Latest / NCCU IMES 愛鳴思 / The Global Footprint of TSMC: Views from the US, Germany, and Taiwan

Transcript

The Global Footprint of TSMC: Views from the US, Germany, and Taiwan · NCCU IMES 愛鳴思

• 0:00
  From data to impact. 從數據到影響力。
• 0:02
  Welcome to the bilingual podcast channel from NCCU IMES, bringing you conversations on social innovation and contemporary issues. 歡迎收聽政大愛鳴思的雙語 Podcast 頻道，為你帶來社會創新與當代議題的對話。
• 0:12
  歡迎來到政大愛鳴思，讓我們一起用數據解構問題，以創新重塑社會。 Welcome to NCCU IMES. Let's use data to deconstruct problems and reshape society through innovation.
• 0:21
  Welcome to the very first episode of the NCCU IMES podcast, where global economics meets real-world stories from the heart of Taipei. 歡迎收聽政大愛鳴思 Podcast 的第一集，在這裡，全球經濟與來自台北的真實故事交會。
• 0:29
  I'm your host for this episode, Zoe Weaver-Lee. 我是本集主持人 Zoe Weaver-Lee。
• 0:32
  Today, we're talking about the tiny silicon wafers powering everything from your smartphone to generative AI data centers: semiconductors. 今天我們要聊的，是驅動從智慧型手機到生成式 AI 資料中心一切運作的小小矽晶圓：半導體。
• 0:41
  For decades, Taiwan Semiconductor Manufacturing Company, or TSMC, has operated quietly behind the scenes, 數十年來，台灣積體電路製造公司，也就是台積電，一直在幕後默默運作，
• 0:47
  producing around 90% of the world's most advanced microchips. 生產全球約 90% 最先進的微晶片。
• 0:50
  But the landscape has shifted. 但局勢已經改變了。
• 0:53
  Facing geopolitical friction, supply chain fragility, and consumer demands, 面對地緣政治摩擦、供應鏈的脆弱性以及消費者需求，
• 0:57
  TSMC is building a multi-million dollar empire of fabs in Phoenix, Arizona; Dresden, Germany; and Kumamoto, Japan. 台積電正在亞利桑那州鳳凰城、德國德勒斯登和日本熊本，打造價值數百萬美元的晶圓廠帝國。
• 1:05
  Some local commentary here in Taipei posits that TSMC is turning into AMC, "American Semiconductor Manufacturing Company." 台北這裡有些評論認為，台積電正在變成「AMC」，也就是「美國半導體製造公司」。
• 1:13
  Yet in Washington and Berlin, policymakers hail these fabs as landmark victories for national and economic security. 然而在華盛頓和柏林，決策者將這些晶圓廠譽為國家安全與經濟安全的里程碑式勝利。
• 1:20
  What do the people living in these countries actually think? 那麼，住在這些國家的人們實際上怎麼想呢？
• 1:23
  Joining me in the studio today are three master's students from our IMES program: Lucy from the United States, 今天和我一起在錄音室的，是我們 IMES 學程的三位碩士生：來自美國的 Lucy、
• 1:29
  Leon from Germany, and Ray from Taiwan. 來自德國的 Leon，以及來自台灣的 Ray。
• 1:31
  Today, we're going to talk about lots of important aspects of this question. 今天我們會談到這個問題的許多重要面向。
• 1:35
  We're going to talk about the silicon shield. 我們會談「矽盾」。
• 1:37
  We're going to talk about differences in work culture where TSMC is investing. 我們會談台積電投資地區之間的工作文化差異。
• 1:41
  We're going to talk about the cost this has for consumers. 我們會談這對消費者造成的成本。
• 1:44
  And we're going to talk about some of the domestic politics at play here in Taiwan that will affect the future of TSMC's global footprint. 我們也會談到台灣國內的一些政治因素，這些因素將影響台積電全球布局的未來。
• 1:51
  Welcome to the studio. 歡迎來到錄音室。
• 1:53
  I'd like to start by asking something very simple to Lucy and Leon. 我想先問 Lucy 和 Leon 一個很簡單的問題。
• 1:56
  Had both of you heard of semiconductors before coming to Taiwan? 你們來台灣之前，有聽說過半導體嗎？
• 1:59
  And if so, what was the conversation like back home? 如果有的話，在你們國家大家是怎麼討論的？
• 2:02
  Yes, I definitely heard about semiconductors before coming to Taiwan. 有，我來台灣之前確實聽過半導體。
• 2:06
  I think the first time I heard about them, I had my first job and I was working at a think tank. 我想我第一次聽到它，是在我的第一份工作，當時我在一間智庫上班。
• 2:12
  We had a guest speaker come in and talk about technological areas where the US had an advantage over China, 有一位客座講者來談美國相對於中國具有優勢的科技領域，
• 2:18
  and she pointed to semiconductors as an example. 她就舉了半導體當例子。
• 2:21
  And I think more recently, with the rise of AI, a lot of folks are also aware that Taiwan has a critical place in that process, especially with regard to its abilities in foundry. 而我想最近隨著 AI 興起，很多人也意識到台灣在這個過程中占有關鍵地位，尤其是在晶圓代工的能力上。
• 2:32
  I think for Germany, it has been a large topic in the public debate, especially in the time after COVID, 我覺得在德國，這在公共辯論中一直是個大議題，尤其是在疫情之後，
• 2:37
  when we saw those global supply chain disruptions, because large German car manufacturers need those semiconductors. 我們看到全球供應鏈中斷，因為德國的大型汽車製造商需要這些半導體。
• 2:44
  And we saw the problem that those were mostly not locally produced in Germany, but imported from East Asia, especially Taiwan. 我們也看到一個問題：這些晶片大多不是在德國本地生產，而是從東亞進口，尤其是台灣。
• 2:52
  But I think in the everyday discourse, most people see it as a kind of abstract topic. 但我覺得在日常討論中，大多數人把它看成一個有點抽象的議題。
• 2:57
  So you don't have much actual technological knowledge to understand the technologies behind the different nanometer labels you often hear in the news. 所以大家其實沒有太多技術知識，去理解新聞裡常聽到的各種奈米製程標籤背後的技術。
• 3:06
  And recently, I would agree, I had a feeling that the debate now centers more on AI than on how central Taiwan is to this ecosystem. 而且我也同意，最近我感覺討論更多聚焦在 AI 上，而不是台灣在這個生態系中有多核心。
• 3:15
  Honestly, I'm not really surprised at all. 老實說，我一點都不意外。
• 3:18
