From the Crows' Nest

Ken Miller (00:10):
Welcome to From the Crows' Nest, a podcast on electromagnetic spectrum operations, or EMSO. I'm your host, Ken Miller, director of Advocacy and Outreach for the Association of Old Crows. Thanks for listening. In this episode, we dive deeper into the topic of micro electronics to better understand the history of chip design and manufacturing and the impact that chips have had on every aspect of our lives. I welcome to the show scholar and author Chris Miller.

(00:34):
Chris teaches international history at Tufts University's Fletcher School. He also serves as Jean Kirkpatrick Visiting Fellow at the American Enterprise Institute, Eurasia director at the Foreign Policy Research Institute, and as a director of Green Mantle, a New York, London-based macroeconomic and geopolitical consultancy. You can visit his website at christophermiller.net and follow him on Twitter @crmiller1. Most importantly for this episode though, Chris recently wrote a book entitled Chip War: the Fight For The World's Most Critical Technology.

(01:07):
Before I bring him on the show, I want to say that we talk about microelectronics a lot in everyday business. It's a topic our listeners want to discuss more on the show. It's in the news. It's been the focus of key legislation before Congress, most notably the CHIPS Act. We all look forward to the next iteration of our favorite smart device, and we obviously talk a lot about Silicon Valley being the epicenter of the tech sector in the US. Obviously though, there are technology corridors in every state as well as every country around the world. Microelectronics is quite frankly ubiquitous and it underpins everything we want to do in business and military. But how often do we really stop to think, do we truly understand the complexity of microelectronic sector to make the right decisions across business, government policy and so on?

(01:56):
I would argue that we don't take the time often enough to dive into these critical topics the way we should, to truly understand them and to be honest assessors of how some of these topics have evolved over the decades. But we have to do that if we want to have a strong economy, if we want to have a productive and successful workforce, if we want to have EMS superiority or military advantage in across any war fighting domain, we need to collectively know more.

(02:22):
So, the book Chip War by Chris Miller takes us on this journey. It's a great book. It makes this topic accessible and really highlights the evolution of the electronic sector over the past 70 years. It also sheds light on the path that we need to follow to maintain our advantage in this sector. With that, I want to welcome my guest, Chris Miller. Chris, welcome to From the Crows' Nest. It's great to have you on the show. Thanks for joining me.

Chris Miller (02:47):
Thanks for having me, Ken.

Ken Miller (02:48):
All right, so as I mentioned in the introduction, this is a great book, made this very complicated issue very accessible and really helped uncover a lot of the evolution of the microelectronic sector. I wanted to start with you to talk about what inspired you to write about this, because it's not something I think that would be readily available, or readily to be inspired to write a book like this, but it's very well done.

Chris Miller (03:12):
Well, thanks. Yeah, I initially came to this topic actually not planning to write a book about semiconductors, but about military history. And the question I wanted to ask at the outset was why was it that in the early days of the Cold War, the US and the Soviet Union had what appeared like comparable capabilities to produce the most important technologies? Nuclear weapons, for example, long range missiles and space capabilities. But over the subsequent decades, the US surged ahead when it came to computing capabilities that were attached to military systems like missile guidance, for example. Whereas the Soviet Union completely failed to do the same thing, despite having great physicists and very smart scientists and putting lots of money into it.

(03:51):
I came to realize that one of the key differentiating factors was that the US was able to build a vast semiconductor industry that was eventually came to serve civilian uses, but really had its origins in military technology. Whereas the Soviet Union failed to do that. And I think that explains a lot of why the Cold War ended the way that it did with the US military victory. But it also led me to realize that semiconductors are a lot more important in shaping the modern world than most of us realize. And it's not just historically, of course, it's when you look at US and China dynamics today.

Ken Miller (04:22):
Yeah. And you mentioned in your book, quote unquote, the fate of nations has rested on the ability to harness computing power and so forth, and it's something that will likely continue to be the case moving forward. So it's really important to understand all the forces at play. So, you start off the book, and what I really liked about it was, it wasn't just like, "Hey, here's an important topic." You really start to go back at the beginning and you start the book around World War II coming out of that. And it was a time where there was a lot of industrial activity, a lot of new military technology coming out, and it was one of the key points where we started to see this shift to this understanding or exploration of computerization and microelectronics. Could you talk about that period as a kickoff to really the global security sector that we have today with microelectronics?

