How China Built a Semiconductor Industry
In his later years, Qian Xuesen, the founder of China's rocket industry, who famously left America after persecution, had reportedly said: > In the 60s, we devoted all our efforts to the “Two Bombs, One Satellite” project, and we gained a lot from it. In the 70s, we didn't focus on semiconductors, and we lost a lot because of that. After the chaos of the Cultural Revolution, the People's Republic of China made three major pushes towards a domestic semiconductor industry. Two of them failed. The third succeeded. In this video, we look at how China built a semiconductor industry.
## Beginnings The founder of China's solid state physics academic community is a Chinese returnee named Huang Kun (黄昆). Born in Beijing in 1919, he traveled abroad to England and Bristol University to study physics - earning a PhD. In the late 1940s, he collaborated with the Nobel Prize winner Max Born to finish a textbook Born had started called "Dynamical Theory of Lattice Physics". It was eventually published in 1954. After the outbreak of the Korean War, Huang wanted to return to China. Huang wasn't the only one. Another major academic was Xie Xide (谢希德), then studying at MIT. Her speciality was in surface and semiconductor physics.
Despite the efforts of America's Truman administration, Huang and Xie made their way home and founded China’s first solid-state physics department at Peking University. They also wrote a major textbook on the subject titled "Semiconductor Physics" Other universities like Fudan University in Shanghai followed suit. But today, Peking University is generally seen as the cradle of China's semiconductor industry.
## The 774 Factory China's first production semiconductor industry comes from the Soviet Union. In 1953, China laid out its first Five Year Plan, covering the years until 1957. The Soviets helped make and carry out this Plan. One of the 156 industrial projects completed with Soviet assistance was the 774 Factory - also known as the Beijing Electron Tube Factory. In the 1950s, the 774 Factory successfully pulled its first Germanium crystal, establishing itself as China's most technologically advanced semiconductor factory. A few years later, a silicon crystal was pulled the same way. A small Chinese electronics cluster in the Jiuxianqiao (酒仙桥) sub-district of Beijing grew up around it.
Then in the early 1960s, the Factory started producing germanium alloy junction transistors. These transistors are an improvement on Shockley's original bipolar junction transistors. Over in the United States, before the introduction of the planar transistor, alloy junction transistors were quite big in the market. You take a piece of pure germanium and fuse beads of indium onto both sides of it. This was done manually. Workers in the 774 factory used a homemade soldering iron to melt the indium balls onto the germanium.
Many of these devices went into China's first transistor radios, which began hitting the world market in late 1963 via Hong Kong. Much later, the 774 factory's assets were sold off to a private equity group to found Beijing Orient Electronics Technology Group. Today, a world-leading OLED company. ## Integrated Circuits The integrated circuit trend that began in the 1950s allowed a second semiconductor cluster in China's south to join Beijing's in the north.
In 1965, a Shanghai component factory cooperated with the Shanghai Institute of Metallurgy of the Chinese Academy of Sciences to produce China’s first integrated circuits - a TTL-type circuit. A key early use case for transistors and integrated circuits were the military and their space programs. Famously, the Minuteman II missile program bought a peak of 15,000 integrated circuits a week in 1965.
And the Apollo Space Program is estimated to have bought 200,000 ICs before 1964. Same as it was in China. Semiconductor production in aeronautics was supported by another Chinese returnee named Huang Chang (黄敞). A Harvard PhD, he worked at Sylvania Semiconductors, and obtained permanent residency in the United States. In 1959 he decided to return to China - inspired by Qian Xuesen's contributions to the Chinese missile program. Many of the computers for those missiles were built with self-produced transistors.
Based on this progress as well as others not mentioned in this video, one might judge China's semiconductor industry in the 1960s to be ahead of many others in the world. ## The Third Front China's semiconductor progress ran into obstacles in 1964. That year, Mao Zedong decided to relocate many of China's heavy industry and military installations into the country's interior - the "Third Front" policy (三线建设). This massive relocation was done in light of worsening relations with the Soviet Union.
