Now it is unrealistic to think that all upstream and downstream manufacturers in the semiconductor industry chain will choose domestic manufacturers and form their own system.
Semiconductors involve too many equipment, including air filtration equipment and water production equipment. Currently, no domestic manufacturer is doing this.
Chip factory construction has high requirements on the environment. The dust concentration in the air needs to be reduced to 0 through air filtration equipment. When the line width of the chip is less than 0.5 mm, chemical contamination will be the main factor affecting the yield.
Therefore, for chip foundries, they need to be equipped with specialized precision instruments to filter substances in the air.
Zhou Xin never even thought that the entire supply chain could catch up with the international advanced level within ten years. The upstream semiconductor materials were an even greater difficulty.
In the semiconductor field, technology is not the biggest difficulty. The most important thing is that products can enter the production environment and be constantly updated and iterated.
Therefore, if Americen had not banned the export of semiconductors, it would have been almost impossible for China's semiconductors to catch up with the most advanced technology. Of course, the chances would not be great after the ban, but the chances would be much greater than before.
This kind of software is somewhat similar to industrial software. If you don’t actually run it, you will never be able to create truly usable industrial software.
After listening, Guan Jianying said: "Newman, I roughly understand what you mean. You only want to make photolithography machines and analog ICs, these two technology nodes. At the same time, their upstream and downstream companies do not rule out cooperation with other companies."
Zhou Xin said: “We don’t have the conditions to do it all ourselves, nor do we have the conditions to find Chinese companies to do it all.
My investment company in China has investigated the development of integrated circuits in China. It is not ideal. There are very few companies in this area, and most of them are concentrated in the packaging and testing process with relatively low technical content.
On the basis of ensuring our own survival, we also want to take care of Chinese semiconductor companies upstream and downstream of the supply chain.
After listening to this, do you feel that my goal is not harsh? "
Guan Jianying thought about it. He has been working in analog IC for more than 20 years. He has been involved in everything from analog IC to chip production, so he has a deep understanding of the entire semiconductor industry chain.
“Newman, the goal you set is not too demanding, but it is not easy to achieve either.
We need to respect objective facts. China lacks talents in the field of integrated circuits, let alone engineers who have actually worked in advanced integrated circuit companies.
In order for us to enter the production stage as soon as possible, we need to find engineers with mature experience.
Fortunately, the analog IC field has many engineers who have rich experience in the analog IC industry, and they are not opposed to working in Shenhai.
But for engineers in the field of lithography machines, we can only find experienced engineers from a limited number of lithography machine manufacturers such as Nikon, Canon, and ASML.
They are foreigners and difficult to dig out. "
Today, Nikon and Canon are still the leaders in the industry, while ASML is a rising star. It is precisely because Nikon is the leader in the field of lithography machines that Nikon engineers are difficult to poach.
Twenty years later, it would not be difficult for Nikon to recruit photolithography machine talents.
Of course, at that time, everything that Nikon's technical staff could poach had already been poached by ASML, and the rest either had average skills or could not be poached at all.
Zhou Xin also understood what Guan Jianying meant, saying: "You don't have to worry too much about these issues.
The technical director I am looking for for the lithography machine is Lin Benjian. He will find some engineers he knows to come over. We will build the framework first, and then recruit a small number of experienced engineers from China. There are still talents in China who have worked on lithography machines before. , but their photolithography machines are laboratory products.
But I believe they will adapt very quickly.
As for the analog IC field, I don’t want to find too many experienced engineers. Only the team leader needs experienced engineers. For team members, we all need graduate students and Ph.D.s from China who have graduated from relevant majors.
We offer them high salaries and then select the best from the best, adopting the model of one old man leading many newcomers. "
After hearing this, Guan Jianying immediately objected: "Newman, I have been working in the analog IC field for many years, and this is absolutely not acceptable.
We need to get an analog IC company up and running as soon as possible. We need to create a hit product. After this product is designed, it will be produced by TSMC or Zhang Rujing's chip foundry, and then it will have good sales results and be recognized by the market.
This is Xinxin Semiconductor’s top priority.
Then slowly carry out iterative optimization around this flagship product, from the first generation to the second generation, and slowly develop new chip lines.
This way we can survive.
There is only one old man in each group, and all the others are fresh graduates. Hit products will never be produced.
Because technology is advancing, products designed this year may not have the advantage next year.
Time is so important.
I do not exclude recruiting new people, including Chinese graduates. First of all, the number is limited, and secondly, they must come in with a learning attitude. It will take at least two years for them to adapt before they can truly become relatively skilled veterans.