  If you look through the lens of the capital markets, TSMC has been a global asset for decades. 如果從資本市場的角度來看，台積電幾十年來一直是全球性的資產。
• 3:25
  Over 70% of TSMC's shares are held by foreign institutional investors. 台積電超過 70% 的股份由外資法人持有。
• 3:32
  And also, I don't know whether you noticed the news that Harvard, their funds, they even raised their proportion of TSMC stock. 另外，不知道你們有沒有注意到一則新聞，哈佛的基金甚至提高了台積電持股的比例。
• 3:43
  So I think it's quite famous. 所以我覺得它相當有名。
• 3:45
  I'm not really surprised at all that foreigners get to know TSMC. 外國人會認識台積電，我一點都不意外。
• 3:51
  Great, thank you. 很好，謝謝。
• 3:52
  Great, thank you all. 很好，謝謝大家。
• 3:53
  Let's talk about the value of this sector here in Taiwan. 我們來談談這個產業在台灣的價值。
• 3:56
  Ray mentioned it a little bit. Ray 剛剛稍微提到了。
• 3:58
  I'm sure all of you have heard the term "silicon shield," or this idea that Taiwan's dominance in the semiconductor supply chain serves as a national security defense mechanism in the face of cross-strait tensions with China. 我相信大家都聽過「矽盾」這個詞，也就是台灣在半導體供應鏈中的主導地位，在兩岸緊張局勢下，能作為一種國家安全防衛機制。
• 4:11
  That is, that this technology is so critical and so centralized that a disruption of Taiwan's production would result in catastrophic economic consequences. 意思是，這項技術如此關鍵、如此集中，一旦台灣的生產中斷，將造成災難性的經濟後果。
• 4:21
  So from your perspectives, kind of asking all of you this, reflecting on the folks from your home countries, 所以從你們的角度，這個問題問大家，想想你們國家的人，
• 4:26
  do you actually think this is a concern? 你們真的覺得這是一個值得擔憂的問題嗎？
• 4:30
  We can start with Lucy. 我們可以從 Lucy 開始。
• 4:32
  Do you actually think... 你真的覺得……
• 4:34
  Maybe, is there such a thing as a silicon shield? 或者說，真的有矽盾這種東西嗎？
• 4:37
  Do you think hollowing out Taiwan's semiconductor industry poses a risk to its security when dealing with cross-strait tensions? 你認為掏空台灣的半導體產業，會在處理兩岸緊張關係時對台灣的安全構成風險嗎？
• 4:45
  What's your general thought? 你大致的看法是什麼？
• 4:48
  Yeah, I think in terms of the silicon shield issue, I'm familiar with the argument that if TSMC, for instance, 嗯，就矽盾這個議題來說，我熟悉一種論點：如果比方說台積電
• 4:56
  invests more in facilities like what it has in Arizona, that might lead to hollowing out of Taiwan's domestic industry. 更多投資在像亞利桑那那樣的廠區，可能會導致台灣國內產業被掏空。
• 5:07
  That being said, the Arizona facilities are not producing chips that are as advanced as the ones that can be produced in Taiwan. 話雖如此，亞利桑那廠生產的晶片並沒有台灣生產的那麼先進。
• 5:15
  I think Taiwan's abilities aren't just due to something like investment of resources. 我認為台灣的能力不只是來自資源投入這類因素。
• 5:20
  They're also owing to talent that Taiwan has invested in for decades, and Taiwan's science park model. 也來自台灣數十年來培養的人才，以及台灣的科學園區模式。
• 5:28
  So I think that there are a lot of Taiwan's capabilities that would not be represented in the Arizona facility at the end of the day. 所以我認為，說到底，台灣有很多能力是亞利桑那廠無法複製的。
• 5:36
  Okay, great, thank you. 好，很棒，謝謝。
• 5:37
  Do you think the silicon shield is real? 你們覺得矽盾是真的嗎？
• 5:39
  Do you think that it really does provide some sort of security for Taiwan to have this advanced... 你們覺得台灣擁有這麼先進的……真的能提供某種安全保障嗎？
• 5:44
  Yeah. 嗯。
• 5:44
  Ray, do you want to go first? Ray，你要不要先來？
• 5:46
  Do you have thoughts on this one as well? 你對這個問題也有想法嗎？
• 5:48
  So from my perspective, and this is my personal view, the silicon shield isn't a magical guarantee of peace, of course. 從我的角度來看，這是我個人的看法，矽盾當然不是什麼神奇的和平保證。
• 5:57
  And so I'd like to discuss it in two sides. 所以我想從兩個面向來討論。
• 6:01
  So on the technology side, Taiwan keeps its most advanced nodes at home: the 2-nanometer roadmap and critical advanced packaging like CoWoS, which is the actual bottleneck for advanced AI chips right now, remain heavily anchored in domestic clusters. 首先在技術面，台灣把最先進的製程留在國內：2 奈米的發展藍圖，以及像 CoWoS 這類關鍵的先進封裝——這正是目前先進 AI 晶片真正的瓶頸——仍然高度集中在國內的產業聚落。
• 6:22
  So overseas fabs are important, but they're built with generational lags compared to what's running here. 所以海外晶圓廠雖然重要，但和這裡運作中的相比，存在世代上的落差。
• 6:30
  And second, on whether someone could just take over and run these fabs: advanced semiconductor manufacturing is not an oil well, of course. 第二，關於有人能不能直接接管並運作這些晶圓廠：先進半導體製造當然不是一口油井。
• 6:41
  You can't just seize the physical fab and pump out chips, right? 你不可能直接奪下實體晶圓廠，就開始量產晶片，對吧？
• 6:46
  Modern fabs rely on real-time software patches, specialized chemicals from Japan, lithography equipment from ASML, 現代晶圓廠仰賴即時的軟體修補、來自日本的特殊化學品、來自 ASML 的微影設備，
• 6:55
  and continuous domestic engineering telemetry. 以及持續的國內工程遙測支援。
• 6:59
  So if the global supply chain or operational support is cut off, a cutting-edge fab halts production within weeks. 所以如果全球供應鏈或營運支援被切斷，一座尖端晶圓廠在幾週內就會停產。
• 7:07
  So the shield isn't just owning the concrete buildings. 所以這個盾牌不只是擁有那些水泥建築。
• 7:11
  It's an irreplaceable, living global network. 它是一個無可取代、活生生的全球網絡。
• 7:17
  Yeah, that's a good last point. 嗯，最後這點講得很好。
• 7:19
  One of the bigger debates that I've seen is whether or not, 我看過比較大的爭論之一是，
• 7:22
