WEBVTT - ASML, the Obscure Powerhouse at the Cutting Edge of Chip Technology 0:00:10.800 --> 0:00:14.840 Hello, and welcome to another episode of the Odd Lots podcast. 0:00:14.920 --> 0:00:21.240 I'm Joe Wisenthal and I'm Tracy Halloway. So Tracy obviously, Yeah, 0:00:21.280 --> 0:00:25.000 we've done a lot of episodes about the semiconductor industry, 0:00:25.120 --> 0:00:28.560 about chips. There's one specific, I guess I would say, 0:00:28.600 --> 0:00:31.200 sub component of the story that people are like, Oh, 0:00:31.240 --> 0:00:32.800 you guys got to do that, you guys got to 0:00:32.840 --> 0:00:34.800 do that, which we have yet to hit so far. 0:00:35.520 --> 0:00:37.760 You say, there's one thing that we have yet to do, 0:00:37.800 --> 0:00:40.320 but I have a feeling like this is the Endless 0:00:40.360 --> 0:00:43.200 Semiconductor series, and as soon as we finished this episode, 0:00:43.240 --> 0:00:46.440 we're going to discover some other hidden component of the 0:00:46.479 --> 0:00:48.800 semi conductor supply chain and that's going to lead to 0:00:48.880 --> 0:00:52.240 another episode. But yes, you're right, um, there is one 0:00:52.360 --> 0:00:57.200 sort of big elephant, big semiconductor thing in the room, 0:00:57.400 --> 0:01:00.600 and that is a company called a s m L. 0:01:01.120 --> 0:01:03.120 Not to be confused with a s m R, which 0:01:03.160 --> 0:01:06.680 I always seem to do a s mL but hopefully 0:01:06.680 --> 0:01:10.000 for like a certain kind of person, listening to an 0:01:10.040 --> 0:01:13.400 hour of people talking about chips is a type of 0:01:13.880 --> 0:01:16.600 for them SMR. So maybe killed two birds with one stone, 0:01:16.600 --> 0:01:18.480 but yes, s m L, you know, one of the 0:01:18.560 --> 0:01:21.680 things that we established in thinking about how these sort 0:01:21.720 --> 0:01:25.319 of the chip ecosystem works. Maybe part of our characterization 0:01:25.440 --> 0:01:30.400 is we twan semiconductor the biggest contract fabrication company in 0:01:30.440 --> 0:01:32.039 the world. I think we sort of think of as 0:01:32.080 --> 0:01:35.000 like they're the final boss right in chims, Like in 0:01:35.040 --> 0:01:37.840 the end, they're like the central bank of chips. Their 0:01:37.920 --> 0:01:41.800 capacity kind of almost dictates chip capacity overall. There's some 0:01:41.840 --> 0:01:46.119 other companies that make chips, including Intel and Global Boundaries, 0:01:46.200 --> 0:01:51.000 but t SMC is the big one. But TSMC has 0:01:51.040 --> 0:01:53.840 to buy equipment from others to you know not no 0:01:53.880 --> 0:01:57.520 one is completely self sufficient in this industry, and t 0:01:57.720 --> 0:02:01.720 SMC is a huge client or a huge purchaser of 0:02:01.840 --> 0:02:06.120 equipment made by this company. It's a Dutch company a SML. Yeah, 0:02:06.200 --> 0:02:08.360 so you mentioned that it's Dutch, And this is the 0:02:08.400 --> 0:02:11.160 other thing I mean, in addition to not really understanding 0:02:11.360 --> 0:02:14.120 what this company does or the type of equipment it's 0:02:14.120 --> 0:02:18.480 actually making for semiconductor manufacturers like t SMC, the other 0:02:18.520 --> 0:02:21.200 thing I don't really get about it is why is 0:02:21.240 --> 0:02:23.480 it a Dutch company? Because the one thing I know 0:02:23.480 --> 0:02:26.640 about it is it has its origins in the US 0:02:27.280 --> 0:02:29.760 UM I think in the you know, like nineteen eighties, 0:02:29.760 --> 0:02:31.920 it sort of came out of the collapse of a 0:02:31.960 --> 0:02:35.520 bunch of US lithography firms or something like that. And 0:02:35.600 --> 0:02:38.520 yet now it's a Dutch company that has this enormous 0:02:38.720 --> 0:02:43.760 role in the global supply chain. It's squarely like kind 0:02:43.760 --> 0:02:46.600 of crowning, like it's sort of like a kingmaker for 0:02:46.760 --> 0:02:51.800 semiconductor technology or expertise. And I don't know, I just 0:02:51.840 --> 0:02:55.120 have so many questions already. I know we haven't even started. Yeah, 0:02:55.240 --> 0:02:57.040 I know, I have no questions too. We'll get started 0:02:57.040 --> 0:02:59.120 just a second. But you know, you mentioned the oddity 0:02:59.160 --> 0:03:03.280 of it big Dutch. There's another element here, and I 0:03:03.280 --> 0:03:05.280 I don't you know, I feel like reluctant to like 0:03:05.360 --> 0:03:11.760 talking like cliches or stereotypes, but I don't really think. No, 0:03:11.880 --> 0:03:15.160 