Chip War (FT's Best Business Book 2022) ft. Chris Miller

[00:00:00] Chris

What you find is that, at points of geopolitical crisis or tension, it’s basically universally been the case that whichever countries control the telecom infrastructure will use it for their purposes. There has, basically, been a guarantee in the past that telecom’s infrastructure, plus advanced computing power, will be used by great powers for espionage and for military purposes when it is possible to do so. 

[00:00:36] Intro

Imagine spending an hour with the world’s greatest traders. Imagine learning from their experiences, their successes, and their failures. Imagine no more. Welcome to Top Traders Unplugged, the place where you can learn from the best hedge fund managers in the world so you can take your manager due diligence or investment career to the next level.

Before we begin today’s conversation, remember to keep two things in mind, all the discussion we’ll have about investment performance is about the past and past performance does not guarantee or even infer anything about future performance. Also understand that there’s a significant risk of financial loss with all investment strategies, and you need to request and understand the specific risks from the investment manager about their product before you make investment decisions.

Here’s your host, veteran hedge fund manager, Niels Kaastrup-Larsen.

[00:01:30] Niels

For me, the best part of my podcasting journey has been the opportunity to speak to a huge range of extraordinary people from all around the world. In this series, I have invited one of them, namely Kevin Coldiron, to host a series of in-depth conversations to help uncover and explain new ideas to make you a better investor.

In the series, Kevin will be speaking to authors of new books and research papers to better understand the global economy and the dynamics that shape it so that we can all successfully navigate the challenges within it. With that, please welcome Kevin Coldiron.

[00:02:08] Kevin

All right. Thank you, Niels, and welcome everyone. If you had to name the single most influential man-made product, what would it be? 

Our guest today is going to argue that it’s the semiconductor chip. He believes that there’s no other item that has had a more decisive influence on international politics, shaping the military balance between countries and forging globalization, than the semiconductor chip. Just to add a little spice to that, you’re going to find out that the most advanced chips, the ones that we rely on for our most advanced technology, can only be made, often by a single company, using an extraordinarily complex machine that itself can only be made by a single company. That creates what our guest calls choke points, vulnerabilities in our ability to produce the technology we need, but also very important levers of power that can and have been used by the US and China against each other.

So, the person who’s going to help us understand this, Is Christopher Miller. He’s an associate professor of history at Tufts University in Boston, and author of a really exceptionally entertaining and eye-opening new book called Chip War: the Fight for the World’s Most Critical Technology. Chris, thanks for joining us today and welcome to the show.

[00:03:34] Chris

Thank you for having me. 

[00:03:36] Kevin

So can you tell us a little bit about your background? I You’ve written three previous books and they’re all about Russia. Tell us, is your family from Russia? What’s your connection there? How did you end up writing those books?

[00:03:55] Chris

I’ve got no pre existing connections to Russia. I just got interested in the country’s history and its economy when I was doing my graduate studies. To me, the history of the Soviet Union provided a fascinating alternative case study to what economics looks like in countries that I was familiar with.

The more I studied it, the more I wanted to understand how did that system work? Since it existed for 70 years, it worked badly, but nevertheless it persisted for a long time. The three books that I previously wrote before publishing Chip War were on different aspects of Russian history, politics, economics.

That actually gave me the initial idea for the book that ended up being Chip War. I started writing Chip War, not intending to write a history of semiconductors, but intending to understand why was it that, during the Cold War, both the US and the Soviet Union were capable of producing the key military technologies of the early Cold War: nuclear weapons and long range missiles, but the Soviets completely failed to produce the weapons systems and the computing components in them that defined military power by the end of the century. 

As recently as Russia’s war in Syria, from 2015 to present, at least 80% of the bombs Russia’s dropped in Syria have been unguided dumb bombs – totally the bombs that would’ve been dropped in World War II because Russia’s really struggled to miniaturize computing power and deploy it to weapon systems in a systematic way.

That seemed like a puzzle that was worth exploring. I realized, as I did that research, that the key explanation as to why the Soviets fell behind in military power while the US surged ahead, and many other things in the role of the semiconductor. That’s how Chip War actually emerged out of my interest in Russian history.

[00:05:49] Kevin

Yeah, that’s fascinating because there are a number of elements that flow through, it seems like, from your earlier books to this one, including the ineffectiveness of (what you call) the copy of that strategy, or basically stealing intellectual secrets and trying to use that as a basis for technology. 

Before we get into that, I want to hear a little bit about your process because I have to say the book was a little bit different than I expected. I was thinking it would be a little more of a dry, technical, geopolitical conflict type book, and all those elements are there, but really it reads a little bit like a spy novel, in a sense. You’re telling the history of the chip industry through the personal stories and plot lines of the people and companies that were involved. You’ve got stories about Morris Chang, who’s the founder of TSMC; Andy Grove, the founder of Samsung and Sony. You’ve got KGB agents in there. 

Tell us about your research process. How did you compile all these different stories and interviews and details? 

