China's chip manufacturing: After ten years of drinking ice, the passion is hard to cool down
[Copy link]
Repost an article [Author | Sister Dandan Source | Cool Play Lab (ID: coollabs)]
The Huawei incident has become a big deal.
Even my mother asked me, "Is Huawei going to be finished?"
When the time came, the old lady asked an ultimate question, "Can our own chips in the mainland work?"
This reminds me of a video that is known as "the most amazing interview on CCTV."
Reporter: How is our steel production capacity so far?
Section Chief Ma: We have a long history of steel development and a large number of industrial workers.
Reporter: So how is our steel production capacity so far?
Section Chief Ma: We implement the fine steel strategy in order to improve equipment level, save energy and reduce emissions, and promote the development of circular economy in the process of adjusting layout.
One is really daring to ask, and the other is really unable to answer.
Once TSMC cuts off supply to us, the true status of mainland China's chips will be that it will lag behind the West by 5-10 years or even more.
Chip production is roughly divided into three major links: chip design, chip manufacturing, and chip packaging and testing.
The chip design tools used are foreign EDA software. Even Huawei HiSilicon, China's best chip design company, has only just begun to "de-Americanize."
The world's largest and best chip manufacturer is TSMC. TSMC has now achieved mass production of 5-nanometer process technology, while the best chip manufacturer in mainland China, SMIC, has just completed mass production of 14-nanometer process technology.
The lithography machines used to manufacture the most advanced process chips have to be imported from ASML in the Netherlands. But now they are no longer available because the United States does not allow it.
The best lithography machine manufacturer in mainland China is Shanghai Microelectronics, which currently produces 90-nanometer lithography machines.
China's best performance is in the packaging and testing stage, but the profit of this stage is not high.
The reality is that we are lagging behind.
No one cares that the chip industry is an extremely money-burning industry, with billions or tens of billions invested but not even a splash.
No one cares that the chip industry is a time-consuming industry and that results may not be seen for several years or even more than ten years.
No one cares about the talent gap in the domestic chip industry, which has reached more than 300,000.
No one cares that the chip industry is a "winner takes all" industry - the boss eats meat, the second one drinks soup, the third and fourth ones lick the bottom of the bowl, and as for the rest, it all depends on fate.
No one cares. Due to the Wassenaar Arrangement of 1996, Chinese engineers cannot enter the core departments of European and American semiconductor companies, and Chinese companies cannot buy the latest second-generation equipment.
Our latest research is based on technology that is already two generations outdated.
No one cares about these things, since foreign chips are available for purchase. It is right to criticize domestic chips.
But even so, there are still a large number of people who stick to this industry and never leave.
Even if you drink ice for ten years, your blood will not cool down.
I know their stories. So I want to say a few words for mainland chips. I am afraid that if I don't say it, it will be too late.
I hope everyone can understand that the gap in this industry is getting bigger and bigger - the top two in market share will make all the money. This will also support them to invest more money in research and development.
If you find that the mainland chip industry is narrowing the gap, I hope you understand that they have paid something heavier than sweat.
The first company to help Huawei was SMIC, whose founder was Zhang Rujing.
Before founding SMIC, Zhang Rujing was already a well-known figure in the industry. He joined the American semiconductor giant Texas Instruments.
During his more than 20 years at Texas Instruments, Zhang Rujing built 9 factories in the United States, Japan, Singapore, Italy and other places.
He also became a recognized "factory building maniac" in the semiconductor industry.
At this time, Zhang Rujing's father, Zhang Xiulun, flew from Taiwan to the United States and asked him a question:
"When will you come to the mainland to build a factory?"
Zhang Xiulun was once a famous steelmaking expert in my country. During the Anti-Japanese War, the 21st Arsenal he and his wife presided over produced 90% of China's heavy machine guns.
In 1949, he brought Zhang Rujing, who was just one year old, to Taiwan.
Now that his son has achieved success in his career, he also hopes that his son can build a chip factory in mainland China.
Zhang Rujing was touched by his father's words.
He was taught from childhood that " we are proud to be Chinese, we are Chinese ." Now, it is time to contribute to the country.
After considering for a few days, Zhang Rujing made the most important decision in his life - quit his job and return to China to start a factory.
Within a month, he returned to Taiwan and founded World Advanced Semiconductor.
