magic iron author
[Editor's note] From December 1947, the world's first transistor was born in Bell Labs in the United States, to 1959, Noyce of Fairchild Semiconductor wrote a plan to build an integrated circuit and invented the world's first silicon Integrated circuits and chips began to develop rapidly, and even became one of the economic pillars of a country. How did American chips come about and rise? What kind of chip story is there in the competition with China, Japan and South Korea?
Venture capital agreement written on a dollar
January 24, 1848, morning.
The American River north of Sacramento, California's capital. A carpenter, Zhan, stumbled across a golden sheet of metal on the river bed while inspecting the sawmill waterways. After careful identification, the owner of the sawmill determined that it was gold with a purity of 23K. News of the discovery of gold spread like wildfire, and the gold rush in the American West began.
The gold rush allowed the United States to harvest more than 3,600 tons of gold and attracted about 300,000 people to participate in the western development. But compared with the gold rush a hundred years later, this wave is at best a small wave. 109 years later, the investment and entrepreneurship boom that started in semiconductors has pushed the agricultural country of North America to the position of global technology hegemon.
The wind started at the end of Qingping, and the waves formed between the waves. The semiconductor giant tide sweeping the world originally started from an ordinary job search help.
It was one day in June 1957. Arthur Locke, a newcomer to the investment bank Hayden Stone in New York, USA, received a letter from his boss, Alfred Coyle. The letter was written by 7 people including Eugene Kreiner. In the letter, they asked Hayden Stone to help find a company that was interested in entering the silicon transistor business. If the company could hire the 7 people, they would guarantee that within a year. Diffused silicon transistors are delivered within.
These seven people are the core employees of Shockley Semiconductor Laboratory, including Gordon Moore, who later proposed "Moore's Law". They wrote to Haydenstone because they thought their boss, William Shockley, was a bad manager.
At that time, Shockley was well-known in the scientific community. He was the co-inventor of the transistor. At the end of World War II, he prompted the US government to stop " Operation Downfall " against Japan, abandon large-scale amphibious landing operations, and turn to Hiroshima. Drop the atomic bomb on Nagasaki and end the war.
Shockley was awarded the Special Service Medal by Robert Patterson, then US Secretary of War.
But what really made Shockley's reputation skyrocket was his status as a co-inventor of the transistor during his time at Bell Labs, earning him the title "Father of the Transistor." But Shockley had no hope of promotion at Bell Labs because of his bad relationship with his colleagues, so he resigned to start the Shockley Lab in Mountain View, California, vowing to make a million dollars and slapped Bell Labs in the face.
In order to recruit the smartest people in the United States, Shockley offered a monthly salary of $750 (more than Bell Labs) and posted job advertisements in the form of code.
Shockley got his wish to recruit a "doctoral line", including Intel founders Robert Noyce and Gordon Moore. But Shockley lacked the ability to deal with people, and knew nothing about management skills. He was "a (scientific) genius and a (operational) waste."
Once the female secretary's hand was scratched by the doorknob, he believed that someone among his subordinates had deliberately murdered, and used a polygraph to find out the "real murderer" behind the scenes. But what made Moore and others determined to flee was not only the chaotic management of the company, but also the disorder and randomness of the development direction.
At that time, the market was at a turning point when transistors were replacing vacuum tubes. The mainstream of the market was expensive germanium transistors. Shockley Lab's vision was to mass-produce cheap junction silicon transistors to replace germanium transistors and win orders from the U.S. military. But after the company's official operation, Shockley continued to allocate production resources to complete academic tasks for himself. The swing of business direction made the company not launch mass-produced products as scheduled after one and a half years of official operation, which greatly disappointed the doctoral staff.
The seven, including Moore, believe they have all the expertise needed to make diffused silicon transistors, and the transistor industry is in a growth phase, so they should be able to find another job.
After the decision was made, Moore and other seven people wrote to Haydenstone for help. By coincidence, the letter went to Arthur Locke, a 31-year-old newcomer in the workplace. Locke made a lot of money by investing in transistor companies not long ago. He did not promise to find jobs for these seven people, but made a bold suggestion: "Why don't you start your own company?" The seven people were startled, Locke After giving them a class on investing and entrepreneurship, they brought in the wobbly Robert Noyce.
