1 00:00:01,560 --> 00:00:05,080 Speaker 1: Welcome to Crash Course, a podcast about business, political, and 2 00:00:05,160 --> 00:00:08,600 Speaker 1: social disruption and what we can learn from it. I'm 3 00:00:08,640 --> 00:00:13,960 Speaker 1: Tim O'Brien. Today's crash Course, chip Wars and the most 4 00:00:14,000 --> 00:00:18,520 Speaker 1: Important company You've Never heard of. Big global collisions are 5 00:00:18,560 --> 00:00:23,880 Speaker 1: always about power and often about money. Politics, religion, nationalism 6 00:00:24,160 --> 00:00:27,520 Speaker 1: and egos play their parts, of course, but money and 7 00:00:27,560 --> 00:00:31,400 Speaker 1: how money gets made goes a long way toward explaining things. 8 00:00:32,080 --> 00:00:35,400 Speaker 1: Once upon a time, trade, wealth and power centered around 9 00:00:35,440 --> 00:00:40,320 Speaker 1: things like silk, silver, gold, spices, sugar, tea, tobacco, cotton 10 00:00:40,640 --> 00:00:44,920 Speaker 1: and the like. After that, the Industrial Revolution was built 11 00:00:44,960 --> 00:00:48,479 Speaker 1: on the back of coal, steel, and of course oil, 12 00:00:49,280 --> 00:00:53,160 Speaker 1: and in our era, the era of technological revolutions built 13 00:00:53,200 --> 00:00:57,760 Speaker 1: on bits and bytes. Semiconductors or chips, are the coin 14 00:00:57,840 --> 00:01:03,480 Speaker 1: of the realm. Semiconductor are tiny silicon wafers crisscrossed by millions, 15 00:01:03,640 --> 00:01:10,080 Speaker 1: sometimes billions, of microscopic transistors that store and process digitized information. 16 00:01:10,959 --> 00:01:15,720 Speaker 1: They allow for massive, accessible, and ubiquitous computing, and they 17 00:01:15,760 --> 00:01:19,360 Speaker 1: are such stuff as dreams are made on. They're pivotal 18 00:01:19,360 --> 00:01:23,640 Speaker 1: components of the burgeoning artificial intelligence industry, and they now 19 00:01:23,760 --> 00:01:28,840 Speaker 1: power everything from cars and appliances to missile defense systems 20 00:01:29,040 --> 00:01:33,279 Speaker 1: and nuclear weaponry. Chips are so central to the consumer world, 21 00:01:33,720 --> 00:01:37,600 Speaker 1: the shape of economies, and national security that they also 22 00:01:37,680 --> 00:01:42,600 Speaker 1: inform geopolitical maneuvering and the headbutting between China and the 23 00:01:42,720 --> 00:01:46,640 Speaker 1: US for global dominance. Who else would you bring on 24 00:01:46,720 --> 00:01:49,680 Speaker 1: to talk about the meaning of the sprawling, fascinating and 25 00:01:49,800 --> 00:01:53,200 Speaker 1: complex collision other than the author of a book titled 26 00:01:53,360 --> 00:01:58,200 Speaker 1: appropriately Chip Wars, The Fight for the World's most Crucial Technology. 27 00:01:59,040 --> 00:02:01,960 Speaker 1: In addition to writing an illuminating book, Chris Miller is 28 00:02:02,000 --> 00:02:04,680 Speaker 1: also a historian at the Fletcher School of Law and 29 00:02:04,720 --> 00:02:09,920 Speaker 1: Diplomacy at Tufts University, and he joins us here today. Welcome, Chris, 30 00:02:10,520 --> 00:02:13,320 Speaker 1: Thank you for having me so talk to us a 31 00:02:13,360 --> 00:02:17,160 Speaker 1: little bit as we start things off. Chris, about how 32 00:02:17,240 --> 00:02:21,400 Speaker 1: you wound up being such an interesting geek and so 33 00:02:21,639 --> 00:02:24,600 Speaker 1: fascinated by the minutia of Silicon Valley. 34 00:02:25,320 --> 00:02:29,680 Speaker 2: Well, when I started writing a book on semiconductors seven 35 00:02:29,720 --> 00:02:32,160 Speaker 2: years ago, I knew very little about them. I've got 36 00:02:32,160 --> 00:02:36,040 Speaker 2: no background in octical engineering or semiconductor of physics. But 37 00:02:36,080 --> 00:02:38,600 Speaker 2: I realized that there were a couple of things that 38 00:02:38,680 --> 00:02:42,160 Speaker 2: made chips unique, and I wanted to understand how they 39 00:02:42,200 --> 00:02:45,560 Speaker 2: came to be One was that I knew that semi 40 00:02:45,600 --> 00:02:49,080 Speaker 2: conductors improved at a rate more rapidly than basically any 41 00:02:49,080 --> 00:02:53,399 Speaker 2: other product in human history, doubling in capability every year 42 00:02:53,520 --> 00:02:55,280 Speaker 2: or two of what's known as Morre's law, and I 43 00:02:55,360 --> 00:02:58,000 Speaker 2: want to understand, well, how is that possible to sustain 44 00:02:58,480 --> 00:03:00,760 Speaker 2: that would have increased for so long. The second thing 45 00:03:00,800 --> 00:03:04,760 Speaker 2: I wanted to understand was how it was that China, 46 00:03:04,919 --> 00:03:08,680 Speaker 2: which I thought of as a manufacturing superpower, spent just 47 00:03:08,720 --> 00:03:12,239 Speaker 2: as much money importing chips as it spent importing oil. 48 00:03:12,280 --> 00:03:14,480 Speaker 2: And it seemed to me you couldn't understand the structure 49 00:03:14,520 --> 00:03:17,959 Speaker 2: of the world economy without making sense of how that 50 00:03:18,040 --> 00:03:21,120 Speaker 2: could be. And the third thing that drove me to 51 00:03:21,440 --> 00:03:24,200 Speaker 2: write a book on this topic was trying to understand 52 00:03:24,240 --> 00:03:27,400 Speaker 2: how it was that such a critical technology was defined 53 00:03:27,520 --> 00:03:30,640 Speaker 2: by just a couple of companies and just really a 54 00:03:30,680 --> 00:03:34,400 Speaker 2: handful of countries, And that seemed counterintuitive. We think of 55 00:03:34,480 --> 00:03:37,920 Speaker 2: competition spurring technological advances, but in the chip industry there's 56 00:03:37,920 --> 00:03:40,680 Speaker 2: some competition for sure. There's also a lot of concentration, 57 00:03:41,040 --> 00:03:43,560 Speaker 2: and all those things seemed like interesting puzzles to explore, 58 00:03:43,840 --> 00:03:45,560 Speaker 2: and that's what first got me interested in chips. 59 00:03:46,200 --> 00:03:48,520 Speaker 1: So had sort of the core of your work been 60 00:03:48,760 --> 00:03:54,440 Speaker 1: more involved with geopolitics and other intangible things before you 61 00:03:54,520 --> 00:03:55,480 Speaker 1: focused on chips. 62 00:03:55,760 --> 00:03:58,240 Speaker 2: Yeah, that's right. My background is in the intersection of 63 00:03:58,240 --> 00:04:01,720 Speaker 2: economic history and geopo politics, and that's what my academic 64 00:04:01,760 --> 00:04:04,920 Speaker 2: work had focused on. And I hadn't previously dug into 65 00:04:05,000 --> 00:04:07,800 Speaker 2: the history of a physical product like chips. But when 66 00:04:07,840 --> 00:04:10,320 Speaker 2: I started to do so, I realized both that you 67 00:04:10,320 --> 00:04:14,360 Speaker 2: couldn't understand how chips emerged without the geopolitical background, but 68 00:04:14,400 --> 00:04:18,599 Speaker 2: also you could understand either the geopolitical situation today with 69 00:04:18,880 --> 00:04:23,120 Speaker 2: US China tensions being the predominant factor, or actually over 70 00:04:23,120 --> 00:04:25,920 Speaker 2: the past half century without putting the development of computing 71 00:04:25,960 --> 00:04:27,279 Speaker 2: power at the center. 72 00:04:28,040 --> 00:04:30,200 Speaker 1: So let's start a little bit with that, about the 73 00:04:30,240 --> 00:04:33,520 Speaker 1: development of computer power and the history of chips themselves. 74 00:04:33,800 --> 00:04:36,080 Speaker 1: There are a lot of companies and names involved here, 75 00:04:36,160 --> 00:04:39,240 Speaker 1: so for clarity and efficiency's sake, I might try to 76 00:04:39,279 --> 00:04:42,200 Speaker 1: circumscribe some of the things we talk about. But let's 77 00:04:42,200 --> 00:04:44,479 Speaker 1: go back to the nineteen fifties for lack of a 78 00:04:44,520 --> 00:04:47,360 Speaker 1: better starting period, because some of this could go back 79 00:04:47,360 --> 00:04:50,760 Speaker 1: to the nineteenth century. But let's talk about the development 80 00:04:50,760 --> 00:04:54,640 Speaker 1: of transistors and this little collection of physicists who first 81 00:04:54,640 --> 00:04:59,240 Speaker 1: started in experimenting in the late nineteen fifties with transistors 82 00:04:59,240 --> 00:05:02,320 Speaker 1: and wound up winning Noble prizes. But tell us a 83 00:05:02,320 --> 00:05:05,719 Speaker 1: little bit about the roots of chips in the transistor era. 84 00:05:06,440 --> 00:05:09,839 Speaker 2: Well, the first transistor was invented in the late nineteen forties, 85 00:05:10,279 --> 00:05:13,560 Speaker 2: and it was clear it'd be useful for amplifying signals 86 00:05:13,560 --> 00:05:16,440 Speaker 2: like radio signals. That's probably why you've heard of a transistor. Radio. 87 00:05:16,520 --> 00:05:18,640 Speaker 2: That was one of the initial use cases. And it 88 00:05:18,640 --> 00:05:21,640 Speaker 2: wasn't really until the mid fifties or late fifties that 89 00:05:21,960 --> 00:05:24,960 Speaker 2: people realized transistors could be used for computing as well. 90 00:05:25,000 --> 00:05:28,400 Speaker 2: They could turn circuits on and off, and in doing 91 00:05:28,440 --> 00:05:31,200 Speaker 2: so produced ones and zeros ones when the circuit was on, 92 00:05:31,800 --> 00:05:35,320 Speaker 2: zero's when they were off. But individual transistors in the 93 00:05:35,360 --> 00:05:39,280 Speaker 2: nineteen fifties were pretty flimsy devices that often broke or 94 00:05:39,360 --> 00:05:42,200 Speaker 2: burned out, and so it wasn't until at the end 95 00:05:42,200 --> 00:05:45,160 Speaker 2: of the nineteen fifties Texas Instruments and Texas and then 96 00:05:45,200 --> 00:05:49,400 Speaker 2: Fairchild Semiconductor in California learned how to put multiple transistors 97 00:05:49,400 --> 00:05:51,320 Speaker 2: in the same piece of silicon, and that was a 98 00:05:51,440 --> 00:05:53,920 Speaker 2: revolution in quality control because you didn't have to worry 99 00:05:53,920 --> 00:05:56,279 Speaker 2: about the wires that connected them. They were just built 100 00:05:56,320 --> 00:05:58,200 Speaker 2: in the same piece of material. 