it is true of like Northern Europe or Europe in 0:03:15.240 --> 0:03:18.919 general as being like this like cutting edge high tech 0:03:18.960 --> 0:03:21.560 hotbed for anything. When I think about tech, I think 0:03:21.560 --> 0:03:24.959 about Silicon Valley, maybe more of the consumer and but 0:03:25.000 --> 0:03:27.720 also like you know, obviously a long history. It means 0:03:27.720 --> 0:03:31.000 Silicon Valley for a reason. And I think about various 0:03:31.000 --> 0:03:33.720 parts of East Asia, and when I think about the 0:03:33.800 --> 0:03:36.720 engineering prowess in Europe. Hey, I don't think about tech 0:03:36.760 --> 0:03:39.120 in Europe that much, and when I do think about 0:03:39.200 --> 0:03:42.520 engineering prowess in Europe, it's typically I'm thinking more on 0:03:42.640 --> 0:03:47.000 these sort of like bigger industrial engineering, so a company 0:03:47.080 --> 0:03:51.000 like Siemens or companies that are really good at public 0:03:51.240 --> 0:03:54.560 public works or trains or whatever. And I don't think 0:03:54.600 --> 0:03:58.840 of Europe as being a hotbed of say, semiconductor innovation, 0:03:59.760 --> 0:04:02.520 and there's probably countless like counter examples. I'm just sort 0:04:02.520 --> 0:04:04.800 of thinking, like it doesn't fit into my mental models 0:04:04.800 --> 0:04:07.160 of this stuff. So it is interesting that it's Dutch. 0:04:07.520 --> 0:04:09.880 Well also, just when you think about the European market, 0:04:09.960 --> 0:04:13.360 like you start thinking about the biggest companies there, and yeah, sure, 0:04:13.400 --> 0:04:17.760 stuff like Semens, LVMH like luxury good makers. But mL 0:04:17.920 --> 0:04:21.520 is absolutely massive, and like just looking at the share 0:04:21.560 --> 0:04:24.760 price chart, it has had a huge, huge run up 0:04:25.040 --> 0:04:27.719 over the past year or so, I mean basically since 0:04:27.880 --> 0:04:31.240 the global pandemic, much like a lot of other semiconductor stocks, 0:04:31.240 --> 0:04:35.640 but I mean amazing run up, huge market cap, and 0:04:35.720 --> 0:04:39.359 yet it's sort of like this company that outside of 0:04:39.400 --> 0:04:43.440 the semiconductor sphere, it doesn't seem to get that much attention. Yeah, 0:04:43.520 --> 0:04:45.159 I mean it's A. It's like a it's like a 0:04:45.240 --> 0:04:48.360 third over three billion dollar market cap. It is. It's 0:04:48.640 --> 0:04:51.159 it's one of the biggest companies in the world. Um, 0:04:51.200 --> 0:04:53.200 but not many people know about it, far from my 0:04:53.279 --> 0:04:55.919 household name. Okay, so we have a million questions. So 0:04:55.920 --> 0:04:57.880 we got to get right into this discussion. And we 0:04:57.920 --> 0:05:00.480 have the perfect guest to tell us a out this company. 0:05:00.520 --> 0:05:03.240 We're gonna be speaking with Chris Miller. He's an assistant 0:05:03.279 --> 0:05:07.480 professor of international history at the Fletcher School at Toughs University, 0:05:07.560 --> 0:05:10.120 and he is the author of a forthcoming book that 0:05:10.160 --> 0:05:12.719 will be out next year entitled Chip War, The Struggle 0:05:12.760 --> 0:05:16.440 for the World's Most Critical Technology, And he can answer 0:05:16.520 --> 0:05:20.400 all of our questions about a SML. Chris, thank you 0:05:20.520 --> 0:05:23.760 so much for joining us. Thanks for having me. Chris, 0:05:23.760 --> 0:05:26.000 what is lithography? You know? I think that this is 0:05:26.040 --> 0:05:28.360 one of these questions that's like the word gets It's 0:05:28.360 --> 0:05:31.560 probably come up on like every episode. And I pride 0:05:31.600 --> 0:05:34.920 myself on never being too embarrassed to like ask a question, 0:05:35.320 --> 0:05:38.200 but I think I actually was too embarrassed to ask 0:05:38.760 --> 0:05:42.040 this on some of the other episodes. I'm like, hm, phothography, 0:05:42.520 --> 0:05:46.080 what phithography? So if you want to make a semi 0:05:46.080 --> 0:05:50.080 connected device, you take a slab of silicon, you cover 0:05:50.240 --> 0:05:54.200 it with chemicals called photo resists, which are chemicals that 0:05:54.279 --> 0:05:58.600 react with light, and then you shoot light raise or 0:05:58.680 --> 0:06:02.200 now extreme all tra violet light rays at the silicon 0:06:02.200 --> 0:06:05.800 wafer UH, and the shapes that you shoot at it 0:06:05.839 --> 0:06:09.719 will form the transistors. So that's that's the simplest version. 