Endgame with Gita Wirjawan - Chris Miller - US-China Chip War: What Could Go Wrong?

Episode Date: May 13, 2024

Join Endgame YouTube Channel Membership! Support us and get early access to our videos + more perks in return: https://sgpp.me/becomemember ---------------------- Click here to get the “Chip War” book at Periplus bookstore ---------------------- Chris Miller and Gita Wirjawan discuss the technological competition between the US and China, with a focus on the chip industry. The conversation highlights Europe's approach to tech, the impact of the US uniting its allies for military and technological containment of China, and the potential risks of China's response. Finally, they also touch on the importance of creating global stability and security in advancing chip technology as the backbone of modern society. About the Guest: Professor Christopher Miller is an associate professor at the Fletcher School at Tufts University and a nonresident senior fellow at the American Enterprise Institute. He is the author the bestselling book, "Chip War: The Fight for the World’s Most Critical Technology” (2022) and serves on the Geopolitics Advisory Council at McKinsey & Company. About the Host: Gita Wirjawan is an Indonesian entrepreneur, educator, and Honorary Professor of Politics and International Relations at the School of Politics and International Relations, University of Nottingham. He is also a visiting scholar at The Shorenstein Asia-Pacific Research Center (APARC) at Stanford University (2022—2024) and a fellow at Harvard Kennedy School's Belfer Center for Science and International Affairs. #Endgame #GitaWirjawan #ChrisMiller ---------------------- Thank you to The Belfer Center for Science and International Affairs at Harvard Kennedy School for providing support for this episode. Visit the link below to know more about research, ideas, and leadership programs for a more peaceful world: https://www.belfercenter.org/ ------------------ Earn a Master of Public Policy degree and be Indonesia's future narrator. More info: [e-mail] | https://admissions.sgpp.ac.id | https://wa.me/628111522504 Visit and subscribe:  @SGPPIndonesia  |  @Endgame_Clips 

When I look at the technological competition today, it seems to me that there are far more factors that are driving it in an unstable direction, a stable direction, and the most important factors that nobody knows the direction in which technological development will take. Chris Miller. Chris Miller. Associate Professor of International History at the Fletcher School of Law and Diplomacy at Tufts University. Professor Chris Miller, author of the Chippewar. Just to give you a sense of how timely Chris Miller's books are, I'll give you the titles of the others. Putinomics, power and money and resurgent Russia, we shall be master and the struggle to save the Soviet economy Mikhail Gorbachev and the collapse of the USSR.

The risk on the Chinese side is that China overestimates China's power. The risk on the side of the United States, as well as other countries in the region, is that they underestimate U.S. staying power. What could go wrong? The strongest earthquake in 25 years hitting Taiwan today, leaving at least nine debt and hundred It's injured. The U.S. share of chip manufacturing has plummeted in recent decades. It costs at least 20% more to build and operate a new fab in the U.S. than in Asia.

Today, Asmel has a monopoly on the fabrication of UV lithography machines, the most advanced type of lithography equipment that's needed to make every single advanced processor chip that we use today. I don't think it's about can China do it. I think it's about can anyone do it. The most interesting country perhaps is India right now. To the extent that there is a disruption, long enough, it's pretty apocalyptic, right? The downside risk is just so large.

Hi, friends, I want to take this opportunity to thank you for being with us ever since we started Endgame some years ago. The conversations have been invariably elevating and animating. At least from my personal point of view, it's been a tremendously rewarding experience. And I'm hopeful that you could be further supportive of us by way of clicking on the subscribe button, watching every episode as much as possible, if not as fully as possible, and also joining us as a member of the Endgame channel. I can only promise you that whatever we're going to be doing going forward will try to make it.

endgame a better experience for all of you. Thank you. Hi friends. Thank you so much. Today we're honored to have Professor Chris Smiller, who is a professor of international history at Tufts University, but he's also a fellow at the Kennedy School at Harvard University. Chris, thank you so much. Thank you for having me. Why don't you tell us a little bit about your upbringing, where you grew up and how you became so good at studying history and became a historian? Well, I, I guess, I grew up outside of Chicago and it was an undergraduate at Harvard and studied history. And I just got fascinated in understanding how the world came to be in the way we know it today. And initially got interested in Russian history, which is what I did my PhD on.

I wrote several books on different aspects of Russian economic, technological history, trying to understand how Russia had developed and the way that it did in a fairly unique path over the course of the last couple of centuries. But over the past decade or so, I came to realize. was that technology was a really important driver, not only of Russian history, but of really all world history. And one of the key areas of technology that I concluded was understudied was semiconductors.

And so that's how I came to spend most of the last half decade focused on the history of the chip industry. How long did it take you to write this book? Well, to research the book took a long time. There was a whole lot of research, both in the history of the industry, but also in how it interlinked with geopolitical and macroeconomic and trade trends. So it took many years to actually do all the background research.

The writing was relatively fast, but I started writing once I more or less knew what I wanted to say. Safe to say a few years? A couple years, yeah. Okay. You wrote about a bunch of things in a book, right? And you alluded to the fact that, you know, China has been buying a lot more chips than oil. and talk about that a little bit. Well, to me, that was one of the puzzles that inspired me to write the book.

