Augustin Langlet

Prof. Thomas Jacobs

EXPOS-UA 2

28 April 2023

When Goliath Gets Jealous

 

Keep your mouth closed.
Guard your senses.
Temper your sharpness.
Simplify your problems.
Mask your brightness.
Be at one with the dust of the earth.
This is primal union.

–      Laozi, “Chapter 56,” Tao Te Ching

 

            “The World Relies on One Chip Maker in Taiwan, Leaving Everyone Vulnerable” says a Wall Street Journal headline. “The World Is Dangerously Dependent on Taiwan for Semiconductors” says a Bloomberg headline. “Cold war 2.0 will be a race for semiconductors, not arms” says a Guardian headline. China has been trying to catch up in its technology to be able to manufacture leading edge semiconductors. Combined with a pandemic supply chain shortage, this has made the US realize its reliance on foreign manufacturing. The Biden administration’s Chips and Science Act passed last year, giving funding for companies to build new factories in America, has been the first major step in their attempts to bring semiconductor production back. As semiconductors continue to be the foundation of all technology, commercial and military, politicians have been scrambling to take control of the industry.

            But wait a second. Just eight years ago, sitting in an elementary school history class in Taipei, I was told we were a small, humble island nation of 23 million people. Our economy was based on manufacturing. We had no Apples or Facebooks. But everything from cheap toys to bikes carried the “Made in Taiwan” label. I was told we weren’t allowed to officially be part of the global community. We weren’t recognized as a country, excluded from the UN and WHO, and we couldn’t do much about it without provoking a war. Amidst all this, how did this island, always overshadowed by the geopolitical field of China, become this crucial to such an important modern technology? And why is it suddenly the US’s obsession?

            Producing semiconductors is extremely hard. The field requires expensive research and development, and the rate of technological improvement creates fierce competition. Famously illustrated by the questionably accurate Moore’s “Law,” being a couple years ahead makes all the difference in this exponentially advancing industry. The supply chain of manufacturing these semiconductors is characterized by the phases of design, fabrication, packaging, and testing. Shaken up by the pandemic shortages and Chinese advancements in the industry, everyone is concerned about the prevalence and crucial link of Taiwan in that supply chain. Taiwan Semiconductor Manufacturing Co. (TSMC) fabricates 92% of high-end chips and 56% of revenue share in the fabrication market, not to mention other Taiwanese companies in various steps of the supply chain (Jie et al.).

            Only 40 years ago, Taiwan had a primarily agricultural and factory assembly-based economy. Still developing its economy, Taiwan had to find valuable and highly technical industries to get into for world-class products that could make it a more stable and significant player in the global economy (“How Taiwan Created TSMC”). Looking for ways to advance its economy, the Taiwanese government employed a more decentralized, competitive approach to researching new industries that differed from the centralized and isolated efforts in a few big tech firms like in Korea and Japan. In 1973, the government set up the Industrial Technology Research Institute (ITRI) to target R&D projects that would spin-off into countless small companies with different expertise and high competition from both domestic and global markets (“How Taiwan Created TSMC”; Feigenbaum 6~7). By 1985, ITRI had successfully taken its firsts steps in bringing Taiwan into tech, but the country remained far behind the leading edge. This was when Taiwanese Texas Instruments’ senior vice executive Morris Chang went back to Taiwan after the government offered him to lead ITRI. At the time, the still young, less complicated, and less competitive semiconductor industry only consisted of traditional Integrated Device Manufacturers (IDMs), vertically integrated companies like Intel that designed and fabricated their own chips from start to finish (Mark Liu 318). But with it becoming increasingly hard to have both successful leading-edge chip designs and leading-edge manufacturing, IDMs began to have high lead times and limited room for growth. Taiwan, at that point, already had a strong assembly manufacturing base and a cheap, expanding talent pool but weak Intellectual Property strength, sales, and marketing sectors. Taking that into account, the only good option Chang saw was a new business model: the pure-play foundry (318). He created TSMC as a fabrication only firm that would dedicate all its resources into the best manufacturing technology while only taking in customer designs. TSMC could focus purely on doing the best manufacturing without relying on successful designs to fill its factories. After a few years, the model became a success, coinciding with the rise of fabless firms, companies that only designed chips without having to invest into expensive and complicated manufacturing. TSMC was a platform of manufacturing, adapting and serving customers as it specialized in perfecting its specific foundry process.

            Today, TSMC’s prominence as the main supplier of high-end chips has become a hugely political contention. Especially after the pandemic chip shortages, everyone has been scrambling to find alternatives to the potential bottleneck of relying on one geopolitically contested country for the supply of such a crucial technology. In Do Artifacts Have Politics, Langdon Winner examines the framework in which to evaluate an innovation’s political impacts that helps explain the situation of the foundry model. In the essay, Winner separates their impacts into two broad categories of those that arise from a particular group or society’s use of a technology and technologies that inherently require certain political paradigms. He argues that fully evaluating technological innovations requires us to answer the question: “does this state of affairs derive from an unavoidable social response to intractable properties in the [innovations] themselves, or is it instead a pattern imposed independently by a governing body, ruling class, or some other social or cultural institution to further its own purposes?” (Winner 131). In order to understand TSMC’s prominence and how the US wants to retain control in the industry, we’ll have to first answer what parts of the foundry model are inherent political tendencies of the technology and what factors depend on the context it is used in?

