Chip Island: How TSMC and Taiwan Triumph

Podcast with Owen Lin 林宏文 on ChinaTalk

While I was in Taiwan this past summer, I had the great honor of meeting a superb journalist who has shaped himself into TSMC’s unofficial historian, Owen Lin 林宏文. Mr. Lin in 2023 reflected upon his over 30 years of reporting on Taiwan’s semiconductor industry to write《晶片岛上的光芒》，slated to be released in 2025 in English under the tentative title Chip Island: How TSMC and Taiwan Triumph.

Chip Island offers a quintessentially Taiwanese perspective on the island’s central role in U.S.-China relations, and a conversation with Lin sounds nothing like either the Cold War–saber rattling you’d hear in Washington or the diatribes about reunification you’d hear in Beijing. A few gems from our chat:

• “Once reporter was asking [Morris Chang] a question about TSMC’s competition with Samsung, and before the reporter could finish, Morris interrupted and said, ‘I didn’t say they were admirable, I said they were formidable.’ He emphasized the word “formidable‘ several times in English.”

• “We know that companies like UMC and TSMC all spun off from ITRI. So the predecessors of these companies, their teams, and their talent — ITRI played a crucial role in all of this.”

• “The production cost of Viagra accounts for only 0.7% of its price. If you reduce manufacturing costs by half, it only decreases the overall cost by 0.35%, which has a minimal impact. In contrast, semiconductor costs account for a much larger portion of the final product price. When TSMC uses the best technology to produce NVIDIA’s GPUs, the cost reduction is substantial.”

• “Both the Taiwanese government and TSMC are extremely important. The Taiwanese government represents our democratic system, chosen by the people’s votes. It embodies the direction we’ve chosen for ourselves after years of struggle. TSMC is also crucial.”

• “Taiwanese people are more dedicated to work and willing to work overtime, but American and German employees probably aren’t as willing to work overtime. Even Japanese aren’t as willing to work overtime now.”

• “Most Taiwanese people agree that TSMC is indeed a ‘Sacred Guardian Mountain’（护国神山）, [but terms like this and] ‘Silicon Shield’（矽盾）weren’t invented by Taiwanese people; they originated from foreign media.”

• “Taiwan used to be like the black-clad stagehands who move props unseen, working behind the scenes while companies like Apple and NVIDIA took center stage. Now, these “stagehands” have become crucial players.”

Our conversation was conducted in Mandarin and can be found on ChinaTalk (Spotify, Apple Podcasts, CastBox).

Lily Ottinger

and

Nicholas Welch

co-hosted with me and posed some insightful queries. ChinaTalk has painstakingly transcribed and translated our conversation into English below for your reading pleasure. Any errors in translation are our own.

Forging the Silicon Shield: TSMC and Taiwan’s Future

Arrian Ebrahimi: Hello everyone, I’m Liu Tianwen（刘天文）, your host for today. My English name is Arrian Ebrahimi, and I’ll be co-hosting today’s program with two other members of the ChinaTalk team — Lily Ottinger, a University of Kansas graduate currently working as an editor in Taipei, and Nicholas Welch, who is pursuing a JD at Duke Law School. Thank you, colleagues, for your valuable support today.

I’m currently pursuing a JD at Georgetown University Law Center in Washington D.C. For the past two years, I studied for my master’s degree at Peking University’s Yenching Academy. Before going to China, I worked at the Semiconductor Industry Association in the United States. That job sparked my interest in China, and since then, the chip industry has been my entry point for understanding Chinese civilization.

I’m sharing my personal background as primer to introduce our special guest today, Lin Hong-wen. This summer, I interned in the legal department of a Taiwanese semiconductor company. To gain a deeper understanding of Taiwan’s semiconductor industry and its international role, I read a book called Chip Island: How Taiwan and TSMC Triumph.

This book explains the history, management methods, and international role of Taiwan Semiconductor Manufacturing Company (TSMC), among other key topics.

During my visit to Taipei this summer, I had the good fortune to chat with our guest today, Lin Hong-wen, the author of Chip Island. We’ll be listening to Taiwan’s leading chip journalist tell the story of TSMC and Taiwan.

Mr. Lin, thank you very much for joining our podcast today. Could you please introduce yourself, and then we can begin our discussion?

Lin Hong-wen: Thank you, Tianwen [Arrian]. I’m also very pleased to meet you all.

My name is Lin Hong-wen [Owen Lin]. I’ve been a news reporter in Taiwan for 30 years, and this is my 31st year covering the semiconductor industry. I started covering semiconductors in 1993.

TSMC was established in 1987, so I began following this company shortly after its founding, writing daily news about it. I’ve also written extensively about the semiconductor industry, not just in Taiwan. Over the past 30-plus years, I’ve traveled all over — to the United States, Japan, and Europe. I even went to Scotland three times, when it was known as Europe’s “Silicon Glen.”

I wrote this book with the main purpose of addressing what I saw as a gap in the existing literature. In 2022, when I started writing, I had read several books related to chips, such as Chris Miller’s Chip War and Yasu Ota’s The Geopolitics of Semiconductors （半導體地緣政治學 by 太田泰彥）. While these books were well-written and explored the semiconductor industry from different angles, I felt they lacked a perspective from the actual happenings in the semiconductor industry, especially in Taiwan.