I'd come to think about global trade as being in no small part about big oil tankers, leaving the Middle East and traversing around the world to deliver oil to the customers that needed it. But if you look at big flows in world trade, the biggest flow is the flow of chips into China. And it's been that way every year for the past decade. And this seemed like something that wasn't in my mental map of what the international trading system looks like and wasn't in most people's. And it was a puzzle because China is the world's manufacturing superpower in almost every type of good. But it imports most of the chips that it needed. And one of the reasons why I wrote chip boards, I want to understand, well, why was that?

Why wasn't that China couldn't produce most of the chips that not only its manufacturing base, but the entire world's manufacturing base relies on? Why has China not caught up with whatever some of the other guys would have done well in, particularly Taiwan. Well, I think there's two reasons why. One is just that the chip industry has taken decades to build and tens of billions of dollars a year in investment. So catching up is an extraordinarily expensive proposition,

even for the world's largest economy. So it's partly a question simply of dollar values, but it's also a question of technological complexity. There's really no industry that is as precise in its manufacturing and as complex in its supply chain as making advanced chips. And so it's the industry that's the hardest of any industry

to reach the cutting edge because the complexity and because the technological frontier is moving forward every single year. So you're not catching up to a stagnant frontier. You're trying to catch up to an industry that's racing forward as rapidly as any other industry

in all of human history. You talked about this in the book. You share with us a little bit about why the Intel is of the world, the Sonys of the world, that would have been way at the front, you know, way earlier than, you know, a TSM would have been, why did they not, they just seem to have lost the plot, right? And I think if it's safe to assume that they probably would have detected, right, what they did wrong.

Why has it taken so long for them to, you know, try to catch up with a TSMC? Well, I think in the chip industry, what you find is that technological, shifts drive business model shifts. And it's hard for companies that made their business work in one paradigm to shift their business model to a new paradigm. And if you look at both Sony in the 80s and 90s and Intel more recently, they found that as business models were changing, it was hard for them to disrupt their existing business and pivot towards something new.

And TSM was founded not intending to be the world's most important ship company. It was founded to put Taiwan on the map at a time when Taiwan was not really on the map at all. And because of that, it had a sort of insurgent business model. It didn't design any chips. It produced chips only for other companies. And at the time, most people thought it was a crazy idea. But it turned out to be a brilliant idea.

And it put TSMC in a position of having the best case study of this business model of any company in the world, which is why in the business of creating chips for their customers, TSM has over half the world's market share today. When it comes to the most advanced shifts, it produces almost all of them. And that's because of this unique business model that his competitors couldn't keep up with. You think Minister Katie Lee, the fact that he was able to convince Morris Chang, would that have been serendipitous?

Or it was just the fact that, you know, he wasn't able to, he wasn't going to go back to Texas Instruments because he wasn't able to convince the people at Texas Instruments. You know, I think a lot of historical turning points like this are a mix of serendipity, but also of structure. I think KT. Lee, as well as several of his peers in the Taiwanese leadership, were trained as engineers. They followed technological advances very closely. They realized that the electronics industry was going to be a driver of growth in East

and Southeast Asia for many decades to come. they realized that they could use this industry to plug into U.S. and Japanese supply chains. And so there was a lot of structural factors that were pushing them towards realizing that a decision like this could be pivotal. But I think it was serendipity that Morris Chang ended up out of a job just at a time when people like KT Lee were looking to make some big bets on the future of Taiwan's technology. And so I think it really was a mix of structure and serendipity to put this decision in the hands of these two gentlemen at the exact time. You know, a few months ago, there was this historic meeting

between President Xi and President Biden in San Francisco. It was coined as the San Francisco consensus on the American side. On the Chinese side, it was coined as the San Francisco vision. And it seems to imply that, you know, I think they're prepared to be competitive with each other, but they're also prepared to be coexisting, right? And I want to just take you through what Professor Graham Allison has always alluded to, which is to sit at a strap, right?

Picture or draw a picture where you think China and the U.S. technologically will be able to coexist in a harmonious manner. Or perhaps draw a picture where you think technologically China and the U.S. are not going to be able to coexist harmoniously from a technological standpoint. Well, I think the APEC meeting that you're referencing is a great case study for analyzing this question, because in addition to Xi and Biden, one of their attendees was Morris Chang, who is the Taiwanese representative to APEC, which I think is an extraordinary piece of evidence about the central role of Taiwan in this question, central because Taiwan produces all of these ultra-critical.

conductors, but also central because Taiwan is this perhaps central flashpoint between the United States and China today. It's the place where it would be most predictable for things to go very wrong. And I think that illustrates the complexity. It's easy for political leaders to say they want a sort of stable, competitive relationship. And I think both sides sincerely do want that. But achieving that is something that's very different than wanting it. And it's difficult to achieve both because.

the US and China aren't the only players, but also because the playing field is multidimensional. It's a geopolitical playing field. It's an economic playing field. It's a technological playing field. And all of these intersects in very unpredictable and hard to control ways. And so when I look at the technological competition today, it seems to me that there are far more factors that are driving it in an unstable direction than a stable direction. And the most important factor is that nobody knows the direction in which technological development will take. And so how can political leaders predict the direction of the diplomatic relationship when they're competing in a context in which they have no idea what innovation is just around the corner. I want to pick up on this.