Under the leadership of Zhang Rujing, World Semiconductor quickly built two factories and became the third largest chip company in Taiwan.
The company's performance is getting better and better.
But at the same time, World Advanced Semiconductor also attracted the attention of TSMC. In the chip industry, when a rising star appears, the best way to destroy it is to acquire it.
So, TSMC initiated the acquisition of World Advanced Semiconductor.
Zhang Rujing had no objection to this acquisition because TSMC offered a high price of US$5 billion, which was almost 8 times the price of World Semiconductor's stock at the time.
But Zhang Rujing put forward a condition, " After the acquisition is completed, World Science's third factory must be built in the mainland."
No one objected to the proposal at the time and the acquisition was completed quickly.
But after that, Zhang Rujing repeatedly approached the head of TSMC to inquire about building a factory in mainland China, but he did not receive a clear response.
TSMC is not prepared to fulfill its promise.
After some consideration, Zhang Rujing decided to resign.
TSMC disagreed and warned Zhang Rujing, "If you leave, your large number of shares in TSMC will be reclaimed."
Then, no need.
He no longer wanted the money, the company, or the stocks. Zhang Rujing resolutely left Taiwan and came to the mainland to build a factory.
This time, Zhang Rujing turned his attention to Shanghai.
At that time, the chip manufacturing process in mainland China was stuck at 0.5 microns and stagnated.
Building a wafer processing plant is too expensive. Even a common 8-inch production line costs 1 billion US dollars.
Against this backdrop, Zhang Rujing played the role of a pioneer.
To show his determination, he moved his wife, children, and even his 90-year-old mother to Shanghai.
When I first came to Shanghai, I was in need of everything: talent, equipment, capital, and technology to build a factory.
But when they learned that Zhang Rujing was going to build a factory in mainland China, in addition to more than 100 of his former overseas colleagues at Texas Instruments and World Semiconductor who were willing to follow him, more than 300 Taiwanese compatriots also followed him to Shanghai .
Everyone has seen Zhang Rujing's sense of mission, his patriotism, and his desire to return to serve his country . This time, he finally has the opportunity.
These things made Zhang Rujing very surprised. He found the backbone for the initial construction of the factory.
In terms of funding, Zhang Rujing used all the resources and reputation he had accumulated over 20 years. In just one year, he miraculously raised 1 billion US dollars.
You should know that the initial investment in Project 909, the largest electronic industry project since the founding of the People's Republic of China, was 10 billion RMB.
One Zhang Rujing is equivalent to a national project.
After solving the problems of talent and funding, technology and equipment became the biggest problems.
After the end of the Cold War, Western countries, under the leadership of the United States, signed the Wassenaar Arrangement , which restricts the export of various commodities or technologies such as advanced materials, electronic devices, and military products.
China is a "banned country" under this agreement .
Even if the price China offers to purchase technology is difficult for other countries to refuse, the United States will intervene and directly interrupt the transaction.
In order to purchase advanced equipment and technology, Zhang Rujing visited the five major churches in the United States to endorse his character.
And repeatedly promised that SMIC's products would only be used for commercial purposes and not in military and other fields, only then it obtained the export license from the United States.
Everything is ready, just waiting to start.
On August 1, 2000, SMIC broke ground and just 13 months later, in September 2001, SMIC officially went into production.
13 months also broke the world record for the speed of building a wafer processing plant.
At the same time, SMIC's production line below 0.25 microns also pushed mainland chips to the nanometer level for the first time.
In order to attract more overseas talents and Chinese to return to China, Zhang Rujing built the SMIC School from kindergarten to high school around the factory.
Moreover, some workers can also go to Shanghai Institute of Microelectronics for further studies - this is something they had never dared to think about before.
At the same time, more than 1,500 housing units were built.
Family, dream, career, all can be found in SMIC. In this atmosphere, SMIC continues to attract overseas talents.
With a sense of belonging, work efficiency will naturally become high.
In just three years, SMIC has four 8-inch production lines and one 12-inch production line.
It is no exaggeration to say that in these three years, the mainland chip industry has been accelerated by 30 years.
By 2003, SMIC, with the concerted efforts of thousands of people, had proudly built six factories, achieved 90-nanometer integrated circuit line width, and became the world's third largest foundry.
Mainland China's chip manufacturing industry has ushered in its most glorious moment.
However, TSMC has been keeping an eye on Zhang Rujing.