  if China does take over Taiwan's semiconductor manufacturing, it could actually keep it running for very long, 如果中國真的接管台灣的半導體製造，它是否真的能讓這些廠長期運作下去，
• 7:29
  considering its lack of know-how, lack of engineers who would be there to keep it running. 考量到它缺乏相關技術知識，也缺乏能讓工廠持續運轉的工程師。
• 7:35
  TSMC almost definitely has a backup plan for what would happen if that were to occur. 台積電幾乎肯定有一套應變計畫，以防這種情況發生。
• 7:40
  One of the things I was struck by when I visited TSMC is the security, the level of security. 我參觀台積電時印象很深的一點是它的保安，那種保安等級。
• 7:47
  I remember one of the things that was really interesting to me is every piece of paper inside TSMC is laced with metal, 我記得讓我覺得很有趣的一件事是，台積電裡面的每一張紙都摻有金屬，
• 7:56
  and they have metal detectors, so you cannot write a note without it being detected in leaving the building. 而且他們有金屬探測器，所以你不可能寫張紙條帶出大樓而不被偵測到。
• 8:02
  So they're very serious about security, they're very serious about leaks. 所以他們非常重視保安，非常重視洩密問題。
• 8:06
  So I would not be surprised that they have contingency plans for things like this happening. 所以如果他們對這類情況有應變計畫，我一點都不意外。
• 8:11
  So that's a good point at the end. 所以最後那點講得很好。
• 8:14
  Lucy, one of the things you mentioned, the US narrative... Lucy，你剛剛提到的一件事，美國的論述……
• 8:18
  I want to look specifically from the US perspective about the hollowing out of Taiwan's semiconductor industry. 我想特別從美國的角度來看台灣半導體產業被掏空這件事。
• 8:24
  US President Donald Trump has mentioned more than once that Taiwan has, quote unquote, "stolen" the US's semiconductor industry. 美國總統川普不只一次提到，台灣「偷走」了美國的半導體產業。
• 8:32
  What do you think is the administration's narrative about TSMC's investment in the US right now? 你覺得美國政府目前對台積電在美投資的論述是什麼？
• 8:37
  And is the public buying it? 大眾買單嗎？
• 8:38
  Yeah, so from what I've observed, the administration's narrative has been that TSMC's investment in Arizona demonstrates that tariffs and industrial policy can bring critical manufacturing to the United States. 嗯，就我觀察，政府的論述是：台積電在亞利桑那的投資證明了關稅和產業政策可以把關鍵製造業帶回美國。
• 8:50
  The idea is that foreign companies that want access to the US market are able to have that if they produce in the US. 這個想法是，想進入美國市場的外國企業，只要在美國生產就能進入。
• 8:57
  I think that while the public generally supports foreign investment that boosts manufacturing jobs, 我認為雖然大眾普遍支持能增加製造業就業的外國投資，
• 9:03
  the latest polling seems to suggest that folks aren't necessarily buying it when it comes to tariffs in particular. 但最新的民調似乎顯示，特別是在關稅方面，大家不一定買單。
• 9:10
  The US is preparing for its midterm elections right now, and affordability has been a huge issue. 美國現在正在準備期中選舉，而生活負擔能力一直是很大的議題。
• 9:15
  Tariffs have been doing pretty poorly with respect to that. 在這方面，關稅的表現相當不好。
• 9:19
  But one other thing that I'd like to note is that the administration, as you mentioned, 但我想補充的另一點是，就像你提到的，政府
• 9:24
  has often framed this as Taiwan stealing, quote unquote, 經常把這件事包裝成台灣「偷走」了
• 9:27
  "America's semiconductor industry," and that's oftentimes framed as being a more competitive relationship. 「美國的半導體產業」，這常常把兩者描繪成比較競爭的關係。
• 9:33
  But I think the history of TSMC in Arizona even demonstrates that this is a much more cooperative relationship. 但我認為台積電在亞利桑那的歷史正好證明，這是一種更合作的關係。
• 9:40
  The initial investment started in 2020. 最初的投資在 2020 年就開始了。
• 9:42
  That's before the CHIPS and Science Act even came about. 那是在《晶片與科學法案》出現之前。
• 9:47
  And then later on, the CHIPS and Science Act came about and brought about billions of dollars of incentives for TSMC to produce. 後來《晶片與科學法案》出現，為台積電在美生產帶來了數十億美元的補助。
• 9:54
  And this, of course, comes on top of the US and Taiwan's long-running cooperative history in the semiconductor sector. 當然，這還建立在美台雙方在半導體領域長期合作的歷史之上。
• 10:01
  Yeah, that's a really great overview of that. 嗯，這個概述非常好。
• 10:04
  And I think just last month, an additional $100 billion was committed, right, to keep expanding those Arizona fabs. 我記得就在上個月，又追加承諾了 1,000 億美元，對吧，用來持續擴建亞利桑那的晶圓廠。
• 10:10
  So it really has been quite significant. 所以規模真的相當可觀。
• 10:12
  But that's really great. 這真的很好。
• 10:14
  Leon, a similar question. Leon，類似的問題。
• 10:15
  We've seen that TSMC's joint venture in Dresden, ESMC, focuses on automotive and industrial nodes. 我們看到台積電在德勒斯登的合資公司 ESMC，專注於車用和工業用製程。
• 10:21
  So those are more so legacy chips, not so much cutting-edge chips. 所以那些比較偏成熟製程晶片，而不是尖端晶片。
• 10:26
  Does the German public see this as a rescue package for Germany's automotive backbone? 德國大眾是否把這視為德國汽車產業支柱的紓困方案？
• 10:31
  Or is there public pushback instead regarding the energy consumption and massive state subsidies? 還是反而在能源消耗和大規模政府補貼方面出現民眾反彈？
• 10:38
  I would definitely say that it's seen quite positively in Germany, given the structural change we now have, especially in East Germany. 我肯定會說這在德國被看得相當正面，考量到我們現在面臨的結構轉型，尤其是在東德。
• 10:46
  While there's also car industry in East Germany, I think it was more based on coal mining and the chemical industry. 雖然東德也有汽車產業，但我覺得那裡過去比較以煤礦和化學工業為主。