            Winner’s most confident example for inherent politics in technologies is plutonium recycling in nuclear energy. With its requirement of centralized investment and “sacrifice of civil liberties” regarding the safety of nuclear materials and technologies, he quotes Russell W. Ayres that proliferation of plutonium recycling technology will bring about “’pressure to eliminate the traditional checks the courts and legislatures place on the activities of the executive and to develop a powerful central authority better able to enforce strict safeguards’” (134). Looking at the foundry model, its capital intensive, horizontally integrated nature has led to a centralization of the process. This explains the prominence and reliance on Taiwanese semiconductors and the high entry barrier. But the rest of the supply chain, which is also capital intensive, has horizontally integrated along with fabrication, creating a specialized global supply chain that has benefited under free trade. For example, TSMC relies on machines only produced by Dutch company ASML while US companies dominate chip design. This is to say, the technology’s complexity and capital intensity inherently make highly specialized firms more competitive and dominate their respective process. So, when the Wall Street Journal headline says the world relies on one chip maker, “Making Everyone Vulnerable,” so is the world reliant on one machine manufacture, a handful of chip designers from the same country, and a handful of mineral suppliers, “Making Everyone Vulnerable.”

            At the same time, there were key political motivations and circumstances in Taiwan that made the foundry model’s success possible. This would constitute an impact from “Forms of Order,” where “many technical devices and systems important in everyday life contain possibilities for many different ways of ordering human activity,” and thus the way the technology is treated changes the trajectory of its impact (Winner 127). It was the specific combination of a strong manufacturing base, lack of sales, marketing, and IP strength, and need for Taiwan to develop more technical industries that prompted Morris Chang to create the foundry model (Mark Liu 318). Furthermore, besides the government making a bit less than half of the initial investment into TSMC, it was largely left on its own. The government’s decentralized strategy towards investment gave plenty of room for startups and small firms to compete and adapt to each other and the global market (Feigenbaum 7).

            Inherently vertically separated and conducive of free trade while birthed in an environment of competitive development, the foundry model’s political characteristics gives a general explanation to the status quo. But the importance of semiconductors for national security, as most modern military technologies rely on chips, has prompted governments to rethink the paradigm created by the foundry model. For China and the West, they’re answer has been to move towards independent production or regionalization. Using embargoes and industrial policy such as the CHIPS act to separate production into parallel markets (Lin 5). But in order to create many regionalized foundry markets that are as effective as Taiwan’s, is it as easy as throwing billions into building factories on American or Chinese soil? If Taiwan’s politics have already proven a layer of unique circumstances that fostered the model’s success, digging deeper into other factors might tell us more about how the unassuming country got to where it is and what the US needs to understand about the industry.

            Looking at less obvious fields, another of Taiwan’s exports that has had outward impact and geo-political implications is its music. As an article from USC’s US-China Institute explains, walking in China’s big cities in the early 2000s, “the artists dominating the airwaves [were] not from the Chinese mainland” (Liu and Song). Listed on Baidu, 12 of the top 25 singles of 2007 in China were Taiwanese, and another seven were from Hong Kong or Singapore (Liu). Starting from the mid-80s when China began allowing Taiwanese music in the mainland, artists like Deng Lijun followed by A-Mei were able to bridge the cultural and political gap to create music that all Chinese speakers could resonate with. After them, new artists like Jay Chou, Wang Leehom, JJ Lin, Rainie Yang and groups like MAYDAY, S.H.E., and Shin each brought their distinctive voices to a greater Chinese audience. Yet again, this small country managed to create something that penetrated and flourished beyond expectations. It did so in the most populous country in the world, and a mortal enemy at that.

            Putting this through the lens of Winner’s framework of inherent and structural politics of artifacts. Taiwanese music from the 80s to 2000s was a fresh and developing field, influenced by strong structural factors. Across the Taiwan Strait, the Chinese government still restricted outside cultural influences and closely overlooked its domestic scene. Despite a period of strict martial law after the Nationalists escaped to the island in 1949, Taiwan was having its rapid development, becoming a part of the world economy, and eventually democratizing. Taiwan’s structural politics built a perfect cultural backdrop for entrepreneurship and new ideas. Taiwanese arts and culture were able to draw from Western and Japanese influences to develop a diverse, mature scene of its own (Liu and Song). This was key to Taiwan’s musical success in China. Deng Lijun brought in a mix of traditional and contemporary love songs, Jay Chou brought mandarin into the realms of R&B, and MAYDAY introduced China to rock while students in China felt their local artist to be “still conservative” (Liu and Song). A culture of openness and entrepreneurship was deeply intertwined with the development of the Taiwanese music industry and is fundamental to the Taiwanese identity. Understanding that unlocks another aspect of the foundry business that can’t be explained on balance sheets.

