The first is the question of capital investment: to build a whole set of production lines by yourself, you must first invest a lot of money. Huawei's annual profit is tens of billions. Take 20 19 as an example, its net profit is 62.7 billion, equivalent to nearly 9 billion dollars. This fund looks quite large, but if it really hits the construction of IDM, it will be a drop in the bucket.
It is impossible to master the OEM technology of various devices at the chip manufacturing end, including EDA software for chip design. Even a small-scale IDM needs tens of billions of dollars, which is a burden for Huawei.
Huawei also has other business lines, especially in the field of communication, which is its foundation. It is impossible to put all the funds into IDM construction, otherwise it may affect its main business. However, without a lot of capital investment, the construction of IDM will be a long and distant thing, and it will also bear fruit in a short time.
Secondly, the talent problem: IDM construction is not enough with money, but also with people. There are not many people in China, and it is rumored that Huawei is in contact with the domestic personnel reserve.
At present, the semiconductor industry in China is still quite short of personnel. In recent years, the rapid development of the Internet has caused a large number of talents to flow to the Internet, resulting in serious brain drain in other fields. The relatively low salary of the physical manufacturing industry triggered a wave of "persuasion", and many people related to semiconductors also ran to the Internet.
Therefore, although the Internet in China is very developed, it is actually not a good thing. First, this development is illusory, but it is developed at the application level. Once the core technology is cut off, it will fart immediately. Second, it will lead to brain drain in many other fields and affect long-term development.
Insufficient development of the whole industry: at present, the overall IDM construction in China is also very lacking, and many core areas are lacking or backward in technology. Mask aligner, 28 nm node, is planned to go offline this year. If it can be mass-produced, it will take 1~2 years, and the R&D planning node of EVU mask aligner is in 2030. In terms of OEM technology, SMIC has only mastered 14nm, and the real 7nm process will not be realized until N+2 process, which will take at least 2-3 years.
If Huawei is allowed to build its own IDM, it is impossible for mask aligner, a device with the most domestic 28nm nodes, to develop itself. At present, this equipment is being developed almost all over the country, and it is impossible to rely entirely on Huawei's own strength. Huawei can do its own OEM, and I think it will progress faster than SMIC.
There are chip design software such as EDA, which is another dimension. If you want to turn Huawei into a hardware-based manufacturer, then suspend it.
Lscssh technology officer's point of view: On the whole, it is quite difficult for Huawei to build its own IDM, and its domestic foundation is weak. Let Huawei build from scratch. This difficulty is by no means unimaginable to outsiders. Of course, the existing gossip does spread that Huawei wants to establish its own IDM, at least for now. I think Huawei really has similar ideas, but it is hard to say whether it really needs to be built or implemented.
This wish is actually not good, and Huawei will not do so.
To understand this problem thoroughly, we must first talk about the basic concepts.
There are two modes of chip enterprise, so it's really not that easy to think about IDM. But let Huawei give up IDM just because of difficulties? Far from it, let's look at the second reason:
Money, people and time all depend on luck. Chips are so expensive to manufacture.
Let's talk about chip design first. Without hundreds of millions of funds, let alone designing chips. Take Xiaomi 20 17 as an example. The expression of S 1 is not satisfactory, and S2 has not yet appeared. After hitting the wall, Reebs sighed with emotion: "The chip industry started at 65.438+0 billion, and the result was 654.38+00 years." When Xiaomi developed the 2855438+0 chip, it also used high salary to dig people, and even some technical data could be bought with money, but it cost 1 billion, but what it got was the backward chip technology of 28 nanometers.
This is just the chip design industry. The chip manufacturing industry is even more terrible. Look at TSMC's annual R&D investment. In 20 19, the investment of R&D was USD 3 billion, accounting for 18% of the income of that year. This is only an investment in improving the chip manufacturing process. Intel is even more horrible, reaching $65.438+0.34 billion, exceeding the sum of all R&D investments of Huawei in 2065.438+09. Even so, Intel still failed to upgrade its 10nm process to 7nm, because making chips requires not only money, but also sophisticated talents, which is time-consuming and laborious, and the most annoying thing is personality and luck. ...
Simply put, chip manufacturing is about throwing money, people, time and luck. But this is not the fundamental reason why Huawei gave up IDM.
The national team is surprised ahead, and the national chip industry chain is about to rise. Actually, I understand the meaning of the topic. What I really want to solve is the problem of "sticking the neck" in the United States, instead of giving Huawei the heavy responsibility of "building a completely independent chip industry chain in China". In fact, a few more Huawei can't afford this burden. In fact, in the establishment and improvement of the chip industry chain, the country has already had a layout:
Speaking of which, everyone must be very interested to see what "02 Special Project" is and keep your eyes open (your Baidu can't see it clearly):
Do you feel tears in your eyes? Did you suddenly find that our country is still very strategic? As early as 2006, more than ten years ago, it has been carefully laid out for the construction of the national chip industry chain.
Some people don't believe that the special project has been established. And the result?
Although most of them still have a certain gap with the most advanced level in the world, we have been working hard, especially after the 20 18 ZTE incident. From the state to the industrial capital, we have clearly realized that the independence of the chip industry chain is the general trend, and "Polish" has disappeared. In just three years, China chip industry-related enterprises can be said to do their best, and every day is a new atmosphere.
