These tiny and intricate silicon wafers can be found in almost every appliance we touch and are powering transformative technologies like artificial intelligence, electric vehicles, and advanced wireless networks. Pretty much everything from microwaves to the stock market runs on chips. The boom in innovation has turbocharged leading chip makers like NVIDIA to astronomical market caps. The VanEck Semiconductor ETF, which tracks semiconductor companies, has climbed 45% since the beginning of 2024 despite recent turbulence in the sector.

這些微小而複雜的硅晶片幾乎存在於我們接觸到的所有電器中，併為人工智能、電動汽車和先進的無線網絡等變革性技術提供動力。從微波爐到股票市場，幾乎所有的東西都依靠芯片運行。創新的蓬勃發展使英偉達（NVIDIA）等領先芯片製造商的市值飆升至天文數字。追蹤半導體公司的 VanEck 半導體 ETF 自 2024 年初以來已攀升了 45%，儘管該行業近期動盪不安。

America has led the world in designing and building the most advanced chips in the market, but that position is being challenged as big manufacturing hubs like Taiwan and Korea rise in prominence. And massive government spending efforts in China to bolster its chip industry also threaten to upend the U.S.'s competitive advantage in the sector.

在設計和製造市場上最先進的芯片方面，美國一直處於世界領先地位，但隨著臺灣和韓國等大型製造中心的崛起，這一地位正受到挑戰。而中國政府為扶持芯片產業而投入的鉅額資金也有可能顛覆美國在該領域的競爭優勢。

Policymakers have responded with subsidies to boost domestic chip makers and bring some components of production back to U.S. soil, but industry leaders say it will be a long road.

政策制定者通過補貼來促進國內芯片製造商的發展，並將部分生產環節帶回美國本土，但行業領袖們表示，這將是一條漫長的道路。

We are somewhere between a decade to two decades away from supply chain independence.

我們距離供應鏈獨立還有十年到二十年的時間。

On this Growth Stories Spotlight, we speak to Chris Miller, economic historian and author of the book, Chip War, the fight for the world's most critical technology, to get his thoughts on the state of the industry and the race to reassert America's dominance over the global semiconductor sector. Chris, thanks so much for joining us today. Thank you for having me.

在本期 "增長故事聚焦 "節目中，我們將採訪經濟歷史學家、《芯片戰爭：爭奪世界最關鍵技術》一書的作者克里斯-米勒（Chris Miller），瞭解他對該行業現狀以及美國在全球半導體行業重新佔據主導地位的競爭的看法。克里斯，非常感謝你今天參加我們的節目。感謝您的邀請。

You give a comprehensive look at the chip sector and the rising tensions between the U.S.

您對芯片行業以及美國與中國之間日益緊張的關係進行了全面的分析。

and China in your book, Chip War. So how would you characterize the state of the industry right now?

在您的《芯片戰爭》一書中，您提到了中國和美國。那麼，您如何描述目前的行業現狀？

I think the industry is simultaneously booming because of all the investment in AI, with data centers being built requiring more and more expensive semiconductors than ever before, but simultaneously being stressed by the U.S.-China political competition, which is restricting market access and adding new limits on who can sell which types of chips. And of course, China is one of the world's largest economies. And so it matters tremendously if U.S. companies are unable to sell certain products to them. And we saw that pandemic era shortages were really a wake up call for executives and policymakers and investors as well, about the importance of expanding or at least trying to bring back some chip manufacturing capabilities to American soil. And part of that effort resulted in the 2022 Chips and Science Act, which earmarked $39 billion in subsidies and gave investment tax credits to strengthen domestic chip production. So how would you say that program is doing and what sort of impact have we seen on the industry?

