The Millennium Semiconductor Survival Guide

"Xinxin Technology has entered into a comprehensive long-term licensing agreement with ARM, which allows Xinxin Technology to use all of ARM's current and future IP within the term of the agreement.

This is an unprecedented agreement, Xinxin Technology has jumped from a customer of ARM to a strategic partner."

In this day and age, no one knows the significance of this incident. The media in Silicon Valley, for the sake of Zhou Xin, reserved a corner for this news in the corner of their magazines.

Portal websites such as Sina, Netease, and Sohu have far more reports on this matter than Silicon Valley Magazine.

The agreement was signed with Xinxin Technology as the main body. Xinxin Semiconductor is a secondary company focused on chip design under Xinxin Technology.

Long-term licensing agreement, how long? The period is twenty years, which is already the longest period under English patent law.

Zhou Xin intends to extend this time for another ten years in 2010.

At the beginning of the Internet bubble, Zhou Xin acquired more than 40% of ARM's shares. Later, it became more and more difficult to acquire ARM's shares.

Fortunately, through Warren Jensen, a 7.6% stake in ARM was acquired from an investment institution at a premium close to 100%.

So far, Zhou Xin's ARM shares have added up to more than 51%.

Holding more than 51% of ARM's shares also allows Zhou Xin to make a final decision on the authorization of ARM's technical architecture.

Guan Jianying took this authorization and returned to China with great enthusiasm.

The value of the mobile Internet has not been reflected, and the handheld computer is only a very narrow business direction. Using the analogy of the business types of later generations, it is equivalent to electronic paper books.

There are big companies doing it, there is a fixed audience, and new players are constantly entering the field, but the problem is the lack of imagination and development prospects.

Guan Jianying doesn't care too much about these, because it is not difficult for Xinxin Semiconductor to survive on the chip design of handheld computers.

At least this field is more than enough to support a chip design company like Xinxin Semiconductor, and the long-term technology license signed with ARM allows them to stably conduct research and development around the ARM architecture.

The only one who feels uneasy is Zhou Xin, who doesn't trust the agreement on paper so much. Even though he is the largest shareholder of ARM and has veto power, ARM and Xinxin Technology have signed a technology licensing agreement for up to 20 years.

Even with so many safeguards, he still has a premonition that the agreement will be useless in the future when the situation deteriorates.

The technical agreement can be a dead letter, the technical authorization can be terminated at any time, and the signed contract can stop the supply on the grounds of safety. Only technical talents, technical data accumulation, manufacturing process reserves, etc. are the reliable foundation.

"How to evaluate the long-term technology agreement signed by Xinxin Technology and ARM?"

"ARM is the largest technology provider in the field of mobile chips. It is mainly responsible for providing design solutions to help chip design companies carry out secondary design based on ARM's technical solutions. Generally, cooperation agreements between ARM and chip designers are signed according to specific cores. protocol.

For example, Samsung signed a technical cooperation agreement with ARM around the architecture of ARM v6. ARM will provide software development tools, technical materials and sales license certification for the v6 model to Samsung.

Around mobile chip architecture licensing, ARM and a large number of semiconductor companies have carried out technology licensing, including a series of well-known companies in the semiconductor field such as Philips Semiconductors, Samsung, and Texas Instruments. As a new player in the semiconductor field, Xinxin Technology is only commendable in that its founder is Zhou Xin, the richest man in Asia.

In addition to being the richest man in Asia, Zhou Xin is also a doctoral candidate in the field of semiconductors. When Zhou Xin established Xinxin Technology in Zhangjiang, he was almost skeptical about this matter.

He thinks that Zhou Xin is just playing tricks to please Hua Guo officials. Judging from this unprecedented long-term technology agreement with ARM, Zhou Xin's relationship with Xinxin Technology is not just a simple game.

This technology licensing agreement is a bit similar to "supermarket procurement". If other semiconductor companies buy raw materials one by one from ARM, they will perform secondary processing on this raw material and process it into corresponding dishes.

Then the agreement signed by ARM and Xinxin Technology is a big sale in the supermarket. You can choose whatever raw materials you fancy. In the future, when Xinxin Technology's chip design capabilities gradually improve, this agreement will bring huge benefits.

Huaguo will also have a company that can truly compete with other semiconductor giants in the field of chip design.

