Due to the U.S. ban, TSMC has stopped producing Kirin chips for Huawei after September 15. At the Huawei mobile phone launch conference more than a month later, Huawei Consumer Business CEO Yu Chengdong also He said frankly, "Kirin 9000 will be the swan song of Kirin chips."
Judging from the overall response in the technology circle, there is naturally more regret and regret for the forced discontinuation of Kirin chips. At the same time, there is also quite a lot of indignation and sympathy for the situation of Huawei's mobile phone business. After all, it has been a long time. , Kirin chips all symbolize the pinnacle level of China's mobile phone chip development.
However, in a tragic atmosphere, in the past month and a half, six chip companies have obtained licenses from the U.S. Department of Commerce and will continue to provide chips to Huawei. Is this change a question? Does this mean there is some kind of turning point in the US's clampdown on Huawei? What impact does this trend have on Huawei's development? We may see the answer from Huawei's actions.
When it comes to chip blockade, everyone may think of Huawei’s Kirin series of mobile phone processor chips, but in fact this is only one of Huawei’s HiSilicon chip product lines. The United States is targeting Huawei The chip sanctions cover all chip design and production links of Huawei HiSilicon. These chip product lines mainly include the following six categories:
1. HiSilicon chips: mainly used in mobile terminal equipment, including Kirin 990, High-end chips such as Kirin 980, Kirin 970, and Kirin 960. In addition to the high-end 9 series, Kirin also has mid-to-low-end series chips in the 6 series, 7 series and 8 series. The latest Kirin 9000 chip uses a 5nm process;
2. Kunpeng chips: mainly for servers field, the Kunpeng 920 chip is completely independently developed by Huawei. It is the world's first 7nm data center ARM processor and is mainly suitable for Huawei's Taishan servers;
3. Ascend chip: processing for the field of artificial intelligence The processor adopts its own DaVinci architecture. Ascend chips are divided into two series: high-end and low-end. The high-end is Ascend 910 (7nm process) and the low-end is Ascend 310 (12nm process). AI chips are the latest step in the era of artificial intelligence. One of the core technologies;
4. Baron chip: HiSilicon’s 5G baseband chip. The latest one is Barong 5000 (7nm process) that supports 5G dual-mode. It is mainly used in Kirin 980 and Kirin 990, among which Kirin 990 also launched a version integrating the Balong 5000 baseband chip;
5. Tiangang chip: HiSilicon's 5G base station chip, the industry's first 5G base station core chip, and also the world's first A chip with super integration, super computing power, and ultra-wide spectrum has brought revolutionary improvements to AAU. The latest Tiangang chip is made of 7nm process. The 2 million 5G chips that TSMC produced for Huawei during the buffer period of the ban mentioned before refer to Tiangang chips.
6. Lingxiao chip: mainly used in routers. The Hi5651 chip is the industry's first 4-core 1.4GHz home routing processing chip. The built-in IPv6/v4 dual-stack hardware processing engine has obtained carrier-grade certification; supported 256 connection nodes, optimized through algorithm enhancement to identify terminal location and interference conditions. This type of chip has lower process requirements, and the 28nm process can already meet the demand.
Due to the U.S. ban, the first five of the above six product lines have basically been discontinued. Among them, Kirin chip (mobile phone SoC chip), Balong 5000 (baseband chip), and Shengteng chip (artificial intelligence chip) directly affect Huawei's consumer electronics business (mobile phones, tablets, smart home and other electronic products), and Tiangang chip directly affects 5G communication equipment Business, Ascend chips also have an impact on Huawei's industrial-grade artificial intelligence products. These three areas are Huawei's core business areas.
In comparison, the server field is important but not urgent, and Lingxiao chips are basically not affected.
As mentioned earlier, the U.S. Department of Commerce has approved six chip manufacturers to continue supplying goods. Everyone in the news media was very happy to see the comments. It seems that this is good news. But it's not.
For example, TSMC, which attracts the most attention, continues to provide chips to Huawei. What it provides is a 28nm process, which cannot meet the needs of the five major chip product lines mentioned above. Moreover, China’s wafer foundry The engineering level can also achieve 28nm. To put it bluntly, this is something that everyone can do. Of course, we can't say that it is useless. After all, it is barely a good thing for the Lingxiao chip. Other products such as those provided by AMD are basically products whose "technical shelf life" is about to expire.
So in my opinion, the significance of this show by the United States is far greater than the actual significance. After all, first of all, this cannot solve the problem of insufficient supply of Huawei's core chips - even if it purchases AMD's chips, it can only put If it doesn't work or won't be used soon, then why buy it? Second, it cannot solve Huawei's basic needs in any key field. For example, for Huawei's consumer business, if Qualcomm agrees to sell chips to Huawei, then it can be considered a solution for Huawei.
