Shulou(Shulou.com)11/24 Report--

On September 12, 1958, Jack Kilby (Jack Kilby) from Texas Instruments successfully integrated five components, including germanium transistors, to make the world's first germanium integrated circuit.

In July of the following year, Robert Noyce (Robert Norton Noyce) of American Xiantong Semiconductor Company successfully invented the world's first silicon integrated circuit based on silicon plane technology.

Jack Kilby (left) and Robert Noyce (right) as you know now, the inventions of these two bigwigs are of great significance. The emergence of integrated circuits has not only strongly promoted the miniaturization of electronic devices, but also laid the foundation for the comprehensive arrival of the chip era.

After the birth of DRAM in the 1960s, with the development of computer technology, the electronic industry began to try to use integrated circuit technology in the field of computer storage.

At that time, semiconductor storage technology was divided into ROM and RAM. ROM is a read-only memory, stored data will not be lost due to power outage, also known as external memory. RAM is a random access memory, which is used to store operational data. After a power outage, the data will be lost, also known as memory.

Today, we focus on the field of RAM.

In 1966, Robert H. Dennard, from the IBM Thomas J. Watson Research Center, first invented DRAM memory (dynamic random access memory).

Based on the "MOS transistor + capacitor structure", the Robert Danard memory has the advantages of low energy consumption, fast reading and writing speed and high integration. Until now, our computer memory, mobile phone memory, graphics card memory, etc., are based on DRAM technology.

In June 1968, IBM registered a patent for the transistor DRAM. However, just as they were preparing to industrialize DRAM, the US Department of Justice launched an antitrust investigation against them.

These investigations have delayed the industrialization of IBM's DRAM, creating opportunities for other companies.

Soon after, in 1969, Advanced Memory System (Advanced memory system) of California successfully produced the world's first DRAM chip (capacity only 1KB) and sold it to the computer manufacturer Honeywell.

After receiving this batch of DRAM chips, Honeywell found some problems in the process. So they went to a new company and asked for help.

The company was co-founded by Robert Noyce (the above-mentioned silicon integrated circuit inventor) and Gordon Moore (Moore's Law) in 1968.

Robert Noyce (left) and Gordon Moore (right) after Intel was founded, its main business was to develop transistor semiconductor memory chips.

At that time, there were two main research directions of semiconductor technology, namely bipolar transistor and field effect (MOS) transistor. Intel itself did not know which direction was right, so two research groups were set up to follow up on the two technical directions.

In April 1969, the bipolar team was the first to make a breakthrough with the introduction of a 64bit-capacity static random access memory (SRAM) chip, C3101. This chip is Intel's first product, and its main customer is Honeywell.

The Intel C3101 field effect management team did not want to lag behind. In July 1969, they launched the 256bit capacity static random access memory chip-C1101. This is the world's first large-capacity SRAM memory.

In October 1970, the FET team made continuous efforts to successfully launch its first DRAM chip (which is also considered to be the first mature commercial DRAM chip in the world)-C1103.

The Intel C1103, with 18 pins, capacity 1Kbit, sells for $10. After its launch, C1103 achieved great success and soon became the best-selling semiconductor memory in the world, serving important customers such as HP, DEC and so on.

With the help of C1103, Intel has also grown rapidly. In 1972, Intel employed more than 1000 people and earned more than $23 million a year. In 1974, the global market share of Intel DRAM products reached an astonishing 82.9%.

Intel's early team came at a time when Intel was making a lot of money in DRAM, while its competitors were on the rise.

In 1973, American Texas Instruments (TI), Mostek and other manufacturers successively entered the DRAM market.

After Intel launched the C1103, Texas Instruments disassembled and copied it, and studied the structure and technology of DRAM through reverse engineering. Later, in 1971 and 1973, they launched 2K and 4K DRAM respectively, becoming a strong competitor to Intel.

Texas Instruments, Intel's old rival Mostec, founded by L.J.Sevin, a former chief engineer at Texas Instruments Semiconductor Center (1969), is also technically strong.

