There is no doubt that the semiconductor industry is absolutely knowledge-intensive. For details, please refer to our previous report: “Starting with ARM, there is not much time left for the Chinese team”, it is recommended to feel the horror of patents in the chip industry through this report: not making a chip, ARM, which only sells chip design IP, has a market value of up to 40 billion US dollars, and has even become a wrestling field at the national level of China, the United States, Japan and Europe.

This article will take memory as an example to discuss the lack of domestic chip patents.

The reason why memory is chosen is that memory is the largest branch of the semiconductor industry, accounting for more than 30% of logic chips (also known as processors). In addition, giants such as Samsung, SK Hynix, and Micron Technology, which monopolize the storage industry, maintain their monopoly position through layers of patent networks, making competitors and customers miserable. Still fresh in memory.

Another reason is that the two major storage bases in China, Changjiang Storage and Hefei Changxin, will enter the stage of substantial mass production. It is rumored in the industry that the second phase of the National Fund will also be strongly supported, and storage will become the most popular in the domestic semiconductor field in the next few years. Areas of focus for the spotlight.

The two most important products of memory are NAND Flash (flash memory) and DRAM (dynamic random access memory), and the monopoly of these two products is astounding.

In the NAND industry, the combined market share of Samsung, Kioxia, Western Digital and SK Hynix exceeds 80%; while Intel plans to sell its NAND business to SK Hynix in the near future. Once the transaction is concluded, the market share of the leading players will reach 97% ; The monopoly of DRAM is not much less, Samsung, SK Hynix and Micron have a combined market share of more than 90%.

500 billion gamble, how to break the chaos of memory chips

Figure 1: DRAM memory and NAND memory market landscape, Source: Gartner, Trendforce

The reason why such a strong monopoly is formed is not only because the industry is extremely expensive, but the patent wall of storage is high and thick, which has become an insurmountable barrier for latecomers. In this report, we will use three cases to intuitively illustrate the patent problem of storage.


visible difference

Consistent with the monopoly position mentioned above, Samsung, SK Hynix, Toshiba and Micron Technology have an absolute advantage in the number of patents in the current memory patent applications, forming a technological blockade for new entrants, which is what we often say “is better than you”. Work harder than you.”

Only SMIC (SH: 688981/HK: 00981) and Zhaoyi Innovation (SH: 603986) are among the Chinese mainland enterprises. Since 2012, the number of large-scale patent applications has appeared, but there are several orders of magnitude with the leading players. difference.

Taking Zhaoyi Innovation, a village-wide hope in the memory industry, as an example, according to its 2019 annual report, “As of the end of 2019, the company has accumulated 1,195 valid patent applications at home and abroad, and obtained 581 domestic patents, 23 US patents, and 3 European patents. Patents.” That is, there are only 26 patents in effect worldwide.

According to the latest IFI Claims report, Samsung Electronics has 76,600 patents (including foundry, panel, mobile phone and many other patents), SK Hynix has 7,934 patents, and Micron has 7,488 patents. pieces; and this is a competition that only involves the number of patents, not the quality.

500 billion gamble, how to break the chaos of memory chips

Figure 2: Global memory patent applicants and their application trends, source: incoPat, Western Securities


The Tragedy of Jinhua, Fujian

In order to break through the situation of being constrained by the oligarchs of South Korea and the United States, three major storage bases have been established in mainland China since 2016, namely: Yangtze River Storage located in Wuhan, led by Zhao Weiguo, chairman of Ziguang Group; Fujian Jinhua located in Quanzhou, led by Taiwan UMC Provide technical support; and Hefei Changxin located in Hefei, led by Zhu Yiming, chairman of Zhaoyi Innovation.

The total planned investment of the three major storage bases exceeds 450 billion yuan; at the same time, a large number of high-paying engineers and technology are recruited from overseas companies.

It can be said that China has brought out the strongest team and the largest financial support, and rolled up its sleeves to participate in the global gamble of memory.

Although the total investment of nearly 500 billion yuan looks amazing, the annual capital expenditure of the memory industry, which is actually the investment of a few oligarchs represented by Samsung, is as high as 40-50 billion US dollars. In other words, 500 billion yuan, if thrown into the global memory industry to burn, is enough to burn for two years at most.

But unfortunately, one-third of the 500 billion bet has already fallen in front of the patent battle. This is the Fujian Jinhua incident that shook the industry before.

In December 2017, when Fujian Jinhua was making great progress, the US memory leader Micron sued Fujian Jinhua and its partner Taiwan UMC in California, USA, on the grounds of intellectual property theft. In October 2018, the US Department of Commerce placed Fujian Jinhua Being placed on a list of entities that cannot buy components, software and technology products from U.S. companies makes Jinhua the second Chinese company to be banned after ZTE.

After Fujian Jinhua was blocked, UMC could only be forced to retreat. The project that had high hopes in the past was immediately hit by shock. It is estimated that Jinhua, which has been shut down for more than a year, will only be left to sell equipment and workshops. Enter the high-yield global tech gamble.

The other two still in the bureau, Changjiang Storage and Hefei Changxin, tried to avoid repeating the same mistakes through cooperation and agreement, as well as self-development.

Taking Hefei Changxin as an example, in May 2019, Hefei Changxin announced that its DRAM technology came from Qimonda, and obtained more than 10 million DRAM-related technical documents and 16,000 patents through cooperation. Since then, Hefei Changxin has cooperated with Polaris Innovations Ltd. , Lanbo Shi signed an agreement to obtain DRAM chip technical documents and patent licenses; Hefei Changxin is said to have hired former Japanese Elpida executives. That is to say, it has basically obtained the technical endorsement of Japan and Europe.