As a result, Lokra came to his own boss, Coyle, and together with Moore and other 8 people, decided to set up a company to produce diffused silicon transistors. After the meeting, Locke took out a stack of $1 bills, and 10 people signed their names on the bills, which was considered an agreement of intent, and then each kept a bill. Of course, the content of the agreement is agreed orally by both parties.
The most bizarre agreement in the history of venture capital was born.
On September 18, 1957, on what the New York Times called the 10 most important days in human history, Robert Noyce and seven other colleagues from the Shockley Semiconductor Laboratory collectively handed over to Shockley. resignation letter. Shockley was furious at the time, berating the eight young men as the " traitorous eight" . The name of the "eight rebels" has become part of the legend of Silicon Valley, and "rebellion" has also been passed down from generation to generation as Silicon Valley culture.
The "Eight Rebels" are: Robert Noyce, Gordon Moore, Julius Blank, Eugene Kleiner, Kim Jean Hoerni, Jay Last, Sheldon Roberts and Victor Grinich
In August 1957, after a month of running around, Loklar made an investment in Sherman Fairchild . Fairchild's father, Fairchild Sr., is IBM's largest individual shareholder. He inherited a large amount of wealth from his father. He is a technology enthusiast and has a successful experience in investing in technology companies, so he hit it off with Locke, Established Fairchild Semiconductor Corporation.
The founding of Fairchild Semiconductor was equivalent to the beginning of the embryo of Silicon Valley, marking the beginning of savage growth and disruptive innovation.
Before the founding of Fairchild, strictly speaking, there was no established venture capital in the United States. Entering the semiconductor industry is generally through corporate transformation. Texas Instruments used to be oil extraction equipment, and then transformed into semiconductors; or, like IBM and Motorola, semiconductors are only one department in their huge business branches.
But Fairchild Semiconductor is completely different. It is the beginning of venture capital. Its entrepreneurial team and investors are completely separated, and the entrepreneurial team also holds shares.
According to the agreement reached, Moore and other eight people will contribute $500 each to get 100 shares, and Fairchild's Fairchild Photographic Equipment Company will provide up to $1.4 million as Fairchild in the first 18 months as an investor. Start-up capital for semiconductors; if Moore et al. cannot make Fairchild Semiconductor make a profit of $300,000 for 3 consecutive years, Fairchild will buy out the company for $3 million. After 3 years, within 7 years, The price to buy out the company rose to $5 million.
It can be seen that this agreement allows Moore and others to open their hands and feet to start a business. If the business fails, each person can still get 226,000 US dollars. If the business is successful, the return is more. The risk of starting a business is basically borne by the rich second-generation Fairchild.
Fairchild took on the greatest risk and the greatest reward. After Fairchild Semiconductor was established, he could buy the entire company for as little as $5 million.
Subsequent development proved that this was a win-win agreement. Fairchild started to make a profit in just 6 months after its establishment, and Fairchild made hundreds of millions of dollars. Moore and other eight people, after two years, each received an income of 250,000 US dollars, which is 500 times the initial investment. Moore's partner and later Intel founder Robert Noyce once said with emotion: "People like me, who thought this life was just a life of working and earning wages, suddenly, we got a share of a start-up company. ."
Most of the eight people, including Noyce and Moore, were from small towns. Before coming to California, Moore studied missile tail flames in a laboratory with a monthly salary of $100. Noyce’s dream as a child was not to pick up his brother’s leather shoes to wear.
The success of Fairchild Semiconductor has brought Silicon Valley into a stage of savage growth.
On the one hand, venture capital has developed rapidly, the mechanism has been continuously improved, and the scale of funds has continued to expand. Arthur Rock took advantage of the success of Fairchild Semiconductor to transform himself into a professional venture capitalist. He went to the East Coast where funds were abundant to raise capital. In the early 1960s, he raised US$86 million by himself alone. Intel, Apple is a successful investment case for Locke. Later, the famous venture capital fund Kleiner Perkins Huaying and Red Shirt Capital developed and expanded in Silicon Valley, and the scale of venture capital in Silicon Valley also ranks first in the world.
On the other hand, stock options have become the hard currency of Silicon Valley semiconductor companies, making it possible for ordinary people to get rich overnight: some secretaries spend a few dollars exercising their stock options and end up making millions; one wears a T-shirt , A young man who looks ordinary may be a tycoon with a fortune of hundreds of millions of dollars. Dozens of stories of overnight get-rich-quit appear every month in Silicon Valley. In this regard, "The Courier" reporter Mike Malone wrote: "Silicon Valley has gathered enormous wealth, unbelievable wealth, excess wealth, and they have rained down on the uncivilized people not much bigger than the kid. engineer."