And thus it was carried out with great haste - involving nearly 500 factories, 1.5 million workers, and 200 billion RMB of investment. In 1970, the military relocated its integrated circuit research in Hebei - Institute 24 - to the remote area of Yongchuan County (永川县) within the mega-city of Chongqing. The Third Front's results are mixed. On one hand, it stimulated the electronics industries of many underdeveloped areas. Some areas reported company growth of 700%. Yet on the other hand, the Third Front slowed development by deliberately stretching the supply chain. Many of the relocated areas were in mountainous regions and lacked proper transportation networks - causing delays.
It also separated R&D and manufacturing, making it difficult for the research produced in the labs to ever make it to the fabs. Moreover, Problems on the factory floors had to be handled by unqualified factory staff, perpetuating poor yields and low production. ## Cultural Revolution And then we have Mao Zedong's Cultural Revolution, which lasted from 1965 to 1975. The Revolution was an immensely damaging period - not only for China itself but also its semiconductor development. Many colleges closed
down. Those that remained open still suffered immense disorder. People with western education or connections - including the aforementioned Huang Kun - were persecuted, removed from positions of power, or sent to the communes to work. The students either worked in factories or studied politics. The universities did not reopen until December 1978, when Deng Xiaoping emerged as China's paramount leader, refocusing the country's direction. This left a 16-year gap between graduating classes in semiconductor technology education - 1966 to 1982. There were a few pockets of progress. The Third Front relocation allowed the
aforementioned Institute 24 to work in relative peace. But overall, China was isolated and steadily fell behind the rest of the world. People of the era express frustration with this long gap in Chinese semiconductor progress.
## Catchup Thus, the late 1970s and early 1980s were a period of recovery and catch-up. Early Chinese efforts on semiconductor industry development were simple and rather decentralized. On the talent front, expand the base by producing more semiconductor graduates and send Chinese overseas to learn in the West.
Secondly, governments on various levels tried to acquire new equipment from abroad to upgrade their old state-owned semiconductor factories or build new ones - spending about 1.3 billion RMB to acquire it. Local governments, state-owned enterprises, and universities funded new semiconductor fabs. 33 new fabs were constructed from 1984 to 1990. It was a bit of a crazy time. It is important to note here that while export controls like CoCOM did exist, the Chinese were little affected by them. Most stuff they got was trailing edge stuff that was not restricted. And the Chinese - like the Soviets - were quite good at evading restrictions, usually by funneling them through Hong Kong. Funds were limited - at the time, Chinese governments were importing equipment to build its now-formidable electronics and computer industries too. So this semiconductor equipment
was old and often came second-hand. Factory workers often had little clue how it all worked. Foreign visitors to the factories noted the lack of good production habits. Factory work was seen as an "iron bowl" job. There were more people in the fab than what was needed, and most did not work efficiently.
A later report by the Chinese Semiconductor Industry Association investigated the outcome of 24 acquired lines and found that just ten of them functioned. The growing Chinese electronics industry - frustrated with weak domestic IC production - quickly turned to imports. Really, only one of China's factories in the 1980s actually performed up to expectations: Wuxi Factory No. 742. ## Wuxi Factory 742 The factory now called Wuxi Factory No. 742 began as the Jiangnan Radio Factory in 1960 (江南无锡电器材厂), producing Soviet-designed diodes in the city of Wuxi.
They later became a state-owned company - Wuxi referring to its local city - and then in 1968 merged with an industrial school. They produced discrete devices for use in military equipment like radios and did it very well, winning several gold medals. In 1980, the Chinese government chose Factory 742 as a core factory for a new project referred to as Project 65 - because it was part of the Sixth Five-Year Plan from 1981 to 1985. In Project 65, the government would provide about 270 million RMB of funds to acquire two production lines from Toshiba - a 3-inch wafer bipolar at a 5-micron process node and a 4-inch wafer memory node. Both then quite far behind the leading edge.