Over the past 30-40 years, the global semiconductor industry has undergone a dramatic transformation. We’ve seen a paradigm shift from vertical integration to vertical specialization, from IDM (Integrated Device Manufacturer) to Fabless and Foundry models.

Taiwan seized a significant opportunity in this trend of vertical specialization. Companies like TSMC and UMC rapidly developed in the wafer foundry business. Additionally, Taiwan has numerous IC design houses and has also captured opportunities in packaging and testing markets.

Taiwan has seized the biggest opportunities in the professional specialization roles that emerged from the disaggregation of IDMs. I felt that Taiwan’s experience was crucial, yet other books didn’t mention Taiwan’s development.

How did Taiwan rise? How did it become the world leader? I believed these stories needed to be told, especially now that every country is investing in semiconductors due to geopolitical concerns. Yet, many don’t understand how Taiwan developed to this level.

Why can TSMC fabricate over 90% of the world’s advanced process technologies? Why do Taiwanese packaging houses account for nearly 60% of the global market share? These are crucial aspects of Taiwan’s development that the world needs to understand.

Every country has a “Chips Act” and is subsidizing semiconductors, but not knowing how Taiwan succeeded is dangerous. All these countries investing might labor in vain and fail in manufacturing.

This is why I wrote this book. It was published in Taiwan in July last year and in Japan this April. I’ve already been to Japan four times since the Japanese publication, as they’re very keen to understand Taiwan’s success story. I’ve had many exchanges with Japanese friends.

The Korean version will be published by the end of this year, and the English version will be out in the first half of next year. I hope to let everyone know that Taiwan’s development experience is very worthy of reference. Of course, Taiwan has also had failed cases, and I don’t shy away from discussing these.

For instance, Morris Chang of TSMC was very successful, but his first tenure was quite disastrous, and he later had to close down those operations. As a frontline reporter, I don’t conceal facts, but I strive to present Taiwan’s development experience fairly and objectively.

Arrian Ebrahimi: Thank you, Mr. Lin. You mentioned earlier that you’ll be releasing Korean, Japanese, and English versions in the coming months and years. What new aspects will you cover in each language version? How will you tailor the content for new audiences?

Lin Hong-wen: The Japanese version came out earlier, so there won’t be significant changes between the Japanese and Korean versions. For the English version, I’ll be making more updates. I believe the English-speaking readers, especially given that my book came out in July last year, will need more current content. I’ve added a lot of new developments, including how AI is driving changes in the semiconductor industry, and the integration of SOC (System on Chip) and HBM (High Bandwidth Memory). These developments have had a significant impact on the industry.

I’ll also include more topics that English readers, particularly American readers, are interested in. For instance, TSMC is now starting to build a plant in Germany, so I’ll include some European developments as well. I think among English readers, there will be many European readers, so I want to incorporate new information for them.

Arrian Ebrahimi: Thank you, Mr. Lin. We’ll discuss the topics of AI chips and memory later. Actually, you’ve already started discussing some of the content, but I’m still curious about you personally. You mentioned that you started as a journalist around the same time TSMC was founded, not long after. Since you studied at National Chiao Tung University（国立交通大学） for your undergraduate degree, journalism probably wasn’t the most likely career path for you, right? I’ve lived in Hsinchu, and most people there dream of becoming engineers. Why did you want to become a journalist?

Lin Hong-wen: National Chiao Tung University should be familiar to many friends in Taiwan and mainland China. Essentially, Taiwan’s Chiao Tung University was re-established in Taiwan after the separation of the two sides of the strait, originating from the Chiao Tung University in mainland China.

Chiao Tung University is one of the top institutions for science and engineering. After studying there for four years, I realized I might not be suited to be an engineer. So after graduation, I became a journalist. I chose journalism because I enjoyed writing.

After becoming a journalist, I found it quite suitable for me. On one hand, I enjoy writing, and it’s a great pleasure for me. I often get to talk with many people and meet industry leaders. Standing on the shoulders of giants, I can see far.

Another important factor is that with my science and engineering background, I found it easier to cover this industry compared to journalism graduates who might face more obstacles. As I mentioned earlier, through my news coverage, I’ve gained many insights. Covering an industry isn’t just about writing industry news; I’ve seen a lot of competition between countries, economic and industrial development, and now geopolitical developments.

Through this industry, I’ve gained knowledge in many other areas. This exploration has been very interesting to me. Every day, I wake up with a pile of things to understand, analyze, and explore. I find this to be a very happy journey of exploration in my life.

Arrian Ebrahimi: Thank you, Mr. Lin. I really admire you. The next question is about both your book and yourself. You mentioned that you’ve been a journalist for over 30 years, and this book is based on your interviews and reports over these years. Before discussing the process of writing this book, could you first tell us how you entered the chip industry and became a specialized journalist in this field? Then, how did you summarize your 30 years of experience into one book?

Lin Hong-wen: When I started covering the semiconductor industry, it was my supervisor at the newspaper who assigned me to it because there was a vacancy. They thought that with my background from Chiao Tung University, I could quickly take on the role.

Initially, when I was covering TSMC and UMC, I found it interesting. As you know, Taiwan’s semiconductor development started in Hsinchu Science Park. At that time, I would go to Hsinchu every day, visiting different companies and chatting with people.