101 00:05:58,880 --> 00:06:01,719 Speaker 1: And their goal in doing that, cris was what like 102 00:06:01,760 --> 00:06:04,200 Speaker 1: what was their holy grail as they sat around trying 103 00:06:04,240 --> 00:06:07,240 Speaker 1: to put the world onto a chip. 104 00:06:08,480 --> 00:06:13,000 Speaker 2: The goal was to make computing more reliable, more inexpensive, 105 00:06:13,240 --> 00:06:16,480 Speaker 2: and to allow the components to be shrunk down in size, 106 00:06:16,760 --> 00:06:19,320 Speaker 2: because before that point you had to wire every individual 107 00:06:19,360 --> 00:06:22,160 Speaker 2: transistor together and so you had a jungle of wires. 108 00:06:22,160 --> 00:06:23,960 Speaker 2: But if you built them all into the chip, it 109 00:06:24,080 --> 00:06:28,200 Speaker 2: was much more straightforward, much simpler, and could be mass manufactured. 110 00:06:28,880 --> 00:06:31,040 Speaker 1: And did they see that as a commercial thing? Did 111 00:06:31,040 --> 00:06:34,320 Speaker 1: they see that, as you know, computing that big industry 112 00:06:34,360 --> 00:06:37,320 Speaker 1: and big government would use, and that very little of 113 00:06:37,320 --> 00:06:39,200 Speaker 1: that would wind up in the hand of individuals. 114 00:06:39,960 --> 00:06:42,560 Speaker 2: You know, The initial use case was for missile guidance systems. 115 00:06:42,839 --> 00:06:45,440 Speaker 2: The Pentagon and NASA were the initial buyers because they 116 00:06:45,440 --> 00:06:49,080 Speaker 2: had big budgets and were willing to spend large sums 117 00:06:49,200 --> 00:06:52,520 Speaker 2: for low production volumes in the early stage. But just 118 00:06:52,560 --> 00:06:54,480 Speaker 2: from the first couple of years after the first chips 119 00:06:54,480 --> 00:06:56,840 Speaker 2: were invented, there were already people like Bob Noyce, who 120 00:06:56,880 --> 00:07:00,479 Speaker 2: was one of the founders of Fairchild Semiconductor, who realized 121 00:07:00,520 --> 00:07:03,920 Speaker 2: that if you projected the technology forwards, you could see 122 00:07:03,920 --> 00:07:07,240 Speaker 2: that costs would come down, that computing power would increase, 123 00:07:07,320 --> 00:07:09,920 Speaker 2: and that if that happened, there'd be many more uses 124 00:07:09,920 --> 00:07:13,440 Speaker 2: for this device, including in civilian and consumer applications. 125 00:07:14,000 --> 00:07:18,040 Speaker 1: So Chris semiconductor innovation was a particularly US phenomenon in 126 00:07:18,080 --> 00:07:23,680 Speaker 1: its early days and into the transformative technology boom that 127 00:07:23,760 --> 00:07:26,680 Speaker 1: chips helped usher in. Why was that what made the 128 00:07:26,760 --> 00:07:30,040 Speaker 1: US such a perfect peatrie dish for all of this 129 00:07:30,200 --> 00:07:34,239 Speaker 1: between say, the late nineteen fifties in the nineteen nineties. 130 00:07:34,640 --> 00:07:37,920 Speaker 2: And there were a couple factors. One was that the 131 00:07:38,040 --> 00:07:40,640 Speaker 2: US government did put a lot of money beyond the 132 00:07:40,680 --> 00:07:43,640 Speaker 2: chip industry in the early stages. The Pentagon and NASA 133 00:07:43,680 --> 00:07:47,160 Speaker 2: were big buyers of chips the nineteen fifty sixty seventies, 134 00:07:47,640 --> 00:07:49,640 Speaker 2: and so that set the industry in motion. But then 135 00:07:49,720 --> 00:07:54,679 Speaker 2: the companies that defined early Silicon Valley managed to scale 136 00:07:54,800 --> 00:07:57,400 Speaker 2: up not by selling to the government, but by selling 137 00:07:57,440 --> 00:07:59,800 Speaker 2: to a vast consumer market, and the US have the 138 00:07:59,800 --> 00:08:02,720 Speaker 2: world world's biggest consumer market. And not only that, they 139 00:08:03,080 --> 00:08:05,440 Speaker 2: learned to take the products that worked in the US 140 00:08:05,480 --> 00:08:08,520 Speaker 2: consumer market and sell them globally, and so the US 141 00:08:08,600 --> 00:08:10,760 Speaker 2: was probably the best place in the world to scale 142 00:08:10,840 --> 00:08:14,520 Speaker 2: and that was absolutely critical to the proliferation of this technology. 143 00:08:14,920 --> 00:08:18,000 Speaker 1: It's sort of an interesting lesson about how innovation occurs, 144 00:08:18,040 --> 00:08:20,560 Speaker 1: isn't it, Because when we talk about innovation, are we 145 00:08:20,640 --> 00:08:24,760 Speaker 1: talking about primarily a private sector phenomenon? Does it involve 146 00:08:24,800 --> 00:08:28,880 Speaker 1: public sector funding? Are there overlaps? You know? What does 147 00:08:28,920 --> 00:08:31,720 Speaker 1: it say about how technological innovation occurs? 148 00:08:32,440 --> 00:08:35,240 Speaker 2: You know, I think I've over the course of researching 149 00:08:35,240 --> 00:08:36,880 Speaker 2: the book, I actually came to think that there's two 150 00:08:36,880 --> 00:08:41,000 Speaker 2: different types of innovation. There's devising a new product, which 151 00:08:41,040 --> 00:08:44,280 Speaker 2: requires understanding the engineering behind it, for example, but then 152 00:08:44,320 --> 00:08:47,560 Speaker 2: there's learning to scale it in an efficient way. And 153 00:08:47,679 --> 00:08:51,440 Speaker 2: in the chip industry, the hardest thing is often learning 154 00:08:51,520 --> 00:08:54,960 Speaker 2: to scale something. Learning to mass manufacture the initial engineering 155 00:08:55,000 --> 00:08:57,520 Speaker 2: is tricky, but it's the companies that can scale most 156 00:08:57,520 --> 00:09:01,120 Speaker 2: efficiently that are often the winners. And that's something that 157 00:09:01,200 --> 00:09:05,480 Speaker 2: I think when we think of innovation, we often overemphasize 158 00:09:05,520 --> 00:09:08,280 Speaker 2: the importance of the individual scientists tinkering away in their 159 00:09:08,360 --> 00:09:11,480 Speaker 2: lab and underemphasize, well, what does it take to actually 160 00:09:11,559 --> 00:09:14,280 Speaker 2: mass manufacture that in a way that's cost effective? And 161 00:09:14,280 --> 00:09:17,040 Speaker 2: that type innovation has been just as important as the 162 00:09:17,080 --> 00:09:19,000 Speaker 2: individual inventions. 163 00:09:18,559 --> 00:09:21,840 Speaker 1: So innovation in the production process itself. 164 00:09:21,760 --> 00:09:22,920 Speaker 2: That's right exactly. 165 00:09:23,280 --> 00:09:26,400 Speaker 1: And of course, like Landmark, US chip makers came out 166 00:09:26,400 --> 00:09:30,440 Speaker 1: of this brew the early leaders like Texas Instruments and 167 00:09:30,559 --> 00:09:35,080 Speaker 1: fair Child, but ultimately one of the big five hundred 168 00:09:35,160 --> 00:09:39,200 Speaker 1: pound gorillas in the whole industry was Intel. You know, 169 00:09:39,240 --> 00:09:44,800 Speaker 1: it eventually lapsed Texas Instruments. Intel inside became a branding 170 00:09:45,640 --> 00:09:48,920 Speaker 1: coup for them in the personal computer era in the 171 00:09:49,000 --> 00:09:51,480 Speaker 1: nineteen nineties. That really took off in the eighties and nineties. 172 00:09:52,240 --> 00:09:55,960 Speaker 1: What made Intel such a world beater? And maybe we 173 00:09:55,960 --> 00:09:57,800 Speaker 1: can also get into a little bit about Gordon Moore 174 00:09:57,800 --> 00:09:59,520 Speaker 1: and Moore's low since you brought that up at the 175 00:09:59,520 --> 00:10:02,400 Speaker 1: top of the sh show, But is there something about 176 00:10:02,520 --> 00:10:05,839 Speaker 1: Intel's own trajectory as a company that's instructive. 177 00:10:06,480 --> 00:10:09,360 Speaker 2: I think what made Intel so extraordinary in the early 178 00:10:09,400 --> 00:10:12,120 Speaker 2: days of Intel was that it reinvented itself a number 179 00:10:12,160 --> 00:10:15,439 Speaker 2: of different times. It was founded as a firm that 180 00:10:15,480 --> 00:10:18,120 Speaker 2: would make memory chips. That was the primary type of 181 00:10:18,160 --> 00:10:20,360 Speaker 2: chip in the sixties and seventies, and it did that 182 00:10:20,440 --> 00:10:24,160 Speaker 2: for some time, but it also on the side began 183 00:10:24,800 --> 00:10:28,160 Speaker 2: investing in microprocessors, a type of chip that processes data, 184 00:10:28,200 --> 00:10:31,160 Speaker 2: which was a real niche market in the seventies. But 185 00:10:31,200 --> 00:10:36,920 Speaker 2: Intel pioneered the first commercial microprocessor and slowly began building 186 00:10:37,000 --> 00:10:39,880 Speaker 2: up that business. And in the nineteen eighties, when the 187 00:10:39,920 --> 00:10:42,520 Speaker 2: memory business was in a slump thanks to a lot 188 00:10:42,559 --> 00:10:47,000 Speaker 2: more intensified competition, Intel decided to completely jettison his memory business, 189 00:10:47,040 --> 00:10:49,199 Speaker 2: get out of the industry that it was founded on, 190 00:10:49,480 --> 00:10:52,320 Speaker 2: and focus entirely on microprocessors, and that was a really 191 00:10:52,320 --> 00:10:55,240 Speaker 2: wrenching change. One employee said that it was like Ford 192 00:10:55,280 --> 00:10:58,680 Speaker 2: deciding not to produce cars. Got a huge cultural shift 193 00:10:59,040 --> 00:11:01,800 Speaker 2: that was real risk because the time the microprocessor market 194 00:11:01,840 --> 00:11:07,320 Speaker 2: was tiny, but Intel that effectively that intelligently microprocessor