Funds and talents gathered in Silicon Valley, making Silicon Valley's semiconductor industry become one of the three pillars of the United States in the 1970s, along with automobiles and steel. However, what the Americans did not expect was that when they were sitting on the golden mountain of semiconductors and counting money, Japan on the other side of the Pacific Ocean had quietly been eyeing them.
Japan discovers America's wealth code
Americans liken semiconductors to "crude oil for industry", while in the eyes of the Japanese, semiconductors are "food for industry."
Since the Meiji Restoration, Japan has been strengthening the role of steel in the basic industry, using it as a pillar to develop production and revitalize the industry, and also shouted the slogan "Steel is the country".
However, after World War II, the Korean War broke out, and Japan was positioned by the United States as the "front line of the Cold War", so it gave great support in technology, and Japan also paid close attention to the development of advanced technology in the United States. The wealth code of the semiconductor industry was quickly discovered by the Japanese. The shrewd Japanese at that time did the math (in the 1970s) and found that the value of steel was nothing compared to semiconductors.
In terms of value per ton, steel is worth 80,000-90,000 yen at most, while the value of chips is generally 250-300 million yen per ton. If it is a large-scale integrated circuit (high-end chip) , the value per ton is soared to 2.5 billion yen. In other words, chips are worth at least 3,000 times more than steel.
It is even more attractive if the added value is increased. One gram of metal silicon is only 1 yen. After being purified into polysilicon, the value rises to 10 yen. It is then drawn into monocrystalline silicon, and cut and polished into silicon wafers. The value of each piece rises to 1000 yen (one gram of monocrystalline silicon). Equivalent to an unpolished silicon wafer) , each piece of silicon is worth 10,000 yen after being made into a chip. In the production process from metal silicon to chips, after each series of processes, the silicon only changes its shape once, and the added value increases by 10 times.
The Japanese came to the conclusion that the chip industry is simply modern alchemy!
In addition, the chip products are small in size and high in added value, and the energy consumption is 1/500 of that of steel. It is indeed a perfect product for Japan, which is short of energy and relies on exports to drive the economy.
More importantly, the Japanese discovered the wealth code of the Americans, and the Americans were still ignorant. For a long time, American companies believed that Japanese companies were nothing more than "toy manufacturers", producing products with low added value and no technical content.
The start and take-off of the Japanese semiconductor industry is inseparable from the Ministry of International Trade and Industry as a "parents who raise their own". Because of the strong support of the Japanese government, it is only because of the strong support of the Japanese government that Japanese semiconductor companies have beaten their Silicon Valley counterparts.
In 1957, the Ministry of International Trade and Industry formulated the "Temporary Measures for the Revitalization of the Electronic Industry", which set the tone for the development of semiconductors. At this time, Fairchild Semiconductor had just been established, and the transistor market had just emerged.
Initially, the Japanese semiconductor industry was weak and unable to compete with the United States, and the Japanese government directly stepped down to support it. In 1972, Japan concentrated on large group companies such as Fujitsu, Hitachi, NEC, Toshiba, and Mitsubishi Electric, and began to implement the four-year plan of "Electronic Computer Development Promotion Fee. Subsidy System", and in 1976, the "Very Large Scale Integrated Circuit Technology Research Group" was implemented. The four-year plan refers to the research and development of semiconductor core technologies. In the plan started in 1972, the Japanese government invested nearly 60 billion yen, of which 20 billion yen was intensively invested in 1973 and 1974. Under the plan started in 1976, the R&D expenditure invested by the government rose to 70 billion yen.
The Japanese government's intensive investment in the research and development of core semiconductor technologies has rapidly shortened the technological gap between domestic semiconductor companies and American competitors, and played a huge role in the rise of the Japanese chip industry. A leading figure in a major Japanese company commented on this: "If we consider that the integrated circuit (chip) industry at that time was an unprofitable industry and we were being overwhelmed, then these measures were very effective. "
The strength of Japanese semiconductor companies has thus greatly increased, and it is recognized that "in some advanced device technologies, it has surpassed the United States."