Every day, I would compile tables, listing the products, investment amounts, and other details of many Taiwanese companies. I was so young at the newspaper that they called me “Table Brother”（表弟）— the brother who was good at making tables.

In this learning process, I began to find the chip industry very interesting. Now, everyone knows how important chips are. They’re in your phone, your computer, or anything connected to the internet. But in the early days, it wasn’t necessarily like that. When I first started reporting, it was mostly about consumer ICs, like toy ICs.

Young people might not know this, but we used to have musical greeting cards for Christmas. When you opened them, there would be a music IC that would play a tune. Taiwan used to make many of these audio ICs. Then, as telephones (not mobile phones, but desktop phones) started to develop, Taiwan slowly moved from these small consumer ICs into PCs.

Now we’ve entered the AI era with GPUs. I feel that over these 30 years, I’ve grown alongside the semiconductor industry. I started knowing nothing and gradually came into contact with many aspects of the industry’s development.

Many Taiwanese companies have gone through this process. As I mentioned, there have been several generations of different ICs. Taiwan has been slowly capturing these opportunities. Not all companies in Taiwan have successfully transformed, but some have. For example, MediaTek initially made ICs for PC-related optical drives, but when smartphones took off, they seized the opportunity.

TSMC’s Global Niche

Arrian Ebrahimi: I understand. Thank you, Mr. Lin. Now we can move on to some topics about TSMC itself. But you just mentioned some areas that TSMC has entered. First, let me ask about a quote from Morris Chang that you mentioned in your book regarding Samsung: “Samsung is a formidable but not admirable competitor”（三星可畏但不是可敬的對手）. Why did Morris Chang say this about TSMC’s competition with Samsung?

Lin Hong-wen: Morris Chang said this in a very specific context. A reporter was asking him a question about TSMC’s competition with Samsung, and before the reporter could finish, Morris interrupted and said, “I didn’t say they were admirable, I said they were formidable.” He emphasized the word “formidable” several times in English.

Morris has never explained why Samsung is not admirable; he didn’t emphasize that part. But he did stress that they are formidable. My interpretation is this: Samsung is indeed formidable. If we were to call any company in this industry a “gorilla,” Samsung would definitely be one. Intel used to be a gorilla too, but they’re struggling now, while Samsung remains strong.

The key point is that Samsung has become number one in many industries they’ve entered. Whether it’s memory, phones, or in the past, TVs and displays, they’ve become number one in many fields they’ve entered.

Their competitive methods... Well, I understand this because I wrote a book about Samsung in 2012, describing them as a “financial crocodile” — a scary crocodile. I wrote that book because at that time, Samsung was beating many Taiwanese industries badly. Taiwan’s position was miserable; Taiwan’s panel industry was suffering, many industries were struggling, all beaten by Samsung.

So I understand why Morris Chang calls Samsung formidable. In fact, he’s never criticized Samsung, but he once shared an experience. Very early on, when he returned to Taiwan and was still at ITRI, before founding TSMC, he was invited by Lee Kun-hee, the previous chairman of Samsung (the father of Lee Jae-yong), to visit Korea. He went with Stan Shih, the founder of Acer, and some others from ITRI. Lee Kun-hee showed them Samsung’s DRAM investments and essentially said, “Don’t invest yourselves, just come work for me.”

Morris, of course, didn’t give an answer. He felt he could do better by returning to Taiwan and wanted to start his own company, not work for Samsung. When he talks about this incident with Samsung and Lee Kun-hee, I think he’s clear that he believed he had the ability and ambition to build a company that could compete with Samsung.

Everyone knows that before returning to Taiwan, Morris was already the general manager of the semiconductor business at Texas Instruments. He was a very important leader in that company. So I think he had high expectations for himself. Of course, when he first founded TSMC, it was tough, but his goal was to build a very important company with great influence. This is my understanding of his comments about Samsung.

Arrian Ebrahimi: Thank you, Mr. Lin. The content you just discussed is very carefully and vividly described in your book. Actually, while you were speaking, I was checking to see if it was around page 10 where you mentioned Morris Chang’s visit to Samsung. I forgot exactly which page it was, but I recommend our audience to read your book and vividly see what experiences you had.

Mr. Lin, you just discussed past events and how we got to the situation where Samsung dominated the entire field. But you’ve already started discussing AI chip trends. I also saw that you recently published an article called “AI Era Drives Semiconductor Landscape Reorganization.” Could you please explain how AI chips are affecting the memory industry and Samsung’s leading role? What impact have they already had?

Lin Hong-wen: Thank you for this question, Tianwen. It’s a very good topic, especially as Semicon Taiwan is currently being held in Taiwan. Interestingly, we were just talking about the changes in the AI era. I think the impact is very dramatic. Everyone only sees NVIDIA taking off because of the AI era, with its stock price soaring, but there are many stories behind this.

This year’s Taiwan semiconductor exhibition is very interesting. Samsung and SK Hynix have sent their most important people to Taiwan. In the past, Samsung and Hynix would send people to attend Taiwan’s semiconductor exhibition, but not their top executives. This time, they all came. Why? Well, there’s another interesting news: in the first half of this year, Korea’s semiconductor exports to Taiwan increased by more than double.