[00:07:08] Chris

One of the things that I realized when I started doing the research was that you couldn’t tell the history of the chip and its impact on the world by just looking at governments. You had to understand the companies, as well, that produce the technology. It’s an interesting challenge for a historian because it’s easy to research governments. You know where to find them. They publish lots of papers. They often have archives that you can go in and look at the documents historically, whereas companies aren’t usually like that.

And so to understand the rise and fall of companies, You’ve really got to interview the people who worked there. My research process involved a fair amount of looking at what governments were doing, but also a whole lot of interviewing people who had worked in the semiconductor industry, from the scientists who were devising the devices to CEOs who ran many of the companies that were discussed in the book, to their customers and their suppliers, as well. 

I conducted over 150 interviews with people in the industry, in the US, and Taiwan, and Japan, and Europe; some of whom were at the creation of the chip industry and were in their nineties, like Morris Chang, for example. There were others who are at the cutting edge today, and are still working at chip firms. Those interviews gave me a granular understanding of what made the firms that succeeded, succeed; and what made the firms that failed because you can get inside their cultures and understand what other business models didn’t work.

It also helped me get to know a number of the personalities involved in the chip industry. I was immediately struck by the extent to which there were some fascinating characters involved. Not only brilliant geniuses, from the perspective of physics or engineering (of which there are, of course, many), but also just really interesting people. {These were] people willing to take big bets to risk it all; large personalities who really wanted to engage in a race to control and produce the world’s most important technology. 

I left with the sense that you couldn’t really understand how Moore’s law was possible – this doubling a computing power every single year, that we take for granted (or every other year). You couldn’t understand that by looking just at government policies, or by corporate research. You had to look at the personalities as well because, ultimately, it was the people who were driving it forward. The ambition that undergirded the key figures was, I think, an extraordinary feature to watch and to understand. I tried to let some of that come through in the book. Hopefully, as readers flip through the pages, they get to know some of the individuals who created these companies and the ambitions that drove them.

[00:09:43] Kevin

So, what I want to do is get us to talk about a couple of those key companies that are the choke points in the industry. Before we do that, I do think it would be helpful for a little bit of high level, if you will, technical detail. People talk about semiconductors, chips, all the time, and often I suspect they don’t really know exactly what they mean (and by people I mean me).

So, can you tell us, at a really high level, what is a semiconductor chip and why is it so critical? Why is it such an important piece of our technology infrastructure? 

[00:10:34] Chris

So, first off, [regarding] terminology, the words semiconductor, chip, and integrated circuit are all used basically interchangeably. So, they refer to the same thing, which is a piece of material, usually silicon, that has lots of little transistors carved into it. A transistor is just a small circuit that’s either on, or off. If it’s on it creates a one. If it’s off, it creates a zero. All of the ones and zeros, undergirding all data and computers, all software, all programs, all of these ones and zeros are not just existing ethereally in the cloud, they’re actually circuits that are on or off. 

The process of getting access to more computing power is, quite simply, the process of carving more of these circuits into silicon chips. The first chips that were commercially available in the 1960s had just a handful of these circuits on them. So, they could remember four ones or four zeros at a time.

Whereas today, if you go to an Apple store and buy a new iPhone, you’ll get 15 billion of these transistors on just the main chip in an iPhone. There are actually many chips inside of an iPhone. This increase in the number of transistors you can put on a chip (and, therefore, the increase in the number of ones and zeros a chip can process or remember) has driven the exponential growth in computing power that we call Moore’s law.

And in order to cram more transistors on chips, you have got to make them smaller and smaller and smaller. The first transistors that were invented (over half a century ago) were visible. They were so large you could see them. Whereas today, on a new iPhone (for example), each of the transistors will be the size of a coronavirus, and they’re carved by the millions with basically perfect accuracy every single day.

[00:12:19] Kevin

Yeah, that’s just absolutely phenomenal – some of the technological stories that you tell about the miniaturization of the process. Okay, so basically, chips are embedded across, virtually, any product that uses technology these days.

You talk about,early on in the book, a typical process for manufacturing a chip, which I think does a great job of highlighting two of the key themes of the book. You say that, just generically, you might start with ( if you wanted to manufacture a chip) a blueprint from a UK based firm that would be designed using software created by US firms (maybe programmed by Israeli engineers).

So you’ve got your design and then you send it over to a facility in Taiwan to manufacture it. That facility buys gasses and silicon from Japan and then carves the design into the silicon (as you said) using very, very precise equipment made by one of five companies: a Dutch company, a Japanese company, or three in the US.

Then the chip, once it’s manufactured, gets packaged up, sent to Southeast Asia for testing, and then eventually sent to China for assembly into a product. So, that’s how a chip is made. When you think about that process, you see that there are these choke points (that I referred to at the beginning); that there are vulnerabilities; that there’s reliance on a few companies in a very long chain. That’s theme number one, I would say. 

Then the second theme is that China (again, this is a generic process), comes in at the very end. They’re utterly dependent (as you say) on firms that are basically either in the US (or allied to the US) to get their technology.