I mean, we've kind of talked about this earlier in the sense that at every level of the supply chain with respect to chip manufacturing, chip design and all that, there's just too many countries, too many players that are involved, right? It just seems to me that it would be stupid to do anything that would unshackle that, right? And in a way that serves as a deterrence for or would respect to anything that's not very good for a lot of people. Is that something that could sustain for a long time? I think that dynamic is certainly present.

If you look at U.S. policymakers and Chinese policymakers, there is a very real sense of the economic risks that would stem from disrupting supply chain. But I also think we've seen both sides over the past five, seven years take a lot of steps that have disrupted supply chains, not in dramatic ways, but in ways that have really shifted investment flows and trade flows over a sustained period of time and created substantial costs for businesses and to some extent for consumers. And so I don't think we should overestimate the extent to which political leaders are indexing on economic efficiency or GDP per capita growth. It would be great if they all did. But I don't think that's how politics works. Is there a greater risk to overestimating than underestimating each other between China and the U.S.? To me, I think the risks are different for each side. I think the risk on the Chinese side is that China overestimates China's power, overestimates

China's influence in the region. I think the risk on the side of the United States as well as other countries in the region is that they underestimate U.S. staying power. the U.S. has all of its flaws, all of its debates out in the open. And so they're very easy to see. With China, it's the opposite. It's all covered up. And so you never know which flaws are real, which ones are hidden. I don't even think China's leaders have a full assessment of the challenges that they face. And I think that's also a source of uncertainty as well, that neither side is really sure what the balance of power actually looks like because there's so much uncertainty.

the chips act 2002 how is that going to translate into the ability of the u.s to actually insource you know the chip manufacturing capabilities from elsewhere like Taiwan well we we know right now that there are currently around 350 or 400 billion dollars of investment projects being built in the United States across the chip industry both manufacturing but also the materials the the broader supply chain, which is a huge increase in investment. It's roughly 15 times the rate of investment in the chip industry from the previous decades average, so a massive increase in investment. And that will create more chip making in the United States relative to the prior trend.

And we're seeing similar programs in Japan and in Europe as well that will have effects in those countries too. However, the amount of investment that's been made in Taiwan over the past three decades is extraordinary. and the position of TSM, not just its prior investments, but also the investments making now that will be online in a couple of years, and it will be cutting edge in a couple of years, are also hugely significant.

So on one hand, the Chips Act is giving the U.S. more margin for flexibility in case of the crisis. On the other hand, Taiwan is going to remain absolutely at the center of the chip industry very far into the future. Both those dynamics are true. And I think they interact with a third. third kinnamic, which is that China is pursuing its own version of a chips act. And I think China is spending significantly more in the United States. Will the US have enough human capital for the kind of precision manufacturing that's needed for the latest type of chips?

You know, in the chip industry, the way you develop human capital is by training employees in a fab. Okay. There's only so much you can learn in university. You can get your PhD in material science or in physics, but unless you've done the manufacturing yourself, you don't really understand how to do it. And so the reason why Taiwan is the best place in the world when it comes to human capital, it's because they've got the largest number of fabs. It's as simple as that. And that's a challenge for the U.S., for Japan, for anyone trying to build up, because you don't have the talent until you build the talent. And it means you do face bottlenecks when you're trying to ramp up production relative to your prior trend.

So what's going to happen in Arizona? Well, I think you've seen the companies... You see probably a lot more Taiwanese first than gradually more and more non-Taiwanese working in the FAPS. That's already what you see. In the TSM plant in Arizona, there's a very large number of Taiwanese who are working there. I think some temporarily, some for a longer period of time.

And it will take, I think, years for the talent pipelines to become fully developed. Would this then be reflected in... the continuation of the United States open-mindedness from an immigration standpoint? Well, that's a key challenge to the United States is making sure that it remains open to skilled immigrants in this sector as well as other sectors. If you look historically, what you find is that many of the key innovations in the ship industry were made by immigrants. The founders of Fairchild semi-conductor. Including Morris Chang.

More Shane would be another good example. The founders of Fairchild, 25% of them were non-native-born Americans. If you go down the list of key innovations over the last 50 years, there are a series of key innovations that were made by Korean Americans or Egyptian Americans or Chinese Americans. Morris Chang is the way the outlier because he's both an example of immigration, United States, but he's also a fairly rare example of immigration from the United States. And so he's the opposite side of the story of an immigrant who he failed to keep. wish we had. You know, in this kind of what, you know, to sit at a strap, what could go wrong?