In fact, it was not just TSMC. When Zhang Rujing left with more than 300 people, Taiwanese people also panicked.
In 2000, the Taiwanese authorities fined Zhang Rujing $150,000 on trumped-up charges and warned him to immediately withdraw his investments from the mainland.
But Zhang Rujing refused to give in and announced that he would give up his Taiwanese household registration.
The Taiwanese government was so angry that it couldn't catch the man and had no choice but to strictly ban all Taiwanese technology companies from entering the mainland.
At the same time, TSMC also started its own operations.
They have been preparing silently, waiting to deal a fatal blow to Zhang Rujing and SMIC.
TSMC founder Morris Chang
In August 2003, SMIC was about to be listed in Hong Kong.
Just a few days before its listing, TSMC sued SMIC in California, USA, and demanded US$1 billion in damages.
TSMC is not suing SMIC for infringing its patents because the two parties have already cross-licensed their patents.
TSMC sued employees of SMIC for stealing TSMC's trade secrets.
Many of SMIC's engineers come from TSMC, which in itself puts Zhang Rujing at a disadvantage - whether intentionally or unintentionally, engineers will use TSMC's menus and other information on certain projects.
No one can explain these clearly.
You know, SMIC's revenue was only 360 million US dollars in 2003. Losing this lawsuit would be like killing SMIC.
The lawsuit dragged on for two years, and SMIC could no longer afford to delay. SMIC chose to settle out of court, paying TSMC $175 million in installments over six years.
Because the lawyer hired at that time was not professional, SMIC was tricked by TSMC again in the "settlement agreement".
TSMC set up a "third-party escrow account."
SMIC must store all its technologies in this account for TSMC to "freely inspect". Its core competitiveness and the most private part are available for others to rummage around at will.
In this way, it becomes easier for TSMC to cause trouble for SMIC.
One year and seven months after the "settlement agreement" was reached, TSMC sued SMIC again, accusing SMIC of violating the "settlement agreement."
TSMC asked SMIC to pay the remaining 130 million yuan of settlement money in one lump sum, and accused SMIC of infringing 82% of TSMC's patents in the 0.13-micron process.
But at that time, everyone in SMIC felt that reaching a settlement was a good thing - it would be beneficial to the company's long-term development.
They are all trying their best to implement the "settlement agreement".
This lawsuit is essentially vicious competition and deliberate sabotage. The biggest reason may be that SMIC's technology at that time had entered the 45-nanometer threshold.
Even the Japanese media at the time said, "TSMC is far ahead in the world, but its development in the mainland market lags behind SMIC. Preventing SMIC from growing bigger is the real reason why TSMC took SMIC to court."
SMIC also responded strongly, firmly denying its infringement and preparing sufficient evidence.
Everyone in the company felt that the accusation was "groundless" and that they could win the lawsuit.
The entire company is operating normally and Zhang Rujing's work is also proceeding normally.
Just one day before the verdict was announced, Zhang Rujin convened a regular meeting of the project team to arrange daily work.
However, SMIC lost again.
The price this time was even more painful - on top of the $175 million, they had to pay an additional $200 million, plus 10% of the shares.
As a result, SMIC has basically lost its competitiveness.
In addition, TSMC also made a requirement: Zhang Rujing must leave SMIC.
The moment he learned the trial result, Zhang Rujing burst into tears on the phone.
On the third day, he resigned.
Before leaving, TSMC asked him to sign a non-compete agreement - starting from 2010, he was not allowed to engage in chip-related work within three years.
After saying "I am responsible for the matter" and telling his colleagues "not to be defeated", the 61-year-old man left SMIC, where he had worked for 9 years.
The light that had just been ignited by mainland chips has gone out again.
After Zhang Rujing left SMIC, a man named Jiang Shangzhou took over SMIC.
A few years ago, it was Jiang Shangzhou who received Zhang Rujing and facilitated the establishment of SMIC in Shanghai.
Apart from his recognition of Zhang Rujing's abilities, few people know that in 1998, Jiang Shangzhou had carefully studied Taiwan's semiconductor industry.
Then came to a conclusion:
"If Shanghai supports the semiconductor industry at this time, then between 2015 and 2020, the technology level and production scale of Shanghai's integrated circuit production lines will likely surpass Taiwan, China."
We have always said that Shanghai does not believe in the Internet.