What does this mean? The AI server chips, as we know, have NVIDIA as the most important customer, and TSMC helps with the manufacturing. Because these AI chips need to be power-efficient, high-performance, and well-packaged, they now package logic ICs (SOC) and HBM together. That’s why Korea’s memory exports to Taiwan have grown so much — more than doubled. NVIDIA needs these AI server products, and they need to package GPUs and HBM together. We call this CoWoS (Chip on Wafer on Substrate) technology.

In the past, system memory and logic chips were separate and could be shipped separately. But now, because TSMC controls the most advanced process technologies and NVIDIA places all its orders with TSMC, the HBM memory must be packaged together.

This has produced a very significant effect. Yesterday, I heard executives from Samsung and Hynix speak. One Hynix executive said he had already been to Taiwan more than a dozen times this year. Why so often? As I mentioned, they need to constantly discuss with TSMC how to package HBM together through CoWoS packaging. There are many areas where they need to collaborate.

As for Samsung, it’s like this: because TSMC and Samsung are competitors in wafer foundry, TSMC tends to support Hynix and Micron, the other two memory manufacturers. Samsung, of course, tends to be excluded. This can’t be helped; it’s the reality of competition.

TSMC’s Origins

Arrian Ebrahimi: I see. Mr. Lin, you’ve already started discussing modern issues, but we seem to have gotten too excited and moved on to current problems without telling our audience where TSMC actually came from. You mentioned that when you started as a journalist, Morris Chang had just left ITRI to found TSMC. It seems my colleague Nicholas has a question about TSMC’s origins.

Nicholas Welch: I’d like to ask how much of Taiwan’s semiconductor strategy can be attributed to fortunate economic planning after World War II. For example, consider the influence of Yin Zhongrong, who had a background in industrial engineering. In the 1950s, he established a model where key economic planning was led by industrial engineers rather than professional economists. This laid a good foundation for the Taiwan government to fully support TSMC in the 1980s and justify this approach on national security grounds. However, when Yin Zhongrong was in charge of Taiwan’s economy, semiconductors hadn’t even been invented yet. Did he foresee that Taiwan would dominate one of the world’s most important industries 30 years later?

Lin Hong-wen: Thank you for your question. It’s an interesting one. Personally, I don’t believe that Yin Zhongrong or those involved in planning Taiwan’s semiconductor development ever imagined that Taiwan would one day dominate the world’s most important foundry business. I’m certain they never considered such a possibility.

When they were promoting industrial development back then, Taiwan was in a very different situation. In the 1950s, after World War II and the relocation of the Nationalist government to Taiwan, the island had virtually no industry. There were only some light industries like textiles. In fact, Taiwan’s economy in the 1950s was mainly based on home workshops producing items like Christmas lights or umbrellas for the U.S. market. Taiwan had very weak economic strength, with no heavy industry or electronics industry to speak of.

At that time, they likely didn’t understand how important integrated circuits would become. I believe the government officials of that era didn’t have a deep understanding of semiconductors. However, they were willing to trust some experts who had returned to Taiwan. They allowed these people to help, including Morris Chang, who was one of them.

Another important factor was that Taiwan’s technology came from RCA through licensing. RCA was an American semiconductor company that was preparing to exit the industry. They likely felt they didn’t want to continue developing, so Taiwan, through someone working at RCA, asked for help. Later, Taiwan sent people to learn from RCA. We called them “knowledge-seeking ambassadors” — about 40-50 people who brought back RCA’s technology.

Following this, the establishment of Hsinchu Science Park and the Industrial Technology Research Institute (ITRI) were crucial steps. Looking back now, it might seem like they were incredibly farsighted and had grand visions. But in reality, it was just a step-by-step process of improvement.

In fact, TSMC only gained widespread attention after 2018, following the start of the U.S.-China tech war. Before that, who really cared about TSMC? Globally, I was probably the most dedicated person following TSMC. Not many people paid attention to it because it didn’t have its own products — it was just a foundry.

Recently, I was in Japan discussing this and showed some statistics. In the 2013 semiconductor revenue rankings, Intel was still at the top of the top 10 companies. TSMC’s revenue had long surpassed Intel’s, but TSMC was never included in these rankings. Why? Because TSMC doesn’t have its own products. Past semiconductor rankings were always based on products. So Intel was first, Samsung second, but TSMC, despite making $70 billion last year, wasn’t listed. My point is that nobody cared about TSMC before. Now, suddenly, because of the U.S.-China tech war or because of COVID, everyone has realized that TSMC is making all the world’s semiconductors.

Arrian Ebrahimi: Lin has quickly taken us from TSMC’s beginnings to its current state. There’s another question about the government’s role during TSMC’s founding process. My colleague Lily has a further question about the government.

Lily Ottinger: Yes, hello everyone. We’ve been discussing the step-by-step development process, but I’d like to talk about Taiwan’s history and successful strategies.

First, Taiwan implemented the “Land to the Tiller” program. After that, what policies do you think were crucial for enabling TSMC’s success? In the step-by-step process you mentioned, which steps do you consider most important?

Lin Hong-wen: For TSMC, I think the most important factors were talent, capital, and technology. Regarding talent, as I mentioned earlier, many people brought back technology from RCA training. Taiwan’s education system, especially in science and engineering, is quite strong and practical. People are very diligent and willing to work overtime. These are characteristics of Taiwanese engineers.