So, those two themes (the choke points and if you will) and China’s vulnerability (and the implications of that) are the things I’d like to explore. So, if we start with choke points, how did we get to a situation where we’ve got these choke points? The industry didn’t start that way.

The industry started with companies that designed and manufactured their own chips. Could you walk us through just how we went from those initial stages to this much more complex world of integrated supply chains with these key vulnerabilities in there? 

[00:15:07] Chris

The key driver of the dynamic of having choke points (and a choke point is just a production process that is very, very difficult to find alternative suppliers for – find some sort of substitute for)… The reason we’ve got so many choke points in the chip supply chain is because making chips is just extraordinarily complicated. The process of moving from four transistors per chip, in 1960, to 15 billion or so, today, has required extraordinary specialization in all of the steps you mentioned: in the software tools, in the chemicals used, in the machine tools used to manufacture chips. 

In order to produce the precision required, specialization has been the only way to move forward. Companies have only been able to produce silicon that is 99.9999% pure by focusing solely on the process of purifying silicon, and the same thing is true for all of the other steps. We’ve needed that precision (enabled by specialization). The effect of it has been to leave expertise, in each part of the production process, in the hands of just a couple (or, in some cases, just one) of companies. I think the best way to understand the increase in precision is to look at what the process of making a chip looked like in 1960 or so when each chip had four transistors in it.

Bob Noyce, who was one of the individuals who founded Fairchild Semiconductor, was known as the person who (more than anyone else) created modern Silicon Valley. He later founded Intel. When he was making his first chips, right around 1960, he would jury rig all of the tools he needed himself.

So, rather than buying tools off the shelf, he would just make them himself. One of the steps of the production process is called lithography (which I’m sure we’re going to talk about a bit). Lithography means shining light in a specific pattern onto a chip, where the light reacts to chemicals and lets you create shapes on the piece of silicon.

The first lithography tools that Bob Noyce used, in the 1960s, were made with a 20 millimeter camera lens that he bought from a local film shop. Today. If you want to buy a cutting edge lithography machine, it costs you $150 million. The machine takes multiple 747s to transport because it’s so big, and it involves the flattest mirrors ever made.

So the transition from a 20 millimeter off-the-shelf camera lens to the flattest mirrors ever made, inside the optics of a new lithography machine, is just one example of the huge jump in precision that has been necessary and possible for producing chips. That’s also why we don’t have lots of people producing the tools used at each part of the process. It’s easy to have a lot of people produce 20 million camera lenses, but producing a $150 million tool, involving the flattest mirrors ever made, is something that you just can’t have multiple companies do. It’s just too hard. 

[00:18:07] Kevin

Yeah. That story, that’s one of the ones I wanted to talk about.

I wanted to talk about two companies and one of them is ASML, the Dutch Lithography company. Maybe we could do that right now. Tell us a little bit about, we went from this world where Bob Noyce is shining light on a chip through basically a homemade process to (I think you’d call it in the book) the most complex piece of machinery ever assembled by humans.

I’d like people to understand a little bit more about the company that makes them and how they do that. Before that (I don’t know how you feel about it), to me, when I read about ASML, in some ways it was quite an uplifting story because it’s [all about] this is what humans can actually do when we cooperate.

They’re pulling together technology, expertise from all around the world. It’s extraordinarily complex, and they’re assembling it together and producing this machine that’s projecting light onto just an ultra fine (or a size that it’s almost mind bogglingly small). I’m just , wow, we can really do great stuff when we get together. So, tell us a little bit about that company and the product they make. 

[00:19:35] Chris

So, ASML is the biggest tech company in Europe, although most people have never heard of it. It was spun out of Phillips (the Dutch Electronics Company), which several decades ago was a fairly large semiconductor producer – no longer is.

Like many other chip makers, Phillips at first was making all of their tools in house, but eventually realized that this was not a viable way to do business. So they spun off ASML several decades ago. At the time there were a handful of companies that were making these types of lithography tools – two in Japan, and Canada, – Nikon, a couple in the United States (while they were already on their last legs), and ASML. 

In the early 1990s there was a big debate, in the chip industry, about what the next type of lithography would be. One of the trends that had developed, over the chip industry, is that the 20 millimeter camera lenses that Bob Noyce had used for the first chips, used visible light to carve on the chips. Visible light has a wavelength of a couple hundred nanometers, depending on which color you’re looking at. That might seem like a pretty small amount but for chips that, today, have shapes measured in just a handful of nanometers, 200, or 300, or 400 nanometers is too broad to do the precise carving that was necessary. So, there has been a steady decrease in the wavelength of light used.

[00:20:59] Kevin

Sorry to interrupt, Chris, just to summarize, what you’re saying is that you’re shining light, essentially through a mask, to create a pattern on a silicon chip. The actual wavelength of visible light, that’s coming through that mask, is wider than what you’re trying to create the pattern on. So, you have to move to a different type of light. You have to move from visible light to extreme ultraviolet light, which has much, much smaller wavelength in order to imprint that pattern.