China's three major Internet giants are BAT (Baidu, Alibaba, and Tencent), with Baidu in Beijing, Alibaba in Hangzhou, and Tencent in Shenzhen.
In fact, Jack Ma started his business in Shanghai. But no one in Shanghai believed what he said, and he couldn't recruit anyone, so he had to return to his hometown of Hangzhou.
In 2008, the Shanghai Municipal Party Committee Secretary specifically asked all relevant staff at a meeting to reflect on "why Shanghai couldn't keep Jack Ma".
And, "it is quite regrettable that Shanghai has lost such a giant enterprise that has grown from a small business."
For a long time, the most outstanding Internet company in Shanghai, China's financial center, is Jiayuan.com.
Because of Shanghai, choose to believe in the semiconductor industry.
In the past "golden twenty years of the Internet", the Shanghai government has invested a lot of money in the semiconductor industry to support the development of mainland chips.
What they reaped the first reward was a carefully packaged scam and the ridicule of the whole world.
On February 26, 2003, leaders from all walks of life in Shanghai attended a press conference. The press conference was hosted by the Shanghai Government Information Office, and the mayor of Shanghai excitedly introduced it to the people of the country:
"This is Shanghai's technological image."
The product released was called "Hanxin No. 1" - a chip that can perform 200 million operations per second. Its performance even exceeded that of Intel's products of the same period.
It took Intel 40 years to develop such a product, but it only took 16 months to develop "Hanxin No. 1".
Professor Chen Jin, the inventor of "Hanxin No. 1", is also called the "new godfather of domestic chips" by many media.
With such achievements, Shanghai will of course continue to make additional investments. At the same time, Chen Jin was also awarded various titles such as "Shanghai Science and Technology Entrepreneurship Leader" and "Dean of the School of Microelectronics of Shanghai Jiao Tong University".
For a time, Chen Jin gained both fame and fortune.
Three years later, Chen Jin was exposed for the scandal of "sanding off the logo on the Motorola chip and replacing it with the word 'Hanxin'".
The research and development process of the so-called "first domestic chip" is just a piece of sandpaper.
After the news was confirmed, everything was in chaos. Chen Jin became a "scum" that everyone wanted to beat, and the mainland chip industry was ridiculed by the whole world.
The American media directly stated, " The Chinese are only capable of copying, they will never be able to truly develop chips."
The city that suffered the most was Shanghai.
Will you continue to support it? No, it's said that "once bitten by a snake, you will be afraid of the rope for ten years". Besides, Internet companies are so profitable, why bother with the chip industry, which is "costly, laborious and thankless"?
Logically speaking, Shanghai should be the province or city that is best at "counting accounts" in China.
But it is still persisting - it is still stuffing all kinds of resources into a place called Zhangjiang High-tech Park.
This time, they got SMIC.
In the first few years, SMIC's development exceeded everyone's imagination. Jiang Shangzhou's judgment seemed to be coming true.
However, the clash with TSMC and Zhang Rujing's departure still made Jiang Shangzhou's wish come to nothing.
He no longer even had the chance to see such a scene - two years after taking charge of SMIC, Jiang Shangzhou died of illness.
At that time, SMIC was in a precarious situation.
At that time, the entire chip industry was in a downward cycle, and coupled with the impact of the financial crisis, SMIC suffered losses every year.
By 2010, SMIC had suffered a loss of 3 billion.
Fortunately, the team left by Zhang Rujing is still there, and SMIC's basic foundation has been stabilized.
However, SMIC, which has lost its competitiveness, is falling further and further behind TSMC - TSMC's annual R&D expenses are even more than SMIC's 20 years.
In order to "survive", SMIC can only "pick up" what TSMC has left. When TSMC upgrades its technology, SMIC will take over the market that TSMC has given up.
When 4G emerged, TSMC started using 10nm technology, while SMIC had just started mass production of 28nm.
It was at this time that Liang Mengsong joined SMIC.
Liang Mengsong is also a former TSMC general. After leaving TSMC, he joined Samsung as the director of the R&D department and upgraded Samsung's manufacturing technology directly from 28 nanometers to 14 nanometers.
TSMC also sued Liang Mengsong on the grounds of "continuously leaking TSMC's business secrets to Samsung."
The final result is similar to that of Zhang Rujing - he cannot continue to provide any services to Samsung in any capacity or in any other way within three years, and after the restriction ends, he cannot work for TSMC's competitors.