demand 195 00:11:07,360 --> 00:11:09,440 Speaker 2: growing and growing, and the reason it did grow was 196 00:11:09,440 --> 00:11:13,440 Speaker 2: because Intel managed to get its microprocessor designed into the 197 00:11:13,480 --> 00:11:16,880 Speaker 2: first PC, and so from the earliest days of the 198 00:11:16,880 --> 00:11:21,200 Speaker 2: computer industry there was a symbiosis between Intel's chips and 199 00:11:21,240 --> 00:11:25,760 Speaker 2: the Windows operating system, and almost every PC had to 200 00:11:25,800 --> 00:11:29,199 Speaker 2: be compatible, which meant that either Intel or its competitor AMD, 201 00:11:29,320 --> 00:11:32,920 Speaker 2: which had similar IP rights, were a duopoly that controlled 202 00:11:32,960 --> 00:11:35,560 Speaker 2: the market for producing chips for PCs. 203 00:11:36,040 --> 00:11:39,040 Speaker 1: Let's just define More's law for our listeners, because it 204 00:11:39,160 --> 00:11:41,800 Speaker 1: was a rule that Intel tried to live by, and 205 00:11:41,800 --> 00:11:45,920 Speaker 1: it became a shibbalith for Silicon Valley. Gordon Moore was 206 00:11:45,920 --> 00:11:48,600 Speaker 1: both an engineer and then later a chief executive. I 207 00:11:48,600 --> 00:11:51,439 Speaker 1: think of Intel, what was Moore's law. 208 00:11:51,679 --> 00:11:54,000 Speaker 2: Well more so really a prediction at first, rather than 209 00:11:54,040 --> 00:11:56,960 Speaker 2: a law. Gordon Moore wrote an article in nineteen sixty five, 210 00:11:57,120 --> 00:11:59,559 Speaker 2: so just seven years after the first chip was invented, 211 00:12:00,000 --> 00:12:03,280 Speaker 2: deserving that the number of transistors per ship was doubling 212 00:12:03,320 --> 00:12:06,439 Speaker 2: every year or two, and therefore the computing capabilities of 213 00:12:06,480 --> 00:12:09,199 Speaker 2: chips was doubling at the same rate, and he predicted 214 00:12:09,200 --> 00:12:11,520 Speaker 2: that would continue for at least a decade through nineteen 215 00:12:11,559 --> 00:12:14,480 Speaker 2: seventy five. And it turns out that, in no small 216 00:12:14,520 --> 00:12:17,840 Speaker 2: part thanks to his work running R and D at Intel, 217 00:12:18,280 --> 00:12:20,920 Speaker 2: Morrislow continues all the way up to this day. We've 218 00:12:20,960 --> 00:12:24,640 Speaker 2: had basically a doubling every two years on a schedule 219 00:12:24,679 --> 00:12:27,800 Speaker 2: from nineteen sixty five to present. And there's no industry 220 00:12:28,120 --> 00:12:30,160 Speaker 2: really in all of human history that's tipped up that 221 00:12:30,240 --> 00:12:31,320 Speaker 2: rate of growth for so long. 222 00:12:32,200 --> 00:12:34,960 Speaker 1: So they had an intense R and D apparatus, and 223 00:12:34,960 --> 00:12:37,239 Speaker 1: they were well managed, and they were willing to cannibalize 224 00:12:37,240 --> 00:12:41,400 Speaker 1: existing businesses to find growth, and they spawned a whole 225 00:12:41,480 --> 00:12:44,559 Speaker 1: or helped spawn a whole cluster of companies in Silicon Valley, 226 00:12:44,600 --> 00:12:49,959 Speaker 1: from Qualcommon Broadcom to Nvidia and and and D and 227 00:12:50,040 --> 00:12:52,560 Speaker 1: maybe one of the only biggies that's not based in 228 00:12:52,600 --> 00:12:56,480 Speaker 1: Silicon Valley is Micron. I think they're in Boise. What 229 00:12:56,600 --> 00:13:00,200 Speaker 1: links all of those Silicon Valley companies together that were 230 00:13:00,200 --> 00:13:05,440 Speaker 1: building out all of this chip capacity and innovation and 231 00:13:05,880 --> 00:13:08,000 Speaker 1: new products that were linked to all of this. 232 00:13:08,520 --> 00:13:10,560 Speaker 2: Well, the first thing that linked them together was their workforce. 233 00:13:10,679 --> 00:13:12,559 Speaker 2: One of the things that you find in the early 234 00:13:12,679 --> 00:13:15,640 Speaker 2: days of industry is that many people would be constantly 235 00:13:15,679 --> 00:13:19,280 Speaker 2: switching jobs between the different companies you mentioned, and that 236 00:13:19,360 --> 00:13:24,079 Speaker 2: let them share know how share technology produced a lot 237 00:13:24,080 --> 00:13:27,360 Speaker 2: of lawsuits between companies and former employees in the process, 238 00:13:27,520 --> 00:13:30,360 Speaker 2: but it ended up being a real catalyst for innovation 239 00:13:30,600 --> 00:13:34,520 Speaker 2: because it meant that these advances quickly disseminated throughout the industry, 240 00:13:34,920 --> 00:13:38,400 Speaker 2: and because the technology was advancing so rapidly, in the end, 241 00:13:38,520 --> 00:13:42,120 Speaker 2: companies didn't worry all that much about their technology leaking 242 00:13:42,200 --> 00:13:45,080 Speaker 2: because if it leaked in the time it took to 243 00:13:45,120 --> 00:13:47,240 Speaker 2: leak your competitors were still a couple of years behind you, 244 00:13:47,679 --> 00:13:49,480 Speaker 2: and so for companies they didn't tell they knew, they 245 00:13:49,520 --> 00:13:51,480 Speaker 2: just had to race ahead as fast as it could. 246 00:13:51,559 --> 00:13:53,319 Speaker 2: So that was one factor. The other factor was that 247 00:13:53,720 --> 00:13:57,319 Speaker 2: as the industry got more complex, this supply chain began 248 00:13:57,360 --> 00:13:59,559 Speaker 2: to split apart. So the earliest firms they did all 249 00:13:59,640 --> 00:14:01,959 Speaker 2: their work in house. They designed their software and house, 250 00:14:02,000 --> 00:14:04,600 Speaker 2: they purified their chemicals and house, they built their tools 251 00:14:04,640 --> 00:14:09,080 Speaker 2: in house. But as the level of manufacturing sophistication increased, 252 00:14:09,600 --> 00:14:13,120 Speaker 2: firms that specialized in one part of the production process emerged, 253 00:14:13,440 --> 00:14:15,920 Speaker 2: and most of those firms emerged in Silicon Valley. And 254 00:14:15,920 --> 00:14:18,840 Speaker 2: so today in Silicon Valley there are a number of 255 00:14:18,840 --> 00:14:23,240 Speaker 2: precision tool makers that manufacture the equipment that makes chips, 256 00:14:23,520 --> 00:14:25,840 Speaker 2: and they sell to the entire industry. Every company that 257 00:14:25,880 --> 00:14:28,680 Speaker 2: wants to manufactured chips has to buy from this small 258 00:14:28,720 --> 00:14:31,720 Speaker 2: set of companies, and that is all still today largely 259 00:14:31,720 --> 00:14:33,000 Speaker 2: headquartered in Silicon Valley. 260 00:14:33,800 --> 00:14:38,040 Speaker 1: So it's a very unique economic ecosystem in the economic 261 00:14:38,160 --> 00:14:41,120 Speaker 1: history of the world. These are sort of rare and 262 00:14:41,200 --> 00:14:46,720 Speaker 1: revolutionary moments when this kind of an ecosystem builds out correct. 263 00:14:46,760 --> 00:14:50,960 Speaker 1: It's not entirely frequent, and when it really comes together, 264 00:14:50,960 --> 00:14:52,000 Speaker 1: it's revolutionary. 265 00:14:52,120 --> 00:14:54,360 Speaker 2: Right, that's right. I think you could look at maybe 266 00:14:54,440 --> 00:14:57,520 Speaker 2: Detrite and Autos as another example in American history, but 267 00:14:57,560 --> 00:15:00,320 Speaker 2: there aren't that many of them. If if you were 268 00:15:00,360 --> 00:15:02,640 Speaker 2: a chip firm that wasn't in Silicon Valley, and there 269 00:15:02,640 --> 00:15:04,720 Speaker 2: have been a couple of successful ones, the barriers to 270 00:15:04,840 --> 00:15:08,000 Speaker 2: entry were much higher and the pathways to success were 271 00:15:08,160 --> 00:15:08,920 Speaker 2: much narrower. 272 00:15:09,400 --> 00:15:11,640 Speaker 1: Another development of this time you've sort of touched on 273 00:15:11,720 --> 00:15:14,400 Speaker 1: it is that a lot of the chip firms began 274 00:15:14,560 --> 00:15:19,760 Speaker 1: outsourcing production. If I'm oversimplifying, correct me. But they kept 275 00:15:19,800 --> 00:15:24,360 Speaker 1: design in house, but fabrication facilities or fabs as the 276 00:15:24,400 --> 00:15:28,800 Speaker 1: industry calls them, began to be separated for lots of 277 00:15:28,880 --> 00:15:33,600 Speaker 1: companies from the design process, and being fabulous for a 278 00:15:33,640 --> 00:15:37,040 Speaker 1: period of time was considered a competitive advantage. 279 00:15:37,080 --> 00:15:40,200 Speaker 2: Correct, Yeah, that's right. And in the early days of 280 00:15:40,200 --> 00:15:42,160 Speaker 2: the industry you couldn't do one without the other. You 281 00:15:42,200 --> 00:15:44,280 Speaker 2: had to know how to design to do the manufacturing, 282 00:15:44,280 --> 00:15:47,600 Speaker 2: and vice versa. But a couple decades ago, new software 283 00:15:47,600 --> 00:15:51,240 Speaker 2: tools emerged that made chip designed essentially a process of 284 00:15:51,280 --> 00:15:55,200 Speaker 2: computer programming, whereas the manufacturing still involved all of the 285 00:15:55,280 --> 00:15:59,520 Speaker 2: chemicals and the tools and the sophisticated production processes and 286 00:15:59,640 --> 00:16:02,400 Speaker 2: diss plans became so different that there's actually not much 287 00:16:02,440 --> 00:16:05,400 Speaker 2: that linked them anymore. And so companies realized that if 288 00:16:05,400 --> 00:16:08,080 Speaker 2: they specialized, they could get very good at whatever they 289 00:16:08,160 --> 00:16:10,880 Speaker 2: focused on and not have to deal with the rest 290 00:16:10,880 --> 00:16:14,720 Speaker 2: of the industry. And for companies that's focused on ship design, 291 00:16:15,080 --> 00:16:17,480 Speaker 2: there's an added benefit, which is that the rest of 292 00:16:17,520 --> 00:16:21,520 Speaker 2: the manufacturing process was hugely capital intensive, but chip design wasn't, 293 00:16:22,040 --> 00:16:23,680 Speaker 2: and so you could have very high margins in your 294 00:16:23,720 --> 00:16:27,080 Speaker 2: products with very little capital costs. And for companies like 295 00:16:27,160 --> 00:16:30,840 Speaker 2: Qualcomm or Broadcom or Nvidia, it's been a very attractive 296 00:16:30,960 --> 00:16:31,640 Speaker 2: business model. 