Huawei Hisilicon launched the first Kirin chip in 2009. Although the response was average at that time, it played the movement of Kirin soaring, and there was no small progress every year: Kirin 925 led Mate7 into the high-end camp; Kirin 955 helped Huawei's P9 sales exceed 10 million ... The self-developed chip became the biggest weapon for Huawei's mobile phone to break away from domestic friends.
However, on August 7, 2020, Kirin series high-end chips were forced to retire early. Yu Chengdong said that the most advanced Kirin 990 and Kirin 1000 series will not be produced after September of 15, and Huawei Mate40 will become the swan song of Kirin's high-end chips. The reason for being out of print is simple: TSMC will no longer work for Huawei due to the US ban.
TSMC has not struggled. It is difficult to find a high-tech process in the world. TSMC actually has a strong voice and just surpassed Intel to become the world's largest semiconductor company a few weeks ago. Therefore, in the face of the US ban, TSMC also mediated, but as long as the United States mentions the name of a company, TSMC executives can be scared into a cold sweat. This company is Jinhua, Fujian.
Fujian Jinhua was founded on 20 16, with the goal of achieving a breakthrough in the field of memory chips. Fujian Jinhua is an IDM integrated process, that is, design, manufacturing and packaging must be done. Once the products land, the whole semiconductor process in Chinese mainland will be upgraded and improved. Jinhua invested 37 billion yuan in the first phase, and also carried out technical cooperation with Tailian Power, the second largest factory in Taiwan Province Province.
R&D personnel fought day and night. After more than a year's establishment, Jinhua has built a 12 inch production line and is ready to put into production. Unexpectedly, the iron fist of capitalism was ushered in.
2065438+200765438+In 2007, Micron Technology of the United States immediately attacked Jinhua on the grounds of stealing intellectual property rights, and Jinhua was not to be outdone. Both parties sued each other in Fuzhou, China and California, USA. Just when the situation was anxious, the Trump administration, which had been eyeing for a long time, launched a blitzkrieg on 20 18129 October: Jinhua, Fujian Province was included in the entity list, and cooperation between American enterprises was strictly prohibited.
After the ban was issued, the R&D support staff of the American company applying materials who cooperated with Jinhua packed and evacuated on the same day, and two other American companies, Corey and Lin Fan, also quickly recalled the engineers who came to cooperate. What's more, ASML in Europe and Tokyo Electronics in Japan also suspended the supply of equipment to Jinhua, because the equipment contained original products from the United States.
When recalling the scene of foreign capital evacuation, Jinhua employees concluded: "These people just gave us time to say goodbye."
Fujian Jinhua official website's production progress card was not updated on the trial production day of 20 18, but the product page directly showed "the page is under construction". Last May 10, the Financial Times reported that Jinhua had begun to seek to lease or sell its own factory. In just one round, the seed player who broke through the storage in China was knocked down at the starting line.
The "entity list" is like a notice of death penalty, which can instantly make enterprises fall into hell. The determination of the US sanctions and the intensity of the crackdown have made TSMC shudder, which is also supported by the core components and core technologies of the United States. Similarly, Samsung, which was eager to grab TSMC cake, also lost the following; SMIC also implied that it might not be able to sign contracts for "some customers".
Why are these companies unwilling to touch the "scales" of the United States? Does the United States really dominate the world in the field of semiconductors? Actually, it is not.
Although the output value of American semiconductor industry accounts for about 47% of the world, it is in an absolute advantage in quantity; However, South Korea, Europe, Japan, China, Taiwan Province Province, Chinese mainland and other "strongmen" also have their own strengths, and the gap with the United States is not an insurmountable gap.
For example, South Korea has an overwhelming advantage in the field of memory chips, with an output value of 654.38+050 billion US dollars, and Shuangqiang (Samsung and Hynix) occupies 65% of the market;
There are three troikas (Infineon, stmicroelectronics and NXP) in the field of analog chips in Europe, which have never fallen out of the top 20 in the world since 1980s.