That’s the kind of…

[00:21:38] Chris

That’s right. So, one of the questions was, what’s the right ‘new type of light’ to move towards? There were different wavelengths that were debated. EUV (extreme ultraviolet, which is a little bit past X-ray in the light spectrum) was one option. Most people thought it was impossible – just too hard to do for two reasons. One, because X-rays go through materials. Developing a set of optics and mirrors that could reflect X-rays was going to be very hard. And two, because it was really difficult to create enough EUV light to actually get the quantities needed to do the carving. You need a fair amount of it to react with the chemicals and carve the patterns for you.

But ASML bet, to the contrary, that it would be possible. They spent three decades and many tens of billions of dollars developing the systems needed to produce light EUV, and then the systems to collect the EUV after it was produced, using these ultra flat mirrors, and direct it towards a silicon wafer.

And so today (ust to walk you through what one of these machines will do), they’ve got tiny balls of tin that are 30 microns (30 millionths of meter wide) dropping at a very rapid speed (several hundred miles an hour) through a canister inside the machine. They pulverized each ball of tin with a laser, and after shooting it twice it explodes into a plasma at several hundred thousand degrees Fahrenheit (so, we’re talking 50 times hotter than the surface of the sun inside of this machine). That plasma emits EUV light at 13.5 nanometers, which is then collected by a set of mirrors. Because of the mirror challenge that I mentioned. X-rays go through most types of materials. Producing these mirrors was one of the hardest parts of the EUV system.

Just the production of the mirrors, itself, is a technological marvel. Getting them flat enough, so that they don’t lose any of the light that they’ve collected but direct it in exactly the direction it needs to go, was another challenge. They’ve got the flattest mirrors ever produced by humans in each of these machines.

Then the light is collected via these optics. It shines through a mask onto the silicon wafer where it eventually carves the chip. That’s how this machine works. It took three decades to develop. Today, if you want to buy one, it’ll cost you $150 million, which is the price of the precision that a EUV tool makes possible.

[00:24:12] Kevin

Yeah, I think that what you say in the book, that the mirror was so pure that if it was scaled up to the size of Germany, that the biggest irregularity would be a 10th of a millimeter, and this is phenomenal. 

Okay, so this is a lithography machine made by this Dutch company. This company is the only company in the world that can produce these machines.

Why are those machines so critical? 

[00:24:46] Chris

So they are the only company that can produce them. They have 100% market share, and I think they will for a very long time. Yet the machines are critical because you can’t make an advanced chip without them. The precision patterning that you can do with light at 13.5 nanometers lets you carve very small shapes at high volume.

If you’re trying to use visible light or even ultraviolet light, you just can’t get the precision you need. Every new chip in an iPhone, every new chip in a PC, data center chips, are all now using this tool to carve ultra small shapes and pack more and more transistors, and thus more and more computing power, onto them.

So, everyone listening to this podcast will currently be benefiting from the precision carving of an EUV machine enabling more computing power that they take for granted. 

[00:25:43] Kevin

So without these machines, there are no iPhones, there’s no (presumably) supercomputers that AI technologies rely on that also use these chips that are produced by the ASML machine.

So, that’s a clear (if you will) choke point. The other choke point, that I think people are probably a little bit more familiar with, is the firm that dominates the manufacturing of these chips, which is Taiwan Semiconductor Manufacturing Company, TSMC, which has been in the news a lot lately because Warren Buffet’s firm has just bought a sizable stake in it.

Tell us about the history of TSMC and why it’s so important. It seems to me, and I could be wrong, but it seems to me that it was innovative on multiple dimensions, but perhaps the most important was just the business model it chose to pursue. Could you maybe just tell us a little bit about how this company started and why it’s so important?

[00:26:51] Chris

Yeah, I think that’s absolutely right. It is a story of business model innovation, leading to technological innovation. You wouldn’t have had one without the other. TSMC was founded in 1987 in Taiwan by a gentleman named Morris Chang, who I think is the most underrated and underappreciated businessman of the last century.

He was born in mainland China, fled the Communist revolution, enrolled as the first Chinese student, the only Chinese student in his class at Harvard, and then went on to (in some ways) single handedly build the manufacturing capabilities that made chip making possible in the US. He was passed over for the CEO job of Texas Instruments in the 1970s – one of the greatest errors TI ever made.

[00:27:35] Kevin

Why did they pass him over? I was curious. You don’t say that in the book. Was there? I don’t know. Why do you think they didn’t make him CEO? 

[00:247:44] Chris

It’s not exactly clear. He was known to have a pretty sharp set of elbows which I think was key to his success as a manager in some ways, but everyone recalled him being a pretty hard driving manager. That may have been part of it. I’ve heard people also suggest that it was because he was Chinese. That was part of it. I don’t know if that’s true, but I’ve heard that suggested as well. I don’t have high confidence in either of those explanations, but certainly it was, in hindsight, an historical error for TI and for Texas.