As a digression, a question that readers may see here is why not develop their own? Although domestic memory companies have some independent research and development, the underlying physical technology of storage is basically the same, and there is almost no way to bypass others. Way to go.

In addition, when Japan replaced the United States as the global storage hegemon, then South Korea subverted Japan. In fact, in the early days, it all relied on the introduction of overseas technology. “Bring it” is not shameful, but it is necessary to be extra careful to be labeled as “stolen”.

500 billion gamble, how to break the chaos of memory chips

Figure 3: Review of the “Lianmei Case”, Source: China Merchants Electronics


Mostek, which “lives” by patents

The third part mentions one company, Mostek in the United States. Chinese people may be very unfamiliar with it. In fact, in the 1980s, Mostek successfully climbed to the top of the world’s memory by defeating the famous Intel.

Later, due to the rise of Japan, it gradually retreated. Finally, Mostek was sold to the French company Thomson at a low price in 1985. Later, with the merger of Thomson and SGS, it was merged into STMicroelectronics.

The magic is that although Mostek’s body no longer exists, there are still a lot of patents in the memory field. STMicroelectronics has earned several times the profits of acquiring Mostek through lengthy lawsuits against other storage companies by relying on these patents.

In fact, it is not STMicroelectronics that makes money through patent litigation in the memory industry. We can jokingly call it “cutting leeks” in the memory industry. Oligarchs often provoke various lawsuits, so that new entrants have poor profits and are often cut at any time.


Revelation and Breakthrough

The support of the national policy for chips has been relatively large. Since 2000, the Chinese government has identified the integrated circuit industry as one of the pillar industries of the national economy, and issued a series of policies to guide and support it. But the real shock to the industry was in September 2014, the establishment of the National Integrated Circuit Industry Investment Fund (referred to as the “Big Fund Phase I”).

Before that, chips were a bitter industry. At that time, the market value of SMIC was less than 100 billion Hong Kong dollars, and China Micro Corporation (SH: 688012) was unknown. Later, the first phase of the big fund invested to support many enterprises, and even had the effect of turning a stone into gold. Of course, this is another story.

Additional support is the establishment of the Science and Technology Innovation Board in November 2018 to open a green channel for start-up semiconductor companies. As of October 2020, among the 188 companies listed on the Science and Technology Innovation Board, there are 19 semiconductor companies with a total market value of 715 billion yuan, accounting for 24% of the total market value of the Science and Technology Innovation Board. It can be said that nearly half of the science and technology innovation board is related to the semiconductor industry.

500 billion gamble, how to break the chaos of memory chips

Figure 4: Sci-Tech Innovation Board Semiconductor Companies, Source: China Semiconductor Industry Association, General Administration of Customs of China

With the strong support of the policy, the industry began to run wild. According to the statistics of industry associations, as of this year, the national investment in the construction of wafer fabs will exceed 1.5 trillion yuan, and various semiconductor industrial parks have sprung up. The capital market is naturally not far behind. The semiconductor industry has started a surging bull market since 2019, and most companies have directly pulled their valuations to the range of the market dream rate.

The semiconductor boom in the primary and secondary markets has also created a group of upstart Tsinghua science and technology men, which has indirectly driven the wages of related practitioners to rise.

But behind the overheating, there are many problems and chaos.

The first problem is excessive investment, especially the repeated construction of low-end production capacity. Recently, the unfinished semiconductor project represented by Wuhan Hongxin has been widely reported by the media recently, allowing everyone to see the chaos in the industry. Excessive investment exists in every industry that is strongly supported by national policies, such as the new energy fraud in 2016.

All the above discussions focus on the issue of “money”. It can be said that the chip industry, which was struggling before, has turned into a sweet treat overnight. Lack of money is no longer a problem, and it should even be said that there is too much money.

But when it comes to technical deep water, the role of capital plummets. Just like the patent issue mainly discussed in this article, it is another problem in the domestic chip industry that has been selectively ignored by everyone, resulting in the development of the industry being constrained by overseas giants everywhere.

The patent barrenness of memory is only a microcosm of China’s chip industry. The road from big to strong made in China must rely on talents and intellectual property rights. This is also the short board part that urgently needs to be strengthened in the development of my country’s entire chip industry.

Figure 5: China’s semiconductors are highly dependent on foreign countries, source: China Semiconductor Industry Association, General Administration of Customs of China

As mentioned above, it is not shameful to learn from the world’s outstanding chip companies, cooperate and exchange, and even use it as a doctrine. The scary thing is that too populist thinks that working behind closed doors and tightening the belt can break the patent barrier. In fact, it is not much higher than generating electricity with love.

Another point is that the best way to avoid a patent war is still your own strength. It is almost impossible to avoid patent blockade, but you can achieve “you have me, I have you” and use force to exchange for peace. Best policy.

From the perspective of entrepreneurs, what we want to say is that the labor-intensive and capital-intensive business models that rely on OEM, resources, real estate and other labor-intensive and capital-intensive business models to make quick money are still deeply rooted in the hearts of the people. There is a long way to go, and the 150 billion Jinhua death is just a wake-up call. As China climbs the technology tree, we will have to face more challenges brought about by the lack of patents.

The Links:   LB064V02-TD01 MSP430F2232IDAR