• 10:53
  But East Germany actually also has a tradition of semiconductor development, 不過東德其實也有半導體發展的傳統，
• 10:57
  as it was one of the leading countries among the Eastern Bloc for semiconductor development. 因為它曾是東歐集團中半導體發展領先的國家之一。
• 11:03
  Given the word "rescue package," I think the limitation to the rescue package is kind of that the automotive industry and the chip industry, which is now coming, are in very different regions. 說到「紓困方案」這個詞，我覺得它的限制在於，汽車產業和現在進來的晶片產業位在非常不同的地區。
• 11:13
  So the car industry is mostly based in southwest Germany, whereas the chip industry is now established in East Germany. 汽車產業主要在德國西南部，而晶片產業現在則設在東德。
• 11:21
  So we kind of have the problem that people can't directly move to new jobs from the car industry into the semiconductor industry. 所以我們面臨的問題是，人們無法直接從汽車業轉換到半導體業的新工作。
• 11:32
  And actually, the state subsidies are not much of a critique, I would say. 而且其實政府補貼並沒有受到太多批評，我會這麼說。
• 11:36
  They are a pretty typical instrument in Germany's industrial policy to enable investment. 補貼是德國產業政策中促進投資相當典型的工具。
• 11:42
  Furthermore, we don't actually have pretty tight fiscal rules now, 此外，我們現在的財政規則其實沒那麼緊，
• 11:46
  as we had a special fund established last year of about 500 billion euros. 因為去年我們設立了一個約 5,000 億歐元的特別基金。
• 11:52
  Okay, great. 好，很棒。
• 11:53
  I'm also going to ask you, I'm going to test your IMES economics knowledge here with this next question. 我也想問你，我要用下一個問題來考考你在 IMES 學到的經濟學知識。
• 12:00
  And anyone can chime in if they have thoughts on this too. 如果其他人對這個問題也有想法，隨時可以加入。
• 12:04
  It doesn't just have to be Leon. 不一定只有 Leon 能回答。
• 12:06
  We know that producing these chips in the United States and in Germany is more expensive than doing so in Taiwan. 我們知道在美國和德國生產這些晶片，比在台灣生產更貴。
• 12:12
  So do you think the customers of TSMC, which are big companies like Apple and Volkswagen, 那麼你們覺得台積電的客戶，也就是像蘋果和福斯這樣的大公司，
• 12:17
  do you think they will pay the cost, or do you think this cost will be passed on to consumers? 會自己吸收這個成本，還是會把成本轉嫁給消費者？
• 12:23
  Regarding Volkswagen, I think you need to differentiate between the higher-priced luxury cars and the lower-priced cars. 就福斯來說，我覺得要區分價格較高的豪華車和價格較低的車款。
• 12:30
  For the higher-priced cars, I think you can easily pass it along to the customers, whereas for the low-priced cars, 價格較高的車，我覺得很容易把成本轉嫁給顧客；而價格較低的車，
• 12:37
  given the recent competition from China, I think that's really not feasible. 考量到最近來自中國的競爭，我覺得真的不可行。
• 12:44
  Yeah, good point about competition from China as well, Leon. 嗯，Leon 提到來自中國的競爭這點也很好。
• 12:47
  And sorry, Lucy and Ray, do you guys have additional thoughts on that point? 不好意思，Lucy 和 Ray，你們對這點有什麼補充嗎？
• 12:51
  Yeah, so far, from what we've seen, TSMC customers like Apple, which you mentioned, 嗯，就目前我們看到的，台積電的客戶像是你提到的蘋果，
• 12:57
  have been willing to pay more for US-produced semiconductors. 一直願意為美國生產的半導體支付更多錢。
• 13:00
  Obviously, we have a pretty limited time sample here, but that's just the data that we have so far. 當然，我們這裡的時間樣本相當有限，但這就是目前手上的數據。
• 13:06
  We've not seen any companies, to my knowledge, say, "Oh, 就我所知，我們還沒看到任何公司說：「喔，
• 13:09
  we have an increase in price for this product because of the increased cost of producing chips in Arizona." 因為在亞利桑那生產晶片的成本增加，所以這項產品要漲價。」
• 13:15
  If we're talking about really big companies like Apple, they have big, sophisticated supply chains. 如果講的是像蘋果這樣真正的大公司，他們有龐大而精密的供應鏈。
• 13:21
  And oftentimes in the past, when they've had an increase in prices, 過去當他們面臨成本上漲時，
• 13:25
  they're able to kind of work along that supply chain to prevent the costs from coming down to consumers. 他們通常能沿著供應鏈去調整，避免成本轉嫁到消費者身上。
• 13:31
  So it's certainly possible that in the future we could see these costs passed down to consumers if there is no way that can be worked into their overall cost model. 所以未來如果這些成本無法被納入他們的整體成本模型，我們確實有可能看到成本轉嫁給消費者。
• 13:44
  But I'd also just note, you know, it's not just the increased costs at the Arizona facility driving up the cost of goods potentially. 但我也想指出，你知道，可能推高商品成本的不只是亞利桑那廠增加的成本。
• 13:53
  There are other things, like shipping instability in the Strait of Hormuz, that could drive these costs up. 還有其他因素，比如荷莫茲海峽的航運不穩定，也可能推高這些成本。
• 13:59
  So I think it's pretty unlikely that we'll see companies say, "Oh, 所以我覺得我們不太可能看到公司說：「喔，
• 14:02
  this cost is going up specifically because of this one issue related to TSMC in Arizona." 這項成本上漲，就是因為台積電在亞利桑那的這一個問題。」
• 14:07
  There are a lot of different potential causes for increases in prices these days. 現在物價上漲有很多不同的潛在原因。
• 14:12
  That's good news. 這是好消息。
• 14:13
  Ray, any additional thoughts? Ray，有什麼要補充的嗎？
• 14:17
  Actually, they already covered the things I wanted to say. 其實他們已經講到我想說的了。
• 14:22
  But I totally agree that we have to differentiate what kind of products we're talking about. 但我完全同意，我們必須區分在講的是哪一類產品。
• 14:31
  However, large customers like Apple and Nvidia are willing, they're willing to pay that premium, 不過，像蘋果和輝達這樣的大客戶是願意的，他們願意支付那筆溢價，
