So from this point of view, since Huawei has obtained the permanent authorization of ARMv8 architecture, they can design ARM processors independently, and Huawei has mastered the core technology and complete intellectual property rights. Simply put, even if ARM no longer licenses the ARM instruction set to Huawei for some reason, Huawei will not be affected.
On the other hand, ARM will definitely not refuse to authorize Huawei because it is related to its business model. Similar to Qualcomm, they all maintain their own development through the authorization of ARM instruction set and royalty income. So, for Huawei, a big customer with annual sales of more than 200 million mobile phones, what reason does ARM have to choose to refuse?
Of course, there is one more thing we have to admit. The so-called plan for a rainy day is to make a good response plan in advance before the crisis breaks out. We should not think that everything will be fine if we get a so-called "permanent license", because no one can guarantee that ARM will refuse to continue to authorize us for some inevitable reasons. Even if Huawei is granted permanent authorization, won't ARM protect its interests through technical review and standard update?
Therefore, in the absence of this symptom, Huawei, Ali and other enterprises should allocate some resources to support the development of this underlying technology, just as Huawei started the research and development of the operating system as early as 20 12. This underlying technology is difficult to achieve results without several years of silent research and development, and once it forms its own technical scheme, it will also be of great help to the future development of enterprises.
I'm Shinji Mushi, like it ~
The first unprofessional part of this problem is that Huawei itself does not have the ability to produce processors. What Huawei Hisilicon has done is to develop mobile phone processor chips, and then hand over the tested solutions to circuit manufacturing service enterprises to produce chips (such as TSMC). At present, all the chip research and development enterprises in the world, except Intel and Samsung, do not have the production capacity, including Apple, Qualcomm and MediaTek, all develop their own chips and then hand them over to manufacturing enterprises for production. After all, a mature mask aligner can't be bought by anyone who wants to buy it.
At present, ARM is the world's top provider of semiconductor intellectual property (IP). What it provides is actually a solution. ARM is not the only RISC reduced instruction set architecture provider in the world, but its solution is the best and its market share is the highest, reaching 95% at present. So what is the so-called ARM public architecture? Actually, it's like cooking. ARM gives you a recipe, which probably shows you a direction, but ARM won't tell you how many dishes to buy, what the proportion of dishes is, and how much seasoning to put. You must explore and try it yourself.
For example, Huawei Hisilicon, after getting the solution provided by ARM, needs to start designing the internal circuit mode according to its own needs, and integrate the design of the front and rear ends. After the overall scheme is developed, it will be handed over to a company with supercomputer simulation strength (such as synopsys) for computer simulation (the cost of simulation is quite amazing). Then, after the simulation is successful, the whole scheme is handed over to the chip manufacturing enterprise (such as TSMC) for tape-out (i.e. small-scale production test), and then finalized for mass production after a series of tests. Once the streaming fails, you need to start from scratch. At present, according to Huawei's technology, it is more mature and professional.
The answer is yes. First of all, Huawei bought the permanent license of ARMv8, which means that at present, ARM has been unable to balance Huawei, so even if ARM has no other license, Huawei can still make chips.
Even if the conditions are harsh, even if the supply of ARM is completely cut off, Huawei will not buy out, so Huawei can buy solutions from other providers first, but it is not as mature as ARM. Even according to the experience accumulated by Huawei at present, we can slowly explore the mystery of the architecture to complete the research and development of the chip, but in this process, we need to invest a lot of manpower, material resources and financial resources, and compared with the very mature ARM architecture at present, its performance is naturally far from it. But as long as you are willing to increase investment, it is certainly achievable.
And at present, Huawei has started to use its own instruction set on the chip in the server field.
But then again, the problem itself is alarmist. ARM itself makes money by patent authorization, and Huawei is already a super customer of ARM. Will you give up in vain? In addition, ARM has set up a branch in China, and China Capital holds 565,438+0% of the shares. In the future, the cooperation between ARM and Chinese enterprises will only get deeper and deeper, and the time of stopping supply may not happen at all.
Almost all mobile phone processors in the world use the architecture of British ARM company, including Huawei. No matter Huawei, Apple or Qualcomm, they all need to purchase the authorization from ARM, and the price of different levels of authorization is different. Generally speaking, the architecture license of ARM is also limited. If Huawei is stopped from licensing, Huawei's Kirin processor may not be able to continue research and development.
Because all the hardware and software in the mobile phone industry are developed based on ARM processors, at this stage Huawei can only continue to use ARM-authorized processors to ensure performance and compatibility. Even if Huawei has the ability to develop its own processor architecture, it is impossible to put it into the market in a short time, or ARM suddenly stops authorizing processors with new architectures at some time. Huawei can only continue to use products with existing architectures at most, and it is difficult to innovate products, which is naturally fatal to Huawei mobile phones.