0:06:09.920 --> 0:06:12.920 Now today, if you buy a new iPhone, the most 0:06:12.960 --> 0:06:16.640 events process around it will have ten billion transistors. So 0:06:16.720 --> 0:06:21.839 you've got to shoot extraordinarily narrow wavelengths of light through 0:06:22.240 --> 0:06:26.880 masks that create these shapes on the silicon wafer, and 0:06:26.960 --> 0:06:30.440 the masks need to be able to project all of 0:06:30.480 --> 0:06:33.000 these shapes onto the wafer. So making this possible at 0:06:33.040 --> 0:06:36.520 the scale of ten billion transistors per chip is is 0:06:36.640 --> 0:06:39.720 what a SML does. Wait, how many per chip? What 0:06:39.839 --> 0:06:42.560 do you say? Ten billion per chip? That's right, a 0:06:42.640 --> 0:06:46.279 new Apple processor in your iPhone will have ten billion 0:06:46.320 --> 0:06:49.360 transistors per chip. Some chips that go into data centers 0:06:49.360 --> 0:06:52.600 will have more than that. But the scale of transistors 0:06:52.600 --> 0:06:55.560 that we produce at any given year is is more 0:06:55.600 --> 0:06:58.560 than the scale of all goods produced by all companies 0:06:58.600 --> 0:07:00.920 and all other industries and all of world history. It's 0:07:00.920 --> 0:07:04.359 a tremendous number. So could you maybe describe where a 0:07:04.520 --> 0:07:08.960 SML sits in the sort of ecosystem of the semiconductor industry. 0:07:09.000 --> 0:07:13.080 So I gather it doesn't seem to have much competition, 0:07:13.200 --> 0:07:16.840 but like who does it actually supply? And also who 0:07:16.880 --> 0:07:19.200 does it not supply? Like are there people out there 0:07:19.240 --> 0:07:22.360 who try to do this on their own. In the 0:07:22.400 --> 0:07:25.560 early days of the chip industry, companies built lithography machines 0:07:25.600 --> 0:07:28.600 in house, so Texas Instruments would have had its own 0:07:28.640 --> 0:07:32.240 Lithography Machine Division IBM. But today the machines are so 0:07:32.280 --> 0:07:35.440 complex and expensive that there's just a couple of companies 0:07:35.520 --> 0:07:39.200 that make lithography machines in general, and just one company, 0:07:39.240 --> 0:07:42.239 a SML, as able to make a u V lithography machines, 0:07:42.280 --> 0:07:45.600 which are the most advanced type. Anyone who operates a 0:07:45.680 --> 0:07:48.520 chip fab sility where chips are made has to buy 0:07:48.600 --> 0:07:51.440 lithography equipment, and so for the most cutting edge chips, 0:07:51.480 --> 0:07:53.480 you've got no choice but to buy from a s mL. 0:07:54.040 --> 0:07:57.200 This is fasting. So whether we're talking about UNTIL doing 0:07:57.240 --> 0:08:00.960 its own chips or TSMC or any one else. And 0:08:01.000 --> 0:08:03.400 we've talked with other people who talked to Stacy Raskin 0:08:03.480 --> 0:08:07.400 of Burnstein and about the nanometer wars and all of them. 0:08:07.600 --> 0:08:12.080 If you're doing cutting edge manufacturing, you are a customer 0:08:12.200 --> 0:08:14.760 of a a s mL. That's right, that's right. For 0:08:14.800 --> 0:08:17.520 the most cutting edge lithography machines, a SML is the 0:08:17.560 --> 0:08:21.720 only supplier for slightly less cutting edge machinery. Uh Night 0:08:21.760 --> 0:08:24.760 kind of Japan is also a competitor of s mls. 0:08:25.000 --> 0:08:27.800 They have a duopoly for anything that's not the most 0:08:27.840 --> 0:08:31.800 cutting edge. So what is it about the technology that 0:08:32.040 --> 0:08:35.840 makes it, I guess so proprietary to a s mL, Like, 0:08:36.000 --> 0:08:38.080 how did they get into a position where they basically 0:08:38.360 --> 0:08:41.000 control it? And what is it that they've been able 0:08:41.080 --> 0:08:44.640 to do that others haven't. So the the challenge with 0:08:44.960 --> 0:08:47.880 EUV lithography in particular and lithography in general is that 0:08:47.920 --> 0:08:52.080 you've got to uh manage a supply chain that is 0:08:52.160 --> 0:08:55.560 extraordinary complex. A SML has got around four thousand suppliers, 0:08:55.600 --> 0:08:59.120 and many of these suppliers are producing equipment that only 0:08:59.160 --> 0:09:01.280 they can produce. So, just to give you a couple 0:09:01.280 --> 0:09:06.800 of examples them, the mirrors within s MLS lithography machines, 0:09:06.840 --> 0:09:10.320 the eu V machines are the flattest structure that humans 0:09:10.360 --> 0:09:12.720 have ever made, the flattest man made structure in the universe. 