At the beginning, when the Japanese government spent huge sums of money on key technology research, the American professional magazine "Electronics" sarcastically said: "This is not about super-large-scale integrated circuits, but a mystery."
However, the Americans soon stopped laughing.
Inside NORAD in the early 1970s
In the duty room of the North American Air Defense Command, the radar operators on the night shift stared blankly at the large screen in the center. On the screen, hundreds of arcs representing the incoming Soviet intercontinental ballistic missiles are slowly rising, and their landing points point to various important military and economic goals located at the core of the continental United States!
The news of the Soviet missile attack was quickly reported, and the "third world war" was on the verge of breaking out. However, the alarm was quickly dismissed due to a false warning caused by a computer malfunction.
But the mistakes continued. In May and June 1980, the computer issued false warnings one after another, reporting that there were as many as 2,000 incoming Soviet missiles.
The U.S. Department of Defense investigation said the accident was caused by “a malfunction of an integrated circuit the size of a coin inside the computer.”
The computers of the U.S. military are purchased from domestic companies in the United States. These three failures have made researchers at Japanese securities companies who are concerned about high-tech stocks keenly aware that the chip industry in the United States is rapidly declining.
In the US civilian market, it is becoming a consensus that the quality of Japanese chips surpasses that of the US. In March 1980, R. Anderson, director of the Computer Systems Division of Hewlett-Packard Company, threw out a point of view that made the American semiconductor industry red at the semiconductor seminar held in Washington: "The memory produced in Japan has a low defect rate during inspection, The defective rate of the production line is also less, and the defective rate of the final product is even less.” Then Anderson compared the failure rate of 16K-bit RAM in Japan and the United States. The data shows that the failure rate of the products of American manufacturers is Japan’s. 6 times.
Anderson warned the U.S. semiconductor industry that since 1977, Japan's memory quality has been surpassing that of the U.S., and that "the best American products are worse than the worst Japanese products."
Anderson's warning is known as the "Anderson bomb", which means that the Japanese semiconductor industry, especially in the memory chip industry, has surpassed the United States.
Eight years ago, Japan spent a lot of money on core technology research and development, and now it is finally ushering in the harvest season.
Reflected in the market, in 1975, Japan's market share in memory chips was only about 3%. In 1980, this figure rose to nearly 40%, and the share soared more than 12 times in 5 years, while the United States fell from 95% to 10%. In the early 60s.
Under Japan's aggressive attack, American chip companies were defeated like a mountain, and financial data was like melting ice cream, a mess.
In 1981, AMD's net profit fell by 2/3, National Semiconductor lost 11 million US dollars, and earned 52 million US dollars the previous year. The following year, Intel was forced to lay off 2,000 employees. The Japanese continued to expand their victories, and the Americans continued to mourn. In 1985, Intel surrendered and announced its withdrawal from the DRAM storage business. This war caused it to lose 173 million US dollars, the first loss since its listing. At Intel's most critical moment, if IBM hadn't helped and purchased 12% of its bonds to guarantee cash flow, the chip giant would likely have collapsed or been acquired, and the history of the information industry in the United States might have been rewritten as a result.
With the strong rise of Japan's semiconductor industry, Japan-US trade frictions have spread from steel, automobiles, to semiconductors. The semiconductor competition between Japan and the United States has become a hot topic in the media confrontation between the two countries. In 1978 Fortune magazine published an article titled "Japanese Spy in Silicon Valley of America". In 1983, "Business Week" magazine published a long-length album titled "Chip (Semiconductor) War: Threat to Japan", which rarely used "semiconductor war" to describe the intensity of Japan-US semiconductor friction.
Defeat the Japanese who keep saying "no"
In the 1980s, Japan had a thriving economy and was invincible in electronics, steel, automobiles, semiconductors, etc. Japanese companies began to sweep the streets in the United States. In particular, Mitsubishi Group purchased the iconic Rockfeller Center in New York for $846 million 51 % of the shares, the United States called "the economic version of Pearl Harbor".
The Japanese, who have been aggrieved by the United States for many years, finally raised their eyebrows. Sony founder Akio Morita and right-wing activist Ishihara Shintaro wrote a book "Japan Can Say No" to protest against the United States. As if feeling not relieved enough, Ishihara Shintaro then released a book "Japan Can Say No".
However, when the Japanese semiconductor industry was immersed in a festive atmosphere, the United States began to plan how to defeat the Japanese from top to bottom.