In 1973, they launched a 16-pin DRAM product, the MK4096, which also challenged Intel's market position (other companies have 22 pins, and the fewer pins, the lower the manufacturing cost).

In 1976, Mostek launched MK4116, using POLY-II (double-layer polysilicon gate) process, with a capacity of 16K. This product has been a huge success, reversing the market competition and increasing its DRAM market share to 75%.

MK4116 unfortunately, it wasn't long before the company hit a low ebb because of a hostile takeover from the capital market, a sharp change in the ownership structure, severe turbulence in management, and a rapid loss of technical staff.

In 1979, the company was acquired by United Technologies Corporation (UTC). Later, it was resold to St Semiconductor.

In October 1978, four Mostec technicians left to start a new storage technology company in the basement of a dental clinic in Idaho.

The company, later known as the storage giant, Micron.

Besides the success and defeat of Nippon Semiconductor, the founder team of Magnesia, Intel faces a greater threat from abroad. More specifically, it comes from Japan.

In the 1970s, the Japanese economy rose at a high speed. In order to occupy a favorable position in the global science and technology industry chain, they have made a careful layout in the field of semiconductor technology.

In 1976, Japan established the VLSI Joint Research and Development body (VLSI:THE VERY LARGE SCALE INTEGRATED, very large-scale integration) through the national system.

The joint research and development body has a total of six laboratories, which specialize in high-precision machining technology, silicon crystallization technology, process treatment technology, monitoring and evaluation technology, device design technology and other fields.

Soon after, the joint research and development body successfully conquered the core semiconductor processing equipment such as electron beam lithography and dry etching devices, as well as leading process technology and semiconductor design capabilities, laying the foundation for the take-off of Japan's semiconductor industry.

In 1977, with the help of the VLSI project, Japan successfully developed 64K DRAM, which caught up with the R & D progress of American companies.

In the 1980s, Japanese manufacturers (Fujitsu, Hitachi, Mitsubishi, NEC, Toshiba, etc.) continued to overtake American companies by virtue of quality and price advantages.

In 1986, the global market share of memory products in Japan rose to 65%, while that in the United States fell to 30%.

In the fierce market competition, the American Intel Company directly announced that it had abandoned the DRAM market (1985). The only one that can survive in the gap between Japanese manufacturers is Motorola.

Global ranking of semiconductor enterprises (1987) mantis catches cicadas, followed by yellow sparrow. At a time when Japanese semiconductor manufacturers are on the verge of unification, subtle changes have taken place in the external political environment.

In 1985, the Cold War atmosphere between the United States and the Soviet Union continued to weaken, and trade frictions between Japan and the United States continued to increase. Under the pressure of huge fiscal deficit, the Reagan administration began to turn its attention to suppressing the Japanese economy.

This year, the United States dominated the famous Plaza Accord, forcing the appreciation of the yen. At the same time, the American Semiconductor Association has also launched an anti-dumping lawsuit against Japanese semiconductors and other products. Later, the two countries reached an agreement to monitor the price of Japanese semiconductor products.

Under the repeated blows, the market share of Japanese semiconductor products plummeted and soon lost its dominance.

With the rise of Korean semiconductors, has the market share conceded by Japanese manufacturers been taken away by American manufacturers?

No.

As the saying goes, "the mantis catches the cicada, the yellow sparrow comes after". While Japanese manufacturers lose power quickly, another competitor of the United States kills out, that is, South Korea.

As early as when Japan launched the VLSI project, the South Korean government was not idle. They set up the Korea Institute of Electronic Technology (KIET) in the Kamei industrial district in Gyeongsangbei do, which attracts semiconductor talents from the United States with high salaries and focuses on the research and development of key integrated circuit technologies.

In addition to KIET, South Korea's chaebol, such as Samsung, LG, Hyundai and Daewoo, have also taken a fancy to the market prospect of semiconductor technology by purchasing and introducing technology patents and processing equipment to digest and absorb them and accumulate technical strength.

In 1984, Samsung Semiconductor built its first memory factory and mass produced 64K DRAM. No one thought that this unknown South Korean enterprise would become a "giant" of the industry in the future.