0:09:13.240 --> 0:09:16.600 And when you go through the list of materials and 0:09:17.200 --> 0:09:20.319 components that you need to produce an e V lithography machine, 0:09:20.360 --> 0:09:22.720 there are multiple parts of the system that are the 0:09:22.760 --> 0:09:26.560 most this or the most that, and managing that is 0:09:26.600 --> 0:09:29.440 an extraordinary complex business. If you talk to people at 0:09:29.440 --> 0:09:32.679 a sm L, they'll say, our biggest engineering challenge is 0:09:33.120 --> 0:09:36.240 not actually engineering any particular part, engineering the supply chain, 0:09:36.320 --> 0:09:38.559 making sure that all of our suppliers are producing things 0:09:38.840 --> 0:09:40.480 so that they all fit together, they all work together, 0:09:40.480 --> 0:09:42.880 they arrive on time. And it's hard enough to do 0:09:42.920 --> 0:09:46.880 that with basic machinery, but when you're trying to manipulate uh, 0:09:46.960 --> 0:09:49.800 individual atoms, which is what SML is able to do, 0:09:50.240 --> 0:09:53.720 it's even more complex. Tracy, I already loved this episode 0:09:53.760 --> 0:09:55.840 so much. I don't know, like how many things like 0:09:55.880 --> 0:09:59.280 I've learned already in five minutes. And then the fact 0:09:59.320 --> 0:10:03.480 that like it's it's also a supply Like Okay, obviously 0:10:03.520 --> 0:10:05.800 there's a chip supply chain, but then the idea that 0:10:05.880 --> 0:10:10.040 the most advanced technology within the chip is actually itself 0:10:10.080 --> 0:10:13.480 a supply chain. I'm just like, I'm already obsessed with this, 0:10:13.600 --> 0:10:16.080 but where do they So Okay, you mentioned that for 0:10:16.480 --> 0:10:18.720 sort of like okay, for the very cutting edge, there's 0:10:18.760 --> 0:10:24.360 just a SML for slightly less cutting edge. Uh, Nikon, 0:10:24.600 --> 0:10:28.240 did you say, Nikon? That's right, Nikon. In Japan, they're 0:10:28.280 --> 0:10:32.680 also in the game. Do other players aspire to be 0:10:32.920 --> 0:10:36.679 cutting edge or is there some barrier that just basically 0:10:36.720 --> 0:10:39.080 makes it so that no one else is really trying 0:10:39.120 --> 0:10:43.520 to get be at that level in the which is 0:10:43.520 --> 0:10:47.439 when investment in EUV began. Nikon made a choice not 0:10:47.600 --> 0:10:51.840 to try to commercialize UV technology. The first physics papers 0:10:51.880 --> 0:10:55.079 on the V actually came out of a Japanese university, 0:10:55.520 --> 0:10:59.280 so there's plenty of optics expertise in Japan that a 0:10:59.440 --> 0:11:01.720 SML was the only company that was willing to bet 0:11:01.720 --> 0:11:05.800 on EUV from the forward and capable of raising the 0:11:05.800 --> 0:11:09.600 funds and assembling the expertise. So right now, if someone 0:11:09.679 --> 0:11:12.800 wanted to replicate what U vs what a SML has 0:11:12.840 --> 0:11:15.199 done with UV, it would take them a decade and 0:11:15.480 --> 0:11:18.880 billions and billions of dollars in investment. And because the 0:11:18.920 --> 0:11:21.920 suppliers that work with a SML have exclusivity agreements with 0:11:21.960 --> 0:11:25.280 a SML. SML has invested in some of his key suppliers. 0:11:26.000 --> 0:11:29.520 It's just basically impossible for anyone to break into this 0:11:29.640 --> 0:11:32.640 without replicating their entire separate supply chain. It would take 0:11:32.679 --> 0:11:52.320 a decade to do. So, maybe this is a good 0:11:52.320 --> 0:11:55.400 place to start talking about the history of the company 0:11:55.480 --> 0:11:57.400 and where it actually came from. And I think that 0:11:57.440 --> 0:12:00.280 will help us like understand some of these dynamics how 0:12:00.280 --> 0:12:03.679 it built up a competitive edge versus it's you know, 0:12:04.080 --> 0:12:08.960 non existent or very few competitors. But my understanding is 0:12:08.960 --> 0:12:13.160 the sort of like sprung up out of US military 0:12:13.200 --> 0:12:16.520 technology of some sort. Can you start like at the beginning? 