In 1977, major American semiconductor companies such as Intel, AMD, Motorola and Fairchild initiated the establishment of the Semiconductor Industry Association (SIA) . Government support for the domestic semiconductor industry.
Intel's official biography records: At that time, Silicon Valley was generally opposed to Washington, and no one was willing to go to the east to lobby the US federal government and Congress, thinking that going to Washington to beg for support was a curse on the boards of Silicon Valley companies. In particular, Jimmy Carter , who was elected president in 1977, was soft-hearted at the center of the trade friction with Japan. As a result, in the late 1970s, the Silicon Valley was precarious, the semiconductor industry was defeated by Japan, more than 20 American semiconductor companies were acquired by foreign capital, and a large number of talents were drained.
However, with the election of Ronald Reagan , who advertised a "strong America" position as president in 1981, SIA's lobbying finally paid off, and the U.S. government began to investigate the Japanese semiconductor industry on the grounds that it would affect national security. Strong counterattack.
The trick of Japanese semiconductor companies to defeat their American counterparts is high quality and low price. Therefore, the primary focus of the US counterattack is to increase the price of Japanese semiconductors.
In 1982, the U.S. Commerce Department said it would investigate cheap dumping by Japanese chipmakers into the U.S.
In June 1985, SIA filed a Section 301 complaint with the USTR over the dumping of Japanese electronic products, targeting the Japanese government.
Under the strong pressure of the US government, Japan and the United States signed a five-year "Japan-US Semiconductor Assurance Agreement" in early 1986. The main contents of the agreement include: Japan expands opportunities for foreign semiconductor companies to enter the Japanese market; in order to prevent dumping in advance, the Japanese government should monitor the prices of semiconductors exported to the United States and third countries.
Under the strong pressure of the United States, the Japanese government quickly confessed. When the United States conducted an anti-dumping investigation, the Japanese Ministry of International Trade and Industry immediately issued export guidelines, requiring Japanese semiconductor companies to automatically reduce exports to the United States and automatically increase export prices to the United States. When the United States asked Japan to disclose more than 1,000 patents of the "Very Large Scale Integrated Circuit (VLSI) Technology Research Group" (1976-1980) and completely abolish Japanese semiconductor tariffs, Japan did so.
In order to implement the first phase of the "Japan-US Semiconductor Assurance Agreement", the Japanese government also regularly sends personnel to companies to inspect and prohibit the export of semiconductors at low prices.
As a result, Japan's semiconductor production has been deprived of its freedom of operation and placed under the surveillance of the Japanese and American governments.
To make matters worse, the "Plaza Accord" directed by the United States in 1985 led to a substantial appreciation of the yen. In just three years (by 1987) , the original 1 US dollar against 240 yen rose to 1 US dollar against 120 yen. The yen doubled in value.
This in turn forced Japanese semiconductor manufacturers to substantially increase export prices. Japanese semiconductors completely bid farewell to high quality and low price.
Semiconductor memory manufacturers in the United States were almost wiped out by Japan in the early 1980s. It stands to reason that when Japanese semiconductor manufacturers increase product prices, the United States can only accept all orders.
The fact is that while suppressing Japanese semiconductor manufacturers, the United States began to support South Korea and Taiwan, China, and reorganized the global semiconductor industry chain.
shy korean
and a small soy milk shop in Taiwan, China
In the early 1980s, when the United States wielded the sanctions stick to suppress Japanese semiconductors, a wealth rumor about semiconductors began to circulate in the Korean chaebol community: a suitcase filled with semiconductor chips worth more than one million US dollars, equivalent to 10 ships of minerals. The huge profits of semiconductor chips have made the chaebols move.
On March 15, 1983, the Samsung Group issued a declaration of "Why do we want to enter the semiconductor business" in South Korea's "Central Daily", identifying semiconductors as the future of Samsung's entire group. Following in its footsteps are well-known Korean chaebol companies such as Hyundai and LG.
In fact, as early as 1975, South Korea officially released the "six-year plan" to support the development of the semiconductor industry and to realize the localization of electronic accessories and semiconductor production. However, after 8 years of hard work, Korean semiconductors still have not improved. The main reason is that the Japanese semiconductor industry across the sea is too strong, and Koreans are far behind Japan in both products and technology.