It is said that the DRAM industry has experienced nearly 40 years of development since the 1980s. If you use one word to describe these forty years, it is "bloodshed".

The reason is very simple, DRAM semiconductor industry, the biggest feature is its cyclical law. People in the industry have concluded that for every year DRAM semiconductor storage makes money, it will lose money for two years, the so-called "one profit and two losses".

Under this strong cyclical law, it is very difficult to survive for a long time. DRAM manufacturers need to have strong cash flow and financing ability to maintain high-intensity R & D expenditure and maintain the stability of the team.

In the loss cycle, DRAM manufacturers need more money to survive. In the boom cycle, you can't be careless. Manufacturers need to be very careful when choosing the timing of capacity expansion. Otherwise, it may lead to oversupply and turn profits into losses.

Forty years ago, there were about 40-50 DRAM manufacturers around the world. Today, there are only three, which shows the cruelty of the competition.

Over the past 40 years, one enterprise has not only survived, but also killed numerous competitors and occupied a dominant position for a long time. This enterprise is the aforementioned Samsung.

Some students may have heard the story of Samsung Electronics. They have adopted a "killer's mace" strategy-counter-cyclical input-that has been included in the textbooks of countless business schools.

To put it simply, counter-cyclical investment is to make use of the characteristics of the cyclical development of the industry, when the industry enters a trough, when competitors are shrinking their scale, do the opposite, increase investment, expand production capacity, and further depress prices, thus causing competitors to aggravate losses or even close down.

In other words, everyone is burned, but I have more money, burn you to death, I will continue to live.

Samsung, relying on the national strength of South Korea, has successively adopted the strategy of "counter-cyclical investment" to kill numerous rivals and become the leader in the field of semiconductor storage.

Next, let's take a closer look at what has happened over the past few decades.

The first "countercyclical input"

Samsung's first counter-cyclical investment took place in the mid-1980s as mentioned earlier.

At that time, the fierce war between Japan and the United States was raging, the DRAM market was generally depressed, and prices plummeted. The price of DRAM chips fell from $4 per chip (1984) to $0.30 per chip (1985).

When Samsung built the factory to launch the 64K DRAM, the production cost was $1.30 per piece. In the face of the cold winter of the industry, Samsung not only did not shrink its investment, but began to reverse investment and expand production capacity.

By the end of 1986, Samsung Semiconductor had accumulated a loss of US $300 million and had a complete deficit in equity capital, which was close to bankruptcy.

During the critical period, the South Korean government stepped in to "rescue the market", investing a total of nearly $350 million and endorsing it in the name of the government, raising $2 billion in individual financing for Samsung.

Later, Japanese semiconductors were overturned by the United States, coupled with the prosperity of the industry brought about by the hot selling period of PC computers, which led to a smooth turnaround of Samsung and ushered in performance growth.

Soon after, the Korean DRAM manufacturers represented by Samsung gradually eroded the market share given up by Japanese semiconductor enterprises and occupied the dominant position in the market.

The second "countercyclical input"

In 1992, an explosion occurred at the Sumitomo resin factory in Japan, resulting in a shortage of raw materials and a sharp rise in memory prices. This year, Samsung took the lead in launching the world's first 64m DRAM.

In 1993, the global semiconductor market began to weaken again. At this time, Samsung repeated its old trick and adopted a second "counter-cyclical investment". They invested in an 8-inch wafer production line to produce DRAM.

In 1995, Microsoft's Windows95 Windows operating system was released, which greatly stimulated the demand for memory and led to a sharp rise in memory prices. Samsung's investment paid off. With the benefit of hindsight, major manufacturers around the world have invested in expanding production capacity.

The good times did not last long. At the end of 1995, after the manufacturers'8-inch wafer factories were put into operation, the production capacity increased sharply and the supply of DRAM exceeded the demand. As a result, the seller's market became the buyer's market again, and prices began to fall again.

In this case, manufacturers are forced to cut production and reduce the scale of investment.