0:12:16.679 --> 0:12:19.520 How like? I guess this goes back to Joe's question, 0:12:19.640 --> 0:12:22.840 what is lithography? Why is the U. S Army interested 0:12:22.880 --> 0:12:25.319 in it? And how did it play into a s 0:12:25.440 --> 0:12:29.439 ML's creation story. When the transistor was first invented in 0:12:29.480 --> 0:12:32.360 the late nineteen forties in Bell Labs in New Jersey, 0:12:32.400 --> 0:12:36.240 it was predominantly used in military devices for its its 0:12:36.240 --> 0:12:39.959 first commercialization, and there was a scientist in a US 0:12:40.040 --> 0:12:42.600 Army lab named J. Lathrop in the nineteen fifties who 0:12:42.640 --> 0:12:44.760 was trying to find out how to minim miniature eye 0:12:44.800 --> 0:12:48.040 transistors produced them smaller and smaller so they could be 0:12:48.040 --> 0:12:51.800 put in smaller devices. One day, he and his his 0:12:51.840 --> 0:12:56.080 assistant realized that they could use photo resist chemicals, these 0:12:56.160 --> 0:13:00.040 chemicals that react with lights to create shapes on on 0:13:00.160 --> 0:13:02.199 the silicon and germanium that they were working with, and 0:13:02.480 --> 0:13:04.760 they turned their microscope that they were using the lab 0:13:04.840 --> 0:13:08.200 upside down. So normally a microscope lets you see something 0:13:08.240 --> 0:13:10.760 small and it expands the image for your eye. They 0:13:10.880 --> 0:13:13.960 did the opposite. They had a shape that was large 0:13:13.960 --> 0:13:15.880 and were able to project that in a smaller version 0:13:15.880 --> 0:13:18.200 by using an upside down microscope, And with that they 0:13:18.200 --> 0:13:20.960 filed the first patent or photo lithography and coined the 0:13:21.000 --> 0:13:24.960 phrase in the nineteen fifties. Over the next couple of decades, 0:13:25.160 --> 0:13:27.880 photo lithography was used both by chipmakers who were making 0:13:27.920 --> 0:13:30.640 their own machines and house and then eventually a couple 0:13:30.640 --> 0:13:34.800 of specialized photo lithography companies emerged in Connecticut and Massachusetts. 0:13:35.160 --> 0:13:38.200 They were defense contractors primarily, but realized they could use 0:13:38.240 --> 0:13:40.960 their specialized optics things that they had honed in spy 0:13:41.040 --> 0:13:44.880 satellites and military equipment like that for the semiconductor industry, 0:13:45.040 --> 0:13:47.720 and so until the mid nineteen eighties, the center of 0:13:47.760 --> 0:13:50.560 the photo lithography industry was in New England, but those 0:13:50.600 --> 0:13:52.880 companies faced hard times in the nineteen eighties. They were 0:13:53.360 --> 0:13:56.120 poorly managed, and the nineteen eighties were a time when 0:13:56.120 --> 0:13:58.960 the Japanese ship industry in general was rising. A Nikon 0:13:59.320 --> 0:14:01.280 as well as can and then the two camera companies 0:14:01.320 --> 0:14:04.240 began investing in photo lithography. For a time in the 0:14:04.240 --> 0:14:08.160 eighties and nineties, they were the dominant companies in the industry, 0:14:08.280 --> 0:14:11.400 which the US was quite worried about. Worried about being 0:14:11.400 --> 0:14:14.240 too reliant on Japan at a time of commercial and 0:14:14.280 --> 0:14:19.000 also geopolitical tension, and so US chip makers began turning 0:14:19.000 --> 0:14:22.640 to a s mL, both to diversify their supplier base, 0:14:22.680 --> 0:14:25.680 but also because a SML was able to produce very 0:14:25.760 --> 0:14:29.240 high quality equipment as well. In the nineteen nineties in 0:14:29.320 --> 0:14:32.080 the mid as well, was the only company willing to 0:14:32.120 --> 0:14:34.520 take the gamble on EUV, and since that point it's 0:14:34.520 --> 0:14:38.240 become the dominant firm in the industry. We just have 0:14:39.160 --> 0:14:42.480 extreme ultra violet. I don't know if we just want 0:14:42.520 --> 0:14:46.440 to make sure we've established what UV stands for? Can 0:14:46.480 --> 0:14:49.120 I just ask so, can you explain again? What's the 0:14:49.160 --> 0:14:53.320 difference between extreme ultra violet and I guess non extreme 0:14:53.480 --> 0:14:56.800 ultra violet. So over the past couple of decades is 0:14:56.840 --> 0:15:00.320 we've tried to make ever smaller device is in ever 0:15:00.440 --> 0:15:04.640 smaller features on silicon wafers. We've begun to use different 0:15:04.720 --> 0:15:09.080 and smaller wavelengths of light, and so extreme multi violet 0:15:09.200 --> 0:15:11.640 has a wavelength of thirteen point five nanometers. Is the 0:15:11.680 --> 0:15:14.680 smallest wavelength of light that we've been able to use 0:15:14.760 --> 0:15:18.240 in in mass productions. So if you rewind several decades ago, 0:15:18.280 --> 0:15:21.440 we were using larger wavelengths of light that we're uncapable 0:15:21.480 --> 0:15:24.440 of producing the small feature sizes on silicon wafers that 0:15:24.480 --> 0:15:27.560 we demand today. One of the reasons why I think 0:15:28.040 --> 0:15:32.200 the chip episodes, well, why we've done so many chip episodes, 0:15:32.240 --> 0:15:34.880 and why they're why they keep resonating. I think there's 0:15:34.880 --> 0:15:38.040 a few things. I mean, one is there's the chip shortage, 0:15:38.200 --> 0:15:40.280 and it should be noted that the shortage is actually 0:15:40.400 --> 0:15:42.960 wore at the is not really at the advanced level. 