The United States has recently more clearly explained the prerequisites for "release" - "as long as these products are not used in 5G technology."
As we all know, Huawei’s most core needs in the consumer business and communication equipment fields are currently focused on 5G-related components and chip products.
Therefore, the products that the United States allows to export to Huawei are actually products that Huawei does not have high demand at this stage or are related to its 4G business - this cannot solve Huawei's top priority.
So it is obvious that the United States is not sincere in this move, so what is its purpose? Just to disgust Huawei? Apparently not.
From the above we can know two things: First, Huawei's most fatal flaw currently is that it cannot produce chips. This situation will cause Huawei's leading technological advantages to stagnate. Second, the recent actions of the United States are not at all intended to ease the confrontation with Huawei, but have other purposes.
The biggest purpose is to exchange space for time.
The space here refers to the absolute advantage the United States has gained by virtue of its decades-long technological strength ahead of China in the semiconductor field. Its sphere of influence in the semiconductor field is much larger than that of China. Time refers to the time it takes for the United States or 5G technology that is absolutely controlled by the United States to develop.
Of course, the United States will worry that China will start independent research and development after Huawei is sanctioned. However, in comparison, the United States is more afraid that if Huawei is not sanctioned, Huawei will step on their shoulders to help China obtain the advantages of the 5G communications era. Core position - So in fact, the United States is also taking a risk. It wants to bet that the U.S. technology community can take the lead in 5G technology over China before China masters advanced lithography machine technology.
In fact, the United States does have a great advantage. The main problems currently faced by Huawei or Chinese semiconductors include not only the large-scale technology and equipment involved in high-end lithography machines, but also chip design tools - Design software, chip architecture, etc., as well as advanced process technology - TSMC's advancement not only relies on EUV, but their own technology research and development of manufacturing processes is also a key factor.
So the choice before everyone in the American technology community is clear at a glance: On the one hand, China needs to catch up with the technical level of a series of lithography machine-related companies such as ASML, Lam Semiconductor, Texas Instruments, etc. (even because of patents There is another way to go) to catch up with the technical level of network architecture and IP cluster-related companies such as ARM and Intel, as well as the technical level of wafer foundry companies such as TSMC and Samsung; on the other side, the United States only needs to catch up Take a look at Huawei’s leading advantages in the 5G field.
How many years will it take for China to catch up with this gap? How long will it take for the United States? American elites hardly need to think too much. They are willing to take this risk. They do not believe that China can win this time, because in their eyes, the Chinese are no longer like they were when they launched the "two bombs and one satellite" decades ago. United as one.
In addition to this ever-present ultimate goal, another purpose of the United States is to further differentiate and cut off Huawei’s influence.
In fact, even if the supply license is approved, the amount of money involved in the products that these companies can provide is not large. Therefore, the so-called easing the economic pressure on U.S. semiconductor companies, in my opinion, has far greater symbolic significance than In practical terms, what can this level of supply alleviate?
However, this behavior sends a signal that the United States will no longer maintain its crude interference in the semiconductor industry. It does not intend to continue to expand and "implicate", but at the same time it puts forward an implicit condition—— "You can no longer flirt with Huawei."
The intention of the United States is obvious: In the future, you will have to adapt to a new era in which Huawei will gradually become marginalized.
In other words, in order to ensure the ultimate goal of "exchanging space for time," the United States has further weakened Huawei's technological power.
This is a conspiracy and a "idle hand", but it may not be useless.
Of course, if we look at this matter with an optimistic attitude, there are other possibilities. For example, the United States is preventing our country from using national efforts to develop semiconductor-related technologies and helping Huawei overcome what it needs. There are various technical difficulties, so it has softened its attitude - the purpose is still to delay the war.
Then we should increase our efforts to promote the technological development of independent chips.
According to the latest news, Huawei is planning to build a chip factory in Shanghai that does not use American technology. According to people familiar with the matter, the factory is expected to start with manufacturing low-end 45nm chips. , aiming to produce 28nm chips for "Internet of Things" devices by the end of 2021, and 20nm chips for its 5G telecom equipment by the end of 2022.
From this information, we can get at least three pieces of information.
First, the country is carrying out a long-term semiconductor development plan.
Why do you say that? The key lies in the location of Huawei's factory. Shanghai is the most advanced region in China's semiconductor industry. In Suzhou and Shanghai, the semiconductor-related industry chain has a long history of development. In the ICT field, almost all major domestic manufacturers such as wafer foundry, EDA software, packaging and testing are concentrated in this area. At the same time, this is also the area that is closest to the advanced technology and management level of international counterparts. Huawei can maximize its progress here. Make technological breakthroughs without worrying too much about resource allocation issues. In addition, Shanghai's high-quality higher education and scientific research talents can also provide assistance to Huawei's development.