And how is this going to pan out as one of the four that turns out peacefully or one of the 12 that turns out, you know, non-peacefully? What would it take for this to go wrong? Well, I think when it comes to the chip industry itself, any definition of going wrong in the geopolitical sphere, a war. blockade, a temporary disruption. They're all hugely costly for the chip industry. Taiwan imports most of its energy. It needs regular shipments of LNG to keep the lights on. And the chip industry is one of the largest users of electricity on the island. So even if the only thing that was disrupted was LNG tankers coming into Taiwan, that's enough. That's enough. And it means that

any geopolitical crisis would have immediate impact on the chip industry, therefore immediate impact on all of the rest of global manufacturing, which is why the magnitude of risk is just so high. How long would the silicon shield, you know, be around until such time the Chinese could figure out to not only compete, but out-compete the Americans? Well, I think when it comes to foundational chips, the lower technology chips that are in cars and in machinery and in many types of consumer goods, China's aren't. out the technology that it needs. And it's right now building the manufacturing capacity that it needs as well.

And so I think by the end of the decade, China is highly likely to have the capacity to produce most of the chips it needs at the foundational level for its own manufacturing base. The uncertainty comes with regard to high-end chips, the types of processors that are in AI systems or in smartphones. And there, China is building some fairly advanced production, but it faces challenges because the West won't ship it some of the system. of the manufacturing tools that it needs and its own base for producing these tools right now is very weak and so that is really the key question is can china indigenize the production of chip

making tools fast enough and it's going to be a very difficult challenge but you know a lot of people tend to underestimate the chinese oftentimes right if if you think it's going to take another six or seven years is it possible for them to actually be able to do it sooner Well, nothing's impossible. I mean, who would have thought that they're going to be making beautiful gadgets like most humanity you are using, right? Who would have thought that they're going to be making beautiful EVs and all that faster, sooner than a lot of people would have thought? What's the thinking there? Well, I think it's easy to try to answer this question and ask about China's capability.

Right. I would step back and ask about anyone's capability, anyone's capability, whether Japan's or Americas or Europe's. When we look at chipmaking tools, we're talking about the most complex tools humans have ever made, with the most precise manufacturing capabilities, with supply chains that involve hundreds of thousands of parts, across a whole array of different precision engineering sectors from optics to mechatronics. And so if you ask me, pick any company or pick any country, could they replicate what ASML has done when it comes to lithography tools overnight, I would say, no way. And so I don't think it's, I don't think it's about can China do it. It's about can anyone do it? And over some time horizon,

the answer is yes, but it's not going to be next year. I don't think it's maybe a year after that. And that's why I think it's going to be some number of years in the future. Now, it's already been six years since it's been illegal under Dutch law to transfer extreme ultraviolet lithography tools to China. And right now, there's not really any evidence that China has made progress in catching up. You could argue that they're doing it secretly. That's unprovable. But I think it's, as a result, very likely that China will at least take a decade to replicate these tools and perhaps longer. The other part of the question is, well, ASML is not standing still, nor is the rest of the chip industry. And we're going to see continued advances in the quality

of chip making tools. I think if China catches up, it's not going to be because they have successfully replicated what ASML can do, for example. I think it'll be because they've found alternative pathways to produce the same computing capabilities. And here it's hard to know what those pathways are. If we knew what they were, we would already be using them. And so there's real uncertainty, core uncertainty about whether alternative pathways exist, but there's no doubt that Chinese engineers are trying to find them

because right now they have no other option. Some of the stakeholders of ASML include the Samsung's of the world, the Intel's of the world, the TSMCs of the world, right? right. Is that some sort of an insurance that, you know, things are going to favor the U.S. from an ASML standpoint? Well, I think so. I think the fact that the chip industry is, although a lot of the manufacturing happens in Taiwan, the key customers of TSM are basically all U.S. firms, Apple, Nvidia, Qualcomm, AMD. The chip industry is still very U.S.-centric from that perspective. And also, if you look at ASML,

There are three key locations for manufacturing and software development are the Netherlands, Germany, and the United States. And so it's a Dutch company, but it's a Dutch company that operates in the transatlantic context and it, as a result, really relies on access to the US market, US components, and US design software for its machines to function. What's your take on Morse law? Is it going to get more exponential or is it just going to be as we have seen in the last few decades?

Well, when Gordon Moore first set out Morse law, law, he had two different definitions. One was the number of components per chip, which was roughly a proxy of computing power per chip. And the second was the average cost per transistor. And he predicted both that we'd have exponential growth in the computing power per chip, which we've had now for over half a century. And he predicted that the cost of computing would fall roughly comparably, and it has. Which is why you can buy a thumb drive with 10 billion transistors for $10 today. Extraordinary. I think at some point we're going to reach physical limits as to how small we can make transistors.

We're going to get there at some point. They're already, instead of your phone, the transistors are already half the size of a coronavirus. So at some point, atomic level limits will really cause fundamental dilemmas. But at that point, I think we're going to have new techniques for producing just as effective chips in different ways, better packaging techniques, better design techniques, new structures for transistors. And so I'm very optimistic about the rate of technological innovation continuing, but it's going to at some point have to take a different form once it becomes simply physically impossible to shrink transistors any smaller.

Right. You know, if you take a look at the valuations of those involved in chip design versus those that are involved in chip manufacturing, such a disproportionate divergence. Why, why, Why is that? I mean, it sounds like from the last few minutes of our discussion, it sounds like the real complication seems to be in the manufacturing site, right? Because you need that precision manufacturing capability. You just need people that would have been working in the FAPS for enough time so that they can actually do it properly. Why is that divergence of valuations?