As a result, only from 1984 to 1987, Samsung's semiconductor business accumulated a loss of 115.9 billion won. There is a lot of quarrel within the company over whether to continue making semiconductors. Samsung is in a dilemma at this time. If it withdraws from the semiconductor business, the 1MB DRAM factory under construction will also sink into the quagmire, and the losses will continue to increase. Helpless, Samsung could only continue to play with the Japanese in tears.
When Samsung and other chaebols were suffering from losses, the South Korean government took decisive action and carried out core technology research in accordance with the Japanese model. In 1985, the Korea National Electronics Research Institute was established. With this as the core, it organized Samsung, Hyundai, LG and other chaebols to participate in semiconductor research and development, covering Core technology from 1M to 64M. The entire R&D investment totaled 177.9 billion won, of which the government invested 125 billion won, accounting for about 70.26%.
The government undertakes the main investment in core technology research and development, so that South Korean semiconductor companies can endure until the moment of outbreak and survive the most difficult winter.
While the Koreans are dormant, China's Taiwan Province is also quietly waiting for the semiconductor dream to blossom.
Taiwan's semiconductor planning time is actually earlier than that of South Korea. It was February 7, 1974, a seemingly unremarkable morning. In a quiet soymilk shop called Xiaoxinxin on Nanyang Street, there are political leaders of the Taiwan authorities and heads of relevant departments: "Minister of Economy" Sun Yuntian, "Minister of Transportation" Gao Yushu, "Secretary-General of the Executive Yuan" Fei Hua, " Fang Baoxian, Director of the General Administration of Telecommunications, Wang Zhaozhen, Director of the Industrial Technology Research Institute, Kang Baohuang and Pan Wenyuan, Director of the Telecommunications Research Institute of the Ministry of Communications. Except for Pan Wenyuan's research work, all seven of them are technocrats.
After a bowl of soy milk, this group of ambitious people finally agreed to use integrated circuit technology as the entry point for Taiwan's transformation of high-tech industries.
Like South Korea, Taiwan is also watching Japan lead the investment through the government to make up for the shortcomings of core technologies and achieve a leap in the technological level of "later forces". Therefore, Taiwan also followed the pattern of cats and tigers. In 1984, the "Industry Research Institute" took the lead in investing 70 million US dollars to start a large-scale integrated circuit research and development program.
However, Taiwan Province lacked large-scale electronics companies such as Samsung at that time, and the investment in R&D was far lower than that of South Korea. The level of semiconductor development was always far behind that of South Korea and Japan.
In 1986, after Sun Yunying's "three visits to the cottage", Zhang Zhongmou returned to Taiwan as "President of Industrial Technology Research Institute and Chairman of UMC". Zhang Zhongmou used to be a core executive of Texas Instruments. He has rich experience in international market business battles and led Texas Instruments to defeat Intel in the DRAM memory market.