297 00:16:31,760 --> 00:16:35,400 Speaker 1: Everybody got rich, that's right. I want to come back 298 00:16:35,400 --> 00:16:38,400 Speaker 1: to this because I also think that that phenomenon in 299 00:16:38,400 --> 00:16:41,200 Speaker 1: some ways ended up planting the US chip industry. But 300 00:16:41,280 --> 00:16:43,840 Speaker 1: we'll come to that after we take a quick break, 301 00:16:44,160 --> 00:16:50,920 Speaker 1: and here for one of our sponsors, we're back with 302 00:16:51,040 --> 00:16:54,520 Speaker 1: Chris Miller, historian and author of Chip Wars, to talk 303 00:16:54,520 --> 00:16:59,720 Speaker 1: about why semiconductors or chips are now so central to economic, business, 304 00:17:00,120 --> 00:17:04,440 Speaker 1: and political battles. Chris, we've been talking about US dominance 305 00:17:04,440 --> 00:17:08,160 Speaker 1: of chip innovation and manufacturing in the post World War 306 00:17:08,200 --> 00:17:12,600 Speaker 1: two era. But something happens in nineteen eighty seven that 307 00:17:12,680 --> 00:17:15,959 Speaker 1: not many people other than specialists were paying attention to. 308 00:17:16,600 --> 00:17:20,960 Speaker 1: A Taiwanese businessman named Morris Chang founds a company called 309 00:17:21,040 --> 00:17:27,119 Speaker 1: Taiwan Semiconductor Manufacturing Corporation. And this ultimately turns out to 310 00:17:27,160 --> 00:17:29,360 Speaker 1: be a watershed moment, doesn't it. 311 00:17:29,359 --> 00:17:34,000 Speaker 2: It absolutely was because it created the first semiconductor foundry, 312 00:17:34,480 --> 00:17:36,679 Speaker 2: which is a business that doesn't design any chips and 313 00:17:36,760 --> 00:17:41,280 Speaker 2: only manufacturers ships for other customers. And when this idea 314 00:17:41,480 --> 00:17:44,000 Speaker 2: was set out nineteen eighty seven by Morris Chang, who 315 00:17:44,000 --> 00:17:47,240 Speaker 2: founded TSMC, it was really a radical idea because there 316 00:17:47,320 --> 00:17:50,439 Speaker 2: were no customers at first. There weren't any firms that 317 00:17:50,520 --> 00:17:53,840 Speaker 2: only designed chips and didn't have any manufacturing facilities. But 318 00:17:53,880 --> 00:17:57,160 Speaker 2: Morschang met that if you built the manufacturing, new chip 319 00:17:57,200 --> 00:18:00,359 Speaker 2: design firms would emerge that would need his service, and 320 00:18:00,400 --> 00:18:05,640 Speaker 2: that's exactly what's happened. And so today TSMC provides most 321 00:18:05,760 --> 00:18:08,480 Speaker 2: or some cases all of the manufacturing for many of 322 00:18:08,520 --> 00:18:13,800 Speaker 2: the world's biggest chip designers, Apple, Nvidia, AMD, Qualcom, all 323 00:18:13,840 --> 00:18:16,600 Speaker 2: of them rely very heavily on TSMC to actually do 324 00:18:16,680 --> 00:18:19,320 Speaker 2: the manufacturing of all their chips, and. 325 00:18:19,359 --> 00:18:21,639 Speaker 1: Like all of the trippy people who have come in 326 00:18:21,640 --> 00:18:24,320 Speaker 1: and out of this industry, Morris Chang himself has a 327 00:18:24,440 --> 00:18:28,159 Speaker 1: very interesting personal history and journey. He grew up in China, 328 00:18:28,320 --> 00:18:31,960 Speaker 1: emigrated to the US, was educated at Harvard and MIT. 329 00:18:32,160 --> 00:18:36,520 Speaker 1: I think he left MIT without getting his PhD. And 330 00:18:36,560 --> 00:18:39,240 Speaker 1: he winds up at Texas Instruments, where he learns a 331 00:18:39,280 --> 00:18:42,640 Speaker 1: lot about the industry. But then he gets passed over 332 00:18:43,160 --> 00:18:46,879 Speaker 1: as president of Texas Instruments, which is again one of 333 00:18:46,920 --> 00:18:50,240 Speaker 1: those sort of interesting moments in business history when someone 334 00:18:50,280 --> 00:18:54,280 Speaker 1: who is both a profit and an innovator is overlooked 335 00:18:54,280 --> 00:18:55,720 Speaker 1: by the people around him or her. 336 00:18:56,480 --> 00:19:00,479 Speaker 2: It's an incredible story of reinvention. Because more Chang had 337 00:19:00,480 --> 00:19:03,199 Speaker 2: spent his entire career at Texas Instruments, but when he 338 00:19:03,280 --> 00:19:06,159 Speaker 2: was passed over, he left and moved to the island 339 00:19:06,160 --> 00:19:08,520 Speaker 2: of Taiwan, which at the time was about as far 340 00:19:08,600 --> 00:19:10,480 Speaker 2: as the center of the tip industry as you could 341 00:19:10,520 --> 00:19:13,800 Speaker 2: get to set up this new business TSMC. 342 00:19:14,440 --> 00:19:16,600 Speaker 1: In fact, he was invited to go back to Taiwan, 343 00:19:16,640 --> 00:19:19,560 Speaker 1: which he had not lived in before, by Taiwan's political 344 00:19:19,640 --> 00:19:23,320 Speaker 1: leadership because I guess they were thinking decades ahead at 345 00:19:23,359 --> 00:19:23,720 Speaker 1: that point. 346 00:19:23,800 --> 00:19:26,760 Speaker 2: Huh, that's right. They saw the trends chips were going 347 00:19:26,800 --> 00:19:29,440 Speaker 2: to get more important, and they wanted a chip industry 348 00:19:29,480 --> 00:19:32,560 Speaker 2: for themselves, and Marris Shechang was their best hope of 349 00:19:33,240 --> 00:19:35,440 Speaker 2: trying to build one, and so they put up most 350 00:19:35,480 --> 00:19:38,840 Speaker 2: of the initial capital for his company and were very 351 00:19:38,840 --> 00:19:41,320 Speaker 2: supportive of TSMC all the way up to the presence. 352 00:19:41,359 --> 00:19:44,600 Speaker 2: And today TSMC is not only Taiwan's biggest company, it's 353 00:19:44,600 --> 00:19:48,480 Speaker 2: the biggest publicly traded company by market capitalization in all 354 00:19:48,480 --> 00:19:51,240 Speaker 2: of Asia. So there's really no company that's more important 355 00:19:51,280 --> 00:19:53,040 Speaker 2: for Taiwan's economy today. 356 00:19:53,480 --> 00:19:55,800 Speaker 1: TSMC is the company that we referred to in the 357 00:19:55,800 --> 00:19:58,040 Speaker 1: opening of the show as the most important company you've 358 00:19:58,080 --> 00:20:01,760 Speaker 1: never heard of. Let's talk a little bit more about 359 00:20:01,880 --> 00:20:06,000 Speaker 1: psmc's business model and what made them such a world beater. 360 00:20:06,840 --> 00:20:10,040 Speaker 1: You know, Morris Chang eventually handed off the reins to 361 00:20:10,200 --> 00:20:13,719 Speaker 1: RICKSI in two thousand and five, and the leadership at 362 00:20:13,760 --> 00:20:17,040 Speaker 1: TSMC really looked at where the industry was going and 363 00:20:17,080 --> 00:20:20,919 Speaker 1: who they wanted to supply to in a way that 364 00:20:21,000 --> 00:20:22,720 Speaker 1: the US companies overlooked. 365 00:20:22,800 --> 00:20:26,359 Speaker 2: Right, Yeah, that's absolutely right, And TSMC, I think deserves 366 00:20:26,400 --> 00:20:30,600 Speaker 2: a lot of credit for being willing to take risk 367 00:20:30,680 --> 00:20:35,240 Speaker 2: on manufacturing when other companies warn't. Repeatedly, Morris Chang and 368 00:20:35,240 --> 00:20:39,879 Speaker 2: TSMC doubled down on a very expensive manufacturing processes, spending 369 00:20:39,880 --> 00:20:42,800 Speaker 2: billions and billions of dollars in capital expenditure as well 370 00:20:42,800 --> 00:20:45,840 Speaker 2: as R and D over the past twenty years, betting 371 00:20:45,880 --> 00:20:49,399 Speaker 2: that if they were able to produce ever more advanced technology, 372 00:20:49,560 --> 00:20:52,320 Speaker 2: the companies that were the customers would be guaranteed to 373 00:20:52,359 --> 00:20:55,720 Speaker 2: buy it. And that proved right, and it's why TSMC 374 00:20:55,800 --> 00:20:58,800 Speaker 2: has more manufacturing capacity than any rivals. It is more 375 00:20:58,840 --> 00:21:02,439 Speaker 2: advanced cailities and rivals. And today when you look at 376 00:21:02,440 --> 00:21:06,119 Speaker 2: the foundry market, TSMC produces over half of the volume 377 00:21:06,119 --> 00:21:08,040 Speaker 2: of the foundry market. When you look at the advanced 378 00:21:08,040 --> 00:21:11,720 Speaker 2: segments of chip production, TSMC's market shares over ninety percent. 379 00:21:12,359 --> 00:21:16,720 Speaker 1: And they also saw that the mobile phone market was 380 00:21:16,840 --> 00:21:20,040 Speaker 1: going to be transformative at a time when the US 381 00:21:20,080 --> 00:21:23,200 Speaker 1: ship industry was still sort of PC centric. That's right, 382 00:21:23,359 --> 00:21:25,399 Speaker 1: I think famously, I can't remember, but there was an 383 00:21:25,400 --> 00:21:29,320 Speaker 1: Intel executive who who basically said, I don't really get 384 00:21:29,320 --> 00:21:31,720 Speaker 1: it what was happening with mobile phones in terms of 385 00:21:31,800 --> 00:21:33,360 Speaker 1: it being an important new market. 386 00:21:33,840 --> 00:21:36,160 Speaker 2: At one point, Steve Jobs came to Intel and said, 387 00:21:36,400 --> 00:21:38,760 Speaker 2: will you produce the first chip for the iPhone? Intil 388 00:21:38,760 --> 00:21:41,760 Speaker 2: did an analysis internally and said, sounds like a low volume, 389 00:21:41,960 --> 00:21:42,960 Speaker 2: low margin product. 