Japan not only has a unique image recognition chip, but also several companies headed by Shinetsu Hitachi have firmly grasped the upstream materials of semiconductors all over the world.
Taiwan Province Province of China is even better than the United States in the field of chip foundry, reaching the level of hundreds of billions of dollars. TSMC and UMC account for 60% of the scale, and the packaging test foundry headed by Sunmoon can also grab 50% of the market;
Relying on the huge downstream market, the field of chip design in Chinese mainland has developed rapidly in recent years. Not only did Huawei Hisilicon, the top ten chip design giant in the world, be born, but its overall chip design scale also ranked second in the world.
From the perspective of book strength, these companies can even "de-Americanize" the chip industry and jointly build a mobile phone without American chips. But the United States banned it on 5 15, but all powerful people were afraid to follow.
The situation of "one super is more powerful" seems to be like a "paper tiger". Under the hegemony of the United States, it may be the current "truth" that many semiconductor manufacturers divide and rule. What everyone is afraid of is actually two sharp swords held by the United States: chip equipment and design tools. These two swords, together with Japanese materials, constitute three powerful cards for the US-Japan semiconductor hegemony: equipment, tools and materials.
So, what is terrible about these three swords held by the United States and Japan? How can we capture powerful technology giants? Only by understanding these answers can we understand Huawei's way out.
First, the equipment: the outer brain made of chips.
For the general industry, the equipment manufacturer is a shovel seller, and basically it will be over with money; Semiconductor equipment manufacturers are different, not only providing equipment to sell shovels, but also selling brains throughout the service, which can be described as the external brain of chip manufacturers.
Chip manufacturing costs are high, and only when the yield is controlled at about 90% will there be no loss. However, we should know that chip manufacturing starts from 1000 processes, which leads to that even if the pass rate of each step is 99%, the final yield will approach 0 after multiple accumulation of 0.9*0.9. Therefore, in order not to lose money, the pass rate of each step must be controlled at 99.99%, or even above 99.999%.
In order to achieve this situation, the complexity of equipment is extremely high. As far as the most advanced EUV mask aligner is concerned, a single device has more than100000 parts, 40000 bolts and more than 3000 wires. Only the hoses add up to two kilometers long. Such a huge equipment weighs 180 tons, and a shipment needs 40 containers, 20 trucks and 3 freighters.
More importantly, even if the equipment is bought back, it is far from being as simple as putting it away and plugging it in as a TV refrigerator. Generally speaking, it takes one year to debug and assemble a high-precision mask aligner. The assembly of parts, the setting of parameters, the debugging of modules, even the tightness of screws and the external temperature will affect the production effect. Even if a subway passes a mile away, most of the equipment will fail collectively.
This is also a "common problem" of all precision instruments. For example, ten years ago, the instruments of Peking University 12 high-precision laboratory with a value of 400 million yuan suddenly broke down. The reason was that the underground 13.5-meter Beijing Line 4 passed through the east gate of Peking University, resulting in vibration from 1 Hz to 10 Hz, and the high-precision laboratory of Peking University had to move collectively.
Therefore, every time semiconductor manufacturing equipment is started for a period of time, it is necessary to contact special original factory service personnel for on-site adjustment. ASML ASML, a mask aligner giant in the Netherlands, once had a customer who wanted to replace optical equipment; As engineers in ASML were unable to go abroad at that time, they invited outstanding employees from their customers to study in the company. It took nearly two months to master the skills of replacing individual parts.
Therefore, semiconductor giants such as ASML and Applied Materials not only sold their equipment, but also set up a huge support team of about 2,000 people in China. Among them, the second largest income of applied materials is service, accounting for more than 25% of the income, and it is growing steadily.
The terrible thing about the equipment factory is that it not only determines the manufacturing process of the factory through "one generation of equipment, one generation of technology and one generation of products", but also firmly holds the factory in its hands through after-sales service. With the technology becoming more and more sophisticated, the right to speak of equipment manufacturers has been further improved.