As part of the transfer process, Toshiba hosted a few Chinese engineers at their plant to learn recipes. The transfer went smoothly and the line entered high-volume production in 1984, producing 30 million wafers that year. After some early learning troubles, the factory raised its yields to over 80% by 1985, producing 70+ million integrated circuits for televisions. By the late 1980s, half of China's televisions were claimed to have been made with chips from 742.
Soon thereafter, the factory realized it needed to upscale its native R&D capabilities. So in 1986-1987, the 742 merged with Institute 24's Wuxi city branch to create Wuxi Huajing Microelectronics Company (無錫華晶電子), or just Huajing. The factory was reputed as Mainland China's leading semiconductor company. They also claimed several R&D milestones. Despite China's isolation, the company received foreign experts, had books and papers from IEEE, and sent workers abroad to study.
In 1987, it became China's first company to put an LSI-class integrated circuit into production and also the first to prototype a VLSI-class 64K RAM chip. Huajing's successes as a smaller factory convinced the Chinese government to make it the central focus of their next big semiconductor policy push. ## Project 908 As early as 1983, Chinese leaders had been aware that this first attempt to catch up in semiconductors had not panned out.
They identified several reasons for this. Investment in semiconductor facilities was done haphazardly, leading to a lack of scale and production. Yet it took years - until 1989 - for the Chinese State Council to approve a new measure to address this. In 1989, Chinese IC consumption was 400 million chips. Domestic production across all
its scattered fabs was just 114 million. Not to mention those chips were horrendously backwards. Policymakers decided to emulate the successful Koreans and make Chinese chaebol of their own. They cleared the decks, establishing five major semiconductor companies in China's northern and southern regions. Two were in Shanghai, one was in Shaoxing City of Zhejiang Province, one more in Beijing, and the last was Huajing. These would be China's semiconductor national champions.
Thus in 1990, we have Project 908. 908's goal was to demonstrate to the world that China can build a leading integrated design manufacturer in semiconductors. It had several parts but its biggest and most significant package would grant Huajing 2 billion RMB to upgrade its existing 5-inch wafer line to a 6-inch wafer line - then considered a leading edge technology. It would run a 1 micron or 0.8 micron node, producing 12,000 wafers a month. That is a lot of money. In 1989, the mainland's per capita GDP was 1,538 RMB
or $400. So 2 billion RMB would be equivalent to the annual production of 5 million Chinese. But despite all those resources, Project 908 was a resounding failure. The new production lines were obsolete before they even activated. In 1997, the fab had capacity for 12,000 wafers a month but was only making just 800 each month. Huajing fell into crisis. ## 908's Failure As a component for a finished good, semiconductors had always been a state-funded thing in China - heavily reliant on government funding. Huajing sprung from a
government factory and Project 908 was a government-initiated project. So naturally 908 and others of its like - there was a 907 project too which suffered similar issues - were government-led initiatives. This made it difficult for Huajing to succeed. Any investment decision - like for a litho machine - required government-sign off, which wasted time.
And local governments fought bureaucratic battles over how to do things and what was to be done. Memorably, 26 agencies locked horns over where to do a single dredging operation. You might imagine how such dynamics would affect a major strategic decision like whether to pursue DRAM or ASIC logic chips. When Samsung was grappling over a similar decision in the mid-1980s, there was a circle of executives debating over which to pursue. But in the end, a single person made the final call to go into memory - Samsung Chairman Lee Byung-chul.
Not the case with Huajing and its various government backers. In the end, they decided to produce both DRAM and ASICs in equal quantities, leading to a situation where neither was being produced very well. Huajing's management were hands-on people, and at first they struggled to handle this titanic project. So they devolved some authority to smaller operating units, but this in turn encouraged empire-building and vastly exploded the company's scope.