It could have been that TSMC stood for the Texas Semiconductor Manufacturing Company rather than the Taiwan Semiconductor Manufacturing Company. The world would look very different today, had that been the case, but it was the Taiwanese government that swept in and offered him, essentially, a blank check, in the mid eighties, to build a new business in Taiwan.

Even in his days at TI, he was already exploring a new business model, as you mentioned. Before that point almost all chips were both designed and manufactured by the same company. Chang realized that as chips got more complex, both design and manufacturing were getting harder and manufacturing was getting more expensive because you needed to buy these ever more precise tools.

As a result of that, he looked into the future and envisioned a world in which companies that design chips wouldn’t have to worry about manufacturing. They wouldn’t need the expertise or the capital investment required because they could outsource manufacturing to a firm that focused solely on manufacturing.

And today we call those firms foundries – companies that manufacture chips but don’t design any. Today there are a number of companies that use this Foundry business model, but there were none when Morris Chang started up TSMC with this in mind. 

It proved to be the right business model because it let him sell chips to multiple different customers, which let him build huge, scale reaping efficiencies both in terms of purchasing equipment and materials, but also in terms of honing his technology. Today there’s a pretty clear relationship between the number of chips you produce and the efficiency at which you produce them. You can learn from each chip you produce and hone your processes.

So, because of the scale that the business model allowed, TSMC was able to learn more about the manufacturing processes, and today it has the world’s most advanced manufacturing processes in no small part because it’s the world’s largest chip maker. 

[00:30:14] Kevin

And when you say it’s the world’s largest chip maker, can you put that in perspective? Give us some sense for how reliant we are on TSMC. 

[00:30:25] Chris

Every one of us touches TSMC’s chips every single day. Almost all smartphones have TSMC chips inside, easily 1/3 of PCs do. Data centers are chock full of TSMC semiconductors. If you go down the value chain to simpler goods using simpler chips like dishwashers and microwaves, they too often have TSMC chips inside- older technology but, nevertheless, theirs. Over1/3 of the new computing power the world adds each year, over 1/3 of the ones and zeros that are used for computing, comes from chips produced by TSMC. 

[00:31:02] Kevin

Where does Intel fit into this picture, because Intel is a company that most Americans know, and I think most Americans think of as a chip producer. They are a chip producer. 

They’re a company that was founded by ‘the legends of the business’ and run by Andy Grove for a long time. Where do they fit into the picture? Are they potentially a competitor for TSMC or are they being left by the wayside?

[00:31:41] Chris

Right now Intel is trying to turn itself around. For most of the decades, after it was founded in the late 1960s, Intel was one of the leading US chip makers, both in terms of technological capabilities, but also in terms of business success. They’ve been immensely profitable over the past several decades, but they’ve missed a couple of key technological transitions in the past two decades.

First, when Steve Jobs approached Intel and asked the company to produce chips for iPhones, Intel turned it down thinking that smartphones would be a niche product with low margins. They missed the huge wave associated with smartphones over the past 15 years. Then when it came to the rise in focus on chips for AI purposes (artificial intelligence), about a decade ago, Intel was late to the game on that too. So, both in smartphones and in data centers (which is where most AI chips go), Intel has been losing market share to competitors because it missed key technological transitions. 

As its size has declined relative to TSMC or to Samsung, Intel’s ability to keep up with new generations of manufacturing technology has also slipped. Although it used to be the most advanced chip maker, in terms of manufacturing capabilities, today it’s very clearly behind both TSMC and South Korea’s Samsung too. So, it’s trying to turnaround right now, trying to catch up, launching a foundry business arm, as well as rolling out new chips for AI and data centers, but there’s a long way to go. The new CEO, Pat Gelsinger, has a pretty tough task ahead of him, in terms of turning it all around. 

[00:33:27] Kevin

Apparently you said he’s done a deal with ASML to give them first crack it using their next generation lithography machine. Is that right? 

[00:33:38] Chris

That’s right.

So, that could give Intel a bit of a leg up. I think the company’s got not just challenges in terms of its next generation manufacturing technology, it’s also got to sort out its business model.Right now it’s pretty clear, I think, to most observers, in the industry, that the foundry model works better then the integrated model that Intel stuck with of both designing and manufacturing its chips in house. It’s one of the few companies in the processor chip space that still has that business model, and it just seems to make a lot less sense today relative to the alternative. 

[00:34:14] Kevin

And you do say in the book that, “Hey, look, it’s easy to bemoan the fact that we’ve offshored production and we’ve lost (aside from Intel) a lot of the capability to produce chips domestically.

On the other hand, the US does still have some world class firms that design chips, what you call fabless firms – firms that just do the chip design but don’t produce them. It’s very unlikely that they would have those competitive positions if they also had to manufacture chips themselves. So, that specialization (that helped TSMC develop this huge, important market leading position) also has allowed US software firms to develop similar positions. 