The strength of equipment manufacturers can be clearly reflected in profits. In the past five years, the head effect of chip manufacturers has become more and more obvious, but the net profit rate of upstream equipment manufacturers has increased greatly: the profit rate of pan-forest has increased from 12% to 22%, and the application materials have increased from 14% to 18%. It doesn't exist for the foundry to ask customers to bully the store.
For this reason, the United States has been using various means to ensure its absolute dominance in the field of equipment for 60 years.
According to the 20 19 ranking of the top semiconductor equipment manufacturers in the world, the industry income of the top five semiconductor equipment manufacturers in the world accounts for 58% of the world. Among them, the United States has three seats; The other two seats, Tokyo Electronics of Japan and ASML of the Netherlands, happened to be supported by the United States.
Specifically, Applied Materials Company (AMAT), Pan Forest Company (Forest) and Kelei Company (KLA) are all American enterprises with deep roots.
Among them, the market share of Pan Lin in etching machine is as high as 50%. Applied materials are not only equally divided with Marco Lin in the field of etching machine, but also account for half of the sub-equipment links such as ion implantation and chemical polishing, even as high as 70%. Corey occupies more than 50% market share in the field of semiconductor front-end inspection equipment and 98% market share in coating measurement equipment.
Mask aligner giant ASML, seemingly a Dutch enterprise, actually has an American heart. As early as around 2000, the mask aligner market was still in the stage of DUV (deep ultraviolet) lithography, and Japanese Nikon was the real overlord. However, in the EUV stage, Nikon was eliminated by the United States.
There is a simple reason. EUV technology is at its peak: the leap from traditional DUV to EUV means that the light source is drastically shortened from 193nm to 13.5nm, which requires 20KW laser to bombard 20 micron tin drops at a frequency of 50,000 times per second to vaporize liquid tin into plasma. This is equivalent to letting ping-pong hit a fly twice at a frequency of 50,000 times per second in a hurricane.
At that time, the most advanced EUV R&D institution in the world was the EUV limited liability company alliance led by Intel and the US Department of Energy. The Ministry of Energy has Motorola, AMD, IBM and three national laboratories, which can be described as a collection of the essence of American scientific research. It can be said that only by joining the EUVLLC alliance can you get a ticket to EUV.
At that time, the United States regarded Japanese semiconductor as an enemy, and naturally rejected Japan's Nikon's request for membership, while ASML promised that 55% of the parts and components would be purchased from American suppliers and subject to regular review. This is the game that entered the United States, from a rising star to an "emperor's flower".
The United States not only opened the door to Asme, but also sent a gift: ASML was allowed to acquire Silicon Valley Group, the leading mask technology company in the United States, Brion, the leading lithography detection and solution company in the United States, Cymer, and so on. ASML's heart of science and technology and R&D's body are branded with the Stars and Stripes. This is not decided by the United States.
In the early years, Tokyo Electronics was only the equipment agent of Fairchild, the originator of American semiconductors, and later jointly produced semiconductor equipment with Thermco of the United States. It was not until 1988 that Tokyo Electronics became a wholly Japanese-owned company, but it also shed the blood of American companies.
Therefore, in June 20 19, in the face of the first round of U.S. ban, TEPCO said, "We will not do business with China customers who are prohibited from applying materials and doing business in Marco Lin", and it is clear that American equipment manufacturers are in a dilemma.
So far, relying on years of "time accumulation" and ultra-high precision "technology", the United States has formed a firm initiative in the field of equipment. However, neither time nor technology can be achieved overnight by latecomers.
Second, EDA (design software): the "golden rope" under the ecological network effect
If equipment is the sword of chip production, EDA is undoubtedly the "golden rope" of chip design. Although it is not fatal, it can make "the Monkey King" tied behind one's back and nowhere to put it to use.
EDA's "golden rope" is divided into three parts: first, it is the "PS software+material library" for chip designers, which can make chip design change from physical labor of drawing lines on drawings decades ago to mental labor of "material arrangement and combination+code typing" in software. Moreover, there are only chips the size of fingernails, and there are also billions of transistors. This kind of engineering quantity is impossible without EDA.