390 00:21:43,240 --> 00:21:45,560 Speaker 1: I mean imagine, like, what did that involve in terms 391 00:21:45,600 --> 00:21:50,200 Speaker 1: of a missed opportunity? Go back to school. I want 392 00:21:50,240 --> 00:21:52,320 Speaker 1: to bring in models and politics a little bit here, 393 00:21:52,640 --> 00:21:56,240 Speaker 1: because during the same period when TSMC is planting its 394 00:21:56,280 --> 00:21:58,679 Speaker 1: stakes in the ground and on its way to becoming 395 00:21:58,680 --> 00:22:02,560 Speaker 1: a dominant force, interesting things are happening in Japan and China. 396 00:22:02,760 --> 00:22:06,600 Speaker 1: Japan had been a world leader in miniaturization. Sony was 397 00:22:06,640 --> 00:22:10,280 Speaker 1: an admired company that created really cool little TV sets. 398 00:22:10,760 --> 00:22:14,040 Speaker 1: They believed in the idea of mobility. They had embraced 399 00:22:14,080 --> 00:22:18,200 Speaker 1: transistors as sort of a foundational technology, but they developed 400 00:22:18,200 --> 00:22:22,119 Speaker 1: a meaningful, but not pivotal presence in the chips market. 401 00:22:22,880 --> 00:22:26,080 Speaker 1: Is there a reason that chip making didn't become bigger 402 00:22:26,119 --> 00:22:28,320 Speaker 1: in Japan during the same period of time. 403 00:22:28,960 --> 00:22:31,760 Speaker 2: No. I think it's really a story of Japanese firms 404 00:22:32,240 --> 00:22:36,280 Speaker 2: missing key market shifts. Because Japanese firms were good at 405 00:22:36,280 --> 00:22:38,960 Speaker 2: mass manufacturing certain types of chips, like the memory chips 406 00:22:38,960 --> 00:22:42,639 Speaker 2: that Intel had produced before it focused on microprocessors. But 407 00:22:43,160 --> 00:22:46,600 Speaker 2: if you look at the key drivers of chip demand today, 408 00:22:46,760 --> 00:22:51,320 Speaker 2: PC's smartphones, now data centers in Ai. Japanese firms have 409 00:22:51,359 --> 00:22:53,119 Speaker 2: not been at the center of any of them. And 410 00:22:53,160 --> 00:22:55,880 Speaker 2: so there are niches where Japanese firms still have very 411 00:22:55,880 --> 00:22:59,080 Speaker 2: strong capabilities, but in the core of the chip market, 412 00:22:59,200 --> 00:23:03,720 Speaker 2: Japanese firm play tangential roles. And that was just a 413 00:23:03,800 --> 00:23:07,440 Speaker 2: series of missed shifts in the business that Japanese firms 414 00:23:07,480 --> 00:23:09,280 Speaker 2: failed to foresee and to capitalize on. 415 00:23:09,840 --> 00:23:12,760 Speaker 1: And at the time of you know, Japan's sort of 416 00:23:12,800 --> 00:23:15,479 Speaker 1: economic peak during this period, it was the second biggest 417 00:23:15,480 --> 00:23:18,600 Speaker 1: economy in the world. Everyone assumed that there was a 418 00:23:18,600 --> 00:23:21,520 Speaker 1: possibility that Japan would overtake the United States as the 419 00:23:21,560 --> 00:23:25,480 Speaker 1: world's biggest economy. And then sort of knocking quietly out 420 00:23:25,520 --> 00:23:29,480 Speaker 1: there on the frontier is China. In nineteen seventy eight, 421 00:23:29,560 --> 00:23:34,280 Speaker 1: the Chinese government begins economic reforms under Deng Hauping that 422 00:23:34,359 --> 00:23:38,080 Speaker 1: eventually turns China into the world's fastest growing major economy, 423 00:23:38,680 --> 00:23:41,680 Speaker 1: and you know, over a three decade period, it eventually 424 00:23:42,440 --> 00:23:45,919 Speaker 1: leap frogs Japan faces off with the United States for 425 00:23:46,040 --> 00:23:51,600 Speaker 1: economic hegemony, and yet is also a relatively insignificant chip maker. 426 00:23:52,240 --> 00:23:56,080 Speaker 1: Tell us a little bit about how you see China 427 00:23:56,119 --> 00:24:02,280 Speaker 1: appearing in this complex waltz of innovation, technology, and political 428 00:24:02,320 --> 00:24:06,680 Speaker 1: and economic power that we're seeing take shape during this period. 429 00:24:06,880 --> 00:24:10,159 Speaker 2: Well right now, for the last decade, China is the 430 00:24:10,240 --> 00:24:14,119 Speaker 2: world's largest importer of semiconductors. And the only reason that 431 00:24:14,200 --> 00:24:17,280 Speaker 2: China imports ships is because it can't produce most types 432 00:24:17,320 --> 00:24:21,520 Speaker 2: of cutting edge ships domestically, and this is a challenge 433 00:24:21,520 --> 00:24:24,080 Speaker 2: that China's been trying to address for some time. In 434 00:24:24,119 --> 00:24:28,320 Speaker 2: twenty fourteen, Chooging Paying identified chips as a core technology, 435 00:24:28,359 --> 00:24:30,560 Speaker 2: as he called it, put tens of billions of dollars 436 00:24:30,640 --> 00:24:33,439 Speaker 2: a year for the last decade into trying to domesticate 437 00:24:33,440 --> 00:24:36,719 Speaker 2: ship making technology. But it's been a really difficult struggle 438 00:24:36,800 --> 00:24:40,400 Speaker 2: for China because it lacks all of the key segments 439 00:24:40,440 --> 00:24:42,479 Speaker 2: you need to make a chip, from the tools to 440 00:24:42,560 --> 00:24:46,320 Speaker 2: the chemicals to the design software. And for Chinese firms, 441 00:24:46,359 --> 00:24:48,439 Speaker 2: it's been very difficult to break into the market because 442 00:24:48,440 --> 00:24:52,120 Speaker 2: this is an extraordinarily complex and concentrated market and at 443 00:24:52,119 --> 00:24:55,480 Speaker 2: the very time when Chinese firms were trying to break 444 00:24:55,480 --> 00:24:58,159 Speaker 2: into the market, the industry itself was getting more and 445 00:24:58,200 --> 00:25:00,760 Speaker 2: more and more concentrated, and indeed the trend for the 446 00:25:00,800 --> 00:25:03,800 Speaker 2: last twenty years is more concentration at every segment of 447 00:25:03,840 --> 00:25:04,800 Speaker 2: the chip supply chain. 448 00:25:05,240 --> 00:25:09,520 Speaker 1: Are there cultural aspects to this around what makes an 449 00:25:09,520 --> 00:25:14,440 Speaker 1: innovative country innovative? There's always been this sort of received 450 00:25:14,480 --> 00:25:17,959 Speaker 1: wisdom that a lot of the Asian economies aren't as 451 00:25:17,960 --> 00:25:22,879 Speaker 1: innovative as the US. They're good imitators, but the nature 452 00:25:22,920 --> 00:25:25,320 Speaker 1: of both the political and the economic structure in those 453 00:25:25,320 --> 00:25:28,640 Speaker 1: countries doesn't allow for innovation, and that Japan suffers from 454 00:25:28,640 --> 00:25:31,600 Speaker 1: that to a certain extent too. Did that hold China 455 00:25:31,640 --> 00:25:34,120 Speaker 1: back as the chip wars started to take shape? 456 00:25:34,680 --> 00:25:36,760 Speaker 2: You know, I think you look at China's neighbors, you 457 00:25:36,800 --> 00:25:40,359 Speaker 2: find Korea and Taiwan absolutely central to the production of 458 00:25:40,359 --> 00:25:44,240 Speaker 2: event semiconductors with really advanced capabilities. So I'd be pretty 459 00:25:44,280 --> 00:25:48,480 Speaker 2: skeptical of a explanation that focused on culture. I think 460 00:25:48,560 --> 00:25:52,359 Speaker 2: the heat factors that Taiwan and Korea have spent the 461 00:25:52,440 --> 00:25:55,720 Speaker 2: last four decades trying to integrate themselves into the chip 462 00:25:55,760 --> 00:25:59,040 Speaker 2: supply chain, and they've each become very successful in one 463 00:25:59,119 --> 00:26:03,240 Speaker 2: specific note Korea and making memory ships, Taiwan and making 464 00:26:03,640 --> 00:26:07,639 Speaker 2: advanced processor ships. But otherwise there's still hugely reliant on 465 00:26:08,280 --> 00:26:13,760 Speaker 2: customers on important tools on materials, on software from more 466 00:26:13,800 --> 00:26:18,320 Speaker 2: advanced economies, and so China's just several decades behind where 467 00:26:18,520 --> 00:26:22,240 Speaker 2: Taiwan our Korea is today, and the additional challenge that 468 00:26:22,280 --> 00:26:24,840 Speaker 2: they face is that the geopolitics is making it even 469 00:26:24,880 --> 00:26:26,120 Speaker 2: more difficult to catch up. 470 00:26:26,760 --> 00:26:30,399 Speaker 1: So Korea and Taiwan had the insight and the foresight 471 00:26:30,760 --> 00:26:33,920 Speaker 1: to see how important pivotal chips would be and bet 472 00:26:34,000 --> 00:26:37,919 Speaker 1: correctly on that future, and other countries that didn't, at 473 00:26:38,000 --> 00:26:40,119 Speaker 1: least when it came to semiconductors, were left behind. 474 00:26:40,359 --> 00:26:42,639 Speaker 2: That's right. And if you think of Korea and Taiwan's 475 00:26:42,680 --> 00:26:46,359 Speaker 2: positions over the last several decades, the US has been 476 00:26:46,800 --> 00:26:49,720 Speaker 2: very supportive in the effort to integrate both of these 477 00:26:49,720 --> 00:26:53,560 Speaker 2: countries into the electronics supply chain to help shore up 478 00:26:53,560 --> 00:26:57,560 Speaker 2: their economies because they were both important Cold War allies, 479 00:26:57,680 --> 00:27:01,120 Speaker 2: and so there's both the strategic thinking of those countries 480 00:27:01,200 --> 00:27:04,800 Speaker 2: businesses and political leaders, but also the US actively supporting 481 00:27:04,800 --> 00:27:05,240 Speaker 2: that effort. 482 00:27:05,600 --> 00:27:09,040 Speaker 1: Okay, Chris, We're going to take another break here from 483 00:27:09,080 --> 00:27:12,520 Speaker 1: a sponsor, and then we will come back to more 484 00:27:12,640 --> 00:27:16,399 Speaker 1: chip wars and the present gnarly landscape that awaits us 485 00:27:16,400 --> 00:27:23,560 Speaker 1: all around this. We're back. We're talking with Chris Miller, 486 00:27:23,720 --> 00:27:30,439 Speaker 1: historian and author about economic and technological warfare. Chris, is 487 00:27:30,480 --> 00:27:33,919 Speaker 1: it an overstatement to call chips the new oil? You know? 488 00:27:33,960 --> 00:27:38,719 Speaker 1: I'm reminded of Daniel Jurgen's seminal book The Prize, which 489 00:27:39,000 --> 00:27:43,400 Speaker 1: was this wonderful weaving of both economic history and political 490 00:27:43,480 --> 00:27:47,720 Speaker 1: history and how the world became, you know, shaped for 491 00:27:47,760 --> 00:27:51,040 Speaker 1: the quest to control oil, and I think of chip 492 00:27:51,080 --> 00:27:54,720 Speaker 1: wars as very much in good company with The Prize. 