By the late 1990s, Huajing consisted of 26 factories, firms, and branches managed by 9 departments. It was not only scattered across the whole semiconductor supply chain from materials to design to wafers to assembly, but also had its own hospital, day care, school system, and community centers. There was also an issue of talents. Taiwan and Korea benefitted from waves of returning talents - particularly electric engineering talents - who trained in the West. At the time, China did not have this. Beyond these internal issues, there was a major external one. Due to its participation in the Korean War in 1950, the People's Republic of China was subject to export restrictions on leading edge United States technology. Same goes with Taiwan, another plausible transfer partner.
Nevertheless in 1994, the Chinese government negotiated a technology transfer deal between Huajing and Lucent Technologies. Lucent - the former Bell Labs - would train engineers and transfer a node process plus IP blocks for an advanced telecom switching technology. On the surface, it seems like a huge boon. It wasn't. The government negotiated the deal, meaning that Huajing had no say on what to transfer over. Huajing's managers knew the company was nowhere near ready and wanted a cheap, low-tech technology to start with before working their way up. But the government officials chose for them, and steered towards a flashy technology Huajing could not digest.
Furthermore, what they got they could not use. Huajing's engineers, mostly self-taught, were good at reverse-engineering existing products. When they got the Lucent IP and technologies, they had no idea how to integrate it into what they were already doing. Thus, the technology transfer process did not take.
And then there was a customer issue. When Taiwan's ITRI spun off its first chip company UMC, they helped source customers - tapping a growing network of Taiwanese electronics companies. They provided product IP. They wrangled Taiwanese business tycoons to invest. By contrast, the 908 project never really considered the problem of competing for and getting customer orders. Why would they? In the old times, the planned economy guaranteed demand and there were no imports. But times were different now. All this means that it was not until 1995 that the first production lines began construction. Production did not start until 1997-1998. By then, the 6-inch wafer technology they received from Lucent was out of date and left unused. A massive debacle.
Huajing had took out a billion RMB of government loans to pay for its high tech fab. In 1997, the company began to struggle from high interest rates, turning a staggering loss of 240 million RMB that year. With the company in crisis, Huajing struck a deal with Taiwan. Several Taiwanese with experience in the semiconductor industry agreed to run it. Due to cross-strait rules, they founded a shell company in Hong Kong called Central Semiconductor Manufacturing Corporation or CSMC - sounds familiar - and took over Huajing. They stabilized the company and then sold it to a Chinese SOE in 2002. Today,
it is Wuxi China Resources Huajing Microelectronics. They make commodity discrete power electronics and trade mostly on the power of their storied name. ## Project 909 By 1995, it was clear that Project 908 was a failure - another black eye on China's ongoing semiconductor journey. Everything went wrong and a new approach was needed. Policymakers, listening to advisors, correctly determined that 908 failed because of its government-led status. The government's central role and
Huajing's former life as a SOE crippled its ability to move fast in a dynamic market. In November 1995 - even as the 908 dumpster fire burned over in Wuxi City - the Chinese government took its third and most ambitious yet crack at the semiconductor apple - Project 909. At the time of its initiation in March 1996, it was China's largest semiconductor state project yet.
909 wanted to raise a national champion from birth - Shanghai Huahong Group (华虹半导体有限公司). It would ameliorate all of the problems faced by China's prior forays into semiconductors, bring in and digest foreign technologies, and dominate the domestic market. 909 was funded with over 10 billion RMB directly from the State Treasury and Shanghai's municipal government - at least, we believe so. To be honest, can never be too sure where the money comes from with these projects. Nevertheless, they had a blank check.
To cut red tape, the company had China's serving Minister of Electronics Industry Hu Qili (胡启立) as its chairman of the board. The mayor of Shanghai was on its lead working group, and Shanghai's various governmental departments were told to give 909 the greenest of green lights. Huahong's core achievement would be the construction of a 8-inch memory wafer fab running a 500 nanometer process, aided by another technology transfer deal with Japan's NEC. This joint venture would be called Huahong-NEC. Again, export restrictions were not an issue. In part because of the technology’s lagging sophistication and then-warming relations between the United States and the People's Republic of China.