By allowing Huawei to build here, we can also see that the country has expectations for Huawei to play a leading role in the semiconductor field in the future. Huawei can indeed bring some changes to industry chain companies.
Second, Huawei’s layout focuses on 5G-related chip products.
This shows that Huawei has made considerable adjustments to its business after recognizing the facts. Afterwards, Huawei's 5G business is expected to shrink its front line and temporarily return to China. However, Huawei’s layout of 5G technology has not stopped. It is doing research and development on the Internet of Things ecosystem, and it has also made plans for the production of chips. For example, it has left plans for independent production of chips for Internet of Things equipment, and the manufacturing process The short-term goal of process improvement is to achieve the supply of 5G communication equipment chips - this shows that Huawei will give priority to ensuring the supply of chips in the field of communication equipment.
Previously, Huawei’s Tiangang chip required a 7nm process, which Huawei cannot break through on its own in the short term. Therefore, I think that while Huawei is saving itself, it may also place its hopes on the domestic lithography machine companies. Independent research and development and technological breakthroughs are also an opportunity for foundry companies such as SMIC.
Third, Huawei may have other plans for mobile phone SoC chips, or may choose to retreat tactically.
From this information, we can’t see anything about mobile phone chips. There are two possibilities here. One is that Huawei still hopes that the United States will allow Qualcomm to sell its high-end chips to Huawei. This is actually not impossible. After all, the US election is about to be settled, and there are certain opportunities for change, but this "permission" cannot solve the problem. Huawei's strategic goal, so even if Qualcomm chips can really be used, Huawei will only serve as a transition.
Another possibility is greater - to jointly research high-end photolithography machine technology and process technology with domestic companies. This is a task that requires longer time and technical investment, but this is This is the trend that the United States is most afraid of. Once relevant Chinese companies make greater progress in this regard, then relevant US companies will be under pressure. The United States may even unblock Huawei before China can achieve a major breakthrough - in order to divide the country. Regarding the research and development of basic technologies - this is also worthy of vigilance.
This requires our focus and persistence on strategic goals.
Judging from the above three news, Huawei's situation has actually improved substantially - China's entire scientific and technological community, including national power, has noticed our "shortcomings" and realized the development The necessity and importance of semiconductor technology.
So although Kirin stopped, Huawei did not stop, and China Science and Technology started mobilization.
If China’s semiconductor industry can compete with the United States in five years, or even come close to it, I think we will all be "grateful" to the United States for its sanctions and sanctions on Huawei in the past two years. It is precisely because of the US's relentless sanctions that we have been "awakened" again.
Two years ago, with the rise of China's emerging technological forces, impetuousness was already prevalent. The prosperity at that time was more like a castle in the air, but as people in the game, we enjoyed it.
It was at this time that the United States forcefully poured cold water on Chinese companies. Although we were seriously injured, we finally woke up and saw many things that we had not seen before. The truth - I think the current "technical bullying" is more similar to "paying tuition" at this level. Although tuition is very expensive, we don't get nothing.
What we need to do now is to seize the opportunity while we are awake again and spare no effort to quickly make up for this "shortcoming" discovered and pointed out by the Americans.
The biggest problem in China at present is that the voices in the market are too confusing. National support for the semiconductor industry has actually begun, but there is still a lot of "noise" in public opinion. In my opinion, the state’s support for the semiconductor industry is not just to help Huawei solve the problem of difficulty in chip production. It is fundamentally about supporting the entire 5G industry. Therefore, it is a major opportunity for many domestic technology companies. However, because of its technological leadership, Huawei is a company that "pays more" at the current stage.
We must recognize the fact that China’s 5G take-off cannot rely solely on Huawei. It is also impossible for Huawei to deploy its business to every aspect related to semiconductors by relying on one company, because the development of semiconductors Huawei is not the only one benefiting from technology.
Therefore, lithography machine companies must work harder and cooperate with the country in taking the research and development route of high-performance lithography machines; EDA software companies must also work harder, chip architecture technology R&D companies must also work harder, and even mobile phone companies should strongly support the basic technology development of domestic companies. Research and development... Because in the future, the pursuit of catching up with the United States will inevitably be based on a cluster model.
In one sentence, the United States is using the power of many companies in Silicon Valley to encircle and suppress Chinese technology. Then Chinese technology will inevitably produce many companies that are comparable to those companies in Silicon Valley. When the time comes, Huawei will face a lot of difficulties. Then there is Silicon Valley, and it may just be a company in Silicon Valley, such as Apple, so China's technology will win even if it doesn't win.
Fortunately, such opportunities currently exist in the Chinese market, as well as such market capacity and development environment. This is definitely a good era for Chinese technology companies.
Perhaps the only thing missing is a very ancient spirit - the road is long and long, and I will search up and down.
There is no need to rush, as long as the direction is right, as long as we persevere and unite as one, I think the United States will have the opportunity to experience "China speed" again.