When Morris Chang came to Taiwan, he discovered these qualities and believed Taiwan could succeed in contract manufacturing. However, he didn’t think Taiwan would be successful in IC design, which is why he focused only on wafer foundry.

As for technology, we already had the RCA technology transfer. Another crucial factor was the development of the PC industry. Although semiconductors and PCs are different industries, many talented individuals in Taiwan were involved in PC development. Taiwan became a significant producer of PCs, including desktops and laptops, with many companies also producing in mainland China. This created opportunities, as PCs required semiconductors, and Taiwan had a semiconductor industry. The PC industry became a crucial carrier for Taiwan’s semiconductor industry, providing an outlet for many chips used in PCs.

Regarding capital, Taiwan’s private sector resources have always been quite strong. There was a period when a lot of Taiwanese investment went to mainland China, so capital and talent flowed there. However, in recent years, capital has been flowing back to Taiwan, so funding has generally been abundant.

Arrian Ebrahimi: Before we start discussing TSMC’s internal issues, I have one last question about the government’s role. You’ve previously discussed the different roles of the Industrial Technology Research Institute (ITRI) and the Academia Sinica. Can you compare them? What are the differences between these two research institutes? Before I went to Taiwan, I was not particularly familiar with them, and I think most mainland Chinese and American listeners might not be clear about these two institutions.

Lin Hong-wen: I think ITRI played a crucial role. As mentioned earlier, the technology licensing from RCA and sending people to learn were all done through ITRI. ITRI is a government-led research institution, but it’s not like many advanced research institutions in the U.S. Taiwan’s ITRI focuses more on commercializing technology.

ITRI was involved in PC and communications technology, which were important areas, but semiconductors were perhaps their most significant contribution. The people we sent to RCA to learn came back and initially worked at ITRI. They started with 3-inch and 4-inch wafers (now we’re at 12-inch). A key point is that they established a commercial production line at ITRI. This wasn’t just in a lab; it was capable of mass production.

I believe ITRI played an important role in Taiwan’s ability to seize opportunities in the global semiconductor vertical division of labor. We know that companies like UMC and TSMC all spun off from ITRI. So the predecessors of these companies, their teams, and their talent — ITRI played a crucial role in all of this.

In fact, I know that many countries around the world are now looking to learn from ITRI about how they did this back then. ITRI was a very important starting point. The science park that came later was another crucial element. The science park housed companies like UMC and TSMC. Essentially, the science park provided a single window for the government, allowing businesses to achieve scale and accelerate production. The regulations also favored these companies in terms of exports.

So when you ask what the government did, I think the Taiwanese government did quite a lot in terms of industrial infrastructure.

How the Silicon Shield is Managed

Arrian Ebrahimi: You used the word “predecessor” very aptly, and I asked this question to emphasize that the Taiwanese government has been supporting its industry all along. I remember Morris Chang complaining in an interview with Chris Miller that he was overemphasizing the role of the Taiwanese government, saying they didn’t help him that much. But when you look at the talent and basic technology contributions from ITRI and Academia Sinica, especially ITRI, the Taiwanese government’s support was really substantial. So thank you for explaining that, Mr. Lin.

Now, let’s start talking about some topics internal to TSMC, especially the different roles of their R&D department and engineering. Which aspect of the company do TSMC’s leaders value more: technological development or operational perspective? To ask a more specific question: Would [TSMC CEO] C.C. Wei be more excited about the R&D department’s contribution to the next generation of nanometer process invention, or about a 1% efficiency improvement announced by various factories? Which news would the president of TSMC be more eager to hear?

Lin Hong-wen: In my opinion, both are extremely important, but their importance varies at different times. In the early days, when TSMC’s technology wasn’t leading, they focused more on catching up technologically. When TSMC was founded, its technology lagged behind the world’s most advanced by about three to four generations. Now, of course, it far surpasses others.

TSMC’s technology leadership is crucial and ongoing. Today, we see TSMC as very powerful because its technology far exceeds that of its competitors. But you need to understand that TSMC’s technology was once behind. It wasn’t until around 2000-2001, at the 0.13 micron (130 nm) node, that they successfully developed their technology and started to show signs of leadership. Even then, they weren’t fully ahead.

It wasn’t until 2009, when Morris Chang returned as CEO, that they aggressively invested in 28nm technology and slowly surpassed others. Now, at 7nm, 5nm, and 3nm, they are leading. So from start to finish, technological leadership has always been extremely important.

However, the efficiency improvement you mentioned is something TSMC has always paid close attention to. In addition to technological leadership, operational excellence is crucial in wafer foundry. Good yield, short cycle time, and low cost all directly affect profitability. TSMC places great emphasis on efficiency improvements. They work on this every day, and all their engineers work overtime for this purpose.

So both aspects you mentioned are very important. Mr. Wei now needs to focus on both. Technology must continue to lead for TSMC to maintain its high profitability and create high value for customers. Without technological leadership, it wouldn’t work because others are always catching up.

However, there’s another aspect that’s even more important than these two: customer service satisfaction. From beginning to end, TSMC has always prioritized serving customers and maintaining high satisfaction. It’s not just about leading in technology or improving efficiency.

For example, if you look at TSMC’s competitors like Intel and Samsung, Intel had leading technology early on, but the problem is that Intel has its own products. Samsung also has its own products. When you have leading technology, do you use it for customers or for yourself? Honestly, this creates a conflict. In Samsung’s case, when they have leading technology, they might use it for their own products first and not give it to customers because they want their own products to be good.