• 14:39
  because the cost of a complete supply chain shutdown is infinitely higher than paying slightly more per wafer. 因為整條供應鏈停擺的代價，遠遠高於每片晶圓多付一點錢。
• 14:46
  Furthermore, if you look at the unit economics of a flagship smartphone or an enterprise AI server, the silicon cost, 此外，如果你看旗艦智慧型手機或企業級 AI 伺服器的單位經濟效益，矽的成本，
• 14:55
  I mean the cost of the chips, is only a fraction of the total retail bill. 我是說晶片的成本，只占零售總價的一小部分。
• 15:01
  So absorbing or passing along an extra $20 or $30 per chip to end consumers is completely manageable. 所以每顆晶片多出的 20 或 30 美元，無論是自行吸收還是轉嫁給終端消費者，都完全可以應付。
• 15:10
  And it still has to depend on the products, which products, what kind of product we are talking about. 而且這還是要看產品，看是哪些產品、我們在談的是哪一類產品。
• 15:17
  Yeah, that's a good point. 嗯，這點說得很好。
• 15:19
  It does make me think that, you know, this current ecosystem that we have now, where we have a fabless-foundry model, 這確實讓我想到，你知道，我們現在這個採用無晶圓廠加晶圓代工模式的生態系，
• 15:27
  is still way more affordable than a vertically integrated model like we had in the '80s and into the '90s. 還是比我們在 80 年代到 90 年代那種垂直整合模式便宜得多。
• 15:34
  So I think you're right. 所以我覺得你說得對。
• 15:36
  Ray, I'd like to go back to you and cover your favorite topic, which is the politics of it all. Ray，我想回到你身上，聊聊你最喜歡的話題，也就是這一切背後的政治。
• 15:42
  So you've discussed Taiwan's perspective on whether these overseas investments erode the silicon shield. 你剛剛談到台灣怎麼看這些海外投資是否會侵蝕矽盾。
• 15:48
  We've already talked about that. 這部分我們已經聊過了。
• 15:50
  But just last year, Taiwan lifted a previous ban on overseas production of advanced semiconductors. 但就在去年，台灣解除了先前對先進半導體海外生產的禁令。
• 15:55
  Before that, it was illegal for a semiconductor company to produce anything more cutting-edge than 3 nanometers abroad. 在那之前，半導體公司在海外生產任何比 3 奈米更先進的製程都是違法的。
• 16:01
  And they lifted that a couple of years ago, or last year. 而他們在幾年前、或者說去年解除了這項禁令。
• 16:04
  But this year, KMT members in the Legislative Yuan have drafted the Chip National Security Act. 但今年，國民黨的立委在立法院草擬了《晶片國家安全法》。
• 16:11
  It's not passed yet, it's been drafted, but this would require government approval for future overseas semiconductor investment. 這個法案還沒通過，目前只是草案，但它會要求未來的海外半導體投資須經政府核准。
• 16:19
  How do you think Taiwan's domestic politics are playing into the future of Taiwan's semiconductor ecosystem overseas? 你認為台灣的國內政治，會如何影響台灣半導體生態系在海外的未來？
• 16:28
  To be honest, as an econ student, I don't follow day-to-day party politics all that closely. 老實說，身為一個念經濟的學生，我並沒有很密切地關注每天的政黨政治。
• 16:38
  I think a lot of people in Taiwan feel the same way. 我想很多台灣人也有同感。
• 16:41
  So when politicians debate oversight bills or overseas investment limits, it's really a political reflection of market and resource anxiety. 所以當政治人物在辯論監督法案或海外投資限制時，這其實是市場與資源焦慮在政治上的反映。
• 16:52
  TSMC alone accounts for over 30% of our entire stock market index. 光是台積電就占了我們整個股市指數超過 30%。
• 16:58
  When semiconductors perform well, our macroeconomic data looks incredible on paper, but in reality, it creates massive concentration risk. 當半導體表現好的時候，我們的總體經濟數據在帳面上看起來非常亮眼，但實際上，這造成了巨大的集中風險。
• 17:10
  Like top STEM talent, land, water, and green power are heavily directed toward one single sector, 像是頂尖的 STEM 人才、土地、水和綠電，都大量流向單一產業，
• 17:19
  while traditional industries and services struggle to keep up. 而傳統產業和服務業則難以跟上。
• 17:24
  So when you hear debates about regulating overseas expansion, I don't view it as partisan rivalry. 所以當你聽到關於規範海外擴張的辯論時，我並不把它看成政黨之爭。
• 17:34
  To me, it's simply a real-world question of national portfolio management: how do we support our biggest global asset without leaving the rest of the domestic economy completely unbalanced? 對我來說，這單純是一個關於國家投資組合管理的現實問題：我們要如何支持最大的全球資產，同時又不讓其他國內經濟完全失衡？
• 17:50
  Yeah, that's great. 嗯，很棒。
• 17:51
  To give our listeners a little bit of a bonus part of this episode, Ray, 給聽眾一點本集的額外加碼，Ray，
• 17:55
  do you want to tell them what is at the front of people's minds as we head into elections in Taiwan? 你要不要告訴大家，台灣在選舉前，大家最關心的是什麼？
• 18:02
  If it's not necessarily TSMC and where they invest, what are Taiwanese people thinking about for elections? 如果不一定是台積電和它在哪裡投資，那台灣人在選舉時都在想些什麼？
• 18:09
  When it comes to controversial issues like elections in Taiwan, actually, I'm a neutral person. 說到像台灣選舉這種有爭議的議題，其實我是個中立的人。
• 18:20
  Like I said, I never vote, to be honest. 就像我說的，老實說我從來不投票。
• 18:24
  And from my perspective, TSMC is just like the news I see on TV every day. 從我的角度來看，台積電就像是我每天在電視上看到的新聞而已。
• 18:32
  What Taiwanese people are really concerned about is like housing prices, their minimum wage, 台灣人真正在意的，像是房價、最低工資，
• 18:38
  such things that are really related to their daily life. 這些和日常生活真正相關的事情。
• 18:44
  So I think sometimes when it comes to politics, Taiwanese people will be really crazy about that, but, you know, 所以我覺得，有時候講到政治，台灣人會非常激動，但是，你知道，
• 18:53