At the corner of the circuit diagram of Intel Pentium processor 20 years ago, the current transistor density has increased by 1000 times.
Secondly, the mystery of EDA lies in its rich IP library. Frequently used functions will be standardized into modules that can be directly called, without the need for design companies to redesign. If the chip design is the chef's cooking, the software is the kitchenware, and the ip is the material package.
In fact, EDA giants often benefit from the monopoly of their IP. For example, Cadence has a large number of analog circuit IP, and it is also the king of analog and mixed signal circuit design; Synopsys's IP library is more inclined to DC synthesis and PT timing analysis, so Synopsys is a leader in the field of digital chips.
Among the top three IP companies in the world, EDA accounts for two, with a total market share of 24. 1%. In the revenue of Xinsi Technology over the years, IP authorization is the second largest business after EDA authorization.
Another important function of EDA is simulation, that is, it helps the designed chip to check for leaks and fill gaps. After all, the cost of a movie (trial production) is as high as millions of dollars, which is equivalent to the profit of a small design company for half a year. A word widely circulated in the industry: design is not simulated, two lines of tears.
According to the statistics of a professor at the University of California, the design cost of a 20 1 1 SoC is about 40 million dollars, but without EDA, the design cost will soar to 7.7 billion dollars, an increase of nearly 200 times.
Therefore, EDA is called the highest level in semiconductors. Although the global output value is only10 billion dollars, it can affect the development of the global integrated circuit market of more than 500 billion dollars and the electronic industry of trillions.
EDA is so efficient and easy to use, so what is the status of autonomy in China? Unfortunately, it is more embarrassing than the operating system.
Huada Jiutian, the largest EDA manufacturer in China, has a global share of almost 65,438+0%, while Xinsi Technology (Xinsi Technology), Cadence (Deng Kai Electronics) and Mentor Graphics (Ming Dow Technology was acquired by Siemens in 2065,438+06) have a market share of over 80%.
As a result, although China's chip design ranks second in the world, chip design will face a "tool crisis" under the command of the United States, and it is difficult for a clever woman to cook without rice. However, since the software has been paid, can't the old version be used?
Unfortunately, I can't.
Because there is an ecological network embedded by EDA vendors, IP vendors and foundries. EDA is constantly updated. The new version corresponds to the updated IP library and PDK file. As a process design package, PDK contains parameters such as current, voltage, material and process in chip process, which is necessary data for OEM production. New EDA, new IP and new technology promote and integrate with each other.
Therefore, using old software will be "out of touch" everywhere: you can't get the latest design IP library when designing, and you can't get the latest EDA and PDK when looking for a foundry. In the long run, technology is becoming more and more backward and there are fewer and fewer partners. But since EDA is only the code of 0 10 1, wouldn't it be nice to find some experts in the cracking team?
Unfortunately, this is almost impossible.
Every EDA software will be embedded with a Flexlm encryption software when it leaves the factory, which will lock the EDA and the installed equipment one by one, including host number, device hard disk, network card, use date and other information. Flexlm's key length is 239 bits, so it is difficult to crack it violently. If you use Intel's high-performance CPU to crack it, it will take about 4000 core years, which means that a 40-core CPU needs 1000 years.
Of course, you can also continue to increase the number of CPUs in a distributed way and reduce time. However, even if the crack is successful, it will be verified by EDA vendors through "modifying time, file size and confirming IP source" when it reaches the door of a brand-new IP library, and then it will be rejected. There is a sour feeling of digging an underground tunnel for a hundred years, but hitting a stone.
The cracking effect is not obvious, which is also contrary to the attitude of intellectual property protection in China. Therefore, it is still necessary to rely on companies like Huada Jiutian to develop themselves. So, how wide is this exit? In fact, it is not difficult to simply write a set of software. The key is to have a lot of rich IP, PDK, and the support of upstream and downstream industries. A single breakthrough may not be effective, but it needs a comprehensive breakthrough of the legion, which cannot be achieved overnight.
