. Fortunately, Li Feng had someone do a detailed investigation before coming here.

However, Li Feng didn't say anything, he just asked, "What about the 7 series?"

East said: "Although the 7 series is our newly launched product, the cost of this is much more expensive. At least 500 million US dollars is required for the initial authorization fee, and the patent fee for each chip is 3 US dollars. Although it is a bit expensive, But definitely worth the money.”

Li Feng then asked: "Then can we, Fengrui, obtain the authorization of the architecture level?"

ARM has three different modes to provide external authorization: architecture-level authorization, kernel-level authorization, and usage-level authorization, which can be easily understood literally.
To make a very vivid metaphor: Suppose I wrote an article, and I told A that you can use it after modification, which is the authorization of the architecture level; I told B that you can quote my article in your article, and then It is a kernel-level authorization; I told C that you can only forward my articles, you cannot change them, you cannot add oil and vinegar, that is to use hierarchical authorization.

For example, Qualcomm, Samsung, and Apple are all licensed at the architecture level, which can be modified, while Huawei is authorized at the kernel level, and many other companies with Chinese backgrounds use hierarchical authorization, that is, the public version.
Of course, don't think that you can produce your own chips with the authorization of ARM!Some people always say that making SoC (system-level chip) is as simple as sticking paper boxes with glue against the schematic diagram. They think that after buying the public version solution from ARM and throwing it to TSMC for production, they can make their own SoC (system-level chip) ),
This kind of thinking is almost like thinking that I will play football after watching Messi's extraordinary videos a few times, and I can join Barcelona to play in the Champions League.

Xiaomi's Rebus has also invested billions of dollars in the surging chip, but it has been two years since the launch of the S1, and the S2 has not heard the sound.
It is said that several tape-outs failed, and each time it was 400 million US dollars, and the light did not light up. How can this be played!Since then, the voice of saying that Huawei is engaged in SoC is much smaller, and it is estimated that we will not see the surging S2 in a while.

Fortunately, this is not the era of mobile intelligence. East thought about it and gave the answer: "In this case, the initial authorization fee for the 7 series is 1000 million U.S. dollars, and the patent fee for each chip is 3 U.S. dollars!"

Li Feng said: "According to our investigation, it seems that the patent fee for each chip of other manufacturers in the later stage is 1 US dollar."

East said without changing his face: "That's because they ship a lot of mobile phones, so they have this preferential policy."

Li Feng said: "Okay, how about if we increase the initial licensing fee to US$1200 million and reduce the patent fee per chip to US$1!"

Li Feng still wanted to increase the licensing fee in the early stage to reduce the share in the later stage, just like the agency of "Miracle MU".

Originally, East wanted to take advantage of Li Feng, "Then let's think about it."
Unexpectedly, Li Feng stood up and said, "Then we'll come over in two days!" After speaking, Li Feng was ready to go outside.

Now ARM is not the future ARM, now 1200 million US dollars can be almost half a year's profit of ARM,

So, seeing that Li Feng was really going to leave, East quickly grabbed Li Feng and said, "Mr. Li, it's okay."

In the end, ARM won the present, and Li Feng won the future.

Regardless of how much money Li Feng spends now, when smartphones become popular, he will find that all these efforts are worthwhile.
It can even be said that at that time, if Li Feng comes to ARM again, the price will not be 1200 million U.S. dollars, but at least 3000 to [-] million U.S. dollars. It is absolutely impossible for each chip not to charge you three or four U.S. dollars. If you want authorization at the architectural level, you don’t even have to think about it.
Isn't this the biggest advantage of rebirth!
The rest of the specific agreement will be drafted by the "patent team" led by Li Feng. Li Feng has time to visit the UK.

As a Beckham fan, I felt sorry for myself when I came to the UK and didn’t watch the Premier League. Li Feng even watched the match between Arsenal and Manchester United. The scene was still very hot.

Especially now that Manchester United still has the heartthrob Beckham and the short-tempered Keane, while Arsenal also has Henry the Great and Ice Prince Bergkamp,

However, the result of the game was that Arsenal had the last laugh, and if Li Feng remembers correctly, the Premier League champion this season is Arsenal, and they still ended with an unbeaten record.

Although Li Feng has nothing to do, he has nothing to do, but Li Feng has been thinking about a question recently through the development history of ARM: Which path will Fengrui take in the future?is it open?Still closed?

In the technology circle, the battle between open and closed has been going on for a long time. Intel and ARM, Microsoft and Apple, Apple and Google are like Apple's "closed" iOS system and Google's "open" Android system.

But it seems that "development" has a bigger chance of winning in the end, just like Intel (CISC Complex Instruction Set) and ARM (RISC Reduced Instruction Set), like a person choosing two paths, one with high energy consumption and high efficiency, and one Low energy consumption, low efficiency, but the final destination is the same, low cost, low energy consumption, high efficiency,

Intel is closed, its chips are all produced by itself, and its only competitor is AMD; while ARM licenses the design of the chip architecture to companies such as Apple, and finally occupies 95% of mobile smart devices in the mobile era ARM beat Intel.

In the PC era, the "open" Microsoft also defeated the "closed" Apple;

In the mobile era, "open" Android has also defeated the "closed" iOS. This "victory" is still a bit early, and it can only be said to occupy 80% of the market share;

It seems that the final conclusion should be that "open" beats "closed", but in fact, the market value of ARM is not even a fraction of that of Intel, and ARM cannot control the downstream industry chain; and Apple only uses 20% of the market share , has obtained 80% of the profits of the entire smartphone industry, and the commercial level of the Android system is far lower than that of the iOS system;
In fact, such disputes are meaningless. Openness and closure, in essence, are technology giants serving their own interests.
Because ARM was beaten by Intel, even its mother couldn't recognize it, so it had to go to the road of development. It took 30 years to get to this point; and Google directly open-sourced Android because it lost the opportunity of the mobile era. , are all necessary strategies.

But for Li Feng, because of next year's Openers Conference, Fengrui has already embarked on the road of "openness".
However, for Li Feng, "openness" means openness, but to what extent should it be developed? This is a question?
After all, the reason why Google made Android open source is not only to catch up with Apple, but also for the mobile portal, and Google's core business is search, not hardware.

　Thank you for calling me a big book friend for your reward!

　　
　
(End of this chapter)