I think there are a couple of reasons. is that the capital intensivity of chip making is extraordinary. TSM needs to spend $20 billion for each FAB, and it needs to do that on an almost annual basis. Whereas for chip design companies, the input cost is basically just your employee salaries. And so that makes it a lot more attractive to invest in chip design companies. The second thing is that what we found historically is that the chip design companies can charge a much higher margin than the manufacturers.

You were saying you had an 85% margin. For Nvidia, that's right. That's right. And so if you look at the price, Nvidia charges TSM for the waferes that are manufactured versus the price it charges its customers for the final chips, Nvidia has a much better business model, which is why its valuation is so much higher. Seems a little unfair.

Well, you know, it's interesting. In some ways, the U.S. is worried about its position in the chip industry. ways the decision to focus on design has from a pure financial markets perspective economics perspective been absolutely the right decision how will this affect talent development going forward to the extent that there's more money in the designing space as opposed to the manufacturing space I would think people that you know students at the universities or those that are thinking about going to universities they probably want to be doing designing more than manufacturing you know I think it's

It's absolutely right. That's what we see. And it's a bit of a challenge for the industry. In the past, 30 years ago, anyone involved in ship design also understood a fair amount of the manufacturing process because you were generally working at a company that did both the design and the manufacturing. And so you had to have this holistic sense. Today for most ship designers, you don't need to know anything about ship manufacturing.

And indeed, the specialization is such that you end up not knowing much at all. And that's okay. For individual companies, I think that's fine. But for the industry as a whole, what it's meant is that if you look at young graduates, they're focused on computer science, which is applicable to chip design, as opposed to electrical engineering or material science. And there's very good reasons for that when you look at the starting salaries or the career trajectories. But it does mean that the U.S. has focused on one segment of the industry at the expense of others. And I think that has been a challenge that the U.S. faces.

Where do you think the talent is going to come from in the next 10 to 20 years? Well, I think that the talent... I mean, if I take a look at the STEM students in at the universities in the US, right? I mean, there doesn't seem to be an increase amongst the US students. Seems to be a pretty consistent increase amongst the international students, right? Is that a presumption? I think that's right. You certainly see that in the data.

My sense is that in the long run, talent will flow where it's rewarded most highly. And in this STEM field, we've seen that be software over the past several decades. But beyond that, in the software space, what we've found is that the people who do the best are not necessarily the most brilliant computer scientists, the people that can marry those skills with business know-how.

And so when I, think about talent I'm not only thinking about who can graduate with a bachelor's in electrical engineering I'm thinking about who can found the next startup and that requires more than just electrical engineering knowledge and so I think we do have to think about the pool of some talent but if you've got only stem talent that's not enough and I think that's really important to think about when we have conversations about creating the talent pool of the industry actually needs where do you see the next as the next logical place or

destination in terms of reshoring of manufacturing capabilities beyond Taiwan, other than the U.S.? Well, I think you see most major economies trying to do some version of their own reshoring. Right. Japan's doing it. Europe's doing it. For certain types of manufacturing, you see it in places like Mexico. I think the most interesting country perhaps is India right now, which has a vast domestic market, an extraordinary talent pool, and a government that's very focused right now on semiconductors on AI and on high value manufacturing more broadly, and is trying to benefit both from the AI boom and from China plus one dynamics simultaneously. And so I think it's really worth watching India's efforts very closely. Is that partly geopolitical to you? You know,

it's no doubt partly geopolitical. And I think when Indians speak to the outside, world, especially to the West, they have every incentive to describe it in geopolitical terms. But it's also about India moving with the value chain, but India building its own industry. And so if you listen to Prime Minister Modi talking to domestic audiences, he's not talking about China. He's talking about made in India for India. Right. Well, they produce, what, 10 to 15 billion worth of iPhones last year. I know they're talking to Tesla. They're talking to, but do you see India playing a more proactive role in designing as opposed to manufacturing when it comes to chips? Well, right now India has a huge pool of people with chip design expertise because basically

every chip design company has an office in India, Qualcomm, Nvidia, they all do. And that leaves India an interesting position because it's on the one hand trying to move up the manufacturing ladder, but it's starting in a position of already having this large pool of designers. And I think be interesting to see how this plays out. If you look historically at Taiwan or at Korea, when they started moving up, they didn't have the design expertise. Their only choice was to get good at manufacturing. India is trying to do that alongside the design expertise, and I think that gives it an advantage that other comparative countries didn't have. Where do you see Southeast Asia in all this? Well, I think there's... What could go wrong, what could go right for South Asia?