Third, materials: the last bastion of craftsman spirit
In 20 19, there was a conflict between Japan and South Korea, and both sides were very stiff. However, after Japan cut off several semiconductor materials in South Korea, it was not long before Lee Jae-yong, the head of South Korea's Samsung, flew to Japan to demand concessions. Later, he even went to Belgium and Taiwan Province Province of China, trying to buy or collect some stocks by a detour.
It stands to reason that South Korea is also a semiconductor power, and Samsung is a big country in the field of design and manufacturing, but it has been embarrassed in front of hundreds of millions of dollars of materials.
Is the material really that difficult? To tell the truth, semiconductor raw materials are very rich, for example, silicon wafers are all made of sand from the earth. However, it is not easy to realize the "material freedom" of semiconductors, and it is necessary to get through the two veins of "purity" and "formula".
Purity is an endless road. China has made its own photovoltaic silicon wafer with a general purity of 6-89, that is, 99.999999%, but the purity of semiconductor silicon wafer is 1 1 9, and it is still improving. More than 3 to 5 digits after the decimal point indicates that the impurity content is different from 1000 to 65438+ million times.
How big is this gap? Suppose that the impurities contained in photovoltaic silicon wafer are equivalent to sprinkling a bucket of sand on the playground; Then the requirement of semiconductor silicon wafer is that the area as big as two football fields can only accommodate one grain of sand.
So, why must the impurity content be reduced to such a low level? Because the size of atoms is only110 nm, even if there are only a few atoms of impurities on the silicon wafer, it will completely block a circuit channel and lead to local failure of the chip. If the impurity content is high, it will even be mixed with silicon atoms, which will directly change the atomic arrangement structure of the silicon wafer and completely change the conductivity of the silicon wafer.
The etched silicon surface and tin particles are like the bright moon rising behind the pyramid.
To achieve such purity, we need the perfect combination of science and technology.
On the one hand, a large number of basic scientific instruments are needed to assist. For example, in the process of material production, there will be metal atoms infiltrating into the equipment itself, which will affect the purity, so it needs continuous improvement. The purity is also difficult to confirm. Just like special gases, special instruments are needed to detect the impurity level (PPB level) of 1 billion. To achieve this difficulty, not only semiconductor companies, but also optical companies such as Olympus need help.
On the other hand, process accumulation is also needed from the laboratory to the factory floor. Material manufacturing not only requires high production equipment, but also the floor mats and mops in the factory are high-grade special supplies. Moreover, the difference of temperature and humidity in the production workshop will also affect the purity of materials, so it is necessary to try again and again to reach the standard.
High purity is only the first step, and the configuration of composite materials (such as photoresist) is an insurmountable gap. If "purity" is an art science, then "formula" is a metaphysical science.
In fact, regardless of purification or configuration, the basic theoretical principles and technologies are not difficult. However, how to choose materials and match them to achieve the ultimate effect requires a high degree of dependence on the rules of experience, which is often called "know-how" in the industry.
The same material, different proportions will have different effects; Just as we use red, yellow and blue to match, different proportions can get different colors. Even the same formula and the same technology will have different and even far-reaching effects under different humidity, temperature and even illumination.
These parameters that affect the material effect cannot be obtained by accurate calculation, but can only be prepared, experimented, observed, recorded and improved again and again in the laboratory and workshop. Sometimes, it may take several years to get the promotion of 10%. However, this 10% increase, although only occupying tens of billions of markets, has affected the trillion-dollar semiconductor industry.
Therefore, whether it is purification or formula, what is actually needed is long-term patient standby and extreme concentration. This reminds people of the Japanese sushi god, who only cooks sushi all his life, while an apprentice can practice for five years by twisting a towel. Although in life, this insistence seems a bit pedantic and ridiculous, in fact, Japanese companies do the best in the field of materials.