Regarding Zhou Xin’s establishment of Xinxin Technology in China before, I was even skeptical that someone said that Zhou Xin was trying to cheat financial support. I just want to say what? Not to mention Zhou Xin's net worth of 100 billion U.S. dollars, even if he only has a net worth of 1 billion U.S. dollars, he will not go back to China to cheat on financial subsidies.

In addition, I think Zhou Xin has high hopes for Xinxin Technology. Because I found out that among his companies, Zhou Xin would name it after his own name only if he paid enough attention to it. "

On Doubt, there is a question about Xinxin Technology's obtaining ARM authorization, and there are very few answers to this question.

The main reason is that it is too professional. At present, most Chinese people do not know what ARM is, let alone what IP authorization is.

It's nothing that the outside world doesn't understand, there are always people who understand in Huaguo officials.

Suspicion did not know that Zhou Xin controlled ARM, but Huaguo officials received the news.

From Huaguo's point of view, Zhou Xin has initially proved his ability and willingness.

They just wait for the mobile phone that Zhou Xin said to come out. If the matrix mobile phone can knead the entire industrial chain together and form a snowball effect, then Huahui will do everything possible to help Zhou Xin clear the internal resistance.

From Shen Hai's point of view, the current progress is enough to satisfy them.

Under the urging of the superiors, Hua Hong Semiconductor regards the prototype testing of the Xinxin lithography machine as the top priority during this period.

The technology of Xinxin lithography machine comes from Nikon, and almost all the accessories come from abroad. At present, it is only the integration of technology and lacks original technology.

Even so, it is already a rare breakthrough for Huaguo in the field of chips.

One is because other Huaguo companies will not invest money in this field, and they will lose money even if they do not make money, and there is no hope in sight.

The only thing that is possible is the official organization’s attack. The official attack has decided not to buy technology from abroad. Pure self-research will take too long. After you manufacture the prototype of the lithography machine, you will be several generations behind foreign countries. .

After the prototype is made, it still needs to be tested. The lithography machine has entered the production process of the chip manufacturer, and the other party has extremely high requirements for the yield rate.

Because chip foundry is an industry that requires very high precision control. Military chips are not cost-sensitive, so it doesn't matter if the cost of self-developed products is high and the yield rate is low.

Civilian chips are not the case. The yield rate is related to the lifeline of chip foundries. A low yield rate means that your cost is high. The cost of building a factory is already high enough. If the cost is still high in the chip production process, how can you make a profit?

It all depends on self-research-backward technology-manufacturers with backward technology are unwilling to use it-cannot enter the production environment-low yield rate, which can be said to be a vicious circle.

After groping for a long time, Huaguo realized that the field of semiconductors must rely on the power of marketization, and it is not in line with objective laws to rely on academic research and development.

In future generations, if ASML did not sell the most advanced lithography machines to Huaguo, there is a high probability that Huaguo's lithography machines will always lag behind.

"We mainly purchased the Nikon NSR-S201A lithography machine technology.

This is the world's first excimer laser stepper model exceeding 250nm. The ArF system for the 193nm process targets the 130nm region.

We optimized it and adopted a lens-based scanning stepper to meet the dual requirements of reducing the feature size and expanding the exposure area of ​​semiconductor chips.

During the testing process, the process of Xinxin lithography machine is close to the limit value of the technology, but there is still a lot of room for improvement in the chip yield. "Lin Benjian said when reporting to Shen Hai.

The Nikon technology purchased by the new core lithography machine is the second-generation technology that was launched in the market in 1995 and then optimized.

The same lithography machine does not mean that the technology is sold to you, and you just make it, and it is difficult for you to achieve the effect of the other party.

A very simple example, with the same teaching materials, the same tutoring materials, and the same teacher, some students can get 100 points in the test, while some students can't even get 60 points in the test.

Nikon's NSR-S201A sold more than 200 units worldwide in the first year after its launch, after long-term production verification and continuous optimization by Nikon engineers and chip manufacturers.

If it weren't for Silicon Valley and Wall Street to choose ASML together, Nikon lithography machines may not be reduced to a market share of less than 1%.

The Xinxin lithography machine can reproduce a sample in half a year, which has far exceeded Zhou Xin's expectations.

After Zhou Xin got the news, he thought to himself that I don't have a system, so Lin Benjian doesn't have a system, right?

Shen Hai's leaders did not ask, why not buy the most advanced technology, such a low-level question.

"When will we be able to catch up with Nikon's yield rate?"