Samsung continues to expand its investment. In 1996, Samsung launched the world's first 1GB DRAM, establishing its own leading position in the industry.

From 1996 to 1998, DRAM continued to be in a downward cycle.

In 1999, the downward trend of DRAM prices eased somewhat. Because of the emergence of the Internet bubble, the DRAM industry has entered a short period of prosperity.

This year, in a highly competitive environment, several major changes have taken place in the memory industry:

In South Korea, Hyundai memory merged with LG Semiconductor to form Hyundai Semiconductor, which was later split from Hyundai Group (2001) and renamed Hynix (Hynix).

On the American side, Magnesia acquired the memory department of Texas Instruments.

On the Japanese side, Hitachi, NEC, Mitsubishi Electric's DRAM business integration, together to set up Erpida (ELPIDA).

In Europe, the semiconductor department of Siemens Group is independent and established Yiheng Technology. A few years later, in 2002, it changed its name to Infineon. Later, in 2006, the Infineon Technology memory Division was split and became Qimonda.

Among the top five global DRAM market shares in 2000, there were two Korean manufacturers, Samsung (23.00%) and Hyundai (19.36%).

Soon after, the Internet bubble burst and the global economic crisis broke out. The PC market has been hit hard, the market demand for DRAM has fallen sharply, and prices have plunged again.

In 2001, the DRAM market halved from $28.8 billion to $11 billion.

From 2002 to 2006, the DRAM market gradually recovered from the trough, and the overall growth situation was good.

In 2006, Samsung developed the world's first 1GB DRAM of 50nm process. Hynix developed the highest speed 200MHz 512MB Mobile DRAM in the world at that time.

During that period, the DRAM market gradually formed a top five pattern, namely: Samsung (South Korea), SK Hynix (Korea), Qimengda (Germany), Magnesia (USA) and Erpida (Japan).

The third "countercyclical input"

In 2007, Microsoft launched the Vista system. The system consumes a lot of memory, and DRAM manufacturers expect a big increase in memory demand, so they increase their production capacity one after another.

But in fact, Vista sales are very poor, did not drive the memory market, leading to overcapacity again.

More sadly, in 2008, the financial crisis broke out, which made the DRAM market even worse. The price of memory has fallen all the way, even below the cost of materials.

In this critical period, Samsung sacrificed its "counter-cyclical investment" for the third time, further expanding production capacity and aggravating the industry's losses.

In the spring of 2009, Qimengda, the third-ranked German manufacturer, declared bankruptcy and European manufacturers officially withdrew from the DRAM market.

Qimengda 2011, DRAM supply once again exceeded real demand, prices plummeted. This time, Erpida failed to survive and declared bankruptcy, marking the complete withdrawal of Japanese manufacturers from the DRAM industry.

Erpida chip then, the top five become the top three, only Samsung (South Korea), Magnesia (USA) and Hynix (Korea) remain in the field of DRAM. Taken together, the market share of the three companies is more than 93%.

The present situation of █ DRAM Technology

There has been no significant change in the market pattern of DRAM memory since 2011. However, the user demand and market environment of DRAM have changed a lot.

In addition to the traditional PC, with the rapid development of mobile Internet and Internet of things, smart phones, wearable devices, Internet of things devices (cameras, etc.) rise rapidly, greatly driving the demand for DRAM.

The development of cloud computing, big data and AI artificial intelligence has promoted the increase in the number of data centers, resulting in a sharp increase in servers and network equipment, as well as an increase in DRAM sales.

These requirements gradually subdivide DRAM into standard DRAM, mobile DRAM, drawing DRAM, niche DRAM and other categories.

Standard DRAM is mainly used in PC, server and so on. Mobile DRAM is mainly LPDDR, which is used in smartphone, tablet and other scenarios. The drawing DDR is used for the video memory (GDDR) of the video card. Niche DRAM is mainly used in LCD TV, digital set-top box, network player and other products.

The strong demand for LPDDR multi-product scenarios has driven up the price of DRAM. Around 2018, the outbreak of demand for digital currencies such as bitcoin ushered in a rare "golden age" for the DRAM market.