[00:35:09] Chris

Yeah, I think that’s absolutely right. If you look at the innovation thatTSMC has enabled in Silicon Valley, there’s a number of great examples of this. We don’t normally think (but perhaps we should) of manufacturing advances in Taiwan enabling innovation in Silicon Valley, but that’s exactly what’s happened in this case. [There are] companies like NVIDIA, for example, which produces chips for AI and data centers. Nvidia is one of the most valuable US chip makers today but it manufactures none of its chips. It was possible to found it precisely because it was able to access TSMC’s manufacturing capabilities. The founders of NVIDIA never wanted to do manufacturing, and that was a viable business model because, when they were founded, it was already the case that they could turn to TSMC and make their chips. 

I think the other example is Apple, which most people don’t think of as a chip maker. In fact, Apple is one of the world’s largest chip designers because it designs in-house many of the chips, including the key processor chip, inside of its devices. Apple has never had to manufacture a single chip because it too can take its designs to TSMC, in Taiwan, and get them all manufactured.

And I think it’s unlikely that Apple would be designing its own chips if it weren’t able to outsource the manufacturing to a specialist like TSMC. 

[00:36:27] Kevin

So we’ve got choke points that I identified. We’ve got the TSMC as the manufacturing choke point. We’ve got ASML as also a manufacturing choke point or producing equipment for manufacturing. Really, the US and software designers, would you also consider them a choke point? Are they absolutely critical to most of the technology that goes out there? 

[00:36:57] Chris

Yeah, if you want to design a chip with 15 billion transistors on it you’re not going to do that by hand. Although, drawing out by hand was how the initial chips were designed and they only had a couple thousand transistors on them. You need software to do it for you. Software doesn’t just let you draw pictures or schematics of where the transistors go on the chip. They also help you model how those chips will work so you can understand their functionality before you have to produce the first one.

That software is designed and produced by three companies (largely), all of which are based in the US. So, the US has been able, in recent years, to cut off foreign companies’ access to software, which makes it, basically, impossible them to design cutting edge chips. 

[00:37:42] Kevin

Yeah, let’s talk about that.

That helps us migrate to that second big theme, which is, China’s position in the world and the vulnerability that  it feels, and that it probably has. You talk about two Chinese firms, Jinhua and Huawei. Both, I think, are good examples of how the US has turned these choke points into leverages of power.

So perhaps we could maybe just briefly talk about them. Let’s talk about Jinhua first, because I think it’s a more straightforward case of what people think of as sort of industrial espionage that could have, potentially, led to a competitive advantage and elimination of a US firm; and then how the US used these choke points to thwart it.

[00:38:38] Chris

Yeah, so Jinhua is a company based in China’s Fujian province. It was looking to get into the market for a type of memory chip called DRAM chips. There are three companies today in the world, two in South Korea and one in the US, that produce cutting edge versions of these chips. Jinhua worked with a couple of micron employees in Taiwan to try to and succeed in stealing some of the… 

[00:39:06] Kevin

Micron is the US firm that designs DRAM chips. 

[00:39:10] Chris

That’s right.

[00:39:11] Kevin

They make the DRAM chip. 

[00:39:13] Chris

That’s right. They have facilities in multiple different countries, including one in Taiwan. So, Jinhua worked with these employees of Micron’s Taiwan plant to acquire some trade secrets as to how Micron produces its chips, and tried to take this to house facilities in China and use them to upgrade their production capabilities there.

The Chinese government intervened pretty clearly on behalf of Jinhua, accusing Micron of violating Jinhua’s intellectual property from the things that Jinhua had stolen from Micron. It was a real challenge for Micron because China is every chip maker’s biggest market. Individual companies really struggle to push back against the Chinese government because they’ve got to be able to sell their products in China for commercial viability.

And so they went to the US government and asked for help. For a long time the US government had really struggled to deal with these types of cases. The Chinese courts had ruled that their patent law said one thing. It’s hard to argue against patent law decisions by courts because it was a judicial process. Even if you think it’s a kangaroo court, how do you prove it? So, for a long time, the US tied itself up in knots trying to find legal justifications and legal standards of proof behind the fact that the Chinese government was supporting its companies as they were engaged in IP theft. 

By the time of the Jinhua case frustration had bubbled up both in the chip industry, but then also in the US government to the point where the US took a new type of action and made it illegal to sell certain types of equipment to Jinhua without which it’s impossible to make advanced chips. It essentially put Jinhua out of business.

Today, Jinhua is no longer producing chips because they can’t acquire the machinery that it needs. This was a real new type of punishment for a Chinese firm. It illustrated, in a really visceral way, the power the US government had to, on its own, unilaterally, decide that US equipment couldn’t be shipped to a given firm. That amounted to a death sentence to Jinhua and any other firm that found itself on the wrong side of US regulators. 

[00:41:30] Kevin

Yes, and that, just to put us in the time, that was the Trump administration that ended up doing that. Is that right?

[00:41:38] Chris

That’s right. That’s right. It was happening at the same time as the “trade war” and dispute over tariffs.