Everyone must have noticed the newly emerging chip giant in the A-share market: SMIC. Two months ago, the market value of SMIC was still hovering around 1000 billion. Today, on July 6, 2020, the market value of SMIC directly exceeded 600 billion. Why?
Simply put, SMIC is the final result of the above-mentioned "02 Special Project" (not exact, because SMIC's current product line adopts international technologies including American technologies). More precisely, SMIC is the most authoritative and efficient accepting institution of the national "02 Special Project" and plays a very important strategic role in the closed-loop integrity of China's integrated circuit industry chain. Any latest achievement of "02 Special Project" can directly or indirectly improve SMIC's chip manufacturing capability, which means the improvement of China's chip manufacturing level. At present, the gap between SMIC and TSMC, the most advanced chip manufacturer in the world, is about 3-4 years.
The national chip industry chain built by many enterprises and universities behind SMIC is carrying a heavy burden for itself, Huawei, the motherland and all of us. In the current environment, I believe that it won't be long before China's chip manufacturing will catch up with the most advanced international level.
Generally speaking, Huawei is excellent enough, and we should not put too heavy a burden on one of its enterprises. In the field of chips, more and more national enterprises should and have rushed to the forefront. Come on, China! Come on, China Core!
Nothing is impossible. Now that technology is so advanced, those who engage in this kind of research and development will definitely take into account today's situation. Huawei has R&D capabilities, and there are many domestic technology companies. It is not impossible for Qi Xin to work together to tide over the difficulties, but the road will be long. I thought that equipment research and development in those countries started from scratch. When you encounter something, you must believe that you can overcome it.
IDM refers to integrated device manufacturing and integrated equipment production mode, which is a design and production mode in the field of chips. To put it bluntly, this model means doing everything yourself, from chip design, manufacturing, packaging to testing, and one company covers the entire industrial chain. This model requires very high strength of the company, and only a few companies in the world have such strength, such as Intel, Samsung, Texas Instruments and other very large companies.
Celebrities in the industry broke the news that Huawei is building an IDM model and is cooperating with Hua Hong Semiconductor and SMIC. Then the news of Huawei's recruitment of lithography process engineers spread, and 16 dug up Huawei's lithography technology patent. Now Shanghai Microelectronics is complaining that Huawei has been digging people frequently recently, and Huawei has obtained the employee address books of several semiconductor equipment manufacturers in Shanghai. In addition to the general manager level, most of them have received calls, and some even directly put down their jobs.
Chip design EDA software three American companies account for most of the shares. ASML mask aligner is two generations ahead of Shanghai Microelectronics, TSMC has developed 5 nm, and SMIC is only 14 nm. In this case, is Huawei realistic in IDM?
My answer is: reality, although challenging, but the possibility of success is very high!
In the field of EDA software packages, whether we cooperate with domestic software companies such as Huada Jiutian or do it ourselves, Huawei's software research and development strength can make it in one year. After 2-3 years of application iteration, the software will be very mature.
In the field of mask aligner, although there are 400,000 parts in mask aligner, it is estimated that there are hundreds of parts with no breakthrough in domestic technology or process accuracy. Core optical lens, extreme ultraviolet laser source and two-stage can make breakthroughs in technical research. Moreover, the principle of mask aligner has already broken through, and the bottleneck lies in the process accuracy of parts. Solving process problems is Huawei's strength. Now Shanghai Microelectronics has broken through the 28 nm mask aligner, and Huawei has hired many engineers from mask aligner, which means it can start directly from 28 nm.
1. Capital: Huawei's cash reserve140 billion yuan. If dividends are not paid in the past two years, Huawei's cash reserves can easily reach 200-300 billion, with an investment of 1000 billion. Money is not a problem.
2. Talent: Huawei has strong software strength and world-class material experts. Experts from Germany and Japan can be used in the field of optics. Mask aligner has recruited many people from Shanghai Microelectronics, and Huawei's powerful IPD R&D technology makes it possible to make a rapid breakthrough in technology.
3. Market: Don't worry.
Huawei has a strong CMM capability maturity model for IPD R&D process and software evaluation system, so once the project is formally established, it should be able to develop 28 nm mask aligner in one year, 7- 14 nm in two years, 5 nm in three years, ASML in five years and TSMC in five years, plus SMIC.
Once Huawei equalizes the United States, ASML and TSMC in EDA software, mask aligner and chips, China's chip industry will crush the United States, China will change from an importer of chips to an exporter, EDA software companies in the United States will go downhill, Intel in Qualcomm will go bankrupt, and TSMC will close down soon. .
The United States has not banned China's software and chips at all. Huawei IDM is like two bombs and one satellite to China's scientific and technological community.
From now on, Huawei will become an all-around player:
1, communication equipment pk Ericsson Nokia
2, enterprise business pk Cisco
3, mobile phone tablet wearable device pk Apple Samsung HP Dell
4. HarmonyOS system pk Google
5, automatic driving system pk Google
6. EDA software mask aligner chip pk Three EDA software companies in the United States Qualcomm Intel ASML TSMC
Basically, it is a one-on-one fight with several American masters.
If Huawei is tied to the EDA mask aligner chip within five years, there will be nothing abroad to ban Huawei in the future. 10-20 years later, Huawei's output value will reach more than 500 billion US dollars, surpassing Apple, Toyota and Samsung and becoming the world's first technology manufacturing enterprise.
The success probability of photomask mask aligner is over 80%, and Huawei is right to engage in IDM! Very supportive, I hope Huawei will succeed!