So technological leadership isn’t actually the most crucial thing for a company like TSMC. The most important thing for TSMC is customer service and satisfaction. If customer service is good, then technological leadership and efficiency improvements are all for the benefit of customers.

TSMC has over 30 important customers, and these customers often change their product plans. Sometimes they urgently need something and ask TSMC to quickly adjust production to deliver goods earlier. So TSMC has to change its factory production process from top to bottom. Therefore, efficiency improvements are ultimately about serving customers.

I know Mr. Wei now arrives at the company at 4:30 every morning. Why? To serve customers. He can still call and talk with European and American customers at that time to understand their needs. So I think serving customers is the key to TSMC’s success.

Arrian Ebrahimi: Mr. Lin, you’re mainly saying that R&D and efficient fabs have made different contributions at different historical moments. At some points, one aspect may have played a more important role, while at other times, both were equally important. It’s a vivid description of alternating importance. But may I ask, why are most of TSMC’s leaders trained in factory management rather than coming from the R&D department?

This relates to another question: Since management and R&D play different roles, and the importance and strategic value of these two departments vary at different times, why did Morris Chang want his successors to take turns leading different departments? It seems that most of them initially came from engineering, but later Chang decided to let them experience every department. Can you explain a bit about this internal management approach?

Lin Hong-wen: Yes, you’re referring to when he had Mark Liu, C.C. Wei, and Chiang Shang-yi rotate through R&D, business, and wafer fab management. I think this was basically Morris Chang’s way of ensuring that a CEO would be very familiar with all three departments. So he had them rotate through different positions to gain different experiences.

Secondly, it was also a way for Morris to test them. Being a CEO isn’t just about ability; it’s also about how one handles different challenges. How do they react when faced with tests? Do they panic or handle situations calmly? I think this was all part of the testing process.

He used this method to let each of them hone their experience and mature their ability to handle situations and face challenges. So I think it’s natural to have successors experience different departments. This is an important process.

Arrian Ebrahimi: Before moving on to the next question, I want to ask about Morris Chang. Although he’s retired and says no one can replace his position at TSMC, it feels like his shadow is still there, right? I’ve heard that some people still call him “Grandpa”（爷爷）. Is this really the case, or is it just rumors?

Lin Hong-wen: I think calling him “Grandpa Chang” is quite natural because he’s in his 90s now, so he could be everyone’s grandpa. I believe he’s still a spiritual leader. Although he’s no longer working in an official capacity, I believe he still has a lot of influence. I’m sure he still provides guidance to younger generations, though he probably doesn’t initiate it. But I imagine many people still seek his advice.

For instance, I think C.C. Wei might consult him if there are important matters. So I believe his influence is still there. Even now, I know that many U.S. congressmen, senators, and state governors who visit Taiwan always hope to meet with him and hear his opinions. I think this is quite natural.

Arrian Ebrahimi: That’s a kind of filial piety haha. Before we continue discussing TSMC’s management, Lily has another question about management. It seems that she wants to compare TSMC’s management model with that of another industry.

Lily Ottinger: That’s right.  Mr. Lin, I saw that you also write articles about biotechnology, so I wanted to ask — the Taiwanese government has launched a state-owned biotechnology company called TBMC, which sounds like the biotechnology version of TSMC. What do you think are the differences between foundry manufacturing for integrated circuits and this type of foundry manufacturing for biotechnology manufacturing? What technical challenges might TBMC face?

Lin Hong-wen: Thank you for reading my articles so carefully. You’re right, I did write about TBMC and TSMC. When TBMC was being funded, they explicitly stated they wanted to emulate TSMC’s model, including partnering with the American Resilience Group. Resilience holds a 27.5% stake, which is the same percentage that Philips held in TSMC initially. They hope to create a successful model for biotechnology manufacturing in Taiwan, similar to what TSMC did for semiconductors.

However, I believe there are significant differences between these two industries. Biotech production processes are less complex, with about 30 steps, while semiconductor manufacturing involves around 150 steps. This difference in production complexity results in different entry barriers. Biotech manufacturing is relatively easier to enter, while high-end semiconductor manufacturing, especially for advanced processes, is much more difficult, leading to fewer competitors.

The investment amounts also differ greatly. A 12-inch semiconductor fab now requires at least $10 billion in investment, which is nearly 300 billion TWD. Many biological factories, on the other hand, only require 1-3 billion TWD in investment. This tenfold difference in investment further highlights the higher entry barrier for semiconductors.

Lastly, the products and cost structures of these industries are very different. For example, the production cost of Viagra accounts for only 0.7% of its price. If you reduce manufacturing costs by half, it only decreases the overall cost by 0.35%, which has a minimal impact. In contrast, semiconductor costs account for a much larger portion of the final product price. When TSMC uses the best technology to produce NVIDIA’s GPUs, the cost reduction is substantial.

IC functionality is also very powerful and continuously improves while prices decrease due to Moore’s Law. A single smartphone chip now uses 16 billion transistors, which is more than twice the world’s population. Pharmaceuticals, however, tend to become more expensive over time because they save lives, and people are willing to pay more for health.

These fundamental differences between the two industries mean that TBMC faces many challenges, and its future success is uncertain.