  we're still more focused on the things that like housing prices and minimum wage, that are really related to our real life. 我們還是比較關注像房價、最低工資這些和我們實際生活真正相關的事。
• 19:05
  Yeah, relating back to something Lucy said earlier, this truly is an international issue. 嗯，呼應 Lucy 稍早說的，這真的是一個國際性的議題。
• 19:11
  I think at the core, in the United States, we're dealing with very similar issues about affordability, about inflation. 我覺得在美國，核心上我們也面對非常類似的問題，關於生活負擔能力、關於通膨。
• 19:17
  So I think these are definitely issues that transcend just Taiwan. 所以我覺得這些議題絕對不只存在於台灣。
• 19:21
  So I'd like to turn now to the practicality of semiconductor production in the United States and Germany versus how it works in Taiwan. 那麼我現在想轉向實務面，談談在美國和德國生產半導體，跟在台灣的運作方式有什麼不同。
• 19:29
  And I'm going to go back to you, Ray, because this, I think, is really interesting coming from a Taiwanese person's perspective. 我要再回到你身上，Ray，因為我覺得從台灣人的角度來談這點真的很有意思。
• 19:36
  What is the work culture like in the semiconductor industry in Taiwan? 台灣半導體產業的工作文化是什麼樣子？
• 19:40
  And then Lucy and Leon, commenting on that, what are some of the challenges that you think TSMC will see implementing this kind of ecosystem in your countries? 然後 Lucy 和 Leon 也針對這點評論一下，你們覺得台積電在你們國家建立這種生態系時，會遇到哪些挑戰？
• 19:52
  Actually, one of my relatives was working at TSMC for around 10 years. 其實，我有一位親戚在台積電工作了大約 10 年。
• 19:57
  He was working in the human resources department, but not at a very high rank. 他在人力資源部門工作，但職位不是很高。
• 20:02
  That was the beginning of TSMC. 那是台積電的早期。
• 20:06
  He told me that he worked really hard there. 他跟我說他在那裡工作得非常辛苦。
• 20:08
  I think that's the reason why he never had kids. 我覺得這就是他一直沒有生小孩的原因。
• 20:11
  Yes. 是的。
• 20:12
  So, however, he still thinks TSMC is a very good company, especially since they pay a very good salary. 不過，他仍然覺得台積電是一間非常好的公司，特別是因為他們給的薪水非常好。
• 20:19
  Almost every worker there can afford a house. 那裡幾乎每個員工都買得起房子。
• 20:22
  It will depend on the culture. 這要看文化。
• 20:24
  If you're Taiwanese, you can do that kind of working hours. 如果你是台灣人，你可以接受那樣的工時。
• 20:28
  You can accept TSMC's values. 你可以接受台積電的價值觀。
• 20:30
  On that point, Lucy, you know, if you're Taiwanese, it's kind of accepted that this is a really valuable, 關於這點，Lucy，你知道，如果你是台灣人，大家多少都認同這是一個非常有價值、
• 20:36
  hard place to get into to work, so it's kind of worth it. 很難進去的工作地方，所以算是值得的。
• 20:39
  Is that the case in the United States? 在美國也是這樣嗎？
• 20:41
  What's kind of the challenge been in the United States? 在美國遇到的挑戰是什麼？
• 20:44
  It seems there's been a disconnect with the Taiwanese work culture, 看起來美國和台灣的工作文化之間有一些落差，
• 20:47
  especially for engineers: you're on shift for a long time, and if there's an issue when you're off shift, 尤其是對工程師來說：你要輪很長的班，而如果下班後出了問題，
• 20:52
  you'll be contacted and you may need to come into work and fix something. 你會被聯絡，而且可能需要回公司處理。
• 20:56
  I'd say that's something that we're not super used to in the United States. 我會說這是我們在美國不太習慣的事。
• 21:00
  I think especially if you work a long shift job, it's expected that when you clock out, you have a bit more free time, 我覺得特別是如果你做的是長班的工作，大家會預期下班打卡之後，你會有多一點自由時間，
• 21:06
  and that really is your time. 而那真的就是屬於你自己的時間。
• 21:08
  So it seems like this has led to some misunderstandings between the staff at the Arizona plant. 所以這似乎導致亞利桑那廠的員工之間產生了一些誤解。
• 21:14
  I'm encouraged to hear about things like, apparently, 我很高興聽到一些事情，像是據說
• 21:17
  they brought some of the US-based employees to a Taiwanese fab so they could see how they work. 他們帶了一些美國員工到台灣的晶圓廠，讓他們看看這裡怎麼運作。
• 21:23
  I think that's a really good thing, if you're kind of explaining by showing and saying, you know, 我覺得這是一件非常好的事，如果你是透過實際展示來解釋，說，你知道，
• 21:27
  there's a reason why it has to run this way and there's a reason why we do it this way. 它必須這樣運作是有原因的，我們這樣做也是有原因的。
• 21:32
  I think that's more of a bottom-up way of bringing people to understanding. 我覺得這是一種更由下而上、讓人理解的方式。
• 21:36
  But I think an additional issue we have in the United States is with regard to STEM talent, 但我覺得美國還有一個額外的問題，就是 STEM 人才，
• 21:41
  especially for chip production and not just something like IC design. 特別是在晶片生產方面，而不只是像 IC 設計這類領域。
• 21:45
  That's not something that we've really focused on. 這不是我們過去真正重視的部分。
• 21:48
  So I think that's a further issue for TSMC in the US. 所以我覺得這是台積電在美國面臨的另一個問題。
• 21:51
  Yeah, great, great last point. 嗯，最後這點講得非常好。
• 21:52
  I heard that, you know, TSMC has a very good partnership with Arizona State University, right? 我聽說，你知道，台積電和亞利桑那州立大學有很好的合作關係，對吧？
• 21:57
  And they kind of have a pipeline program, which I think is really helping with the talent issue. 他們有一個人才培育計畫，我覺得這真的有助於解決人才問題。
• 22:03
  But I think that is pretty applicable across the board. 但我覺得這個問題其實普遍適用。
• 22:07
  Anywhere that semiconductors are being produced, the talent shortage is a real problem, also here in Taiwan. 任何生產半導體的地方，人才短缺都是個真實的問題，在台灣這裡也是。
• 22:15
  So yeah, that's a great point as well. 所以，對，這點也講得很好。