493 00:27:55,240 --> 00:28:00,400 Speaker 1: It's also trying and successfully I think, doing so an 494 00:28:00,400 --> 00:28:06,200 Speaker 1: analysis of how integral chips are to our economic futures 495 00:28:06,280 --> 00:28:09,560 Speaker 1: right now. But am I being too flip by calling 496 00:28:09,640 --> 00:28:11,440 Speaker 1: chips the new oil? Now? 497 00:28:11,600 --> 00:28:14,840 Speaker 2: I think that's right, And you certainly see basically every 498 00:28:14,880 --> 00:28:19,320 Speaker 2: major government around the world focusing on chips either as 499 00:28:19,359 --> 00:28:21,520 Speaker 2: a strength that it wants to capitalize on or a 500 00:28:21,640 --> 00:28:25,440 Speaker 2: vulnerability that they want to address because they realize the 501 00:28:25,480 --> 00:28:28,320 Speaker 2: strategic importance of some my conductors and economic terms and 502 00:28:28,359 --> 00:28:32,840 Speaker 2: techological terms, but also in military and geopolitical terms. And 503 00:28:32,880 --> 00:28:37,080 Speaker 2: that's why, from Joe Biden to Shijingping, every major leader 504 00:28:37,119 --> 00:28:39,480 Speaker 2: is talking a lot more about chips today than they 505 00:28:39,480 --> 00:28:40,440 Speaker 2: would have ten years ago. 506 00:28:41,120 --> 00:28:44,600 Speaker 1: The head putting that's occurring now between China and the US. 507 00:28:44,760 --> 00:28:49,200 Speaker 1: It has some of its roots in manufacturing changes, manufacturers 508 00:28:49,200 --> 00:28:52,560 Speaker 1: moving to China where it was cheaper labor, the offshoring 509 00:28:52,600 --> 00:28:57,080 Speaker 1: of American industry to China, and now you have this, 510 00:28:56,720 --> 00:29:01,400 Speaker 1: you know, this real battle for political and technological dominance. 511 00:29:01,440 --> 00:29:03,720 Speaker 1: Do you think that was inevitable? Do you think this 512 00:29:03,880 --> 00:29:08,280 Speaker 1: kind of competition between China and the US was inevitable 513 00:29:08,360 --> 00:29:10,640 Speaker 1: or could it have followed a different path? And we'll 514 00:29:10,680 --> 00:29:12,880 Speaker 1: loop this back into what's going on with chips. 515 00:29:13,480 --> 00:29:18,560 Speaker 2: I think the economic competition was inevitable, But the US 516 00:29:18,640 --> 00:29:22,480 Speaker 2: and Japan had economic competition during the nineteen eighties and nineties, 517 00:29:22,520 --> 00:29:25,720 Speaker 2: and that didn't result in nearly as tense relations to 518 00:29:25,720 --> 00:29:27,480 Speaker 2: we have with the US and China today. But the 519 00:29:27,600 --> 00:29:29,560 Speaker 2: challenge with the US and China today is that on 520 00:29:29,600 --> 00:29:32,880 Speaker 2: top of the economic competition, there's a military and geopolitical 521 00:29:32,920 --> 00:29:37,520 Speaker 2: competition over the South China Sea, over the Taiwan Straits, 522 00:29:37,520 --> 00:29:40,040 Speaker 2: and over the broader question of who gets to set 523 00:29:40,080 --> 00:29:42,440 Speaker 2: the rules in East Asia. Because the US drew the 524 00:29:42,480 --> 00:29:46,280 Speaker 2: map of East Asia after World War Two, it's got 525 00:29:46,520 --> 00:29:49,120 Speaker 2: the most important allies in the region, and China wants 526 00:29:49,160 --> 00:29:51,160 Speaker 2: a different system and So if you didn't have the 527 00:29:51,200 --> 00:29:54,480 Speaker 2: geopolitical tensions, the military arms race, then I think the 528 00:29:54,520 --> 00:29:56,960 Speaker 2: economic issues would be a lot easier to deal with 529 00:29:57,240 --> 00:29:58,959 Speaker 2: than they have been. But the fact is, right now 530 00:29:59,040 --> 00:30:02,680 Speaker 2: they're deeply intertwined because both countries think that their economic 531 00:30:02,760 --> 00:30:06,760 Speaker 2: futures will also feed into their ability to compete militarily. 532 00:30:07,640 --> 00:30:11,479 Speaker 1: The other irony or historical twists in all of this is, 533 00:30:11,520 --> 00:30:15,640 Speaker 1: of course, China doesn't recognize Taiwan as an independent country. 534 00:30:15,880 --> 00:30:20,400 Speaker 1: It believes that Taiwan is essentially its island, and the 535 00:30:20,560 --> 00:30:26,120 Speaker 1: territorial dispute around Taiwan has also shaped the US's dialogue 536 00:30:26,120 --> 00:30:29,280 Speaker 1: with China. And in the midst of all that, you 537 00:30:29,360 --> 00:30:32,560 Speaker 1: have Taiwan shaping up as one of the world's hedgemons 538 00:30:33,000 --> 00:30:38,080 Speaker 1: around chip manufacturing. So, irony of ironies, this contested territory 539 00:30:38,120 --> 00:30:43,240 Speaker 1: also is home to a very productive and world changing 540 00:30:43,600 --> 00:30:49,640 Speaker 1: chip making empire. Is that something that was also never 541 00:30:49,720 --> 00:30:52,200 Speaker 1: going to work out in a linear way. The fact 542 00:30:52,200 --> 00:30:55,840 Speaker 1: that this industry developed on China was also going to 543 00:30:55,880 --> 00:30:59,560 Speaker 1: just add fuel to the fire between China and the US. 544 00:31:00,240 --> 00:31:02,640 Speaker 2: Well, I think no one in Beijing or in Washington 545 00:31:02,720 --> 00:31:05,200 Speaker 2: predicted or planned the fact that Taiwan would be the 546 00:31:05,240 --> 00:31:08,080 Speaker 2: center of the ship industry today. I think in Taiwan 547 00:31:08,240 --> 00:31:10,800 Speaker 2: there are a lot of people who hoped that investing 548 00:31:10,840 --> 00:31:13,640 Speaker 2: in chips would give them an additional measure of security 549 00:31:14,360 --> 00:31:17,360 Speaker 2: against China, and that's proven to be a good bet 550 00:31:17,640 --> 00:31:20,440 Speaker 2: on their part. But it's certainly the case that Taiwan's 551 00:31:20,440 --> 00:31:24,640 Speaker 2: status does intensify the tensions because when the US wants 552 00:31:24,680 --> 00:31:28,720 Speaker 2: to cut off a US chip design firm from selling 553 00:31:28,720 --> 00:31:32,160 Speaker 2: to China, almost all of those advanced chips are actually 554 00:31:32,200 --> 00:31:36,160 Speaker 2: manufactured in Taiwan, and so the US imposes export controls, 555 00:31:36,240 --> 00:31:38,920 Speaker 2: the Taiwanese have no choice but to follow. And so 556 00:31:38,960 --> 00:31:41,600 Speaker 2: we have a situation in which China thinks Taiwan is 557 00:31:41,640 --> 00:31:46,160 Speaker 2: its territory, but Taiwan is imposing US export controls against 558 00:31:46,200 --> 00:31:47,480 Speaker 2: China's high tech sector. 559 00:31:48,400 --> 00:31:53,600 Speaker 1: And as China's economic heften might become more apparent. In 560 00:31:53,640 --> 00:31:58,000 Speaker 1: the US, you've had different politicians and administrations inveighing against China. 561 00:31:58,520 --> 00:32:01,240 Speaker 1: Trump as president and it made a lot of hay 562 00:32:01,400 --> 00:32:03,680 Speaker 1: in veighing against China, but in a lot of ways, 563 00:32:04,520 --> 00:32:09,520 Speaker 1: that wasn't really followed up with strategic measures that blunted 564 00:32:10,480 --> 00:32:13,360 Speaker 1: China in any meaningful way. And then you come to 565 00:32:13,400 --> 00:32:18,040 Speaker 1: Biden and the Biden administration, and in short order he 566 00:32:18,360 --> 00:32:21,480 Speaker 1: and Congress passed the Chips Act. And let's talk a 567 00:32:21,520 --> 00:32:23,720 Speaker 1: little bit about the Chips Act. It's a two hundred 568 00:32:23,760 --> 00:32:28,160 Speaker 1: and eighty billion dollar federal measure that's meant to shore 569 00:32:28,280 --> 00:32:34,280 Speaker 1: up US semiconductor manufacturers who've fallen behind in the great 570 00:32:34,400 --> 00:32:37,800 Speaker 1: arms race that is chips making. But it also imposed 571 00:32:37,880 --> 00:32:40,680 Speaker 1: restrictions on the kind of chips that could go into China, 572 00:32:41,320 --> 00:32:45,640 Speaker 1: and it is a kind of act of economic warfare, 573 00:32:46,000 --> 00:32:49,240 Speaker 1: I think, And you may have a different view of it, obviously, 574 00:32:49,320 --> 00:32:52,200 Speaker 1: but what do you think about the significance of the 575 00:32:52,280 --> 00:32:53,080 Speaker 1: Chips Act. 576 00:32:53,440 --> 00:32:55,320 Speaker 2: Well, I think if you look at the origins of 577 00:32:55,360 --> 00:32:57,960 Speaker 2: the Chips Act and the export controls, I think what 578 00:32:58,040 --> 00:33:01,719 Speaker 2: you find is that there's actually been a bipartisan consensus 579 00:33:01,760 --> 00:33:05,520 Speaker 2: and moving in this direction since really the late Obama administration, 580 00:33:06,120 --> 00:33:07,920 Speaker 2: and so the changes in the top of the White 581 00:33:07,960 --> 00:33:12,200 Speaker 2: House haven't actually dramatically changed policy on this direction. And 582 00:33:12,240 --> 00:33:16,040 Speaker 2: that's because there's a broad sense across both political parties 583 00:33:16,120 --> 00:33:20,120 Speaker 2: in the US and across the bureaucracy that if China 584 00:33:20,200 --> 00:33:23,440 Speaker 2: has access to the most advanced computing capabilities, it will 585 00:33:23,560 --> 00:33:27,680 Speaker 2: deploy these to defense and intelligence uses, and China's not unique. 