Podcast: The global battle for tech supremacy
Chris Miller, author of Chip War, joins the Investor Download to discuss the battle raging for chips and tech supremacy.
Profily autorov
You can listen to the podcast by clicking the play button above. You can watch most recordings of the podcast on the Schroders Youtube channel.
You can also subscribe, download, rate and review the Investor Download via Podbean, Apple Podcasts, Spotify and other podcast players. New shows are available every other Thursday from 5pm UK time.
The full transcript is available below:
[00:00:09.10] - David Brett
Chris, a very warm welcome to the show. Thank you so much for joining us.
[00:00:12.23] - Chris Miller
Thank you for having me.
[00:00:14.16] - David Brett
Where did we find you today?
[00:00:16.20] - Chris Miller
I'm at my home office today, right outside of Boston, Massachusetts.
[00:00:20.19] - David Brett
A lovely part of the world.
[00:00:23.03] - Chris Miller
Yes, indeed.
[00:00:24.21] - David Brett
I just want to... Before we get onto your book, which is what we're here to talk about, Chip War, I just want to get to how you became involved in writing this, because from what I understand, by trade, you're a Soviet Union historian, and yet you've ended up working on a book that details chips and the chip war that's going on. How did you come about that?
[00:00:47.11] - Chris Miller
Well, I got interested in computing and in chips by looking at the history of the Cold War, when the Soviet Union and the United States were pouring money into developing more advanced weapon systems, missiles that could guide themselves. And I realised that at the core of the Cold War arms race was a computing arms race that depended on better and better chips. And as I did that research, it became clear that that race hadn't landed, that, in fact, we were living in the midst of a race to deliver better and better computing systems, both for companies that wanted better products, but also for countries that were doing the exact same thing they done in the Cold War today by trying to take the most advanced computing systems, the most advanced chips, and deploy them for strategic applications. And so the history was actually very live for the present, and that's why I decided to write Chip War.
[00:01:39.07] - David Brett
Yeah, as is the case in most things in life, a lot of these technologies stem from the military and use on the battlefield.
[00:01:48.14] - Chris Miller
Well, that's right. The first chips that were invented were essentially used in guiding either spacecraft or missiles. And there's been a long history of advances, both technological advances, but also manufacturing process advances that were either directly funded by the US military or devised for use in Defence Department programmes. I think most people, when they thought of chips, they thought of the chip inside of their PC or the chip inside of their iPhone and hadn't really realised the extent to which there was this deep relationship with the defence industrial base, too.
[00:02:21.08] - David Brett
Yeah, I just wanted to... Just for people who may not know the history of chips, obviously, you mentioned there about space exploration and missiles. But going back to the 1960s. Can you talk a little bit more about the transistors we saw there and Moore's law in particular?
[00:02:37.02] - Chris Miller
Well, before chips, computers relied on individual transistors, little switches that flip on and off and produce the ones and zeros that undergird all software, most data storage. And these transistors were powerful, but they were also flimsy, and they often had defects that caused them to break. And in the late 1950s, a number of researchers realised that if you put multiple transistors on the same block of semiconductor material, a piece of gallium or silicon, you could simultaneously make them smaller and also make them much, much more reliable. And that was the basic intuition that drove the first chips. And we've essentially been improving on these blocks of silicon with transistors carved into them from the late '50s all the way up to the present. And the advances have improved at an exponential rate. The number of transistors you can squeeze into a chip doubles every two years, roughly speaking. And that's why your smartphone today is around a million times more powerful than the computer that power the Apollo spacecraft as it flew to the moon, because we're able to put more and more computing power on tiny pieces of silicon.
[00:03:43.05] - David Brett
From what I understand, back in the '60s, there were about four transistors on a chip, and now it's up to, well, 12, 15 billion, something like that. Is that correct?
[00:03:49.24] - Chris Miller
That's right. A new smartphone, if you pop it open, just the primary chip that runs the operating system could have 10 or even 20 billion microscopic transistors carved into it. And to make them so small, the size of a virus today. So there's nothing we manufacture at that microscopic scale, but by the billions and billions every year. And the fact that we're able to do so is why chips are today applied to almost every type of device.
[00:04:17.02] - David Brett
Is it fair to say that we might not have had globalisation were it not for these semiconductors and transistors on the chips?
[00:04:24.17] - Chris Miller
Well, I think that's right. If you look at world trade today, what you find is that some of the biggest flows in world trade are actually chips that are being traded across borders. China, for example, spends more money each year importing chips than it spends importing oil. There's no other flow in global trade that's larger than the flow of chips into China and that's just one example of the ways that both chips and then the types of goods that chips have enabled, like phones, computers, are at the centre of globalisation as we know it. That also dates back many decades because it was the chip industry that, I think, pioneered the method of having offshore production and having different components and devices manufactured in multiple different countries. We take that for granted today, but that was a real innovation when it was first pioneered in the '50s, '60s, and '70s, and the chip industry was at the forefront.