我認為，由於對人工智能的大量投資，該行業正在蓬勃發展，數據中心的建設需要比以往任何時候都更多、更昂貴的半導體，但同時中美政治競爭也給該行業帶來了壓力，它限制了市場準入，並對誰能銷售哪種類型的芯片增加了新的限制。當然，中國是世界上最大的經濟體之一。是以，如果美國公司無法向中國銷售某些產品，這將是一件非常重要的事情。我們看到，大流行病時期的芯片短缺確實給企業高管、政策制定者和投資者敲響了警鐘，讓他們意識到擴大或至少嘗試將一些芯片製造能力帶回美國本土的重要性。這項努力的部分成果就是《2022 芯片與科學法案》（

Well, I think it's undeniable when you look at the data that there's been a huge boom in investment in chip making facilities in the U.S., something like a 10 times increase over the last couple of years in investment relative to the prior trend. So that is evidence of success. That's exactly what Congress wanted when they passed the Chips Act. But I think it's also important to keep the scale in context because capital investment in this industry is so huge that even a major multi-billion dollar program like the Chips Act only moves things slowly. And so just to put some numbers on the table, the Chips Act, as you said, allocated $39 billion in capital expenditure over the next couple of years to subsidize chip manufacturing. But the largest chip company in the world, Taiwan's TSMC, spent almost that much money last year in capital investment in just a single year. And so that illustrates just how large the amounts are that you have to spend if you want to change where manufacturing is happening in this industry.

我認為不可否認的是，從數據來看，美國芯片製造設施的投資出現了巨大的增長。這就是成功的證據。這正是國會在通過《芯片法案》時所希望的。但我認為，保持一定的規模也很重要，因為這個行業的資本投資是如此巨大，即使像《芯片法案》這樣價值數十億美元的大型項目，也只能緩慢推進。是以，我只想列舉一些數字，如你所說，《芯片法案》在未來幾年內將撥出 390 億美元的資本支出，用於補貼芯片製造。但全球最大的芯片公司，臺灣的臺積電，去年一年的資本投資就花了差不多這麼多錢。由此可見，要想改變這個行業的生產現狀，必須投入的資金有多大。

And you talked about that speed because that has been some criticism of this, the speed of getting money out there. You see a number of initiatives experiencing delays in the first year. Taiwan's one of those that delayed a second production facility.

你談到了速度問題，因為這一直是一些責備的聲音，即資金到位的速度。你看到很多項目在第一年都出現了延誤。臺灣就是其中之一，延後了第二個生產設施的建設。

But you say the volume of investment tells a different story there. So why is that? And what would you say to investors that are maybe concerned about a return on that investment?

但你說投資額說明了不同的情況。為什麼會這樣？對於那些擔心投資回報的投資者，你有什麼要說的嗎？

I think, first off, in terms of the delays, I think the delays that we've seen have not really been about delays to government incentives because we've more or less known which companies are going to get how much, plus or minus a little bit. That's basically been clear from the early stages because the Commerce Department said they wanted to devote most of the money to leading edge manufacturers, of whom there are only three, and the rest to lagging edge. I think the delays that have existed have been because construction is hard, because workforce challenges exist, and those are difficult problems to work through. And they're being worked through now, but they just take time and they take money. I think when it comes to return on investment, from the government's perspective, their job isn't to think about return on investment. Their job is to think about insuring against the worst case scenario around losing access to manufacturing in Taiwan. And that's what the CHIPS Act is trying to do. For companies, their job, of course, is to think about return on investment. And they've had to balance the fact that they're getting new subsidies from the government in the US, but also costs are somewhat higher in the US.

我認為，首先，就延遲而言，我認為我們所看到的延遲並不是政府激勵措施的延遲，因為我們或多或少已經知道哪些公司將獲得多少資金，或多或少。這一點基本上從早期階段就已經很清楚了，因為商務部表示，他們希望將大部分資金投入到領先的製造商身上，而領先的製造商只有三家，其餘的資金投入到落後的製造商身上。我認為，之所以出現延誤，是因為施工困難，勞動力短缺，這些都是難以解決的問題。現在正在解決這些問題，但這需要時間和資金。我認為，說到投資回報，從政府的角度來看，他們的工作不是考慮投資回報。他們的工作是考慮如何在最壞的情況下避

So the long-term costs of running a facility need to be balanced against the subsidies. And I think all of the companies involved have been trying to find the right balance between the cost dynamics and also ascertain, are their customers willing to pay a bit more for that type of supply chain diversification? And we've seen US subsidies, but they pale in comparison to China. Beijing is basically doubling down and setting aside over $47 billion just this year to boost their chip industry. And we've seen a lot of talk about export controls and tariffs being used as the main policy tools to contain China's technological rise. But in your opinion, have they been effective and how has the industry really responded to that?