After 2019, due to pre-capacity expansion and destocking factors, memory prices fell more. The price of the cryptocurrency market collapsed and the smartphone market entered a mature period, which made the market demand weak and DRAM entered the trough again.

According to the data released by relevant institutions, it is a period of improvement in the DRAM market from the second half of 2020 to May 2022.

Since June this year, the DRAM market has plummeted. Sales fell 36% in June and 21% in July, a total collapse. According to the agency's forecast, the decline in the fourth quarter will further expand.

DRAM market plummeted next, let's look at the development of DRAM in recent years from a technical point of view.

For a long time, DRAM chips have been used to improve the storage density in a miniaturized process.

Every process upgrade of DRAM requires a lot of investment.

Take the update of 30nm to 20nm as an example, the number of lithography masks required by the latter has increased by 30%, and the number of non-lithography process steps has doubled. The requirement for clean room plant area has also increased by more than 80% with the increase in the number of equipment.

In the past, these costs could be offset by more chip output in a single wafer and a premium on performance. However, as the process continues to shrink, the gap between increased costs and revenue is gradually narrowing.

Around 2013, when the manufacturing process entered the 20nm, the manufacturing difficulty increased greatly. After 18/16nm, continued downsizing in the two-dimensional direction no longer has the cost and performance advantages.

As a result, DRAM chip manufacturers began to take a different approach and began to study the scalability in the Z direction. In other words, start promoting 3D packaging.

As an industry leader, Samsung is the first to achieve 3D DRAM from the perspective of packaging. They use TSV packaging technology to stack multiple DRAM chips, thus greatly improving the capacity and performance of a single memory stick. Later, various manufacturers have followed, 3D DRAM has become the mainstream.

In terms of product standards, the industry generally adopts the product standards developed by the solid State Technology Association (JEDEC), that is, the familiar DDR1-DDR5.

Image source: global Semiconductor Watch DRAM Big three, all have the mass production capacity of DDR5 / LPDDR5. Samsung is tinkering with DDR6 and is said to have completed the design by 2024.

In the chip process, the current expression of DRAM is different from that before. In the past, it was called directly 40nm and 20nm. Now, because the circuit structure is three-dimensional, linear measurement is no longer applicable, and there are terms such as 1x, 1Y, 1Z, 1 α, 1 β, 1 γ and so on.

According to the industry, the 10nm~20nm series process includes at least six generations, 1X is about equivalent to 19nm 15nm 1Y is equivalent to 18nm Magi 1Z is about 16-17nm Magi 1 α, 1 β, 1 γ corresponds to 12-14nm (below 15nm).

Image source: global Semiconductor Watch Samsung, SK Hynix and Magnesia entered the 1Xnm phase during 2016-2017, the 1Ynm phase from 2018-2019, and the 1Znm phase after 2020.

At present, the major manufacturers continue to approach 10nm. The latest 1 α nm is still in the 10+nm stage.

█ the past and present of China's DRAM Industry

Finally, let's take a look at the development of domestic DRAM industry.

China is not only one of the important markets of semiconductor memory in the world, but also a "must-compete place" for global semiconductor memory manufacturers.

However, seeking truth from facts, the development of our own DRAM industry lags far behind our competitors.

The start of domestic DRAM industry can be traced back to the 1990s.

At that time, Japan's NEC set up two joint ventures in Chinese mainland to produce DRAM.

The first is Shougang NEC, a joint venture between NEC and Shougang in 1991.

Since 1995, Shougang NEC has used a 6-inch 1.2 micron process to produce 4m DRAM (later upgraded to 16m). Later, when the global price of DRAM fell sharply in 1997, the NEC of Shougang was hit hard and never recovered. Later, Shougang NEC became a contract manufacturing base for NEC overseas and withdrew from the DRAM industry.

The second is Huahong NEC, a joint venture between NEC and Huahong Group in 1997.

Since September 1999, Huahong NEC has adopted 8-inch 0.35m process technology to produce 64m DRAM memory chips. After 2001, as NEC withdrew from the DRAM market, Huahong also withdrew from the DRAM industry.