However, ARM also needs to consider its own interests. In 20 16, ARM was acquired by Softbank and later set up a branch in China. In addition to Huawei, ARM has many domestic partners. If it is easy to stop the authorization of partners, it will have a great impact on ARM, especially there are not many big customers like Huawei in the world. If ARM stops licensing, it will not only harm its own interests, but also seriously damage the reputation and reputation of the industry. These are definitely not weapons.
In addition, ARM is not invincible. If ARM stops authorizing Huawei, then Huawei will definitely take other paths, and it is even possible to completely develop a processor with its own architecture. In fact, this is the biggest threat to ARM. Today, with the global division of labor and industrialization of chips, ARM has no sufficient reason to stop authorizing Huawei.
Several opinions are reasonable. However, looking at the problem from another angle, the United States is now rampant around the world with its strong strength and is ruthlessly threatened by sanctions and supply cuts. Also euphemistically called: extreme pressure makes a global industrial chain fall apart and suppresses the rising China by any means. So be sure to prepare for the worst and plan ahead.
Besides, China is a huge market with a population of 654.38+0.4 billion. In the past, under the background of globalization, global resources were integrated and exchanged with each other, always choosing the best. Now, under the situation that the United States deliberately blocks and cuts off supply, we should seize the window of opportunity created by the American government and launch our own systems and chips. It was not a big problem to worry that there was no market for chips and no ecology in the system. Because this is the only way out of the last fight, and it will be supported by the vast majority of people.
Looking at the 70 years along the way of New China, as long as the enemy blocks something, we can always create something in the end, which is determined by China's will and toughness. There is no insurmountable difficulty, and time is on our side.
First of all, we should analyze this issue from a comprehensive perspective, not just from the standpoint of the United States or Huawei. Personally, I think Huawei must focus on Asia, Africa and Latin America in the future, because if the United States completely bans chips, although there is a gap in Huawei's current technology, the demanders in Asia, Africa and Latin America often need low-end machines, and the chip threshold of low-end machines is low, so Huawei's focus on developing Asia, Africa and Latin America based on 4G can have a good space and environment.
Chip wars are more like a protracted war. If the United States doesn't sell chips to China at all, who will American chips be sold to? At present, China's chip purchases account for about 50% of the total global purchases. If the chip is not sold to China, it will be a disaster for any chip manufacturer. This is the essence of this problem. For a simple example, like Intel's i3, i5, i7 and i9, each generation of products needs a process of design, research and development and sales. In this process, if i3 sells 6.5438+0 million pieces every year, the enterprise will estimate the sales volume according to the sales volume of i3 in the i5 era, calculate the research and development cycle and the time of product upgrading, and so on, so that the enterprise can maintain a relatively stable and long-term output and output between i3 and i9. However, assuming that the sales volume of i5 suddenly halved, enterprises will lose the ability to predict the sales volume of i7 products, which is not only the loss of profits. The market attaches great importance to the competitive environment. If i7 sets the pre-sale target at 200,000, and the trade war returns to normal, the demand for 800,000 yuan in the market may have been taken away by others. So what is the estimated sales volume of the i9 generation? This is the essence of chip wars. Enterprises need a huge design team and constantly pay the designers. If the project is halved in i5, will the staff of these teams continue to maintain or dissolve? If these employees are dismissed, they will be poached by other companies, and it will take a lot of time to establish a stable R&D team. This is just a research and development problem, and subsequent sales networks, dealers' agents, warehousing and transportation will cause more problems. In other words, if the sales volume of i5 is halved, it will seriously affect Intel's forecast of the sales volume of i7 and i9, and then a series of problems will appear, forming a vicious circle, which is the fundamental problem of the enterprise.
Therefore, for Huawei, this will be a protracted war. If we can focus on Asia, Africa and Latin America, even if there is a gap in chip technology, we can maintain relatively stable sales. At the same time, we will speed up independent research and development of chips and improve their technical capabilities. Affected by the trade war, the longer this process, the stronger the speed and ability of Huawei to catch up. And you have to understand that once the estimated chip sales in the United States are shaken, it will inevitably lead to a hole in future sales, and once the hole is taken away by others, then the United States, which was originally firmly in the leading position, will not be worth the candle.
So the chip and the mobile phone are like a highly cooperative system. Whoever leaves this system and wants to come back will have to pay a higher cost and price. So if the United States doesn't sell chips to China, who will his chips be sold to? How to fill the hole after that? This is the core problem that the United States must face in the chip war.
It is an ironclad fact that technology will eventually yield to the market. Throughout the history of scientific and technological development, many of them have become the ultimate winners because of the opening of the backward technology market. X86 was developed by IBM. At that time, this technology was not advanced compared with other companies, but after authorization and opening up, it attracted many companies to manufacture (commonly known as compatible machines) and made a lot of money. In the 286 era, IBM quit and no longer authorized (jealous of others making money). That's why Intel developed 386 itself, and it will be out of control in the future. Another example is Ethernet. In the field of communication, Ethernet technology is backward and has many defects. Even the founder of Ethernet predicted that Ethernet would be eliminated in the future. When a compatible manufacturer developed a 100M fast Ethernet, IEEE didn't even pass the approval of 100M fast Ethernet technical standard, but Ethernet became the de facto standard because of its cheapness and open technology.