TSMC and Taiwan’s Startup Ecosystem

Arrian Ebrahimi: Thank you, Mr. Lin. You mentioned the entry barriers, which leads me to my next topic: TSMC’s support for Taiwan’s startup ecosystem. Before I ask you about this, I’d like to mention the job-hopping culture in the American tech industry. As someone who grew up in the Western United States, I’ve seen my high school classmates jump from company to company, gaining industry knowledge until they create their own startups, which may fail, leading them to join large companies before trying again. How would you compare Taiwan’s and America’s startup cultures, especially regarding job-hopping?

Lin Hong-wen: Taiwan’s situation is more similar to Silicon Valley. Job-hopping is very common in Taiwan because our industrial structure is dominated by small and medium-sized enterprises (SMEs). Taiwan has 1.5 million SMEs, of which 1.4 million have fewer than 5 employees. This is quite different from Japan and Korea, where large conglomerates dominate and people often work for one company their entire lives.

The key difference is that Taiwan has a strong entrepreneurial spirit. In SMEs, people often have ownership stakes, making them more willing to work hard because they’re working for themselves. This entrepreneurial spirit is similar to Silicon Valley.

However, the form of entrepreneurship in Taiwan differs from Silicon Valley. Taiwan specializes in professional manufacturing within the global supply chain. Many Taiwanese companies do contract manufacturing for large American companies like Google, Amazon, and Facebook. Taiwan has seized opportunities in industrial specialization.

For 30 years, the hottest news in Taiwan’s tech industry has been about “concept stocks” related to companies like Apple, Tesla, HP, and Dell. We focus on which Taiwanese companies are doing contract manufacturing or supplying components for these global brands. This structure of specialization continues today with companies like NVIDIA.

Arrian Ebrahimi: Mr. Lin, you’ve mentioned that Taiwan’s small companies are often supporting American big companies. Is it true that Taiwan’s startups are mostly in the chip industry, revolving around TSMC? Is their only goal to provide inputs and technology that TSMC needs?

Lin Hong-wen: That’s not entirely accurate. Taiwan doesn’t just have semiconductor companies. We have a significant PC industry with companies like Foxconn, Quanta, Compal, and others. We also have companies in various related fields, such as Largan Precision, which makes lenses for iPhone cameras, and many companies in networking and communications.

Taiwan also has a large industrial computer sector. While it’s true that many companies support TSMC, providing equipment, materials, and components that TSMC doesn’t develop in-house, this is just one part of Taiwan’s diverse tech ecosystem.

For example, with the rise of AI servers, heat dissipation has become crucial. Many Taiwanese companies are developing solutions for this challenge. So, Taiwan’s startups cover a wide range of technologies and industries, not just semiconductors.

TSMC and Taiwan’s International Status

Arrian Ebrahimi: I’d like to ask a potentially sensitive question: Which is more important, TSMC or the Taiwanese government?

Lin Hong-wen: That’s a profound question that kept me up last night. It touches on a crucial point for Taiwanese people. Before 2018, when the U.S.-China conflict intensified, the world didn’t pay much attention to Taiwan. Many Western observers were focused on China’s rapid growth over the past 30 years.

Both the Taiwanese government and TSMC are extremely important. The Taiwanese government represents our democratic system, chosen by the people’s votes. It embodies the direction we’ve chosen for ourselves after years of struggle.

TSMC is also crucial. It’s unprecedented for a Taiwanese company to reach such technological heights and play such a vital role in global chip production. For a small island of 23 million people to produce a company that influences the entire world and ranks among the top 10 globally by market value is a great source of pride for Taiwan.

Arrian Ebrahimi: This year, TSMC has accepted invitations from many countries to expand production abroad. It seems that the United States, Germany, and Japan are all grateful for TSMC’s choice. How should Taiwan seize this opportunity to enhance its international status?

Lin Hong-wen: I often give lectures in Japan, and I always say that I’m able to publish books thanks to TSMC’s success. Because TSMC is so successful, everyone pays attention to it, and the books I write sell well because everyone wants to read about it. But more importantly, I think Taiwan as a whole should capitalize on TSMC’s success.

People are thinking about how TSMC is now investing in the U.S., Japan, and Germany. In the past, many Taiwanese businesses invested in mainland China or Southeast Asia. We called it “westward expansion” or “southward expansion.” But those expansions were aimed at lowering costs, taking advantage of better production conditions in those places to earn excess profits.

Now, investing in the U.S., Japan, and Germany is fundamentally different because the costs in these places are much higher than in Taiwan. In the U.S., it’s several times higher. TSMC employees sent to Arizona have their salaries doubled, and they’re still paid less than American workers. I believe it’s the same in Germany because their original production conditions are that expensive.

So now, Taiwan isn’t just trying to make money by lowering costs; we’re trying to create value, which is much more difficult. I think Taiwan needs to handle this well. If TSMC can successfully produce in the U.S., Japan, and Germany, if these factories can operate smoothly, it will involve different corporate cultures. As I mentioned, Taiwanese people are more dedicated to work and willing to work overtime, but American and German employees probably aren’t as willing to work overtime. Even Japanese aren’t as willing to work overtime now.

There are also language barriers. To be honest, our English isn’t that good. It’s not our mother tongue, so when we go to Japan or the U.S., there are challenges in communicating with local employees. We see all these challenges.