According to SEMI's data, in 20 19, Japanese companies accounted for 66% of the global semiconductor materials market. Among 19 main materials, there are 14 in Japan, with a market share of over 50%. Among the four core materials, silicon wafer, photoresist, electronic special gas and mask adhesive, which account for 2/3 of the output value, Japan has three items accounting for 70% of the total. In the field of the latest generation of EUV photoresist, three Japanese companies applied for more than 80% patents in the industry.
After Japan has an advantage in material production capacity, it binds customers to death with services.
Many semiconductor materials are extremely corrosive and toxic. A supplier of special gas once described that once the gas leaks, the entire population of Xiamen can be wiped out with only one bottle. Therefore, chip manufacturers can only hand over the transportation, storage and testing of materials to the "mother's family" material suppliers.
On the other hand, although the material is small, it is powerful. Tens of thousands of dollars of materials in semiconductor manufacturing are not up to standard, and most of the billions of products on the production line will be scrapped, so manufacturers will only choose certified and long-term cooperative suppliers. New players have little chance to get on the table.
For material companies, the more they use downstream, the more feedback they get, more case support and more verification opportunities to upgrade the process and improve the ratio, thus further widening the gap with the pursuers. For the latecomers, the business situation can be described in one sentence: one step can't catch up, and every step is in vain.
Japan can achieve this achievement, in fact, it can not be separated from the direction planned by Japan's "Holy Management" kazuo inamori in the 1980s: under the condition that advanced countries in Europe and America are unwilling to transfer technology, the Japanese have no choice but to carry forward their inherent "improvement and improvement characteristics"; All kinds of enterprises should be thorough in their respective professional fields and the technology should be the ultimate, no less than any enterprise in any country in the world.
This craftsman spirit enabled Japan to resist the United States and become the overlord in the field of small-scale materials.
Fourth, where to break through?
When we are doing industrial research, we have a strong feeling that under the pressure of the United States, China seems to be in a desperate situation that can be traced back to infinity:
After discovering the chip card neck, we have a rising Huawei HiSilicon in the field of chip design, but later we found that we still need to break through in the field of OEM. When SMIC attacked the OEM manufacturing of chips, it found that it was necessary to make a breakthrough in the equipment link; When Zhongwei Company and North Huachuang counter-offensive equipment gained something, they found that the core part of the equipment was knocked out. When the parts also made progress, it was found that the chip material was still stuck in the neck.
When we continue to trace back to the source step by step, we find that everything has returned to the basic science that was emphasized countless times before taking office.
Looking back, without the modern binary established by 1703, the machine language after 200 years would be impossible; Without the physical rectification effect discovered by Braun in 1874, there would be no invention and application of transistors more than half a century later; Plasma physics and gas chemistry are the necessary foundations of key equipment such as etching machine.
Among American universities, there are seven in the top ten in global physics, six in global mathematics and five in global materials. The powerful dominance of basic science has become the source of strength for American semiconductor companies.
Behind the strength of basic disciplines is the financial support for basic disciplines in colleges and universities, which has been established by 1957. Lead the application to the ground through super technology projects.
When the United States and the Soviet Union competed for hegemony, the Soviet Union launched the world's first artificial satellite, which stimulated the rulers of the United States and became an important turning point in the development of American science and technology:
On the one hand, in order to keep "America ahead", the government began to send money directly to research institutions. National Science Foundation (NSF) funding for basic research in universities soared from $7 million in 1955 to $200 million in 1968. In 20 18, NSF spent as much as $4.2 billion on basic research. 50 years of basic research funding, the United States federal government paid half.
In particular, NSF provides scholarships to thousands of graduate students in basic disciplines every year, among which 42 Nobel Prize winners have been born.
On the other hand, the United States launched a super project to bring research and development results to the ground. 1958, NASA was established, and the Apollo moon landing and space shuttle program that challenged the limits of human science and technology began.
In the process of studying the space shuttle that needs 2.5 million parts (as a contrast, mask aligner has about 6,543,800 parts, and a car has only 654.38+10,000 parts), a large number of cutting-edge technologies have found their place; These cutting-edge technologies that were "unpopular" at that time were transformed into killer civilian products (such as artificial hearts and infrared cameras born from space shuttle parts) when conditions were ripe.