586 00:33:27,800 --> 00:33:30,360 Speaker 2: Every country does that, and in so far as that's true, 587 00:33:30,440 --> 00:33:33,320 Speaker 2: it will give China new capabilities that it currently doesn't 588 00:33:33,360 --> 00:33:37,640 Speaker 2: have and erode the technological gap that still exists between 589 00:33:37,680 --> 00:33:40,520 Speaker 2: the US and Chinese militaries. And so when the national 590 00:33:40,560 --> 00:33:43,400 Speaker 2: security state thinks about chips, it's thinking about the military 591 00:33:43,440 --> 00:33:46,400 Speaker 2: balance first and foremost. But of course chips aren't only 592 00:33:46,520 --> 00:33:48,800 Speaker 2: used in military equipment. Most chips that are sold end 593 00:33:48,880 --> 00:33:51,440 Speaker 2: up in civilian uses. Something like ninety eight percent of 594 00:33:51,480 --> 00:33:53,760 Speaker 2: chips end up in smartphones or PCs rather than in 595 00:33:53,800 --> 00:33:56,920 Speaker 2: defense equipment. The problem, or the challenge is that the 596 00:33:57,000 --> 00:34:00,520 Speaker 2: exact same chips that are used, for example, to a 597 00:34:00,640 --> 00:34:03,640 Speaker 2: military AI system can be used to train a civilian 598 00:34:03,680 --> 00:34:07,280 Speaker 2: AI system, and there's no way of differentiating. And so 599 00:34:07,320 --> 00:34:09,959 Speaker 2: the US is genuine, I think when it says we're 600 00:34:10,040 --> 00:34:14,200 Speaker 2: only trying to stop China from using these for military purposes, 601 00:34:14,560 --> 00:34:17,040 Speaker 2: but in doing so, it's also stopping China from using 602 00:34:17,040 --> 00:34:19,600 Speaker 2: many of these ships for civilian purposes because there's just 603 00:34:19,640 --> 00:34:20,879 Speaker 2: no possible differentiation. 604 00:34:21,800 --> 00:34:23,440 Speaker 1: How is China responded to that? 605 00:34:24,120 --> 00:34:27,520 Speaker 2: Well, actually China hasn't responded much thus far. China's already 606 00:34:27,600 --> 00:34:30,200 Speaker 2: been over the last decade, pouring tens of billions of 607 00:34:30,239 --> 00:34:33,120 Speaker 2: dollars a year into a ship industry, and that continues. 608 00:34:33,160 --> 00:34:35,239 Speaker 2: But that's not a response to US policy. That was 609 00:34:35,280 --> 00:34:38,600 Speaker 2: actually something that predated US policy by almost a decade. 610 00:34:38,800 --> 00:34:42,399 Speaker 2: The only real response that China's undertaken is to ban 611 00:34:42,600 --> 00:34:45,640 Speaker 2: the use of certain ships from one US chip maker, Micron, 612 00:34:46,080 --> 00:34:48,840 Speaker 2: in Chinese products, which will have a moderately negative effect 613 00:34:48,880 --> 00:34:51,440 Speaker 2: on Micron, But it's actually a pretty small move relative 614 00:34:51,480 --> 00:34:54,080 Speaker 2: to the scale of what the US is doing to China. 615 00:34:54,880 --> 00:34:59,200 Speaker 1: If chips supply from Taiwan got cut off, if China 616 00:34:59,360 --> 00:35:05,000 Speaker 1: invaded Thiaiwan and made TSMC its very own prize after 617 00:35:05,040 --> 00:35:09,120 Speaker 1: an invasion, what would be the impact on the global 618 00:35:09,120 --> 00:35:10,520 Speaker 1: economy and on the US. 619 00:35:11,400 --> 00:35:14,640 Speaker 2: Well, if there was a war, an invasion, if missiles 620 00:35:14,680 --> 00:35:17,680 Speaker 2: were flying, the chip making facilities would shut down and 621 00:35:17,719 --> 00:35:20,920 Speaker 2: they wouldn't restart. So I think it's highly unlikely that 622 00:35:21,400 --> 00:35:24,560 Speaker 2: China could invade Taiwan and grab the facilities in a 623 00:35:24,600 --> 00:35:27,280 Speaker 2: functioning matner. Just wouldn't work that way. So the question 624 00:35:27,400 --> 00:35:29,600 Speaker 2: is what would the world look like if all of 625 00:35:29,600 --> 00:35:33,640 Speaker 2: Taiwan's production went offline? And the answer is that we'd 626 00:35:33,880 --> 00:35:37,279 Speaker 2: face great depression levels of disruption to global manufacturing because 627 00:35:37,280 --> 00:35:40,480 Speaker 2: it's not only smartphones and PCs and data centers and 628 00:35:40,520 --> 00:35:43,719 Speaker 2: telecoms infrastructure that rely on the most advanced ships, of 629 00:35:43,719 --> 00:35:46,319 Speaker 2: which ninety percent are produced in Taiwan. So good luck 630 00:35:46,320 --> 00:35:48,960 Speaker 2: producing a smartphone anywhere in the world. But what we 631 00:35:49,080 --> 00:35:52,560 Speaker 2: learned during the chip shortage of the past several years 632 00:35:52,600 --> 00:35:54,960 Speaker 2: is that actually everything relies on ships. A new car 633 00:35:54,960 --> 00:35:58,000 Speaker 2: will have on average one thousand chips inside, and so 634 00:35:58,160 --> 00:36:00,719 Speaker 2: ballpark ten or twenty percent of those are produced in 635 00:36:00,760 --> 00:36:03,560 Speaker 2: Taiwan on a typical car. Well, good luck replacing that 636 00:36:03,600 --> 00:36:06,799 Speaker 2: supply at a time when the entire global production of 637 00:36:06,880 --> 00:36:09,560 Speaker 2: chips is going to shrink by double digit rates. And 638 00:36:09,600 --> 00:36:11,960 Speaker 2: so if you not only add up the losses and 639 00:36:12,000 --> 00:36:14,759 Speaker 2: smartphones and PCs, but also in cars and dishwashers and 640 00:36:15,080 --> 00:36:17,920 Speaker 2: airplanes and microwaves, you find there's not a lot of 641 00:36:17,920 --> 00:36:21,320 Speaker 2: goods we can manufacture without access to semiconductors, and Taiwan 642 00:36:21,400 --> 00:36:24,480 Speaker 2: produces so many semiconductors that the impact would just be 643 00:36:24,600 --> 00:36:25,640 Speaker 2: catastrophically bad. 644 00:36:26,080 --> 00:36:29,279 Speaker 1: You know, we have a strategic petroleum reserve that we 645 00:36:29,520 --> 00:36:33,320 Speaker 1: use to you know, influence oil markets when they're shortages 646 00:36:33,440 --> 00:36:37,040 Speaker 1: or to have an inventory to get the US pass 647 00:36:37,080 --> 00:36:40,840 Speaker 1: shortfalls and energy supplies. Does that argue for the idea 648 00:36:40,840 --> 00:36:46,720 Speaker 1: that we should have a strategic chips reserve inventories alongside 649 00:36:46,760 --> 00:36:49,600 Speaker 1: this push for greater innovations so we don't get the 650 00:36:49,800 --> 00:36:52,520 Speaker 1: US doesn't get put in the position of having a 651 00:36:52,560 --> 00:36:54,879 Speaker 1: shortfall of chips ever again, you. 652 00:36:54,840 --> 00:36:57,399 Speaker 2: Know, I think it's an idea we ought to explore more. 653 00:36:57,480 --> 00:37:01,239 Speaker 2: It's trickier than petroleum because every barrel of oil is 654 00:37:01,480 --> 00:37:03,880 Speaker 2: more or less the same as another. There's different grades, 655 00:37:03,880 --> 00:37:06,560 Speaker 2: but the differences are pretty minimal, and if you put 656 00:37:06,600 --> 00:37:09,000 Speaker 2: a barrel in storage for two or four six years, 657 00:37:09,000 --> 00:37:12,719 Speaker 2: it still works just as well. We're chips. There's thousands 658 00:37:12,760 --> 00:37:14,719 Speaker 2: of different types of chips, and they are not interchangeable. 659 00:37:14,800 --> 00:37:17,359 Speaker 2: You often have to completely redesign a product to put 660 00:37:17,360 --> 00:37:19,719 Speaker 2: a different ship in it. And because of Moore's law, 661 00:37:19,840 --> 00:37:22,040 Speaker 2: a chip that was made five years ago is no 662 00:37:22,160 --> 00:37:26,200 Speaker 2: longer that great. It's often pretty unimpressive for the capabilities 663 00:37:26,239 --> 00:37:30,399 Speaker 2: that you need, and so a reserve stockpile is much 664 00:37:30,440 --> 00:37:33,040 Speaker 2: harder to execute well, and the cost would be higher 665 00:37:33,239 --> 00:37:35,920 Speaker 2: because you get the degradation over time and you have 666 00:37:36,080 --> 00:37:38,440 Speaker 2: all the different types of products you need to manage. 667 00:37:38,480 --> 00:37:41,600 Speaker 2: But in principle, I think you're right that given the risk, 668 00:37:41,680 --> 00:37:43,960 Speaker 2: we ought to be thinking more hard about are there 669 00:37:44,000 --> 00:37:47,200 Speaker 2: certain types of chips that we can stockpile while minimizing 670 00:37:47,239 --> 00:37:49,200 Speaker 2: the costs opposed by these complexities? 671 00:37:49,600 --> 00:37:52,240 Speaker 1: And do you think US manufacturers, you know, this amazing 672 00:37:52,280 --> 00:37:54,360 Speaker 1: crop of companies we've talked about that are based in 673 00:37:54,400 --> 00:37:57,719 Speaker 1: Silicon Valley for the most part, are in a position 674 00:37:57,920 --> 00:38:01,920 Speaker 1: now to be a disputed world leaders Again, are they 675 00:38:01,960 --> 00:38:06,040 Speaker 1: structured in the right way? Does the billions of dollars 676 00:38:06,080 --> 00:38:09,239 Speaker 1: the Biden administration is pouring in the industry, is that 677 00:38:09,320 --> 00:38:14,080 Speaker 1: going to create a seismic shift in terms of innovation 678 00:38:14,160 --> 00:38:16,520 Speaker 1: and competition in the chips business? 