Chris Miller (05:13):
Yeah. One of the key challenges of the period was how do you fit computing power into small devices? And for a missile guidance system for example, there was a huge premium placed on size. You needed a small size, you needed low power consumption, you needed as much computing as you could get, and that really was a spur to the miniaturization that turned computers from something that fit in the size of a room, which is what they were in the 1940s and '50s, into something that today fits in the size of a smartphone.

(05:44):
And the military was really the key driver there, because the military was willing to pay top dollar for very small production runs. And that was a critical factor in making it possible to produce the earliest chips. And if you look from the 1950s, with the invention of the first integrated circuits all the way up to the present, what you find is that the defense department and DARPA in particular, have been key drivers of almost every single revolution in microelectronics technology over that time period. And it's because they were willing to test the frontier with ideas that weren't commercially viable when they were first pioneered, but then were picked up by commercially focused firms and turned into technologies that were used across consumer devices. You've got a really close relationship between defense demands and then the ability of US firms focused on civilian markets, to capitalize on the technologies that emerged.

Ken Miller (06:36):
And one of the things I've learned, one of the challenges that I've learned and talked about with this topic is that, and you addressed at the beginning, you mentioned it here, was this idea of you miniaturize it, but you have to produce it to scale, and you also have to have the production run that actually helps you make the profit. And early on, the government is really, you mentioned, able to pay top dollar for smaller production runs and stuff, but today ... and that really drove a lot of the development. But today maybe you almost see a little bit of the reverse going on where, with the commercial sector, but it's really hard for the government to afford. Or how does it look today where the challenge to produce to scale, and the large enough production run to actually help companies make money to advance military technology?

Chris Miller (07:18):
Yeah, it's always the problem today, is that small production runs are even more expensive than they were in the past because the economies of scale, both in designing an advanced chip and then in fabricating it are such that it's really hard to do something near the leading edge that isn't selling hundreds of thousands or millions of units. And iPhone chips and PC processors do sell in that volume, whereas defense systems don't.

(07:42):
And I think part of the challenge that the defense industry has dealt with over the past decade or so, is being more active in trying to take commercially design systems and apply them to military systems. And that's part of the solution. But I think more than that, is actually finding ways to drive down the cost of smaller batch production. And we see a lot of focus now on, for example, how do you drive down the cost of chip design? Which has gotten horrendously expensive at the leading edge. And that's going to be a key question because if you look forward, one of the key, I think drivers of innovation in this sphere, is going to be more types of chip designs to capitalize on having designs that are optimized to specific use cases. And that's only going to be possible to do if we're able to find ways to make design cheaper, so it's economically plausible to do it in more cases.

Ken Miller (08:26):
And then you have the challenge of manufacturing the chip, and it seems that over the course of decades, it's really the manufacturing of the fabs are, you only have a few that are really responsible for the large majority of manufacturing of the fabs. So, not only do you have more expensive chip design, but you have less availability to fabricate and manufacture the chip. So, how do you see that tension continuing to play out moving forward?

Chris Miller (08:55):
Yeah, it's a big challenge today, whether you're looking at the leading edge, the most leading edge, or even more generally, fabrication capacity for small batch production is limited. And the foundries are all focused primarily on their main commercial customers, not smaller volume customers, or researchers. I think one of the things that makes me more excited about the recent CHIPS Act funding as well as some of DARPA's programs, is that there's a focus now on making sure there is fabrication capacity that's open and accessible to universities, to researchers and to startups or smaller scale firms. I think that's really important because if you can prove that your chip will work in small batches, and it'll be much more plausible to take it to TSMC, or take it to one of the big foundries and convince them to produce it in larger volumes.

(09:46):
And so, we need to make sure we've got the sort of lab to fab transition more accessible to smaller batch production. And that's especially important I think, in the defense realm where batch sizes are always going to be smaller. Since you're not going to sell a hundred million iPhones, you're going to put it in a smaller number of systems.