"It depends on the progress of our cooperation with Hua Hong Semiconductor, and it can't just be manufactured in the laboratory. We need to really enter the production process." Lin Benjian said.

In Lin Benjian's view, even if Xinxin lithography machine is sold to Hua Hong Semiconductor for free, the yield rate must be increased.

The full name of HP CD is half-pitch CD, which is a technical indicator specially used to describe the lithography resolution in the field of lithography machines.

For example, for 193i lithography machine, the limit value of HP CD is equal to 38nm. And the light source of EUV NXE 3400B can reach the limit of 13nm.

The node CD is completely different from the HP CD. It is a concept of a semiconductor device. The technology node mentioned by the online media is node CD.

Generally speaking, a node CD is about equal to 1/2 HP CD.

In actual manufacturing, it is impossible to use the limit value of HP CD to manufacture, and it will be relaxed a lot.

What this means is that even with the same lithography machine, the light source will affect the process.

"Hua Hong Semiconductor will fully cooperate with the work of the Xinxin lithography machine, and we will do everything possible to cooperate."

"In fact, Shenhai is conducting research on how to develop high-tech industries, including myself, who have done a lot of research on this topic.

The trend of thought is different at different stages. In 1980, I was sent abroad by an organization to study business management.

The trend at that time was large enterprises, and there must be large companies with sufficient internal scale and economies of scope. Only such a framework can be innovative, and can breakthroughs and developments be made in technology.

This was the view of Joseph Schumpeter, including Alfred's later book, Size and Scope, which further added force to this notion.

In other words, industrial competitiveness is almost entirely dependent on internal cultivation capabilities. At that time, foreign semiconductor companies also adopted vertical manufacturing methods.

From chip design to chip manufacturing, to packaging and testing, everything is done by one company.

The reason why there are 907, 908, and 909 projects is based on this kind of thinking.

Hua Hong Semiconductor is a product of Project 909. After researching Samsung at the end of 1995, Comrade * believed that we should also have our own large enterprise. Later, he even said that we should also develop the semiconductor industry in Huaguo.

Many of our leaders are very concerned about the development of Huaguo's semiconductor industry, and the 909 project is also planned at the national level.

We have seen the success of Samsung, so we want to build Hua Hong into a Samsung model.

But until today, the decentralized model of Silicon Valley seems to be not bad, even showing stronger vitality than large enterprises.

When I went to Yanjing to give lectures not long ago, the professor of economics at Yanjing University who was in charge of the lecture put forward a different point of view, that is, it is unhealthy to overemphasize internal creation and management of economic capacity.

It ignores the subtlety and historical peculiarities of industrial evolution. Schumpeter's view is a naive solution to the atomized competitive model of neoclassical theory and policy. The vertically integrated manufacturing model will instead cause new losses and oversimplify the risk of the problem itself.

The relative merit of firms depends on whether the innovation is systemic or autonomous. When innovation is systemic, an enterprise-like structure will be more conducive to rapid technological progress. This is because systemic innovation requires simultaneous change in many different stages of production, and co-ownership of complementary stages reduces the transaction costs of persuasion and coordination.

In contrast, it is plausible that networks, including typically extreme 'markets', are more advantageous when innovation is autonomous.

When a technological change in one stage of production can proceed without requiring a corresponding change in other parts of the system. In this case, the firm's advantages in persuasion and coordination are offset by the economics of specialization and networking capabilities of the innovative small business cluster.

The systemic or autonomous nature of innovation is neither completely exogenous nor completely driven by technology.

The structure of the organization helps to shape the pattern of innovation, which in turn affects the subsequent organizational structure. In short, the theory of organizational structure is also part of the theory of evolution of social institutions.

Whether the model of large enterprises is better or the model of small and medium-sized enterprises is better, we don't know which one is better.

What we pay attention to is to test the truth through practice, so Huaguo chooses to explore the model of small and medium-sized enterprise groups in Zhangjiang.

Pin your hopes on Hua Hong and Xinxin Technology, and hope that you can play a good synergy and drive a group of small and medium-sized enterprises in the semiconductor field to grow together. "

This chapter wants to explain why Huaguo chose to start the University of Malaysia project at that time, and why scientific research-oriented lithography machines were not feasible.

The appearance of Zhou Xin gave Huaguo a new route choice in technological innovation.

The number of words in this chapter is only 4k, but the information density is extremely high, which is more tiring than writing 6k.