是以，營運設施的長期成本需要與補貼相平衡。我認為，所有相關公司都在努力尋找成本動態與確定客戶是否願意為供應鏈多樣化支付更多費用之間的平衡點。我們看到了美國的補貼，但與中國相比，這些補貼顯得微不足道。僅僅在今年，中國政府就加倍投入，撥出了超過 470 億美元的資金來推動芯片產業的發展。我們已經看到很多關於出口管制和關稅的討論，它們是遏制中國技術崛起的主要政策工具。但在你看來，這些政策是否有效？

Well, I think to answer that question, we've got to differentiate between the most cutting edge semiconductors and the more mainstream foundational semiconductors. For the second category, China's got all the technology that it needs to produce these types of chips. And so export controls, limits on technology transfer to China really won't work in stopping China's progress. And that's where we've seen the bulk of China's money going is to building out factories producing mainstream chips. And I think the challenge here is just that China might flood the market in terms of certain segments of the industry, and that it's going to produce chips at very low prices because Chinese firms don't need to make money. And that's going to take market share from Western firms that have traditionally sold to customers in China.

我認為，要回答這個問題，我們必須區分最尖端的半導體和更主流的基礎半導體。對於第二類產品，中國擁有生產這類芯片所需的所有技術。是以，出口管制、限制向中國轉讓技術並不能阻止中國的進步。我們已經看到，中國的大部分資金都用於建立生產主流芯片的工廠。我認為，這裡的挑戰只是中國可能會在某些細分行業充斥市場，而且會以非常低的價格生產芯片，因為中國企業不需要賺錢。這將搶走傳統上向中國客戶銷售的西方公司的市場份額。

So this is going to have, I think, significant impacts on the industry, especially market share in China. And that's where the US government is trying to impose some tariffs on Chinese chips coming into the US to make sure that Western firms face some protection from the really massive subsidies that as you say, China's imposing. When it comes to the most advanced chips, that's where

是以，我認為這將對整個行業產生重大影響，尤其是在中國的市場份額。是以，美國政府正試圖對進入美國的中國芯片徵收關稅，以確保西方企業能得到一定的保護，免受中國鉅額補貼的影響。說到最先進的芯片，那就是

China doesn't have the technology right now. And in order to produce the most advanced chips, you need machines from the US, from Japan, and from the Netherlands to actually manufacture the chips. And those machines are currently illegal to sell into China. And that's where China's really facing challenges. And if you look at the production capabilities in China, they're about five years behind the production capabilities in Taiwan, which is the world's most advanced chip maker. And the US wants to keep it that way, wants to keep that gap so that US firms have an edge over their Chinese competitors. Now, you mentioned the foundational chips that are most likely going to be flooding the market. Do you think the industry has a plan to respond to that?

中國現在還不具備這樣的技術。為了生產最先進的芯片，你需要來自美國、日本和荷蘭的機器來實際生產芯片。而這些機器目前在中國是非法銷售的。這就是中國真正面臨挑戰的地方。中國臺灣是世界上最先進的芯片製造商，如果你看看中國臺灣的生產能力，他們比中國臺灣落後大約五年。美國希望保持這種狀態，希望保持這種差距，這樣美國公司就能在與中國競爭對手的競爭中佔據優勢。現在，你提到了最有可能充斥市場的基礎芯片。你認為業界有應對計劃嗎？

And if so, what do you think are some of the potential responses to protect that market share?

如果是這樣，您認為保護市場份額的一些潛在對策是什麼？

Well, I think there are really only certain categories where that's likely. When it comes to higher end chips, more sophisticated technologies, chips that need to be closely integrated with software, these are places where Western firms probably have a pretty good chance at fending off Chinese competitors because it's not solely about cost. It's really only at the bottom of the market and super commoditized products where cost really matters. And China's going to produce at lower cost, even if it's not more efficient, it's just the subsidies will enable lower prices. And firms that are competing in that segment need to worry, need to think hard about what their strategy is going to be. And I think more technological differentiation is in many cases going to be the answer. If you've got a technological edge, you don't have to compete on cost. And I think that's what Western firms are increasingly focusing on. I want to look a little bit at the supply chain because the tensions between the US and China are also reshaping how companies are approaching that aspect of their business.