• 22:17
  I think it's not necessarily just tech talent as well. 我覺得不一定只是科技人才。
• 22:20
  I think it also extends into things like welding and construction. 我覺得也延伸到像焊接和營造這類領域。
• 22:24
  So there are a lot of issues there as well. 所以那方面也有很多問題。
• 22:26
  Yeah, very true. 嗯，非常對。
• 22:27
  Leon, what about Germany? Leon，那德國呢？
• 22:28
  What's the challenge there? 那邊的挑戰是什麼？
• 22:30
  I think Germany is commonly known for its relatively strict labor laws. 我覺得德國以相對嚴格的勞動法聞名。
• 22:35
  I think that was also very prominent when Tesla established its Gigafactory near Berlin and kind of clashed with the reality of German labor law. 我覺得這在特斯拉於柏林附近設立超級工廠時也非常明顯，有點像是和德國勞動法的現實正面衝撞。
• 22:45
  So the working hours are very much limited, it's pretty hard to lay off employees, and also, 工時受到很大的限制，要解雇員工也相當困難，而且，
• 22:51
  I think in Germany you have the fact that people less personally identify with their employer, 我覺得在德國，人們對雇主的個人認同感比較低，
• 22:58
  as I think you may see in Taiwan if you're working at TSMC. 不像在台灣，如果你在台積電工作，可能就會看到那種認同感。
• 23:02
  But I think the major difficulties in Germany are more like infrastructure and bureaucracy, because in Germany, 但我覺得德國的主要困難比較在於基礎建設和官僚體系，因為在德國，
• 23:10
  infrastructure funding has been very low for the past decades. 過去幾十年的基礎建設經費一直非常低。
• 23:14
  So I think we need to build up that infrastructure first to attract foreign investment. 所以我覺得我們必須先把基礎建設做起來，才能吸引外國投資。
• 23:20
  And also German bureaucracy is limiting those investments, as it's a very hard and very time-consuming, 另外，德國的官僚體系也限制了這些投資，因為要在德國設立子公司，是一個非常困難、非常耗時、
• 23:27
  long process to establish a subsidiary in Germany. 很漫長的過程。
• 23:31
  Great. 很好。
• 23:32
  So staying on this topic of public relations, Lucy and Leon, 那麼延續這個關於公眾觀感的話題，Lucy 和 Leon，
• 23:35
  I'd like to ask you specifically about your home countries' perspectives on AI and data centers. 我想特別問問你們國家對 AI 和資料中心的看法。
• 23:41
  So we know that AI is inherently connected to semiconductors. 我們知道 AI 本質上和半導體緊密相連。
• 23:44
  Semiconductors are the hardware that make AI do its thing. 半導體就是讓 AI 能運作的硬體。
• 23:48
  So how will this sentiment affect incoming semiconductor investment? 那麼這種民意會如何影響即將進來的半導體投資？
• 23:52
  Yeah, I think the biggest area for debate with this in the US right now is related to the conversation around data centers. 嗯，我覺得在美國，目前這方面最大的爭論領域，和資料中心的討論有關。
• 23:58
  This has become really a key issue for politicians in both parties leading up to the November elections that we're going to have in the US. 在我們即將到來的 11 月選舉前，這已經成為兩黨政治人物的關鍵議題。
• 24:07
  So I don't know exactly how this pans out, but I imagine that the decision makers at TSMC, 所以我不確定最後會怎麼發展，但我想台積電的決策者，
• 24:13
  or really any company that is investing in the semiconductor sector in the US, might be looking at those milestones to see how it develops. 或者任何在美國投資半導體產業的公司，可能都會觀察這些關鍵節點，看情勢如何發展。
• 24:22
  I think for Germany, actually, data centers aren't that much of a discussion topic right now, but for AI it's definitely the case. 我覺得在德國，資料中心其實目前不是太熱門的討論話題，但 AI 絕對是。
• 24:31
  I think in general, Germans tend to be more critical towards new technologies. 我覺得整體來說，德國人對新技術往往比較抱持批判態度。
• 24:36
  So with AI, also in public media, you pretty much hear about the risks of AI and ethical questions arising, 所以對於 AI，在公共媒體上你也幾乎只會聽到 AI 的風險和隨之而來的倫理問題，
• 24:44
  but not so much actually the opportunities of AI use. 而不太會聽到使用 AI 的機會。
• 24:48
  So in Germany, also, data privacy laws and everything are very good at regulating pretty much any part of life. 而且在德國，資料隱私法之類的規範幾乎把生活中的每個層面都管得很好。
• 24:57
  So I think that might also limit the opportunities for AI in Germany. 所以我覺得這也可能限制 AI 在德國的發展機會。
• 25:01
  And Ray, you mentioned that when your uncle was working at TSMC, despite the hard work, he still valued it, Ray，你提到你的舅舅在台積電工作時，雖然工作很辛苦，他還是很珍惜，
• 25:09
  because TSMC is an admired place to work at. 因為台積電是一個令人嚮往的工作地點。
• 25:12
  If you tell anyone in Taiwan that you work there, it's considered very prestigious, right? 在台灣只要跟別人說你在那裡工作，就會被認為很有聲望，對吧？
• 25:17
  So as a graduate student who's about to be on the job market, and you're seeing TSMC invest abroad, 所以身為一個即將進入就業市場的研究生，看到台積電到海外投資，
• 25:23
  what do you think that feels like for Taiwanese workers who see this industry as a big part of finding employment after graduation? 你覺得對那些把這個產業視為畢業後主要就業出路的台灣勞工來說，是什麼樣的感受？
• 25:32
  Well, it's a really two-sided reality for young professionals in Taiwan. 嗯，對台灣的年輕專業人士來說，這真的是一體兩面的現實。
• 25:38
  So for engineers, supply chain specialists, and bilingual management graduates, overseas expansion creates exciting global mobility. 對工程師、供應鏈專家和雙語管理背景的畢業生來說，海外擴張創造了令人興奮的全球流動機會。
• 25:49
  You aren't just limited to facilities in Hsinchu or Tainan. 你不再只侷限於新竹或台南的廠區。
• 25:55
  You have opportunities to work in Phoenix, Dresden, or even Japan, managing multinational teams early in your career. 你有機會到鳳凰城、德勒斯登，甚至日本工作，在職涯早期就管理跨國團隊。