0:15:42.960 --> 0:15:45.160 It's a lot of cheap chips, etcetera. But we're starting 0:15:45.200 --> 0:15:50.480 to the chip shortage that relates to automobiles, etcetera. Has 0:15:50.520 --> 0:15:53.200 sort of brought people a lot of awareness about lack 0:15:53.280 --> 0:15:57.200 of domestic US manufacturing capacity. I think another reason people 0:15:57.240 --> 0:16:00.800 care about chips is obviously just the general explosion of 0:16:00.880 --> 0:16:04.600 chip demand. Even where there isn't an acute shortage, there's 0:16:04.680 --> 0:16:07.200 chips h and everything. And then I think the other 0:16:07.240 --> 0:16:10.400 thing that makes it sort of an interesting story right 0:16:10.400 --> 0:16:13.200 now is that, at least in the US and probably 0:16:13.240 --> 0:16:16.400 elsewhere around the world, there is a rethinking about the 0:16:16.480 --> 0:16:21.400 role of state capacity in and um state investment into 0:16:21.480 --> 0:16:23.800 certain space. And of course, as we all know, the 0:16:23.880 --> 0:16:26.080 chip industry and as you just talked about, the chip 0:16:26.120 --> 0:16:29.320 industry overall really was sort of born out of defense, 0:16:29.360 --> 0:16:32.520 so like sort of the ultimate in UH state investing 0:16:32.680 --> 0:16:38.320 and government spending, and at various times throughout US history 0:16:38.360 --> 0:16:40.880 at least we seem to go in waves of how 0:16:41.000 --> 0:16:43.200 much the government wants to get in to protect the 0:16:43.280 --> 0:16:46.040 chip sector, to invest in the chip sector, to build 0:16:46.160 --> 0:16:50.000 and bolster a homegrown chip sector. You mentioned the sort 0:16:50.040 --> 0:16:53.480 of stress and tension with the Japanese or reliance on 0:16:53.560 --> 0:16:56.760 Japanese companies in the eighties and nineties that seemed to 0:16:57.120 --> 0:17:00.160 produce a wave of um sort of defensive and use. 0:17:00.400 --> 0:17:03.720 Perhaps it could be characterized. Talk to us about how 0:17:03.760 --> 0:17:07.080 a SML fits into that in terms of you know, 0:17:07.119 --> 0:17:09.159 when we talked about to say, the history of t SMC, 0:17:09.800 --> 0:17:12.159 that was clearly in part it was a very like 0:17:12.240 --> 0:17:15.840 public sort of private venture. There was the government backed 0:17:15.840 --> 0:17:18.680 it up under the condition that it could raise private 0:17:18.960 --> 0:17:21.520 foreign money as well. Talk to us about the role 0:17:21.560 --> 0:17:25.240 of like public money in the creation of a SML. 0:17:25.880 --> 0:17:29.040 So a sm L emerged first as a division of 0:17:29.040 --> 0:17:31.840 of Phillips, the Dutch electronic company UM and it was 0:17:32.000 --> 0:17:36.280 spun out in at a time when the Europeanship industry 0:17:36.400 --> 0:17:39.240 was relatively small as a player on the world stage. 0:17:39.240 --> 0:17:40.720 It was the U S and Japan at the time 0:17:40.720 --> 0:17:43.280 that were the biggest players, and there were a variety 0:17:43.320 --> 0:17:46.879 of Dutch and European Union programs to support R and 0:17:47.000 --> 0:17:49.960 D at s mL. But for a SML in particular, 0:17:50.400 --> 0:17:53.320 actually the most important government support was from the U 0:17:53.359 --> 0:17:56.399 S government because in in the nineteen nineties, when the 0:17:56.480 --> 0:18:00.240 investments and EUV were being made, Intel, which at the 0:18:00.240 --> 0:18:03.880 time was the the industry leader in hip making, decided 0:18:03.880 --> 0:18:05.800 to take a big bet on eu V being the 0:18:05.840 --> 0:18:11.119 next lithography technology and established a consortium of a number 0:18:11.119 --> 0:18:13.960 of private ship firms and a number of US national labs, 0:18:14.480 --> 0:18:17.359 uh the Lawrence Livermore for example, that would work together 0:18:17.480 --> 0:18:21.720 to produce prototype UV machines. And so the technology that 0:18:22.200 --> 0:18:25.199 we use today an SML Systems really comes from this 0:18:25.280 --> 0:18:28.320 work with US national