Well, in some ways, a lot already has gone right in that Southeast Asia is a, a a big player in the chip industry, especially in the assembly test and package space. Malaysia in particular, but also the Philippines, Vietnam, Thailand to some extent, Indonesia to some extent, Singapore. A little bit Malaysia a lot more, I think, yeah. Yeah. I think China plus one dynamics are something that are naturally causing international investors to look at Southeast Asia. I think the challenge for Southeast Asia is to not just attract China

plus one money, but attract China plus one money that's looking to move up the value chain. Because assembly and packaging facilities are somewhat labor-intensive. They're not super value additive. And so if you're looking at high-value manufacturing, you need to think what's next on the technological trajectory. And if it's just simple packaging, that's really just the first wrong on the ladder. And I think every country in Southeast Asia has to think more carefully than ever about how to keep moving up in a context in which it's not just other countries in Southeast Asia. the entire world right now that's trying to move up the value chain. So would you suggest that those in Southeast Asia that are going to be able to show more

open-mindedness to the U.S., they're likely to be beneficiary of this chip contest? I think that's right. That's right. I think in addition, it's not just about trying to win one plant or two plants. It's about how do you build the ecosystem? And the ecosystem is about... The human capital. The human capital.

The human capital, having chip design alongside chip assembly and manufacturing. Thinking about designing entire systems, not just designing chips themselves. That's how you have an ecosystem that grows over time. And I think it's easy for political leaders and bureaucrats to think, what's the incentive I need to win this specific investment? But you need the broader ecosystem to create an environment in which you're not just attracting foreign firms, but you're actually building your own firms. So that's ultimately what you want to be doing.

Well, there's that rule of law also that needs to be, you know, upheld in not just Southeast Asia, but many developing economies that want to be the destination of this reshoring activity, right? So what you're suggesting is that the reshoring rhetoric is not unreal. It's quite real, then. I think it is. Okay. It's reshoring, it's real, but it's also not self-sufficient. for any given country, which is why we've seen a surge of U.S. investment into India and

Korean investment in the United States and U.S. investment into Israel. This is not a reversal of globalization. In fact, we see larger FDI flows in the chip industry now than I think really ever before. You know, if you take a look at the VC investments into AI, 2021, around 37, 38 billion, dollars would have been made in in the US 25, 26 billion in Asia. Europe, only 8 billion. Why? I mean, you know, for a layman like me, it just seems like Europe seems to be continuously falling behind in this digital revolution.

Is it something that's by design or something else? Well, I do think it's partly by design, not deliberate design, but by policies that are designed in a way to not focus on innovation, but focus on regulation. When I talk to my friends in Europe, they're often proud that they've created the first AI regulation framework. I think that's fine, but I think it's better to have created the first AI companies and then worry about regulating them as a step too. And I do think Europe's approach to tech issues in general is more about control. rather than about fostering the next big technological advance. That is where the policy design, I think, does really matter. I think it's partly the structural,

which is that if you're a firm looking to raise capital or looking to scale, the natural place to do that is not in Europe and the United States. We've got a larger market and vastly larger capital markets that are able to support new firms taking risky bets. And so if you've got a good idea, California is the place to do it, not in Europe. But it doesn't mean that you can't have that in Hamburg or Milan, right?

I mean, why is there not a competitor to Google from Europe, competitor to Twitter in Europe, competitor to meta in Europe? And if you take a look at the GDPR framework, it just seems to indicate that, you know, they can afford to be very strict because they have not been a beneficiary of the digital revolution. I think that's fair. I think it does, though, get back to the question of scale.

If you started in Milan and you scale the Italian market first, that's great, but that's a small fraction of scaling in the U.S. market. And so... We've got 500 million people in the whole union. Well, 500 million people, but divided among 27 different countries, which all have their own slightly different regulatory frameworks and cultural norms, which does have a level of complexity. to scaling up. What I'm trying to get at is, you know, you go back to what Kissinger would have

tried to construct, right? As for China to be closer to the US in the early 70s as to help contain the Soviet Union, you're a student of Russia history. What's seeming to people like me in Southeast Asia is that, you know, we're seeing Russia getting closer to China, which is counter to the logic with which somebody like Kissinger would have postulated in the early 70s, right? Unless the Russians and the Chinese are actually adopting what Kissinger was thinking about in the 70s, right, to try to help to contain the US. So wouldn't the Europeans be vested in making sure that they're participatory in this chip war

narrative in a much more proactive manner? Aside from SML. Yeah. Yeah. Well, we have seen Europe past the EU Chips Act, which I think will have the effect of providing a lot more chips for European manufacturing base. That's why TSM, for example, is building a big new plant in Germany. But I think your question is right that if you look at the next generation chip companies that are being founded, most of them are being founded in the U.S. rather than in Europe. And they're often being founded by Europeans in the U.S. That's, I think, is part of Europe's challenge. So do you foresee a future where we're going to see a lot of, or a lot more Europeans getting

involved in chip designing, chip making in Europe? You know, you certainly see some of that today, but I don't know that I'd predict more of it. I think what we've seen thus far in the last couple of years of AI development is that there are big returns to scale here as well, both in terms of getting your products in the hands of many users and in terms of having a very important. large compute base to take advantage of, which means partnering with one of the big three cloud computing firms, or all U.S. firms. And those dynamics aren't going to go away, which I think means the U.S. is going to be playing this very large role in providing compute

and providing chips for a long time into the future. You know, this compute capability in the latest kinds of GPUs, if you listen to the Jensen Huangs of the world, I mean, he's talking about 10xing, 20xing, 30xing in the newest GPU. And I keep thinking about the energy that's required, right? And there's a lot of talk about sustainability nowadays in many places around the world. I just can't see how the planet is going to attain carbon neutrality at the rate that, you know, there's going to be massively more data centers that need to be built massively much more energy that's going to be required.