And I think a lot of people at the time saw those processes as interlinked. There was a tech dispute, and a trade dispute, and they were both about economic issues. But I think, actually, they were driven by really different factors. The concern about semiconductors, and China’s advances and semiconductors, was emerging from really deep within the US national security state – the intelligence agencies, the military were very, very concerned. Not just about the commercial ramifications of China’s chip industry and its advances, but also what this would enable China to do on the intelligence front and in terms of military technology. 

[00:42:21] Kevin

So in the case of Jinhua, I think, if you read through it and listen to your description, you’re like, “Okay, that was a clear case of theft of intellectual property.” There was punishment meted out, and you could argue whether or not it was too extreme, but it does seem like a more clear-cut case of black and white. Whereas, Huawei, it seems to be more complex. 

Huawei is the Chinese telecom equipment maker that’s been in the news a lot and you tell quite a balanced story about Huawei. In some sense it felt to me like they were following the playbook that, in large part, Japanese firms had done, that Korean firms had done. It initially started off as a low cost producer, gained some market share, reinvested in their business, and moved up the technology chain. So, not really growing based on theft of intellectual property, although there were accusations that they were doing some of that. 

Eventually they became so (if you will) embedded in the system that it became uncomfortable for the US and its allies to have a Chinese firm so embedded in the technology infrastructure, the technology ecosystem that it wasn’t sustainable. So, tell us about what happened to Huawei. 

Do you agree with that description that I gave? And how did the US use its levers of power to influence them?

[00:44:01] Chris

Yeah, I think that description is exactly spot on. Huawei was popular among telecoms companies because it provided reasonably good equipment at low prices. That’s what it was quite good at. The US government had a number of different criticisms of Huawei about selling equipment to Iran that was violating sanctions, about espionage.

The key challenge, with Huawei, was (as you say) that it was too big and too embedded in networks for the US to trust that it wouldn’t be used for espionage in the future, or for coercion in the future. If you study the history of network technology, going back to the days of the telegram, what you find is that at points of geopolitical crisis or tension, it has basically universally been the case that whichever countries control the telecom infrastructure will use it for their purposes.

So in World War I, the first day of the war, the British cutoff Germany’s access to its telegraph network, making it very difficult for Germany to send telegrams across the sea. In the Cold War both the KGB and the US security agency got privileged access to advanced computers and the ability to tap their own telephone networks for spying purposes.

So, there’s basically been a guarantee, in the past, that telecom’s infrastructure plus advanced computing power will be used by great powers for espionage and for military purposes when it is possible to do so. So, the US government was looking at Huawei’s growing market share in the world’s telecom infrastructure (the types of equipment that you need to connect yourself on the internet) and asking itself whether it was plausible to imagine there were ways to guarantee the functioning and the privacy of these networks amid an intensifying geopolitical competition, or whether it was just inevitable that if China was a geopolitical adversary, that Huawei would be used in the interest of the Chinese government.

It concluded that the answer was yes, it was unavoidable, and therefore the only real remedy was to block it out. So, that’s what we’ve seen over the past several years is a blocking of Huawei from US, Japanese, Australian, and many European cell networks. 

[00:46:24] Kevin

So, this seems to me an extraordinarily important point because I think, if you follow the logic through, what we’re saying is that we will not accept (when I say we, I’m talking about the West in the broadest possible sense) a Chinese firm integrated into our ecosystem in the same way we would a Japanese or Korean firm. That basically means, if you follow that logic through (and feel free to disagree with me), that there’s going to have to be two almost fully separate telecommunications and technology infrastructure systems out there that can’t be integrated. They’re basically saying, “Hey, globalization we’re drawing a line under it.”

[00:47:10] Chris

I think the West is saying to China, for these critical types of technology, “You’re welcome to use our system, but we’re not going to use your system.”

So in some places like chip making equipment, China is still using Western systems because it doesn’t have its own. In places where China’s got the same capabilities as the West, like telecom equipment, then yes the inevitable result is two different systems; a decoupling, if you will, of the tech infrastructure.

[00:47:41] Kevin

Yeah. There’s various implications. I know, your time is short here, but if I’m China, I’m definitely not accepting that deal. I’m not saying, “Well sure, we’ll use your systems where we have to.” So, that brings us to the point that influences the title of your book, Chip War, that there is a serious potential for conflict here, in particular with TSMC. It sits at only a seven minute flight from China, and it’s critical to the world’s technology. It seems to me the most likely flashpoint for a conflict between the two countries, which I don’t think is a particularly new or controversial statement.

But what I wanted to ask you was this, earlier in the year I interviewed James Fok, who wrote a book called Financial Cold War, and his thing was, look if we want to prevent conflict between China and the US (in financial terms), we should use the concept of mutually assured destruction.

In other words, instead of building two competing systems, let’s integrate the systems so deeply that breaking them apart would be too disastrous for both countries to want to do. So, I’m reading the last part of your book and the last part of your book is you outline various scenarios for potential conflict between China and the US.