In 2004, China began the second attempt of DRAM industry. It is SMIC that has taken action this time.

At that time, SMIC invested in Chinese mainland's first 12-inch fab (Fab4) in Beijing, mass-produced 80nm process in 2006, and produced DRAM for Qimengda and Erbita OEM.

The good times did not last long. In 2008, SMIC withdrew from the DRAM business due to SMIC's business adjustment. The second attempt failed.

In 2015, China's DRAM procurement volume was about US $12 billion, accounting for 21.6% of the global DRAM supply. The current situation of heavy reliance on imports prompted the third attempt of DRAM business in China.

The most representative of this attempt are the three major memory bases in Wuhan, Hefei and Xiamen. With the help of industrial policies at the national and local levels, these bases have invested a lot of capital (more than 250 billion yuan) to develop semiconductor storage technology and train talents.

At present, the representative domestic enterprises in the field of DRAM are Hefei Changxin, Fujian Jinhua, Ziguang National Core, Zhaoyi Innovation, Beijing Sicheng, Dongxin Semiconductor, South Asia Science and Technology (Taiwan of China), Huabang Electronics (Taiwan of China), Liguang (Taiwan of China) and so on.

Hefei Changxin is the leading enterprise of DRAM memory chip in China. Their DRAM technology mainly comes from Qimengda, a bankrupt German DRAM manufacturer, and Erpida, a Japanese manufacturer.

On September 20, 2019, Hefei Changxin announced the commissioning of Chinese mainland's first 12-inch DRAM plant and launched the first 8G DDR4 manufactured by 19nm process, which is a historic breakthrough.

According to the agency, the production capacity of Hefei Changxin is expected to reach 125,000 pieces from 2022 to 2023.

Jinhua, Fujian, you should have heard of it. A few years ago, they were sanctioned by the United States government, which made a lot of news.

In May 2016, Fujian Jinhua cooperated with UMC to produce niche DRAM. In December 2017, Magnesia accused Fujian Jinhua and Liandian of embezzling their memory chip technology. In January 2018, Fujian Jinhua also filed a lawsuit against Magnesia for patent infringement. In October 2018, Fujian Jinhua was included in the list of export control entities. In November 2018, the US Department of Justice sued Liandian and Fujian Jinhua for stealing trade secrets.

After a lot of trouble, Liandian can no longer handle it. At the end of January 2019, UMC announced the withdrawal of Fujian Jinhua DRAM project. In November 2021, UMC and Magnesia reached a settlement. At present, there is no complete final result of the review by Fujian Jinhua.

█ conclusion

Well, after writing so much, what I see here is true love.

In a word, DRAM memory is not only an important part of computer, mobile phone and other products, but also an indispensable "part" of digital infrastructure.

At present, the domestic DRAM memory has basically solved the problem of whether there is or not. The next step is to solve the problem of improving the rate of quality products, as well as the problem of capacity climbing. In terms of financing capacity, supporting industrial chain and echelon of talents, we still need to continue to strengthen and move forward cautiously.

It is expected that we can break the "top three" pattern as soon as possible and occupy a more important position in the field of DRAM.

Reference:

1. "who is going to break this DRAM technical dilemma? Wang Kaiqi, Global Semiconductor Watch

2. "American Romance of DRAM", Xinguang Society

3. "the Development of Storage Technology", Xie Changsheng

4. "the process of domestic substitution of DRAM chips is tortuous and the future is bright", Xiangcai Securities, Wang Pan, Wang Wenrui

5. "another big bet by storage companies", semiconductor industry observation

6. Memory Chip Industry Research report, Guoxin Securities

7. "domestic storage awaits a revolution", Fu Bin, fruit shell

8. "there is nothing more comprehensive about semiconductor storage than this one."

9. "Science and Technology Chapter 035-Flash memory of Semiconductor Storage", Wumijin, Zhihu

10. Baidu encyclopedia and Wikipedia related entries.

This article comes from the official account of Wechat: fresh Jujube classroom (ID:xzclasscom), author: Xiaozaojun

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.