What's your take on this? You know, in the context of humanities wanting to be carbon neutral by 2050, 2060? I would say a couple of things. I think first, this is the positive news. We've always underestimated the scope for improvements in energy efficiency. We've always underestimated it. And I would be very hesitant to make that same mistake again.

I would bet on efficiency gains. I would bet on a very large number of companies who are trying very hard to make their systems more efficient, being a very powerful force in the future. And you can go back to 2010 or the year 2000 and find plenty of articles about how the first data center boom or even the invention of the internet was going to lead to a swamping of electricity demand as everyone signed online for the first time. It didn't happen because we became much more skilled at being a efficient. But I think you're right that we're going to have to adjust to data center demand.

And for many countries, it's going to be a challenge because most of the big data center operators have green power targets on their books that they're sincerely trying to meet. So, for example, I was, as I mentioned previously to you, I was in Vietnam not so long ago, where there are international data center firms looking very seriously at the Vietnamese market and asking what's the green power availability? And if it's not available, it's a huge problem. And I think you see many countries thinking much harder about how can they match their data center demands

with their green power demand, supply, and it's difficult for many geographies to do. I think next to that, many of the places that have plentiful potential green power, like the Middle East, builds a lot of solar in the Middle East, power, a lot of data centers, have a very complex regulatory landscape to deal with.

Will U.S. or European or other countries' firms be comfortable keeping all their data in Saudi Arabia? Even if power is cheap and green, I'm not so sure. And so that's another part of the regulatory landscape that you've got to think about when it comes to where you build your data centers. And it means that it's not always easy to put your data center right next to where your hydropower is or where your very sunny climate is because you've got to think about the regulatory aspect as well.

I mean, I can speak for Southeast Asia. I mean, if Southeast Asia were to be modern all across, you know, with the exception of Singapore and Brunei, which are already adequately electrified at modern levels, the region is going to need at least $2 to $3 trillion. I mean, where's the money going to come from? It's only going to come from the G7 countries, right? And the way you look at liquidity around the world, it's mostly concentrated in the G7 countries,

which fortunately, unfortunately, also have the technological warwithal for whether it's chip making, AI and all that good stuff, right? And I see AI as something that's going to be proactively played by China in the US in a big way in the next 10 to 20 years. and I see that as a space that's going to require disproportionately much more energy. And that unfortunately undermines, I think, the narrative of sustainability to some extent. I think Sam Altman, the Open AI CEO, attracts a lot of attention saying that the world might need $7 trillion of infrastructure. At least. And a lot of that will be energy infrastructure, not just chip infrastructure.

So I want to go back to the, to sit at the same. trap, right? What do you think the Chinese could do or have to do in order for the U.S. and China to be able to coexist, compete harmoniously? And I want to give you some background in terms of history. We in Southeast Asia have always seen the Chinese as being non-militaristic in the last 2,000 years, with the exception of what happened in the 13th century, when it was actually the Yuan dynasty, which would have been Mongol. And a lot of us in Southeast Asia don't see Chinese or China as a real threat, right?

But a lot of us also think that if things are not managed by the big guns, but big powers, properly, things could go wrong for the little people in Southeast Asia. What's the best way to embrace the future or assure the future for all of us? The way I think about this question is, what has changed for the last 10 years to bring us from a situation where there was competition, but no one was worried about war? Today, where Japan has decided to double its defense spending as a share of GDP.

What clear signal do you need that the security environment has changed than that? I think you ask yourself, what's changed? There's been a series of territorial disputes that had existed for a long time, but have become newly intensified, newly military. whether it's between China and Japan or the Sinkaku Islands, whether it's in the South China Sea, Spratleys, whether it's around Taiwan. We see much more military tension in all of those, or whether it's between China and India, perhaps most interestingly of all, all those spheres have become, they were existing problems that became militarized recently.

And if there was a way to reverse that, then I think we'd end up in a situation where we had peaceful competition that we won't worry could spiral out of control very rapidly. And so I think it's incumbent upon all of the countries involved, but I think especially of China, which has been in all of those cases the driver of increased militarization to wind that back, to say we might disagree about where the boundary lies in the South China Sea, but let's resolve that in terms of litigation or negotiations rather than by sending our maritime militia around to test the boundary in all of these different locations. And to me, that step alone, which would be a big step for China,

but could bring us back to where things were in 2014, which were competitive, but not tense in a way that they are today. See, what worries people like me is that in a much more multipolar world, there seems to be less ability to multilateralize, right? And the onus is really on us in Southeast Asia to get or act together so that we can help support anybody's ability to multilateralize. And I think any kind of flashpoint or potential flashpoint in the South China Sea could only be resolved if there's greater wherewithal to multilateralize the conversation. Unfortunately, we're sort of like in a corner here where we're sort of forced to bilateralize, right? Right. So put that in the context of this technological warfare or potential technological warfare.