labs. It was largely funded by 0:18:28.359 --> 0:18:31.760 industry but using the scientists there. And at the time 0:18:32.080 --> 0:18:34.879 there was some interest in trying to turn the technology 0:18:34.920 --> 0:18:37.440 over to a US company to producing commercialized on the 0:18:37.440 --> 0:18:40.200 grounds that it came largely out of US national labs, 0:18:40.240 --> 0:18:42.760 but there was no US lithography firm at the time 0:18:42.800 --> 0:18:45.919 that was seen as a credible candidate to commercialize. If 0:18:45.920 --> 0:18:48.919 the options were NICON or a SML. Given the tensions 0:18:48.920 --> 0:18:52.280 with Japan, a SML was seen as the least risky option, 0:18:52.680 --> 0:18:55.240 and also they had a long track record of producing 0:18:55.359 --> 0:18:58.640 quality machines. And so we've got this strange situation now 0:18:58.680 --> 0:19:02.160 where a lot of the core technology in this machinery 0:19:02.200 --> 0:19:05.480 that's assembled in the Netherlands actually comes out of California. 0:19:05.600 --> 0:19:08.360 And indeed SML has actually bought a number of companies 0:19:08.359 --> 0:19:10.760 over the course of the past couple of decades in 0:19:10.800 --> 0:19:13.320 California as well. So there's a lot of US technology 0:19:13.840 --> 0:19:16.840 in a SML Systems partially funded by the US government. 0:19:17.040 --> 0:19:19.760 Could you imagine something like that happening today? Like, I 0:19:19.840 --> 0:19:23.520 just think the environment is so different, and the idea 0:19:23.560 --> 0:19:25.879 of like the U S government funding a technology and 0:19:25.920 --> 0:19:28.880 then deciding like, well, okay, I guess the best company 0:19:29.000 --> 0:19:32.119 to actually make this stuff is over in the Netherlands, 0:19:32.119 --> 0:19:33.960 so we'll just let them do it and give up 0:19:34.040 --> 0:19:37.560 like a key component of a highly competitive supply chain. 0:19:37.600 --> 0:19:43.320 It just seems so so unlikely in the current environment. Yeah, 0:19:43.359 --> 0:19:45.240 it's it's an interesting question. On the one hand, you 0:19:45.280 --> 0:19:47.840 do hear a lot of conversation in Washington, d C. 0:19:48.000 --> 0:19:50.760 About joint R and D project with Allied countries, and 0:19:50.840 --> 0:19:52.800 in some ways this is a perfect example of this. 0:19:52.960 --> 0:19:55.200 I think. The other thing is that a SML is 0:19:55.240 --> 0:19:58.040 a Dutch company, But if you look at the components 0:19:58.080 --> 0:20:01.560 of their EUV machines, for example, they're sourcing from all 0:20:01.600 --> 0:20:05.040 around Europe, around the US and really worldwide. So to 0:20:05.119 --> 0:20:08.000 describe them as as a Dutch company misses the fact 0:20:08.040 --> 0:20:11.080 that you can't produce an EUV system was for example, 0:20:11.119 --> 0:20:13.639 the light source, which is produced by an S and 0:20:13.720 --> 0:20:16.080 L subsidiary in San Diego, so that there are a 0:20:16.119 --> 0:20:19.400 Dutch company S, but they're really a global supply chain 0:20:19.480 --> 0:20:22.280 that's focused on on the U S and Europe. So 0:20:22.359 --> 0:20:25.679 this is interesting because you mentioned that, Okay, at the 0:20:25.800 --> 0:20:29.560 time that the technology was sort of they decided a 0:20:29.640 --> 0:20:32.800 SML would be the most credible entity to commercialize this 0:20:33.119 --> 0:20:37.040 sort of US funded technology. There was this view that, Okay, 0:20:37.160 --> 0:20:41.840 it was better them than a Japanese player, in part 0:20:41.920 --> 0:20:45.679 because we already had anxiety about our reliance on Japanese 0:20:45.720 --> 0:20:49.160 chips at the time for other other chips, including DAM. 0:20:49.200 --> 0:20:52.160 How much are the same dynamics essentially in play when 0:20:52.160 --> 0:20:56.800 people think about the geopolitics of chips. Obviously one of 0:20:56.840 --> 0:20:59.640 the things that you know, we talked about anxiety about 0:20:59.680 --> 0:21:03.399 how my we rely on Taiwan. There's perhaps some anxiety 0:21:03.560 --> 0:21:07.240 about the domestic homegrown chip industry. Although China seems to 0:21:07.280 --> 0:21:10.840 be a several years behind in terms of mainland chip technology. 