Ken Miller (10:01):
And that seemed to be one of the forces at work early on, where the ability to take a risk and on a new fab design, without as much concern about the capacity to fabricate a large batch production that you mentioned, with government supporting it and everything. Would you say that the CHIPS Act takes a few of the best practices from earlier times and tries to update them and to address current challenges?

Chris Miller (10:30):
I think the R and D focus side of the CHIPS Act will do some of this, and I hope it does a lot of this because one of the things that's worked really well over the course of US government investment in microelectronics, is making sure we've got clear pathways from turning an idea into a prototype and a prototype into a system that can be rolled out in practice. And DARPA and other organizations, the defense department more generally, has done a lot of work over the past 70 years in funding these types of pathways and making sure that those are still available, even as the cost of fabrication goes up and even as more fabrication capacity is no longer in the US, is really very important. And that's where I think there's the most scope for the government to play a big role on the R and D, side is just making it easier to do R and D.

Ken Miller (11:18):
Now, one of the things I really enjoyed about the book was the introduction of a number of different actors, people that really shifted the debate, shifted the sector in one direction or another over the course of decades. And a lot of these, the people that you talk about in the book, they've seeped into all of our minds and all of our lives, that we never really attached, oh, this happened back then in this context. And it was very interesting to see how we've been affected by microelectronics sector for decades without sometimes without even knowing it. You spend a great deal of time, and we won't go into all of them, but I wanted to know, what are some of the key individuals or groups of people that really were game changers along the way? You mentioned the quote unquote Traitorous Eight, that included Gordon Moore that we all familiar with Moore's Law and so forth. What are some of the major actors that really created the sea change over the decades?

Chris Miller (12:13):
I think the person that stands out for me the most is someone who is not as publicly known as Gordon Moore or Bob Noyce, the other co-founder of Intel or Jack Kilby, one of the inventors integrated circuit, but a gentleman named Morris Chang, who in my view is one of the most underestimated business people or entrepreneurs of the last hundred years. He was one of the key figures in Texas Instruments, in building up that company into a real semiconductor powerhouse, especially in the 1970s into the 1980s, and played a really fundamental role in TI's ability to learn how to produce chips at scale. And TI was one of the companies that first learned how to do that.

(12:54):
And then he was passed over for the CEO job at TI in the 1980s, a great error of 20th century business history, I think. And then was given an offer by the Taiwanese government to set up a new company there. And then that company became TSMC, which is today the world's largest chip maker. And so, Morris changed career, I think, both sheds light on how it was that the US saw the emergence of the semiconductor industry in the US by honing manufacturing technology and by taking chips and applying them to vast consumer markets, but then also how the US found itself falling behind in certain metrics, especially in terms of advanced fabrication over the past decade or so.

(13:35):
And so, I think the career of Morris Chang has both profoundly transformed our lives, and everyone, every day touches a chip that his company made. It's extraordinary, the influence that he's had, but also the challenge in the fact that today, 90% of the most advanced logic chips can only be produced in Taiwan by his company, which of course brings up all sorts of complex geopolitical risks.

Ken Miller (13:59):
Right. Well, since you mentioned that, I mean, I wanted to talk with you a little bit about how the geopolitical nature of this topic, obviously a great deal of capacity rests with Taiwan, and as well as other Asian countries, Korea and Japan, and of course, China obviously plays a key role in this. One of the things I found interesting with China was its dependency on importing the manufacturing of chips. You hear in the news, you think that they hold a different place in the debate than maybe is real. Could you talk a little bit about some of the dynamics in the Asia Pacific region? It's a loaded question, but we can go on from what are some of the things that we have to keep in mind moving forward about the relationships over there, obviously with the Taiwan strait and US China relations that are going to shape this sector moving forward?

Chris Miller (14:53):
Yeah, well I think this is the big question hanging over the industry, and there's a couple dynamics at play. One is that Taiwan remains absolutely central and really irreplaceable in the short, medium term as a supplier of chips. And it's not only the most cutting edge logic chips, although it's important as a crucial supplier of those, everything from smartphones, to PCs, to data centers can barely function without chips from TSMC, but also lagging edge as well. Over one third of the computing power we add each year comes from Taiwan. So, it's a huge number of chips in everything from dishwashers to data centers.