[00:05:16.17] - David Brett
You mentioned China there spending more on chips than oil. I think in your book, you said processing power is the new oil. But from what you're saying, it's almost more important than oil.
[00:05:27.18] - Chris Miller
I think that's right and it's also more concentrated in its production. Saudi Arabia produces, depending on the year, 10 or 15% of the world's oil. Taiwan produces around 90% of the most advanced processor chips. So the amount of concentration in the hands of not just a single country, but also a single company, in the case of the Taiwan Semiconductor Manufacturing Company, is truly extraordinary. And so I think you can argue that companies like TSMC are among the most important companies in the world because the chips that they make are in everything, not just in computers and phones, but in cars and in washing machines and in coffee makers and in medical devices. The list goes on and on and on. The rest of the world economy depends on these absolutely critical companies.
[00:06:11.10] - David Brett
Yeah, I'm surprised the reason you're, but is it ASML? They own 100% of the lithography machines that require, without which the chips wouldn't be able to be produced.
[00:06:23.09] - Chris Miller
That's right. You've seen in the chip industry, the supply chain of companies that's needed to produce a single chip gets longer and longer every year. And the reason is that companies have got to specialise to do their specific part of the production process. And so you've got a series of points in the supply chain where there's just one or two companies that have the type of expertise needed to manufacture at virus-sized scale. And ASML, the Netherlands, is the best example of this. They've got 100% market share in the production of absolutely critical chip-making tools that are used in the manufacture of every chip that is training an AI system for example, and almost all the key processors inside of modern smartphones. And that level of concentration, again, is much higher than in many other segments that we often think of as being critical.
[00:07:12.24] - David Brett
I mean, such concentration in any supply chain is dangerous. Are policymakers or are businesses doing something about that, trying to rectify that situation?
[00:07:23.02] - Chris Miller
Well, the dilemma is that there's a good economic reason why you've got this level of concentration, because it creates efficiencies. Take ASML, for an example. They're capable of producing machines that can use rays of extreme ultraviolet light to carve nanometer scale carvings into silicon. That's billionths of a metre and they're able to do it in a way that's economically viable, so it can be used in the manufacture of your smartphone. It's not a surprise that there aren't dozens of companies that can do this. The amount of investment involved, the amount of R&D involved is just mind bogglingly large and so the concentration that exists in ASML in the Netherlands, in TSMC in Taiwan, it's rational to have concentration, from an economic perspective. But it's irrational from a perspective of supply chain resiliency, because you have single points of failure in some cases, that the entire world economy resolves around. So that's why the last couple of years we've seen in Europe, in Japan, in the United States, a number of big programmes designed to incentivise the production of chips in other locations with the aim of adding a bit of resiliency and it's expensive because we're pushing against the direction in which markets would naturally go.
[00:08:37.08] - Chris Miller
But I think political leaders have concluded it's necessary given the amount of concentration that we currently have and the way that chips are produced.
[00:08:55.17] - David Brett
Your book is entitled Chip War. I guess if there's going to be a build up to a proper chip war, it may have started with trade wars and I just wondered what effects you think the trade war, certainly the US with China, is having on this chip manufacturing and the trade of technology.
[00:09:15.04] - Chris Miller
Well, I think the trade war in chips, which has been going on now for almost half a decade, is partly about trade, but it's also partly about war, because from the perspective of governments, they're interested in which companies are selling which types of chips, but they're also interested in the strategic capabilities that are enabled by these chips. And so if you look at what are the parts of the chip industry that have been the primary focus of governments, it's been on the tools and the chips that enable artificial intelligence. And they're focused there because they believe, and I think they're probably right to believe this, that AI is not just going to be good at creating chat bots. It's also going to be used to improve the way intelligence agencies operate or to help military systems operate increasingly semi-autonomously. And the fact that it's hard to know whether a given AI system will be for civilian use only or also have military applications. The US, as well as some of its allies, have been trying to impose a bit more control over the types of chips and tools that make AI possible so that they can ensure that the military capabilities enabled by those systems don't benefit their competitors.
[00:10:22.11] - David Brett
Do you think it's working, the trade wars?