Perhaps it was the same experience that made him choose to join SMIC.
298 days after Liang Mengsong joined SMIC, SMIC mastered the 14-nanometer process technology.
At the same time, phased progress has been made in the process technology equivalent to 7 nanometers.
This also means that SMIC can provide products for the "first tier" of performance machines.
Huawei also officially cooperated with SMIC at this time.
Just when SMIC was thriving, the 61-year-old man who was forced to leave SMIC had already started the third business of his life.
When he signed the non-compete agreement, he was told that he could work in the LED and solar energy fields in related industries.
After leaving SMIC, he set up a new factory to produce chips for LED products.
Production in the new field is not difficult. There was even news that Zhang Rujing was preparing to retire and engage in some work related to education or charity.
And Zhang Rujing completely disappeared from people's sight.
In 2014, a new company called Shanghai Xinsheng was established in Shanghai Lingang New City.
The material with the highest chip cost is silicon. There are two main paths for the development of silicon wafer technology: one is higher purity and the other is larger size.
The larger the silicon wafer, the more chips can be cut and the lower the cost.
Shanghai Xinsheng, which makes large silicon wafers, was later acquired by Shanghai Silicon Industry Group.
On April 20, 2020, Shanghai Silicon Industry Group went public, and its stock price soared 180% on the same day.
The founder of Shanghai Xinsheng is Zhang Rujing.
That year, he was 67 years old.
Judging from the results, Zhang Rujing is a failed entrepreneur. He started three businesses and was forced to leave three times.
From a practical point of view, Shanghai's 20 years of investment has been a failure. The development of the mainland chip industry has been a failure - there is a visible gap between us and the world's first-class level.
But among them, everyone who has contributed to this industry deserves to be remembered.
They started from scratch and developed the mainland chip industry to what it is today.
Admittedly, the industry is not yet strong.
But they gave up their favorable environment and a potentially brighter future because they all had an obsession in their hearts:
China cannot be without chips! We cannot let others strangle us in chips.
At least, we're on our way.
On May 6, 2020, an authoritative third-party semiconductor research organization released the global top ten semiconductor sales rankings. Huawei HiSilicon made history and made it into the list for the first time, ranking 10th.
On May 15, 2020, the National Integrated Circuit Fund and the Shanghai Integrated Circuit Fund announced an investment of 16 billion yuan in SMIC.
SMIC has also made breakthroughs in "N+1" and "N+2" technologies, with performance roughly equivalent to that of TSMC's 7nm chips.
What about Shanghai, which has been providing chip transfusions to the mainland for 20 years?
If you open the map of mainland China's chip industry, you will find that SMIC, Spreadtrum, RDA Microelectronics, and VeriSilicon... these newcomers in the mainland chip industry are all in Shanghai.
They also account for nearly 50% of the output value of the mainland chip industry.
During the golden 20 years of the Internet industry, Shanghai chose to stick to it. It used 20 years of blood transfusion to leave the last torch for the mainland chip industry.
Now, the flames are spreading across the prairie.
To this day, many people still remember Jiang Shangzhou's words, "Shanghai's integrated circuit production line technology level and production scale may surpass Taiwan, China."
The last 20 years did not belong to mainland chips or Shanghai.
What about the next 20 years?
The first thing that came to my mind was the sentence "When the royal army conquers the Central Plains in the north, don't forget to tell your father during the family sacrifice."
Salute to all those who have worked hard for the mainland semiconductor industry and the Chinese semiconductor industry!
PS:
At the end of this story about mainland chips, there is a little easter egg - in 2018, Mr. Zhang Rujing started his fourth entrepreneurial journey in his life.
He founded the first CIDM model semiconductor company in China.
IDM is a model in which a company independently completes the three links of chip design, manufacturing, packaging and testing. This model requires huge amounts of capital and talent.
CIDM is a shared IDM. Compared with TSMC's production-focused model, it is the best choice for the mainland at present.
Zhang Rujing also said that this would not be the last time he started a business in his life.
We believe.
At the same time, we also believe that the story about mainland chips has just begun.
*References
Boss Dai, "The History of Chinese Chips", published in "Fantong Boss Dai"
20190942 "Zhang Rujing: The Journey of "Core", published in the "Financial People Weekly" column
| 
提升卡
变色卡
千斤顶
京公网安备 11010802033920号