Zhang Zhongmou of Texas Instruments
张忠谋发挥成熟老道的预判能力，认为纯粹的晶圆代工将是一个新的行业风口，可以助力台湾半导体产业起飞，台积电由此成立。
张忠谋的预判离不开美日半导体战争的背景，因为解决美国半导体企业的共同的制造技术课题，以恢复在 20 世纪 80 年代中期被日本半导体企业超过的美国半导体产业竞争力，是美国挠破头焦虑的事。晶圆代工恰好可以给美国人补上制造的短板。也就是说，台积电可以承接美国半导体的制造需求。
台积电的努力很快得到了美国积极的回应，包括IBM在内的美国企业毫无保留地授权给台积电制造专利，弥补了台湾核心技术研发投入不足的短板，日本半导体企业的死敌英特尔更是直接给了台积电一个大单。在英特尔的广告下，英伟达、高通、苹果等芯片设计大厂纷纷下单台积电。在美国的扶持和自身努力下，台积电迅速发展起来。
台积电的晶圆代工模式，针对的是日本半导体企业的IDM（垂直一体化）模式，随着芯片制程的进步，芯片制造厂设备投资越来越成为一个沉重的负担，IDM模式越来越难以适应行业发展，使得日本制造的芯片成本渐渐高于台积电的制造成本，找台积电代工的美国芯片企业终于找回了市场竞争力。
在晶圆制造上扶持台湾打击日本的同时，美国对韩国的扶持同样不遗余力。
在美国的压力下，日本不敢对半导体产业进行扶持；日美半导体协定捆住了日本半导体企业的手脚，日本的半导体产品失去了自由定价权，半导体协议的价格监督机制，让日本的半导体产品永远比韩国对手的贵上一截，想都别想发动价格战打击韩国对手；加之日元大幅升值，汇率损失又捅了日本半导体企业一刀，在行业低谷期，企业要么减产要么停产，无力进行投资。
如此一来，对韩国半导体企业来说，等于日本拱手让出了市场份额。
国际市场上，美国继续收紧日本半导体企业脖子上的绳索：对日本出口的芯片征收100%惩罚性关税，对三星仅象征性收取0.74%的反倾销税。随后，三星的DRAM“双向型数据通选方案”获得美国半导体标准化委员会认可，成为与微处理器匹配的内存，日本则被排除在外。
在美国的打压下，日本半导体产业特别是DRAM存储器领域快速衰落。
2002年，日本DRAM的市场份额已经萎缩到大约10%，韩国的份额则超过40%。日本半导体五巨头开始砸锅卖铁（重组半导体业务），变卖家产，日立和NEC剥离DRAM业务在1999年组成尔必达，意图复兴日本半导体，但坚持13年后破产，被美光科技收购。
五巨头之一的富士半导体无奈之下，将工厂改建成蔬菜大棚，种植无公害生菜，每颗卖400日元，利润竟然高过生产DRAM芯片。
经过近20年的角逐，日本半导体产业彻底衰落，美国重回霸主宝座。
回溯从晶体管发明到如今的70余年间，半导体产业起于加州硅谷方寸之地，扩散至全球亚美欧三大洲。在这70余年里，美国牢牢占据产业链的上游，半导体从一个普通产业，一跃成为其全球霸权的一根重要支柱。
美国半导体维持全球霸权，一方面离不开其明显的竞争优势，即技术创新和商业模式创新。技术创新在硅谷发展早期，可以说是刻在基因里的。仙童半导体时代，由于员工离职率极高，而创办半导体企业的门槛又较低，员工离职往往会带走生产工艺等核心技术，成为原雇主的致命威胁，因此原企业要生存，唯有通过快速技术迭代。
同时，新成立的半导体公司要打开市场，最有效的门票就是产品创新，仙童半导体通过集成电路，有了与德州仪器叫板的实力，英特尔通过CPU，有了与IBM谈判的筹码。技术创新是美国半导体企业的核心竞争优势，这一优势通过专利壁垒稳固下来。相反，制造并不是美国半导体企业长袖善舞的地方。当年的日本，后来的韩国和中国台湾，都是通过在制造上发力加上核心技术研发突破，迅速跻身一流行列。
除技术创新外，美国半导体企业在商业模式上也创新不断，包括建立世界上第一个完善的风投机制，Win联盟，苹果的软硬件闭环生态，行业标准的制定等等。商业模式创新使美国半导体企业把持了行业话语权，而技术创新则是商业模式创新的基石，两者互为表里，形成美国芯片产业的竞争优势，也是实现芯片霸权的白道。
When the "white way" is difficult to work, the U.S. government will personally escort it and use counter-market and anti-globalization "underworld measures" such as sanctions and suppression, tariff barriers, and reorganization of industrial chains to eliminate competitors. Before the 1980s, U.S. chips dominated the global market, relying on the white way; from the 1980s to the present, the United States used both black and white, and over time, in order to maintain chip hegemony, the United States used "underworld tactics" It may go beyond the "white way", as evidenced by the recent introduction of the "Chip and Science Act".
For the challenger of the US chip hegemony, the key to breaking the game is the breakthrough of core technology. The premise of Japan and South Korea's strong ascendance is the breakthrough of core technology under the leadership of their own governments, which has strong practical significance for the current situation. After the breakthrough of core technology, it will become irreplaceable in the global industrial chain and will not be replaced .
"Secrets of the Integrated Circuit Industry" by Yukio Shimura
"Japan's Semiconductor Industry Development and Japan-US Semiconductor Trade Friction" by Feng Zhaokui
"Amazing achievements, Silicon Valley Range Rover, the cradle of high technology", by Michael Malone
"South Korea's Science and Technology Development Model and Experience", author / Li Donghua
Moxian Liu: "Seeing the Frontiers of High Performance Computing from IBM's "Roadmap of the Strongest Quantum Computing"
Bryan McMahon: "AI is leading a new scientific revolution"
👇 Click "Watching" to share insights