[00:10:25.18] - Chris Miller
Well, there's certainly a race underway between what I call the existing supply chain, which connects the US, Netherlands, Taiwan, and China, which is trying to build up its own domestic supply chain. If you look at China's import numbers, they're still importing vast quantities of chips from abroad, from Taiwan, from Korea, from the US. Which are imports that China wouldn't want to do if they could produce them all domestically. So we know from the fact that there's still vast imports coming that China is not as self-sufficient as it would like to be. But we also shouldn't underestimate the extraordinary engineering capability in China, which is trying very, very hard to catch up. And so we're going to see this race play out between the Western supply chain and Chinese competitors over the coming years and probably decade.
[00:11:12.16] - David Brett
The US also announced the Chips Act. I just wondered if you could just give us a brief overview of what the Chips Act is and what it's trying to achieve.
[00:11:20.06] - Chris Miller
The Chips Act was really driven by concern over concentration, and in particular, concentration of the production of advanced chips in Taiwan. So today, as I mentioned, our 90% of advanced processors, and almost 100% of the chips that are used for training AI systems are manufactured in Taiwan. The US Congress, which passed the Chips Act a couple of years ago, was concerned about this concentration given the intensification of geopolitical risks around Taiwan as China's military goes stronger and stronger every year and so that was really the impetus behind the Chips Act, was to have a bit more diversification in terms of where chips are manufactured. And so we've seen over the past couple of years, major new investments announced and construction beginning, both by US firms who are getting support from the Chips Act to build new plants in the US, but also, interestingly, from foreign firms like Samsung of South Korea and TSMC of Taiwan, who are building big new plants in the United States as well, in response to the Chips Act, which is going to help fund a portion of these new facilities. And so this is an effort to create some more diversity in the chip supply chain.
[00:12:31.06] - Chris Miller
But I think it's in its early stages still and I think Taiwan is still a very attractive place to manufacture chips. And we should assume that Taiwan is going to play a very, very big role, an absolutely central role, I think, in the chip industry for the foreseeable future.
[00:12:46.10] - David Brett
How successful do you think the Chips Act will be?
[00:12:50.15] - Chris Miller
Well, if you look at the data on investment in the chip industry in the US, there's a huge step change. Relative to the prior trend, there's been something like an order of magnitude increase of investment. So there's no doubt that the Chips Act has catalysed a lot more construction of chip-making plans in the United States. However, there's a lot of investment happening in Taiwan and in Korea, and in China as well and so I think the chip industry is going to still largely remain focused in East Asia, even though you're going to have more capacity coming online in the US and also in Japan and Europe as well.
[00:13:27.02] - David Brett
If part of all these acts and the trade wars is the gap between, most dominantly, the US and China, but China and the rest of the world. China has already shown itself to be resourceful in the way that it can facilitate and motivate its population to come up with new techniques, new technologies. How dangerous is it that the US is trying to force a decoupling with China? Could it work out bad for them?
[00:14:00.12] - Chris Miller
Well, I think actually what you see is China is trying to decouple from the US. So right now, China is coupled in the sense that it imports all these chips from Taiwan, from Korea, from the US and there's been a strategic goal that the Chinese government has been pursuing for almost a decade now to buy fewer chips from abroad. And so I think the question is, to what extent will this decoupling effort by China work? Will it be able to become completely self-sufficient and reduce its imports of chips from the West? And I think we're seeing some success in certain types of chips. But I also think we see the aggregate volume of chips remain quite large in terms of China's imports. And this is a ongoing challenge that China faces, and it's chosen a very, very difficult industry to become self-sufficient in, because today, no one else in the world is self-sufficient. US isn't self-sufficient. Japan is not self-sufficient. Taiwan is not self-sufficient. And so as China pursues increasing self-sufficiency, I think it's recognising that the costs involved and the R&D barriers involved are really extraordinary, which is why the rest of the world has chosen not to be self-sufficient, chosen instead to rely on this existing international supply chain.
[00:15:09.11] - David Brett
It feels like despite the fact this chip story goes back over half a century, possibly even further, it feels like we're only just entering almost like part two or act two of this story. Is that the way you see it?
[00:15:21.10] - Chris Miller
I think that's right. I think AI has just supercharged all of the dynamics that we've been discussing, both in terms of increasing demand demand for chips, and we see that reflected with companies like NVIDIA, but also in terms of raising the stakes, because the key driver of advances in AI technology has been the ability to apply more processing power, in other words, better chips to AI problems. And so if you talk to big AI labs or big tech companies developing AI systems, you'll hear them relentlessly focused on getting access to the largest number of the most advanced chips that they can because they think their most likely bottom-up as they try to build out bigger and better AI systems, is going to be access to the chips that they need. And so the chip war is, I think, ongoing, both in its initial iteration as a conflict between governments to deploy chips to strategic applications but also between companies, is to try to race to build bigger and better AI capable data centres.