The technology spillover of the space shuttle is not an isolated case. Superconducting magnets used in hospital nuclear magnetic resonance equipment were born in the research and development of American particle accelerator "Tevatron". Super science and technology projects in the United States have become experimental fields, training grounds and folk transformation springs for the achievements of basic disciplines.
In fact, mastering the source technology through basic research and then establishing industrial hegemony step by step is not only the patent of the United States, but also the choice of various industrial powers, and it is a truly feasible way to face American repression. Princes would rather have seeds. Avoid the endless trap of "domestic substitution breakthrough" and realize "basic research spillover".
In fact, Japan has also experienced the difficulties and repression we face.
In the late 1980s, the United States launched a surprise attack on Japan's semiconductor industry: political ban, commercial repression and tariff oppression did everything possible, especially cultivating a "new younger brother" Korea to squeeze Japan's semiconductor industry. In a few years, Japan fell from the throne of the world's largest semiconductor power. The semiconductor departments of Panasonic, Toshiba and Fujitsu, three models that Japan Semiconductor is proud of, have been sold one after another.
Faced with American repression, Japan chose to March into high-tech materials, exchanging time for space and ingenuity for self-confidence.
1989, South Korea made efforts to subsidize memory chips, and the Japanese Ministry of International Trade and Industry formulated the "Basic Plan for Research and Development of Silicon Polymer Materials" with an investment of1600 million yen, focusing on subsidizing silicone enterprises headed by Shinetsu Chemical.
1995, on the eve of the second storage price war in South Korea, Tokyo Huaying (Tok) realized the commercialization of KrF photoresist, breaking the monopoly of IBM in the United States for more than 10 years, and in the following five years, its product technology became the industry standard and the world leader.
In 2005, Samsung took the leading position in memory chips, and Japan Relief Printing Co., Ltd. acquired the photomask business of DuPont Company of the United States for 76,543.8 billion yen, becoming the photomask leader.
In the days when South Korea fully expanded its production capacity and fought with other downstream semiconductor factories, Japan stepped onto the throne of material hegemony step by step. Grabbed a hegemonic sword from the Americans who seemed to have no solution advantage.
But is Japan's success only because it has changed to an upstream battlefield? Obviously not. In the past 30 years, Japan has won 16 Nobel Prizes in three major natural sciences, six of which are in the field of chemistry, and these are the solid foundations for Japan's rise.
How about basic research in China? In 20 18, China's basic research expenditure only accounted for 5% of the total annual R&D expenditure, the highest since 10. In the same period, the proportion of basic research in the United States was 17%, and that in Japan was 12%. In various domestic school forums, posts urging junior students to transfer from basic disciplines to applied disciplines such as financial computers are endless.
So some people laugh that Lujiazui knows more about integrated circuits than Zhangjiang.
In July this year, a myth broke out that more than 90 people from the Chinese Academy of Sciences left their jobs collectively. It is true that everyone has the freedom to choose a job, but what needs to be warned is the reason why everyone makes a choice. The long period, weak transformation and low income of basic subject research make researchers look pale and powerless in the face of rising house prices and tens of billions of profits.
Ren Zeng lamented: In the past, the national policy of industrial development was to spend money, and it didn't matter if the money was spent. Our country is used to repairing bridges, roads and houses ... as long as it costs money. But chips can't cost money. Mathematicians, physicists and chemists must be hit. ...
64 years ago, the United States was awakened by a satellite launched by the Soviet Union. While adding "short-term confrontation", Americans are brewing "long-term innovation", thus opening a breakthrough in many fields and being in a leading position; Now, the ban has also awakened us. Many industries in China are only superficially large, and what is urgently needed is the strength in the bones.
The pain of these crises is always regrettable. In the past few decades, the reality that backwardness will be beaten reminds us again and again that to win the next era, we must achieve innovation and breakthrough in basic technological capabilities.