Summing up experience is that whoever can control the market with technology is the standard, regardless of whether he is backward in technology or not.
Are you tired of not selling this or that consumer electronic products? If you don't sell it, you're not afraid of others starting a new stove? Computers before smart phones are in full swing, and the chips and operating systems are the only ones in the United States, with no branches. But the United States never talks about these things and even turns a blind eye to pirated systems. Most Americans sell on sale, and those surnamed Liu will drive away the surname Ni.
If ARM is no longer authorized, Huawei will definitely find a way to start a new stove. With Huawei's urine and the assistance of the state, it may be successful. Will ARM be happy when it has multiple opponents?
Samsung specializes in this matter. In the past few years, Huawei was miserable and hit it off with BOE. Under the heavy money, BOE's screen appeared. Is it better for Samsung to find its own competitors?
ARM can't break Huawei's back road. If it is no longer authorized, it will only hurt both sides.
Of course, Huawei can't produce processors. Unless ARM authorizes Huawei, what's the point of authorization?
But ARM can't do this, because it doesn't benefit ARM at all, but it breaks its own financial path.
What kind of company is ARM? He doesn't make mobile phones or produce chips, and his only source of income is to authorize his instruction set and architecture to chip designers. Huawei sells more than 200 million mobile phones a year. If ARM doesn't authorize Huawei, will it be hard for itself?
We are always afraid of being stuck in the neck by others, and such worry is necessary. However, as the saying goes, shopkeepers bully customers, and customers can bully shopkeepers when they are older. When a manufacturer like Huawei has occupied enough market share in the industrial chain, his relationship with ARM becomes interdependent, rather than unilaterally requesting authorization from ARM.
Needless to say, ARM established a joint venture company in China last year, with China investors accounting for 565,438+0%. ARM has made up its mind to cover the China market and Huawei, a big customer, and it is impossible to be unauthorized.
First of all, let's talk about three different modes of ARM's external authorization:
1. Architecture/Instruction Set Level Authorization
This authorization can greatly transform the architecture of ARM, and even expand and reduce the instruction set of ARM. Apple is a typical example. Based on the use of ARMv7-A architecture, it has expanded its own swift architecture of Apple.
2. Kernel-level authorization (ip core authorization)
This can be based on a kernel and its own peripherals, such as USART GPIO SPI ADC, and finally form its own MCU, such as Samsung, Texas Instruments (TI), Broadcom, Freescale, Fujitsu and Calxeda.
3. Use hierarchical authorization
This authorization is only for use, and the structure inside is unclear, which is equivalent to giving you a black box and only telling you how to use it.
Apple, Qualcomm and Huawei are all companies authorized by arm, but Apple and Qualcomm have long been granted permanent authorization at the level of ARM architecture/instruction set, and they all simplified the instruction set of ARM as required, which is why Apple's A 12 and Qualcomm's 855 are better than Huawei's Kirin 980 in performance.
Secondly, Huawei obtained the lifetime license of arm processor on 20 18, which is the first license at the architecture/instruction set level, and has the ability to independently develop the architecture of ARM instruction set, but after all, it started late and is not very mature. If under political pressure, ARM cancels the permanent authorization to Huawei and compensates Huawei for a large sum of money as liquidated damages, which is not a bad thing for Huawei and China. After all, there is nothing that China people can't do together in times of difficulty, except that the architecture is not as big as the ecosystem of ARM instruction set. If Huawei wants to be its own instruction set ecology, it will take a long time, and its performance will fall behind at first.
Thirdly, ARM is a listed company, and its main business and income depend on ip authorization. Huawei is also a big customer and will not cancel its processor authorization. After all, the company should put the interests of shareholders first!
In short, whether ARM will temporarily release Huawei's authorization requires Huawei engineers to work hard, accumulate R&D experience and develop their own processors one day! ! !
If ARM is no longer authorized, do you think Huawei can still design Kirin processors? Welcome to comment and leave a message, please click on the attention if you like!
Because of this, the IP authorization of ARM to China was gradually transferred to the joint venture company established by ARM in China, and the Chinese side held 5 1%.
In the short term, what you are worried about should not happen. After all, with the severance payment of 100%, ARM is basically abolished. ARM itself will not sit still because of political games.
If it does happen, we don't have to be afraid of unilaterally interrupting the cooperation agreement, so we don't have to abide by this contract. To put it bluntly, authorization is based on sincere willingness to cooperate. If compliance is no longer necessary, we can continue production without paying the licensing fee. After all, everyone has been exposed to the contents of the basic instruction set, so it is enough to give or not.