679 00:38:17,280 --> 00:38:20,040 Speaker 2: Well, in chip design, the answer is easy because US 680 00:38:20,080 --> 00:38:23,120 Speaker 2: firms already are the world's leaders. Companies like in Vidia, 681 00:38:23,200 --> 00:38:25,239 Speaker 2: the world's first trillion dollars chip company. But in Vidia 682 00:38:25,239 --> 00:38:27,879 Speaker 2: doesn't manufacture any chips, it just designs them. Most are 683 00:38:27,880 --> 00:38:30,960 Speaker 2: manufactured in Taiwan. And so the challenge the Hips Act 684 00:38:31,000 --> 00:38:34,200 Speaker 2: is can you get more manufacturing of chips in the US. 685 00:38:34,239 --> 00:38:36,759 Speaker 2: And I think it's going to be hard. It's more 686 00:38:36,800 --> 00:38:38,680 Speaker 2: expensive to make chips in the US than it is 687 00:38:38,680 --> 00:38:43,560 Speaker 2: in Taiwan or Korea. We've underinvested in the infrastructure and 688 00:38:43,600 --> 00:38:46,839 Speaker 2: the workforce you need for several decades. And these are 689 00:38:46,880 --> 00:38:49,959 Speaker 2: all problems that are solvable, but only over a long 690 00:38:50,000 --> 00:38:52,919 Speaker 2: time horizon, And so we shouldn't expect the Chips Act 691 00:38:52,920 --> 00:38:55,480 Speaker 2: a loan is going to get us independent of our 692 00:38:55,760 --> 00:38:58,640 Speaker 2: reliance and importing chips from East Asia. That's just a fantasy. 693 00:38:58,680 --> 00:39:01,480 Speaker 2: We're going to be importing chips from Taiwan, from Korea, 694 00:39:01,920 --> 00:39:04,600 Speaker 2: from elsewhere in East Asia for many, many years to come. 695 00:39:05,360 --> 00:39:09,279 Speaker 1: How do you see China ultimately responding to all of this? 696 00:39:09,440 --> 00:39:13,640 Speaker 1: The Biden administration is up the anti recently on restrictions 697 00:39:13,680 --> 00:39:18,480 Speaker 1: of ship sales to China, particularly around applications that involve 698 00:39:18,560 --> 00:39:22,440 Speaker 1: generative AI. You know, what are China's options and how 699 00:39:22,440 --> 00:39:23,960 Speaker 1: do you see this playing out there? 700 00:39:24,480 --> 00:39:26,360 Speaker 2: I think the first option of alable to China is 701 00:39:26,400 --> 00:39:30,000 Speaker 2: to make its foreign policy less bellicos to reduce the 702 00:39:30,040 --> 00:39:32,120 Speaker 2: restrictions that are imposted on it. That's not going to happen, 703 00:39:32,120 --> 00:39:36,440 Speaker 2: I don't think under presidency. Second option is to keep 704 00:39:36,480 --> 00:39:39,239 Speaker 2: pouring money into the chip industry and hope that more 705 00:39:39,280 --> 00:39:43,120 Speaker 2: money will help China solve its chip industry problems. You know, 706 00:39:43,120 --> 00:39:45,040 Speaker 2: I think that has some chance of success, though the 707 00:39:45,080 --> 00:39:47,759 Speaker 2: track record is fine but not great, and I think 708 00:39:47,800 --> 00:39:50,600 Speaker 2: it'll be a real hard slog If China wants to 709 00:39:50,760 --> 00:39:53,560 Speaker 2: keep pushing forward in its effort to build a completely 710 00:39:53,600 --> 00:39:55,800 Speaker 2: self sufficient cutting edge supply chain, it's going to be 711 00:39:55,840 --> 00:40:00,080 Speaker 2: a long way away. And the third option is to 712 00:40:00,120 --> 00:40:02,680 Speaker 2: give up and recognize that China will have less advanced 713 00:40:02,680 --> 00:40:04,399 Speaker 2: ships than the rest of the world. And I don't 714 00:40:04,440 --> 00:40:07,080 Speaker 2: think that's something that President she is going to admit 715 00:40:07,200 --> 00:40:09,680 Speaker 2: for a very long time, but that might be the 716 00:40:09,880 --> 00:40:12,319 Speaker 2: reality because the trend in the chip industry is not 717 00:40:12,400 --> 00:40:15,239 Speaker 2: of catching up. Most people think, oh, technology is going 718 00:40:15,280 --> 00:40:18,200 Speaker 2: to naturally disseminate, people will catch up, But the rate 719 00:40:18,239 --> 00:40:20,319 Speaker 2: of progress in the chip industry is so rapid that 720 00:40:20,360 --> 00:40:23,080 Speaker 2: actually the trend is of falling behind. The US fell 721 00:40:23,080 --> 00:40:27,160 Speaker 2: behind Taiwan, Japan fell behind Taiwan, Europe fell behind Taiwan. 722 00:40:27,280 --> 00:40:29,920 Speaker 2: Korea is struggling to keep up to Taiwan. And so 723 00:40:30,120 --> 00:40:33,400 Speaker 2: why should we assume that China, which is several generations 724 00:40:33,440 --> 00:40:36,520 Speaker 2: behind Taiwan in terms of chip making capabilities and can't 725 00:40:36,520 --> 00:40:39,200 Speaker 2: access the chip making tools and can't access the ship 726 00:40:39,239 --> 00:40:43,759 Speaker 2: making software, why should we assume catchup is the base case. 727 00:40:43,800 --> 00:40:46,160 Speaker 2: It seems we should assume China's a normal country, and 728 00:40:46,239 --> 00:40:48,600 Speaker 2: normal countries have tended to fall behind. 729 00:40:48,520 --> 00:40:49,920 Speaker 1: Which means and that China is going to have to 730 00:40:49,920 --> 00:40:52,000 Speaker 1: rethink it's posture towards the rest of the world. If 731 00:40:52,040 --> 00:40:56,560 Speaker 1: it needs that supply of chips to keep coming into 732 00:40:56,560 --> 00:41:00,400 Speaker 1: the country in a way that promotes growth. They can't 733 00:41:00,480 --> 00:41:04,240 Speaker 1: really get there by always carrying around a big stick, right. 734 00:41:04,480 --> 00:41:07,279 Speaker 2: That's the dilemma that imposes in China over the longer run. 735 00:41:07,320 --> 00:41:09,319 Speaker 2: You know, it's not a problem for this year, next year, 736 00:41:09,760 --> 00:41:11,840 Speaker 2: but it is a problem for the late twenty twenties 737 00:41:11,880 --> 00:41:15,560 Speaker 2: if Chinese firms have access to much less capable chips 738 00:41:15,560 --> 00:41:19,759 Speaker 2: and therefore face limitations, inefficiencies, higher costs when they try 739 00:41:19,800 --> 00:41:21,359 Speaker 2: to train and develop aisystems. 740 00:41:22,000 --> 00:41:23,960 Speaker 1: You said at the top of the show, Chris, that 741 00:41:24,880 --> 00:41:28,080 Speaker 1: you really didn't know anything about chips seven or so 742 00:41:28,200 --> 00:41:30,120 Speaker 1: years ago when you first decided to make it a 743 00:41:30,160 --> 00:41:34,040 Speaker 1: focus of your life. So what have you learned over 744 00:41:34,080 --> 00:41:36,920 Speaker 1: that long journey you've had in digging into chips and 745 00:41:36,960 --> 00:41:39,680 Speaker 1: the chip wars? What do you know now that you 746 00:41:39,760 --> 00:41:40,959 Speaker 1: didn't know seven years ago. 747 00:41:42,320 --> 00:41:44,480 Speaker 2: I think the thing that really struck me is the 748 00:41:44,520 --> 00:41:49,000 Speaker 2: appreciation I've developed for advanced manufacturing. I thought of computing 749 00:41:49,040 --> 00:41:52,200 Speaker 2: as something that happened in digits behind my screen, and 750 00:41:52,239 --> 00:41:55,520 Speaker 2: I didn't really realize the extent which there's an extraordinary 751 00:41:55,560 --> 00:41:58,960 Speaker 2: set of companies capable of moving materials that basically the 752 00:41:59,000 --> 00:42:02,799 Speaker 2: atomic level, whose capabilities are necessary to produce all the 753 00:42:02,840 --> 00:42:05,000 Speaker 2: computing that you and I rely on. We take these 754 00:42:05,040 --> 00:42:07,560 Speaker 2: companies for granted, and we shouldn't, because the capabilities are 755 00:42:07,800 --> 00:42:10,799 Speaker 2: really extraordinary, and the countries in which they're located are 756 00:42:11,080 --> 00:42:13,680 Speaker 2: key to shaping not just the future of computing, but 757 00:42:13,719 --> 00:42:15,320 Speaker 2: the balance of power in the world stage. 758 00:42:16,600 --> 00:42:19,000 Speaker 1: Chris, we've run out of time. Thank you for being 759 00:42:19,320 --> 00:42:22,920 Speaker 1: such a great teacher and author and guests all rolled. 760 00:42:22,680 --> 00:42:25,040 Speaker 2: Into one well, thanks so much for having me. 761 00:42:25,920 --> 00:42:28,399 Speaker 1: Chris Miller is the author of Chip Wars. There's also 762 00:42:28,480 --> 00:42:31,480 Speaker 1: a historian at the Fletcher School of Law Diplomacy at 763 00:42:31,480 --> 00:42:35,200 Speaker 1: Tufts University. He can be found on Twitter at c 764 00:42:35,440 --> 00:42:39,680 Speaker 1: R Miller One. Here at crash Course, we believe the 765 00:42:39,760 --> 00:42:44,759 Speaker 1: collisions can be messy, impressive, challenging, surprising, and always instructive. 766 00:42:45,400 --> 00:42:49,280 Speaker 1: In today's Crash Course, I learned that manufacturing really matters 767 00:42:49,560 --> 00:42:52,680 Speaker 1: It doesn't matter how fancy or complex the industry you're in. 768 00:42:53,440 --> 00:42:55,440 Speaker 1: If you don't know how to make things and how 769 00:42:55,480 --> 00:42:58,319 Speaker 1: to make them well and make them competitively, you're in 770 00:42:58,360 --> 00:43:03,000 Speaker 1: for some trouble. Ask us semiconductor manufacturers, Just ask China, 771 00:43:03,520 --> 00:43:05,480 Speaker 1: and just talk to Taiwan when you have a moment. 772 00:43:06,160 --> 00:43:09,040 Speaker 1: What did you learn? We'd love to hear from you. 773 00:43:09,040 --> 00:43:11,960 Speaker 1: You can tweet at the Bloomberg Opinion, handle at Opinion 774 00:43:12,280 --> 00:43:16,640 Speaker 1: or me at Tim O'Brien using the hashtag Bloomberg Crash Course. 775 00:43:17,600 --> 00:43:20,480 Speaker 1: You can also subscribe to our show wherever you're listening 776 00:43:20,560 --> 00:43:23,200 Speaker 1: right now and leave us a review. It helps more 777 00:43:23,239 --> 00:43:26,600 Speaker 1: people find the show. This episode was produced by the 778 00:43:26,640 --> 00:43:32,160 Speaker 1: indispensable Anamasarakas, Moses On Dam and Me. Our supervising producer 779 00:43:32,239 --> 00:43:35,600 Speaker 1: is Magus Hendrickson, and we had editing help from Sage Bauman, 780 00:43:36,000 --> 00:43:40,560 Speaker 1: Katie Boyce, Jeff Grocott, Mike Nitze, and Christine Vanden Bylert. 781 00:43:41,200 --> 00:43:44,479 Speaker 1: Blake Maples does our sound engineering, and our original theme 782 00:43:44,520 --> 00:43:48,680 Speaker 1: song was composed by Luis Gara. I'm Tim O'Brien. We'll 783 00:43:48,680 --> 00:43:51,080 Speaker 1: be back next week with another Crash Course