• 26:04
  On the other hand, for graduate students in humanities, social science, and traditional business roles, the boom can feel isolating. 另一方面，對人文、社會科學和傳統商管領域的研究生來說，這波熱潮可能讓人感到被邊緣化。
• 26:14
  The semiconductor expansion drives up living and housing expenses in major city areas, 半導體擴張推高了主要都會區的生活和居住成本，
• 26:22
  while entry-level salaries in non-tech sectors remain relatively low. 而非科技產業的起薪仍然相對偏低。
• 26:29
  So that's why bridging the gap between high-tech wealth and broader economic development is probably the most pressing issue in our generation. 所以，縮小高科技財富與更廣泛經濟發展之間的落差，可能是我們這一代最迫切的議題。
• 26:40
  Do you think your parents would like it if you started working at TSMC? 如果你開始在台積電工作，你覺得你爸媽會開心嗎？
• 26:44
  Of course. 當然。
• 26:45
  They've got a really good salary there. 那裡的薪水真的很好。
• 26:48
  Yes, yes, I can imagine. 是啊是啊，可以想像。
• 26:49
  So, guys, we've covered quite a lot today. 好，各位，我們今天聊了很多。
• 26:52
  Before we wrap up, I'd love for the three of you to share your final thoughts on some of the future aspects of Taiwan's global semiconductor footprint. 在結束之前，我想請三位分享你們對台灣半導體全球布局未來發展的最後想法。
• 27:01
  If you have some optimism, that's great, but of course, if you have some more realism, some more pessimism, 如果你抱持樂觀，那很好，但當然，如果你的看法比較務實，甚至比較悲觀，
• 27:06
  I'm open to hearing that as well. 我也很樂意聽聽。
• 27:08
  Lucy, do you want to start? Lucy，你要先開始嗎？
• 27:10
  Sure, I'll start us off on an optimistic note. 好，我先用樂觀的角度開場。
• 27:13
  I think that even with all the investment that we've seen in the Arizona facility, in the Dresden facility, 我認為即使我們看到在亞利桑那廠、德勒斯登廠、
• 27:20
  in the facility in Japan, it seems that no one makes semiconductors as advanced as Taiwan. 日本廠的這些投資，似乎還是沒有人能做出像台灣一樣先進的半導體。
• 27:28
  I think that really reflects the level of complexity in semiconductor supply chains, 我覺得這真正反映了半導體供應鏈的複雜程度，
• 27:34
  and the fact that just close to us in Taipei, there are so many companies along the supply chain that specialize in very, very specific goods. 而且光是在我們台北附近，供應鏈上就有非常多專精於非常、非常特定產品的公司。
• 27:43
  And it's not just, you know, the tech aspect of it. 而且，你知道，這不只是技術層面而已。
• 27:46
  It's about Taiwan's logistics sector as well, and the ability to actually, you know, 也關乎台灣的物流產業，以及實際上能夠，你知道，
• 27:51
  put these materials into a process that creates functioning chips. 把這些材料投入製程、做出能運作的晶片的能力。
• 27:56
  So I think that Taiwan's role in the sector is irreplaceable as a leader. 所以我認為台灣作為領導者，在這個產業中的角色是無可取代的。
• 28:02
  Great, thank you, Lucy. 很好，謝謝 Lucy。
• 28:04
  Ray, any final thoughts? Ray，有什麼最後的想法嗎？
• 28:06
  My takeaway is that building deeper global trust is essential for Taiwan to secure its footing on the world stage. 我的結論是，建立更深厚的全球信任，對台灣在國際舞台上站穩腳步至關重要。
• 28:14
  And by expanding overseas operations in the US, Germany, and Japan, 透過在美國、德國和日本擴展海外營運，
• 28:19
  we demonstrate our commitment as a reliable and indispensable global partner, 我們展現了作為可靠且不可或缺的全球夥伴的承諾，
• 28:25
  while learning to navigate international governance and local collaboration. 同時也學習如何應對國際治理與在地協作。
• 28:31
  And ultimately, long-term resilience and geopolitical standing aren't built on isolationism, 而最終，長期的韌性和地緣政治地位，並不是建立在孤立主義上，
• 28:38
  but on fostering mutual trust and deeply integrated international partnerships. 而是建立在培養互信與深度整合的國際夥伴關係上。
• 28:45
  Thank you, Ray. 謝謝 Ray。
• 28:47
  Leon? Leon 呢？
• 28:47
  I think it's beneficial for both sides, in my opinion. 我覺得這對雙方都有好處，這是我的看法。
• 28:52
  So for Germany, the USA, but also for Taiwan. 對德國、美國，也對台灣都是。
• 28:55
  I think also Germany has one of the highest rates of engineers and is leading in many sectors of academic research. 我覺得德國也是工程師比例最高的國家之一，並且在許多學術研究領域處於領先地位。
• 29:05
  And now I've also seen that Germany is connecting more and more with Taiwan's universities, 而現在我也看到德國和台灣的大學、
• 29:12
  with Taiwan's academic research in the semiconductor field. 和台灣在半導體領域的學術研究，連結越來越多。
• 29:16
  I think in Germany, it brings more resilience in our supply chain, and I think for Taiwan, it's a really big geostrategic asset. 我覺得對德國來說，這為我們的供應鏈帶來更多韌性；而對台灣來說，這是一項非常重要的地緣戰略資產。
• 29:24
  Great, thank you all for sharing that. 很好，謝謝大家的分享。
• 29:26
  I totally agree with the final points of optimism. 我完全同意最後這些樂觀的觀點。
• 29:28
  I think there's actually a lot to be hopeful for at the moment. 我覺得目前其實有很多值得期待的地方。
• 29:32
  Of course, there are some big geopolitical issues at play here that will have to play themselves out, 當然，這裡還有一些重大的地緣政治問題正在發酵，必須等它們自行演變，
• 29:38
  and hopefully policy can help that along in a positive way. 希望政策能以正面的方式推動這一切。
• 29:41
  But in the meantime, I want to thank you all for joining me and discussing this today. 但在此同時，我想謝謝各位今天和我一起來討論這個議題。
• 29:45
  I'd like to thank our listeners for joining us for the very first episode of the NCCU IMES podcast. 也謝謝各位聽眾收聽政大愛鳴思 Podcast 的第一集。
• 29:50
  Stay tuned here for future episodes, and we hope to see you back again soon. 請繼續鎖定我們未來的節目，希望很快再與你相見。