我認為只有某些類別的產品有可能出現這種情況。當涉及到更高端的芯片、更復雜的技術、需要與軟件緊密集成的芯片時，在這些領域，西方公司可能有相當大的機會擊退中國競爭對手，因為這不僅僅是成本的問題。實際上，只有在市場最底層和超級商品化的產品中，成本才真正重要。而中國將以更低的成本進行生產，即使其效率並不高，但補貼將使價格更低。在這一領域競爭的企業需要擔心，需要認真思考他們的戰略是什麼。我認為，在許多情況下，更多的技術差異化將是解決之道。如果你有技術優勢，你就不必在成本上競爭。我認為這正是西方企業日益關注的問題。我

And we've seen companies like Apple maybe considering decoupling their supply chains.

我們看到，蘋果等公司可能正在考慮將其供應鏈脫鉤。

Is that how the industry is responding right now? Or is that a trend we could see more of in the future? I think the way I'd describe the trend is not decoupling, but is rather creating a China business for the Chinese market and a non-China business for the non-China market. The reality is that China is the second largest economy in the world and

這就是業界目前的反應嗎？或者說，這是我們未來可能看到的一種趨勢？我認為我所描述的趨勢不是脫鉤，而是為中國市場創建中國業務，為非中國市場創建非中國業務。現實情況是，中國是世界第二大經濟體，而且

Western firms are going to want to keep selling into China to the extent that they're allowed to do so and able to compete with domestic Chinese firms. It's just too large to ignore. But I think many companies are making sure that they've got a fairly separate supply chain for the non-Chinese market, at least separate in terms of critical components, and that they've got enough resilience that they can produce goods outside of China if they needed to. And so you mentioned

西方公司會希望繼續向中國銷售產品，只要他們被允許這樣做，並且能夠與中國國內公司競爭。中國市場太大了，不容忽視。但我認為，許多公司都在確保他們為非中國市場建立了相當獨立的供應鏈，至少在關鍵部件方面是獨立的，而且他們有足夠的應變能力，可以在必要時在中國以外生產產品。所以你提到

Apple. I think it's really striking the extent to which Apple used to produce almost every iPhone inside of China. And now it's really rapidly ramping up its manufacturing operations in India.

蘋果我認為，蘋果公司過去幾乎每部 iPhone 都在中國生產，這確實令人震驚。而現在，它正在印度迅速擴大生產規模。

And I think most people underestimate the amount of investment and amount of progress that

我認為大多數人都低估了投資的數量和取得的進展。

Apple has made. And this is about diversification. It's not that Apple's going to stop selling iPhones in China or stop making iPhones in China, but they want options. They want more flexibility in case of future regulatory risks, either from the Chinese side or from the U.S. side.

蘋果已經做到了。這就是多元化。這並不是說蘋果公司將停止在中國銷售 iPhone 或停止在中國生產 iPhone，而是他們需要選擇。他們希望有更多的靈活性，以應對未來來自中國或美國方面的監管風險。

You mentioned diversification. So is that how this is affecting investment patterns then for these big tech players?

你提到了多元化。那麼，這對這些大型科技企業的投資模式是否有影響？

Yeah, I think you certainly see that in the data. For example, the collapse of foreign direct investment in China has many causes, but one of them is that technology firms are investing more elsewhere in India, in Vietnam, elsewhere in Southeast Asia, because they want that diversification.

是的，我想你肯定能從數據中看到這一點。例如，中國外商直接投資的崩潰有很多原因，但其中一個原因是，科技公司正在印度、越南和東南亞其他地方進行更多投資，因為他們希望投資多樣化。

And they realize that although Southeast Asian India can't today replicate just the depth of the

他們意識到，儘管東南亞印度今天還無法複製《世界人權宣言》的深度

Chinese ecosystem when it comes to electronics assembly and basic manufacturing, there's really substantial buildup happening in India and in Southeast Asia. Labor costs are in some case actually lower than in China, and diversification really matters. And so that's driving a trend of investment dollars flowing into these countries where in the past they would have flowed into China.