In each one of them, it seemed to me, both sides would lose more than they would gain. So, maybe we’re at a point of mutually assured destruction in technology. Do you agree with that? Disagree with that? Do you think deeper integration is actually a more enlightened way forward?

It sounds like that’s not what we’re going to end up with. How do you think about that?

[00:49:57] Chris

I think that the problem with having a lot of faith in mutually assured economic destruction is that historically it hasn’t always worked. Britain and Germany were major trading partners right up until the outbreak of World War I, and then immediately they flipped and Britain launched a very aggressive blockade of Germany trying to starve the German people. So, in a matter of months it went from big trading partners to trying to cut off German food supplies. It was the same thing with the US and Japan in World War II. The US was one of Japan’s biggest trading partners. Then the war started and a blockade went into force.

I think there are some places, historically, where you can say mutually assured economic disruption worked, but there are also a lot of high profile cases where it didn’t work. Just this year, I think the failure of Germany’s strategy to integrate deeply with Russia, in economic terms, has not provided the results that were hoped for.

So, I hope it’s true that we’re in a state of mutually traded economic destruction, but I worry that it’s not true. That worries is one of the things that inspired me to write Chip War, and to understand what are the real risks, if in fact the interdependence we’ve got right now turns out not to be a stabilizing force, but turns out to put us in an even more vulnerable position; if in fact something disastrous does happen in the Taiwan straits.

[00:51:20] Kevin

I start the podcast with my admission that I use the word Chip, but I don’t always know what it means. So, I started with an admission of my ignorance and, maybe I’ll end with the same thing in that a lot of the stories you tell about the individual companies, in the book, are that you have market power, and then your technology is disrupted.

Is there anything on the horizon that could disrupt the chip? Is there technology that could replace it or change it in such a way that these current choke points disappear? Or is that just a technologically illiterate question? Does that not make any sense? 

[00:52:08] Chris

No, I think that’s a great question and I’ll give you two answers that are somewhat hypothetical today but could be quite relevant in 10 years time. The first is looking at quantum computing which relies on a very different underlying technology and still remains somewhat far off from commercialization and practical application. When it is realized it will really transform the ways we compute. It won’t make traditional digital chips irrelevant, but it will, I think, change the value chain in terms of what we pay for, what types of chips we use in different use cases, things like that. I think when quantum comes it will be quite meaningful in reshaping the industry.

Second is the forward march of Moore’s law, which has worked for 60 years, is not guaranteed. It’s a prediction, actually not a law. That could well break down in the future. If it does that will also reshape the chip industry. It will make some choke points even more important, but it also might make some choke points less important.

If in fact Morris law slows, it will be harder to differentiate the cutting edge from the lagging edge. So, that could also really profoundly shape the industry. Now I’m an optimist on Moore’s law. I think we’ve still got at least a decade of progress, and who knows beyond that point. There are some people who are more pessimistic. If the pessimists are right, and Moore’s law is ending sooner, that will also have a major impact on the shape of the industry.

[00:53:41] Kevin

Well Chris, thanks for joining us and enlightening us about things that we use every day, but we really don’t understand, often, how important they are geopolitically. 

The book is called Chip War. It’s eminently readable, it’s fun and really eye opening. Thanks for writing it Chris, and thanks for joining us today.

[00:54:08] Chris

Thanks so much for having me. 

[00:54:11] Kevin

Okay. With that, I’ll pass it back over to Niels. 

[00:54:16] Niels

Thank you so much, Chris and Kevin, for a super insightful conversation looking at the world of semiconductor chips and their importance for the world that we live in. Chris had some really great stories to share, even if we only scratched the surface of what is in his book. So, I really love this conversation. 

I found the discussion about the choke points within the chip manufacturing process fascinating. Of course, the background and history of a firm like ASML, which is much less in the news compared to TSMC, and how they’ve spent more than three decades perfecting the art of lithography after spinning out of Phillips is just amazing. The whole explanation that Chris provided as to how they create this special light needed in order to print on these tiny chips is just mind boggling. 

Of course, the history of TSMC and its founder who, apparently, was passed over as new CEO of Texas Instruments in the 1980s, was very interesting, and the fact that Warren Buffet, in Q3 of this year, bought $4 billion worth of shares is also pretty newsworthy. 

Make sure you go and follow Chris and Kevin’s work, as well as getting a copy of their books because, as you can tell from today’s conversation, some of these ideas and topics are not being discussed enough on mainstream media. From Kevin and me, thanks so much for listening and we look forward to being back with you on the next episode, and in the meantime, take care of yourself and take care of each other. 

[00:55:42] Ending

Thanks for listening to Top Traders Unplugged. If you feel you learned something of value from today’s episode, the best way to stay updated is to go on over to iTunes and subscribe to the show so that you’ll be sure to get all the new episodes as they’re released.

We have some amazing guests lined up for you, and to ensure our show continues to grow, please leave us an honest rating and review in iTunes. It only takes a minute and it’s the best way to show us you love the podcast. We’ll see you next time on Top Traders Unplugged.