(15:27):
China is the other factor. China is still relatively dependent on both imported chips, which they buy from the Taiwanese, South Koreans, United States, but also chip-making tools and software, which is largely imported from the US, Japan, the Netherlands, and a couple of other countries. However, China's been spending tens of billions of dollars a year over the past decade or so, to try to domesticate some of this technology. And the big question going forward is, will this work, or to what extent will it work? And if so, what will that mean economically and geopolitically?

(15:59):
And so, I think we've got an interesting dynamic with China pouring money in, the US tightening controls, trying to restrict the transfer of technology to China, and China growing its military power year after year with the primary goal of being able to better threaten Taiwan, the world's most important producer of chips. So, you really can't understand, I don't think, the geopolitical balance in Asia, or even the military balance in Asia, without understanding the deep into relationship with the world's chip supply. And I guess the final irony to note is that over the past decade, both the US military and the Chinese military have been reliant on chips produced by TSMC. We know that there are chips the US military buys from TSMC and we've got great open source evidence of ways that the Chinese military has been able to access TSMC produced chips. So, both of the world's militaries, as they engage in what can only be described as an arms race in the Taiwan straits are relying on computing power produced by a single firm on the island that is most in contest right now.

Ken Miller (16:59):
And one of the attributes of competition is increasing cost and risk of your adversary, or avoiding cost and risk on your own part. I think when you look at Taiwan's capacity to manufacture chips as well as the trade relationship between US and China, I mean, there's a lot going on in terms of what steps either country would be willing to take, that may or may not increase the risk to chip manufacturing, chip design, freedom of navigation around the Pacific. It seems to be underpinning so much of how we're thinking about moving forward, or how other countries are approaching the issue.

Chris Miller (17:39):
Yeah, I think that's right. And I think if you think about military scenarios in the Taiwan straits, and especially the scenarios that are less than a D-day style invasion, blockade scenarios or limited Chinese strike scenarios, all of them would deeply implicate the semiconductor supply chain. And in the US response to any sort of more limited scenario, the economic cost of any escalatory move would weigh very heavily on the US President's decision-making, which is why in most of the potential military scenarios in the Taiwan straits, you can't really understand the dynamics without thinking about semiconductors.

(18:15):
I think the other facet that's really important is if you look at the way the US military is thinking about how to address the fact that its position vis-a-vis China has deteriorated in the last decade, the strategy is very clearly to offset China's quantitative advantages with qualitative advantages. And the primary qualitative advantage the US has had historically is ability to deploy computing power to military systems. And the question right now is A, can the US keep or even grow its edge vis-a-vis China in terms of these foundational technologies? And then B, can they be deployed in a fast enough time horizon to defense systems, to make a difference on the battlefield?

Ken Miller (18:53):
One of the things that you discuss later in the book, there seems to be a distinction between ... obviously, when we talk globalization and global economy, there's a lot of business decisions. And then of course, you have the security strategic decisions be made by governments. And how do you see that playing out, particularly when you look at the United States or in some of our allies and the ability to influence business decisions to the betterment of the strategic values that we need to protect around the world versus some of our adversaries who don't necessarily have that distinction between commercial and public sector. The government can come in and basically shape the business decisions they feel are necessary. How does that influence the evolution of this?

Chris Miller (19:40):
Yeah, it's an interesting question. I mean, I think one of the trends that we're seeing in all countries is more politicization of the chip industry. Certainly we see this in places is like China, where the government has always been deeply involved and is even more involved. But even in the US, or Europe, or Japan, there's more government involvement today than there was a couple years ago. And for companies, this is uncomfortable. I think for government, this is a bit uncomfortable too, but it's not surprising. I don't think, given A, as you say, adversaries are doing this even more so. And B, unlike a decade or so ago, where most companies in the chip industry preferred to think of their technology being primarily, or in some cases even solely, relevant to iPhones and PCs and consumer devices. Today, it's impossible to think about semiconductors about considering their defense military ramifications.