說到電子組裝和基礎製造，中國的生態系統在印度和東南亞確實有大量的積累。在某些情況下，印度和東南亞的勞動力成本實際上比中國還低，而且多樣化也非常重要。是以，這推動了投資資金流向這些國家的趨勢，而在過去，這些資金會流向中國。

And I want to return to the U.S. because Intel is such a critical domestic producer of these chips.

我想回到美國，因為英特爾是這些芯片的重要國內生產商。

And we've seen sort of the turn or at least attempted turnaround story. And I thought you did such a great job in your book kind of talking about the history of that company and how this isn't really the first time they've had to pivot. So I just wanted to get your thoughts on Intel and some of the changes that are happening with that company.

我們已經看到了這個轉折或者至少是試圖轉折的故事。我認為你在書中做得很好，講述了該公司的歷史，以及這並不是他們第一次不得不轉折。所以我想聽聽你對英特爾以及該公司正在發生的一些變化的看法。

Well, I think Intel is trying to deal with three simultaneous challenges. Challenge one is to get their manufacturing capabilities back at the cutting edge on par with Taiwan's TSMC. And there it looks like they're having some real progress relative to where they were a couple of years ago.

我認為，英特爾正試圖同時應對三個挑戰。挑戰一是讓他們的製造能力回到最前沿，與臺灣的臺積電看齊。在這方面，與幾年前相比，他們似乎取得了一些真正的進展。

The second challenge is that one of their key markets is data centers. And data center dynamics have changed dramatically over the past couple of years because of AI. It's benefited firms like

第二個挑戰是，他們的主要市場之一是數據中心。由於人工智能的出現，數據中心的動態在過去幾年發生了巨大變化。這讓以下公司受益匪淺

Nvidia, which produce slightly different types of chips than Intel. And that's been a challenge for

英特爾和英偉達生產的芯片類型略有不同。這對

Intel because it's found its market share in data centers declining because it doesn't have the right chips focused on that AI. And then the third challenge is that Intel's trying to build up its manufacturing scale because you can't manufacture efficiently in the chip industry without scale.

英特爾發現自己在數據中心的市場份額在下降，因為它沒有合適的芯片專注於人工智能。第三個挑戰是，英特爾正在努力擴大生產規模，因為在芯片行業，沒有規模就無法高效生產。

And to do so it's opening a business to manufacture chips for external customers.

為此，它正在開設一家企業，為外部客戶生產芯片。

And that's a very expensive business to start up because you need to spend the money building the manufacturing capacity today and get the customers hopefully down the road. And I think investors are looking at this latter part of the business, understanding the huge cost that will be involved in the capital expenditure next couple of years, and not yet seeing the revenue and getting nervous about the financial implications of that. And so these are three challenging problems to deal with simultaneous. They're problems that have been created over the last decade. And so I think Intel's management team is right when they say they need some time to turn such a big ship around, especially because the time cycles in this industry are so long. It takes multiple years to build a single factory. And so I don't think you can measure the success of a turnaround in just a couple of quarters. You really need multiple years to ascertain whether, for example, Intel's new foundry business to manufacture chips for external customers, whether that's going to actually work. And so time is what Intel's management is asking for.

這種業務的啟動成本非常高，因為你需要花錢在今天建立生產能力，並希望在未來獲得客戶。我認為，投資者正在關注業務的後半段，瞭解未來幾年資本支出所涉及的鉅額成本，但尚未看到收入，是以對財務影響感到緊張。是以，要同時解決這三個具有挑戰性的問題。這些問題都是在過去十年中產生的。是以，我認為英特爾的管理團隊說他們需要一些時間來扭轉這樣一艘大船的局面是正確的，尤其是因為這個行業的時間週期是如此之長。建立一家工廠需要多年時間。是以，我認為不能僅用幾個季度來衡量扭虧是否成功。比如，英特爾為外部客戶生產芯片的新代工業務，你真

And I think the market is trying to assess how much time they're willing to give them.

我認為市場正試圖評估他們願意給他們多少時間。

Chris, thanks so much for your insights today.

克里斯，非常感謝你今天的見解。

Thank you again for having me.

再次感謝您的邀請。

Thank you.

謝謝。