What could draw the picture? What could go wrong for some of us? Well, I think what you what you see right now is the US bringing together its allies, Japanese, Taiwanese, and the Europeans in an effort to to couple the military containment of China, which is what the US has been pursuing. with a technological containment as well. On the thesis that if you believe that AI will be used in every sphere, including defense and intelligence spheres, that giving China access to the most advanced AI chips

would be unwise in terms of America's ability to retain its military edge. That's the theory of the case from U.S. policymakers. And I think that is inevitably going to have a response in China. It's already catalyzed China to double down on its efforts, to build a self-sufficient ship industry. But there's some risk that it also induces China to take more risky steps in the international stage as well. And that's what has to be most carefully managed, I think,

because the U.S. policy will backfire if it catalyzes China to act in an even more risk-taking way in some of the disputes that we talked about. I don't know we've seen evidence of that yet, but I think we need to watch it carefully to make sure that that doesn't happen. Chris, if we take a look at the number of Chinese studying at all the Americans, universities, we're looking at about 300 to 400,000s. We take a look at the number of Americans

studying at all the universities in China. It's only a few hundreds. Is that part of the reasons why we tend to see misunderstanding more than understanding between the two giants? You know, I think it's more complex than just misunderstandings. I think Americans and Chinese don't understand themselves well today, but I think they understand each other better than they did 50 years ago. And nevertheless, 50 years ago, Nixon and Mao were able to find a pathway out of their political disputes. I think at the core of the competition today is not misunderstanding. It's actually proper understanding that there are strategic interests in both sides that conflict. That's how both Chinese leaders and U.S. leaders see things. And although exchanges can, on the

margin, I think, help reduce undue fears of the other side. I think some of the fears are legitimate and understandable and probably accurate. And so unless we can diffuse the political disputes, I'm not a believer that more exchanges can lead to more understanding, which will on its own resolve the tensions. Is it possible that the Chinese are going to be superior in the chip technology in the next or in the near foreseeable future? Do you have some sort of intuition that they could be? You know, it's certainly possible. The way I think about this question is probabilistically, what's the likelihood that that result materializes?

And I think the competition right now is described in the following way. On the one hand, you've got China, which is starting out five or so years behind the cutting edge technologically, has companies that have privileged access to the domestic market, 20% of world GDP, but have restricted access to the rest of the world market. On the other side, you've got the U.S. plus Japan, plus Taiwan, plus to a large extent Korea, plus the Europeans, which are technologically ahead of China in most segments of the chip industry supply chain, selling to their own markets 70% of world GDP, and there's only 10% that are swing players. And so if you ask me from that basic summary of the situation, which side is likely to win, I would bet on the side that's starting out ahead and

has a larger market. Now is it impossible for the smaller side to catch up? No, not impossible, but I don't think you bet on it as the most likely outcome. You know, Southeast Asia is a region that I think has been very peaceful and stable for the last 2,000 years. It's been able to show greater wherewithal to embrace multipolarity than many other regions, right? We're vested in staying peaceful and stable, right? And we just, we just want to see a much better future for all of us. And I don't know, I'm just curious. What final comments would you have in the context of chip, you know, contest or chip rivalry between the U.S. and China that you think the Chinese and the Americans ought to consider going forward for the benefit of humanity?

Well, I think that the whole world has a vested interest in the stability that you talked about because the downside risk is just so large. When it comes to the chip industry, the challenge is that the exact same technology that is primarily used for civilian devices, smartphones, PCs, consumer electronics, it's also catalyzing military advances and intelligence advances. And governments realize this because they're developing these systems. both the Chinese and the United States and all the leading militaries in the world

are investing very heavily in trying to deploy more advanced technology in defense and intelligence systems. And so the challenge they face is pursuing their competition in defense matters, which governments inevitably will, but doing so in a way that doesn't impose undue costs and innovation in the rest of the economy,

which we all rely on. And what makes it so difficult today is that there's no longer a separate supply chain for defense electronics or for the intelligence community. Everyone is relying on chips made at TSM, for example. And so we're watching the geopolitical competition play out in the civilian electronic supply chain

because it is the world's only supply chain for many types of advanced microelectronics. And that's why this is so difficult to just say, let's confine it to the military sphere, confine it to the geopolitical sphere. It's impossible to separate because everyone is trying to use more civilian electronics in their military and intelligence system.

And so that's why this is such a tricky issue to diffuse because competition is just so deeply embedded in the entire civilian supply chain. Well, on that account, I mean, you've alluded to this in the book that most of the GDP is related to our ability to use chips for our toasters, microwave ovens, our cars, and all that, right? To the extent that there is a disruption, long enough, it's pretty apocalyptic, right? I think the best example of this is during the pandemic. The world actually produced more chips each year of the pandemic, but there were shortages of certain types of chips that were using cars. The world car industry is estimated to have sold half a trillion dollars fewer cars in 2021 and 2022 as a result of chip shortages, even though we produced more chips each of those

years. Imagine a world in which we produce 20% fewer chips globally. The disruption of manufacturing would be vast, which is why it really is the case that it's not just Southeast Asia that has a vested interest and stability. It's global prosperity. That's right. Thank you so much, Chris. Thank you for having me. You've been very kind. Your time. That was Chris Miller, professor of international history at Tufts University and a fellow at the Kennedy School of Government, Harvard University. Thank you.