Lastly, the government reserved a market for Huahong-NEC. The Ministry of Electronics Industry researched the market and selected a rapidly growing product in China - IC cards. In 1997, the State Council reserved that market for Huahong-NEC. ## 909's Fate Did Huahong succeed? Huahong-NEC constructed its fab within two years. Per the requirements of its
technology transfer deal, NEC would provide an early sales channel, so the venture started selling 64-megabit DRAM sold under NEC's brand name. In 2000, its first full year of operations, the company turned a profit of 350 million RMB. The company improved its yields from 50% at the start to nearly 90%. But there was a catch. Japanese workers from NEC were the ones actually running the fab, not Huahong personnel. And they did so in such a way to give Huahong managers no insight into what was going on inside their own fab! By contrast, when Taiwan's ITRI struck their legendary technology transfer deal back in 1974, they laid out a complete transfer process - soup to nuts. The deal was not considered complete
until Taiwan finished a replica fab of their own, on Taiwanese soil, staffed with Taiwanese. So the transfer was not complete. Without that, Huahong could not upgrade the fab. When the company experienced a large operating loss during a memory crash in 2002, Huahong canceled NEC's participation and took full operational control. They then struck a deal with another company called Jazz Semiconductor, a startup aerospace foundry spun off from Rockwell Semiconductor.
Jazz's founding CEO was a Chinese-American returnee named Shu Li. This weird triple-venture then acquired foreign IP to develop China's first contactless smart IC cards for transportation and payment: The Shanghai Public Transport "One Card". They also switched to a foundry model, which probably more contributed to a rebound in revenue in the coming years. Jazz was later acquired by the Israeli foundry Tower and sold their interest in Huahong in 2011. And after a bunch of mergers, Huahong is today China's second largest semiconductor
foundry - first place loser, in my book - with much of their production at the lagging edge. ## Circular 18 Huahong's profits might have been up and down. But China was starting to learn. In 2001, China entered the World Trade Organization, liberalizing their semiconductor markets. Now that policymakers had their national champion, they can build a broad industry behind it by attracting foreign investment. In June 2000, the State Council passes a new policy called Circular 18, or “Several policies for encouraging the development of software industry and integrated circuit industry" (鼓励软件产业和集成电路产业发展的若干政策) Circular 18 offered significant tax incentives for every semiconductor company in China. And
most importantly, it did not matter whether they were domestic or owned by foreigners. In 2000, Richard Chang sells his foundry company Worldwide Semiconductor to TSMC. Soon thereafter, Chang goes to China, raises over a billion dollars, and starts SMIC using basically the same business model. SMIC has gone on to be China's leading semiconductor foundry, and a force to reckoned with. China's first true
blue semiconductor national champion to go alongside Hua Hong and Nexchip. At the same time, there is a massive flurry of semiconductor FDI activity from foreign companies entering the Chinese market. Applied Materials, AMD, Philips, Motorola, ASML and more announce billion dollar investments into China. The industry is up and running. ## Conclusion For additional reading, I really recommend this dissertation by Susan K Mays, "Rapid Advance: High Technology in China in the Global Electronic Age". Mays, who seems to be at UT Austin now, did some incredible work here and I just want to call that out.
Anyway. There is an impression that China did not care much about semiconductors in the past. And now that they are facing export bans, the West has suddenly motivated them to care. Perhaps there is some truth to that. But the People's Republic of China has been grinding on semiconductors since the 1960s. They have thrown billions into it and adapted their approach and strategy multiple times over the decades.
It seems to me like they have always cared, with a capital C. To note, it is hard to ignore the fact that so much of China's accelerating progress today comes from its existing ties to the global semiconductor world at large. The absorption of new knowledge - and equipment - from abroad.
As those ties and connections start to change and maybe even sever, we enter uncertain parts of the playbook. I wonder what happens next to China's semiconductor industry.
2024-07-04 17:26