            As important as political policy and capital expenditure are, the foundry model’s global presence has been equally shaped by Taiwan’s culture. A decentralized and open entrepreneurial spirit allowed for the now bustling tech sector to first emerge. That’s a harder thing to say for a developing China, for example, where the government actively wants its successful businessmen to fear them, like in the case of the sudden disappearance of Jack Ma after commenting on some policies at a panel. But coming from a small developing country, Taiwan’s entrepreneurship also came with a humble receptiveness, ready to adapt and serve market needs as efficiently and cheaply as possible. After all, what those history classrooms taught was to humbly adapt and make the most out of the circumstances we were given, lest not to disturb a sensitive paradigm, political or not. Despite doing pioneering work, Taiwanese entrepreneurs weren’t necessarily setting out to create or disrupt new markets. To illustrate this receptiveness and adaptation, the main player left from the age of IDMs is Intel, which, despite its world-class talent, still struggles to compete with the combination of foundry and fabless firms that TSMC has brought around. During a recent panel, Morris Chang quotes Taiwanese CEO of NVIDIA, Jenson Huang, to explain the difference: “TSMC has learned to dance with 400 customers; Intel has always danced alone” (Chang 1:40). Being receptive, collaborative, yet entrepreneurial has defined what a platform company like TSMC should be like, and it is inextricably linked to a certain Taiwanese cultural mindset. It even translates into work culture. In the same panel, Chang describes the difference between Taiwanese and American technicians where, when a plant breaks down at 1am, it will be fixed at 9am the next morning in the US while a technician in Taiwan would be in there fixing by 2am (“How Taiwan Created TSMC” 8:15). What makes this even more critical is how the talent gap is becoming the top concern and bottleneck for the industry’s development (Lin 4). Whether it’s a receptive entrepreneurial spirit, or a certain work culture, there is more than just politics at play here.

            With the geopolitical circumstances the US is trying to deal with, it must be conscious of how its approach will change the trajectory and structural politics of the foundry model. It must understand that a mix of political and cultural circumstances allowed the foundry model to thrive under a division of labor and free trade where each country specializes in what they’re best at yet relies on each other to bring about the pervasiveness of chips we see today. The US is not Taiwan, neither is China, nor the Netherlands, Korea, or Japan. Embargoes and onshoring of plants in the US already threaten to raise the costs of chips (Lin 5). With our current trajectory, Morris Chang already exclaimed that “in the chip sector, globalization is dead” (Chang 3:51). But the semiconductor industry as we know it cannot function alone. The degree to which US officials want to follow China into an arms race for chips and the degree to which it wants to do it alone is up to them.

           

Works Cited

Chang, Morris panelist. Miller, Chris, panelist. “張忠謀×《晶片戰爭》米勒 半導體世紀對談：支持美國採取政策，延緩中國半導體技術發展” [Morris Chang x “Chip War” Chris Miller; the Semiconductor World Panel: Supporting US Policy to Slow Down China’s Semiconductor Advancements] Youtube, uploaded by 天下雜誌 video｜深度人物專訪, Mar. 16 2023. https://youtu.be/CVG7xse9Rj8.

Feigenbaum, Evan A. “Historical Context of Taiwan’s Technological Success.” Assuring Taiwan’s Innovation Future, Carnegie Endowment for International Peace, 2020, pp. 5–9. JSTOR, http://www.jstor.org/stable/resrep20992.6.

“How Taiwan Created TSMC” YouTube, uploaded by Asianometry, Dec. 6 2020. https://youtu.be/9fVrWDdll0g.

Lin, Judy. “Fab talent crunch: Taiwan's secret sauce for producing excellent semiconductor engineers” DIGITIMES Asia, Dec 1, 2022, http://cc.ee.ntu.edu.tw/~cliu/news/fab_talent_crunch.pdf.

Liu, Marc and Yi Song. “Taiwanese Pop: Music Makes The World Go Round?” US-China Today, Jan. 23 2008. https://uschinatoday.org/features/2008/01/23/taiwanese-pop-music-makes-the-world-go-round/.

Liu, Mark. “Taiwan and the foundry model.” Nat Electron 4, 318–320 (2021). https://doi.org/10.1038/s41928-021-00576-y.

Jie, Yang, et al., “The World Relies on One Chip Maker in Taiwan, Leaving Everyone Vulnerable.” Wall Street Journal, Jun. 19 2021. https://www.wsj.com/articles/the-world-relies-on-one-chip-maker-in-taiwan-leaving-everyone-vulnerable-11624075400.

Winner, Langdon. “Do Artifacts Have Politics?” Daedalus, vol. 109, no. 1, 1980, pp. 121–36. JSTOR, https://www.jstor.org/stable/20024652.