[00:16:22.01] - David Brett
Just one final question on the chip. Is there a danger that companies will focus too much on chips and maybe forget about other stuff that matters? Because the technology just seem so prevalent now.
[00:16:32.11] - Chris Miller
Well, I think most companies have been working through a decade in which they didn't think enough about chips, both whether it's manufacturers who faced shortages during the pandemic or tech companies who have been scrambling the last 24 months to get access to all the AI processors that they need. And so for now, I think most companies are focused on making sure they've got the computing power that their applications require. And the trend is that they need more and more computing power every single a year.
[00:17:01.08] - David Brett
A long, long way to go in this story then on the chips and the chip war. Obviously, we're doing this in conjunction with the FT, the Business Book of the Year Awards, which you won in 2022. And we just asked a few questions about you as an author. Actually, this is one specific for you because I know, I think when you started out to looking into this, you were looking into Soviet Union and weapons and stuff like that. I just wonder, was there a certain point at which you decided to turn your attention to chips and away from what you were originally researching?
[00:17:41.22] - Chris Miller
Well, it was really when I realised that the history I was studying in the 1950s and '60s and '70s was being repeated in some ways today. And then, as now, chips were central to the military balance, central to the strategic balance. But today, because most chips end up in smartphones and PCs, people have forgotten about the deep interlinkages. And it seemed to me that you actually could understand the China-US relationship or the trade tensions without putting chips really at the centre of your analysis.
[00:18:12.24] - David Brett
What do you think makes an award-winning book? I don't know that I got a great number of data points from which to extrapolate.
[00:18:23.02] - Chris Miller
I guess what I really liked about researching and writing Chip War was two things. One was learning about spheres that at first seemed unconnected. And then actually, as I realised when I was doing the research, we're actually integrally interconnected because the supply chains and the demand for computing brought them together. So I didn't start planning to write a book that began with the Soviet Union in the Cold War, and then went through manufacturing operations in Malaysia, then ended up in the Taiwan Straits. But I had a lot of fun learning about those interconnections and following them as I did the research. I think the second thing that I loved was the characters involved. And I had a chance to interview well over 100 people who worked in either academia or worked in industry, worked in government around ships, and learned so much from the stories that they told and the trajectories of their careers. And so it was fun to have a chance to relay a small fraction of those stories in the course writing the book.
[00:19:26.07] - David Brett
Were there any major challenges when you were writing the book? Was there any points you found more difficult than others?
[00:19:33.08] - Chris Miller
Well, I think that the hardest thing was bringing the story all together because it did stretch over three quarters of a century and touched most of the world's continents, finding the right way to narrate the trends was a challenge, but a fun challenge. And I think in particular, a lot of the business dynamics and the technology transitions, they can get really opaque really quickly. And so it was fun to use characters to illustrate those stories that I understood them, and also, hopefully readers understood them, too.
[00:20:04.05] - David Brett
I love the book, so it definitely connected with me. Did winning the award have any impact on the reception of your book at all?
[00:20:14.09] - Chris Miller
Well, I think it probably increase the number of people who'd heard of it. I'm very appreciative of that. I think the book was a bit of foresight, but a lot of luck in terms of hitting at a time in which a lot of people were realising that they needed to understand the chip industry. And so I think the award helped with getting a lot of people who wouldn't have naturally picked a book about chips off the shelves to understand that it was actually critical for understanding US, China relations, for understanding artificial intelligence, for understanding how supply chains functioned. And so that was fun to watch people who otherwise wouldn't be interested in chips get interested through the book.
[00:20:56.15] - David Brett
One final question. What one piece of advice would you give any aspiring writers?
[00:21:04.11] - Chris Miller
Well, I think what I was always surprised by is when I would have interviews or conversations where I wasn't expecting to learn something in an entirely new category and just by asking questions and listening to stories, realised that there was an entirely new angle that I hadn't first thought of. And so I spent a lot of time doing interviews that when I started the interview, I wasn't exactly sure whether or how it would end up being useful. And then being told or explained by my interview subject as to exactly how this or that facet was really a major factor in changing the industry. And so following the evidence trail and following the interview trail was something that ended up being very useful in the research and useful in ways that I often couldn't have expected in advance.
[00:21:50.19] - David Brett
Well, it came together in a fabulous book. Chris Miller, thank you so much for joining us.
[00:21:55.17] - Chris Miller
Thank you.
[00:21:56.14] - David Brett
Thanks, Chris. That was great. Thanks so much for spending some time with me.
Profily autorov
Témy