But regardless, I think the Taiwanese government should use this opportunity to help more people understand Taiwan and its situation. As I mentioned, Taiwan still doesn’t have international status. We can’t even join the United Nations. We don’t have official diplomatic ties with most countries, and our representatives often had to meet foreign officials unofficially to avoid protests from China. When we participate in the Olympics, we’re called “Chinese Taipei.” We can’t even use the name “Taiwan” without facing suppression and protests.

Taiwan has been on the periphery of the world, just trying to survive. 

I think this is a feeling many Taiwanese have in their hearts. It’s hard for us to assert anything because as soon as we do, fighter jets fly over and missiles are pointed at us. Even big brothers like the U.S. tell us what to do. Taiwan is basically a little brother, and our international status is very low, almost non-existent.

But we can only expand our influence through economics and technology. We can make more investments and seize important opportunities in the global industrial division of labor. This is what everyone in Taiwan is working on now.

Nicholas Welch: I’d like to inquire about Taiwanese people’s views on the term “Silicon Shield” （矽盾）, beyond just those in the media. Do average Taiwanese citizens actually use this term?

Lin Hong-wen: Regarding the concept of a “Silicon Shield,” most Taiwanese people agree that TSMC is indeed a “Sacred Guardian Mountain”（护国神山）. TSMC has reached unprecedented heights for a Taiwanese company, leading in technology and playing a crucial role in global chip production. With a market value among the world’s top ten, it’s a source of great pride for Taiwan, a country of just 23 million people, to have produced a company with such global influence.

The term “Silicon Shield” wasn’t invented by Taiwanese people; it originated from foreign media. Since the intensification of US-China tensions in 2018, this concept has gained significance. People realize that TSMC’s production of chips for major global companies in Taiwan means that war cannot occur casually here. A conflict would disrupt the entire global electronics industry, affecting not only economic but also national security, as chips are crucial for missiles, fighter jets, drones, and other security applications.

While average Taiwanese may not discuss the term “Silicon Shield” explicitly, the sentiment behind it is widely understood. Taiwan harbors no hostility towards anyone. We lack international status, so we strive to contribute positively to everyone’s success through our efforts. Our chips enable companies like Apple and NVIDIA to sell more products globally, helping their businesses grow and flourish.

Taiwan primarily focuses on contract manufacturing, with few of our own products. Our industry’s core is built on mutually beneficial partnerships. Taiwan isn’t aggressive; even our fighter jets are primarily for deterrence. In conclusion, while the average person may not deeply contemplate the “Silicon Shield” concept, using semiconductors for self-protection is a common perspective currently.

Arrian Ebrahimi: Mr. Lin, you’ve just explained the Taiwanese mindset. Now, I’d like to compare the perspectives of different nationalities on international relations, particularly Taiwanese and Americans.

After spending a summer in Taiwan, my observation is that average Taiwanese people seem to care mostly about foreign trends that directly impact Taiwan’s internal development. In contrast, Americans tend to perceive certain countries as threats to national security, engaging in unfair trade practices, or exhibiting undemocratic behavior. It appears Taiwanese people focus primarily on their island, while Americans consider the entire world order. What do you think of this view, and how does Taiwan’s worldview influence its international strategy?

Lin Hong-wen: Your perception is understandable and natural. I largely agree with your observations, but it’s important to understand the underlying reasons. As I mentioned earlier, Taiwan has been continually suppressed and lacks international status. We have diplomatic relations with only a handful of small countries in Central and South America and Africa.

In the past, our overseas representatives were never officially received due to the lack of formal diplomatic ties. Meetings with foreign diplomats had to occur outside official offices to avoid protests from China. Taiwan has been on the world’s periphery, while the US is at its center. This fundamental difference in position shapes our respective worldviews.

In Taiwan, our primary concern has been survival. We focus on how others’ actions might impact us and how we can address resulting challenges. The US, being at the world’s center, naturally considers global issues and seeks to maintain world order, sometimes intervening when situations don’t align with its interests.

Your perspective is valid because we come from different backgrounds, cultures, and living conditions. Israelis would likely offer a different viewpoint, as would Japanese, despite our shared Asian heritage and strong work ethic.

Taiwan has transitioned from being a completely overlooked peripheral nation to becoming the center of the global semiconductor industry, which has surprised many. I often use the analogy of stagehands in theater — Taiwan used to be like the black-clad stagehands who move props unseen, working behind the scenes while companies like Apple and NVIDIA took center stage. Now, these “stagehands” have become crucial players.

Taiwan’s fundamental role hasn’t changed; we’re still primarily engaged in contract manufacturing and production within the global supply chain. What has changed is the critical importance of our role. While global trends have shifted, Taiwan continues to perform its core functions, albeit now in a more pivotal capacity.

Arrian Ebrahimi: Thank you, Mr. Lin. Your insights on the differences between countries and cultures, and Taiwan’s evolving international role, provide a natural conclusion to our discussion. We greatly appreciate your time in sharing TSMC’s story and Taiwan’s narrative. I recommend that all listeners purchase your book, Chip Island: How TSMC and Taiwan Triumph, which will soon be available in their native languages. Thank you very much, Owen Lin.

Listen to the interview in Chinese on Spotify, Apple Podcast, or CastBox!