(20:34):
And so long as that's the case, and so long as the military balance between the US and China remains uncertain, I think we're going to have to assume that governments are going to be more involved in semiconductor industries in the supply chain more generally because it's a critical national security issue. I think the chip industry is slowly adjusting to this new reality. But it is a bit of a cultural shift relative to the past couple of decades in which most people in the commercially focused chip industry could pretty safely ignore government and not have to think very much about it.

Ken Miller (21:06):
So, underpinning a lot of the discussions of course is the need for really smart, innovative minds, really skilled people to fabricate design and fabricate chips. Obviously, you need a skilled work workforce. And so, I want to spend a few minutes just talking about the workforce dynamic because the workforces of each of the respective countries that are involved in this around the world, they bring a various perspective to the conversation. Talk a little bit about some of the advantages and challenges of going on in the workforce, US and some of our allies, versus some of our adversaries.

Chris Miller (21:45):
Well, it is a big challenge for the US, especially in parts of the chip production process where the US has been playing a less substantial role than the past. And especially when you look at the fabrication stage of making a semiconductor, the US market share here has declined really dramatically over the past three decades, and as a result, the workforce has declined as well. I think if you look at the pipelines of new workers into this field, you find that there are gaps there as well. If you chart, for example, the number of electrical, electrical engineering graduates from PhD programs versus computer science graduates, you'll find a huge shift in recent decades towards computer science. And that's not a bad thing per se, but we need to make sure we've got the electrical engineers and the material science experts and others that are needed for semiconductor production too.

(22:34):
So one of the things that I think that CHIPS and Science Act will do is put a bit more focus on making sure we have pipelines in place, from picking up first year students in college, showing them where to get the internships that they need, and then helping them understand career paths ahead of them, so that we get the workers with the expertise that is needed in the chip industry, and especially in the fabrication step, which is where there's going to be a lot more demand, I think, in the future than was previously expected.

Ken Miller (23:01):
One of the passages that I really, really liked was you were talking about Sony's Akio Morita, and he made a great statement where ... and you quote him saying, "The United States has been busy creating lawyers while Japan has been busier creating engineers," and basically looking how Japan, and I think a lot of Asian cultures look ... have a much more long-range view of thinking sometimes about career development, workforce development, and so forth. What are some of the lessons that we can learn from them in terms to build up US workforce or generally, or how do we address some of the challenges moving forward based on what we've learned about the sector through your research?

Chris Miller (23:47):
Well, one of the things that I came to realize as I did the research for this book and that I hope the book helps make clear for people who are not semiconductor experts, is really how transformative and cool this technology is. I think for most of US society, when people think of tech, they think of social media or consumer internet firms, and obviously Facebook and TikTok are important, I guess, but far more interesting to me, and I think when you actually dig into it to most people, is the hardware on which all of the internet relies and which social media wouldn't exist. And so, what I wanted to do in this book was to explain, "Hey, this is really important. It's really cool. There's fascinating technology involved."

(24:29):
And I think as we help people understand how critical microelectronics are in their daily lives and how fascinating the underlying science and technology is, I hope that does reopen people's definition of technology. So, they're not just thinking about Facebook, but they're thinking about actually the chips on which all of Facebook's ... all the likes and posts actually depend.

Ken Miller (24:50):
Well, that is all the time we have for this show. Chris, it's great to have you on From the Crows' Nest to talk about this, and we look forward to having you back on from time to time, maybe to keep us informed about how this sector's moving forward. Thank you for joining me.

Chris Miller (25:05):
Great. Well, thanks for having me, Ken.

Ken Miller (25:06):
That will conclude this episode of From the Crows' Nest. I want to thank my guest, Chris Miller, for joining me. Again, his book is Chip War: The Fight For The World's Most Critical Technology. You can visit his website at christophermiller.net and follow him on Twitter @crmiller1. Also, don't forget to review, share, and subscribe to this podcast. We always enjoy hearing from our listeners, so please take some time to let us know how we're doing. That's it for today. Thank you for listening.