Freshly graduated college students, if they are not majoring in microelectronics.
They have never even used Cadence Virtuoso. It takes at least three months to half a year to teach them to use Cadence Virtuoso. They only know how to operate it and know what analog circuit design is and how to do layout viewing and circuit simulation.
Even if you have a PhD in microelectronics, R&D in industry and R&D in laboratories are completely different concepts.
Their understanding of basic concepts in analog IC design needs to be corrected. Many novices tend to ignore some basic concepts in circuit design, such as noise, distortion, linearity, bandwidth, etc.
Because of their lack of practical experience, they will ignore the impact of layout and layout design on circuit performance, and the handling of power and ground planes during design. They will focus too much on the performance advantages of one aspect and ignore other important performance indicators.
They will pay too much attention to the speed of the circuit, which may lead to performance degradation in terms of power consumption, noise, etc.
These questions are very common for newcomers. Twenty years ago, we were still using SPICE developed by Dr. Lawrence Nagel.
Lawrence Nagel is also an alumnus of yours, a Ph.D. from the University of California, Berkeley.
I still remember using SPICE in the early days, then Calma GDS, and most recently Cadence Virtuoso.
When I was still using SPICE, I would make many similar mistakes. Everyone slowly grows from a newcomer to a mature engineer
But at that time, I was working for a mature analog IC company. They had mature products, a mature team, a talent echelon, and a complete training mechanism.
In other words, they have the capital to tolerate mistakes.
But we don't, and I know you have a lot of money, a lot of money, an amount I can't even imagine.
But an enterprise cannot rely solely on your single blood transfusion. Such an enterprise will be matured and it will be difficult for it to have vitality and a future.
Therefore, we need to find mature engineers in the early stage to create a competitive product, and then recruit new people to cultivate reserve strength.
Rather than doing reserve strength training at the beginning. "
SPICE, Calma GDS, and Cadence Virtuoso mentioned by Guan Jianying are all EDA tools used to design analog circuits. There is a chronology between them.
Among them, Cadence Virtuoso is still the mainstream EDA tool today, but Cadence packaged it into something called Cadence Design System, which covers this and implements richer functions.
Zhou Xin said: "Okay, I understand what you mean. I know that Uncle Guan, you have worked in the analog IC field for many years, from a front-line employee to a chief engineer to a vice president. You have your own philosophy on the management of an analog IC company." .
We haven't hired people so quickly yet. Let's build the company and framework first. Are you willing to help me? "
Compared with Lin Benjian, Guan Jianying has just been kicked out of the company he co-founded. To put it simply, when he was frustrated, he agreed very quickly:
"It's okay for me to help you, but my personality is relatively straightforward. As the boss, it's okay for you to have your own ideas, but I will also have my own ideas and persistence.
If we don't have the same ideas, there needs to be sufficient communication between us to reach some kind of agreement. "
Guan Jianying continued: "I never thought that my ideas and my experiences would always be right.
I entered the integrated circuit industry more than 20 years ago, starting with analog ICs. At that time, the industry still adopted the IDM model, which was vertically integrated manufacturing.
Companies such as Intel, Samsung, Texas Instruments, and Infineon all handle chip design, chip manufacturing, packaging and testing, and product sales all by themselves.
Only upstream materials, equipment, EDA and IP cores use products from other companies.
This model is so good. In my opinion at the time, it was a perfect model. From product design to manufacturing, everything was handled by one person, and the product iteration cycle was also shorter.
It is also conducive to research and development, and the response time for back-end research and development results to be verified is very fast.
At that time, no one ever thought about outsourcing chip manufacturing to other companies.
Until Zhang Zhongmou showed up with TSMC, he told manufacturers that you design chips and we will help you with chip manufacturing.
We will bear this part of the risk for you, you just need to tell us your needs.
Just give me the chip design, I will never make my own chip brand.
You know that the risk of the IDM model lies in heavy assets. A 12-inch production line costs US$2 billion.
This requires that the chips you manufacture must be able to be sold.
This has very high requirements on design, manufacturing, management and operation.
Because production lines are subject to depreciation, as technology advances, the original advanced production lines will gradually be eliminated. A production line built with US$2 billion may be scrapped in five years under the influence of Moore's Law.
The IDM model has extremely high requirements on your utilization rate, yield rate and product profit.
That’s why TSMC’s model can be successful. They help manufacturers bear a large part of the risk.
This way the risk is lower and the profit is also lower. "
Today’s first chapter, please vote for me. Is there too much semiconductor content recently?
But the next plot is about to start
(End of chapter)