0:21:11.280 --> 0:21:13.920 How much does it still sort of benefit everyone this 0:21:14.000 --> 0:21:18.640 idea that this crucial component player is not part of 0:21:18.840 --> 0:21:22.080 either in US or Asia. That's an interesting question. I 0:21:22.080 --> 0:21:25.480 think certainly if you're a Chinese customer of SML, you're 0:21:25.560 --> 0:21:28.159 pleased that it's not a U S company, But the 0:21:28.240 --> 0:21:30.760 reality is that if if the US wanted to use 0:21:30.800 --> 0:21:34.119 actually poor controls to constrict SML sales to China, that 0:21:34.160 --> 0:21:36.879 wouldn't be very difficult to do. A SML already doesn't 0:21:36.920 --> 0:21:40.800 send it's eu V machines to China. In theory, that's 0:21:40.840 --> 0:21:44.200 because of Dutch restrictions. In reality, it's because of US 0:21:44.240 --> 0:21:47.760 pressure on the Netherlands to impose these restrictions. And there's 0:21:47.800 --> 0:21:52.160 discussion um in in Washington and Japan elsewhere about whether 0:21:52.200 --> 0:21:54.840 there ought to be stricter limits um the type of 0:21:54.840 --> 0:21:58.080 lithography machines you can sell to China, and legally there's 0:21:58.160 --> 0:22:00.800 there's nothing that would really stop the US from posting 0:22:00.800 --> 0:22:05.720 those restrictions unilaterally. Is the concerning that if those machines 0:22:05.920 --> 0:22:09.919 were shipped to China, that they would be able to 0:22:10.200 --> 0:22:15.320 UH that would accelerate China's semiconductor capabilities, or that literally 0:22:15.359 --> 0:22:18.040 having them in Chinese hands would then maybe allow them 0:22:18.080 --> 0:22:21.760 to be more easily sort of deconstructed and reverse engineered 0:22:22.320 --> 0:22:25.360 and that would be a big knowledge transfer. No, it's 0:22:25.400 --> 0:22:28.439 the former. I think if you just receive an SMaL machine, 0:22:28.480 --> 0:22:30.960 you have no idea how to produce it. Okay, Okay, 0:22:31.119 --> 0:22:33.600 it's that the more advanced cocoric machines you have, the 0:22:33.640 --> 0:22:36.480 more advanced shipmaker you have, the stronger the Chinese ecosystem is. 0:22:37.240 --> 0:22:40.960 So how big of an impediment is not having access 0:22:41.040 --> 0:22:44.840 to a sml S eu V technology to Beijing's like 0:22:44.960 --> 0:22:50.520 overall semiconductor development drive, Like, is it such an essential 0:22:50.520 --> 0:22:53.080 piece of technology that it basically means they're on a 0:22:53.119 --> 0:22:58.200 completely different footing to something like TSMC. That's right. For now, 0:22:58.240 --> 0:23:02.280 there's there's no viable of producing the most advanced chips 0:23:02.280 --> 0:23:06.639 with the smallest features without using UV. There are some 0:23:06.640 --> 0:23:09.000 some scientists who think there might theoretically be ways to 0:23:09.000 --> 0:23:12.040 get around it, but for the next decade there's there's 0:23:12.080 --> 0:23:14.280 just no choice but to use a SMLS machinery if 0:23:14.320 --> 0:23:18.240 you want to produce the smallest chips. So what are 0:23:18.320 --> 0:23:20.560 you know, let's talk a little bit more about a 0:23:20.760 --> 0:23:24.440 s MLS constraints. Everyone this year is becoming aware of, like, 0:23:24.600 --> 0:23:28.159 you know, constraints, and there's only so much boundary capacity 0:23:28.520 --> 0:23:31.960 in the world at any given time. The entities that 0:23:32.040 --> 0:23:34.680 wanted to buy cheap chips that go into cars sort 0:23:34.680 --> 0:23:36.920 of got shut out because they canceled their order for 0:23:36.920 --> 0:23:38.600 a while. And now they're scrambling and it might be 0:23:38.680 --> 0:23:41.399 years before they could catch up again. So we know 0:23:41.560 --> 0:23:45.680 that that's constraints. How strained is a s mls own 0:23:46.080 --> 0:23:49.000 capacity to grow and where do they face Is it 0:23:49.160 --> 0:23:51.680 just in the complexity of the supply chain? Is it 0:23:51.800 --> 0:23:55.280 in raw materials? Like what are their constraints? It's mostly 0:23:55.280 --> 0:23:58.600 in the supply chain complexity. So SML last year's shipped 0:23:58.680 --> 0:24:02.439 thirty one EUV machine sans, so we're talking getting one 0:24:02.520 --> 0:24:04.240 or two more machines out of the out of their 0:24:04.280 --> 0:24:07.240 production process is something that's hard to do because each 0:24:07.280 --> 0:24:10.760 of their suppliers is similarly constrained. And the ability to 0:24:10.840 --> 0:24:14.320 ramp up manufacturing, you know, this isn't high volume manufacturing 0:24:14.320 --> 0:24:17.240 when you're producing thirty one machines a year, and because 0:24:17.280 --> 0:24:20.040 their supply chain has so many specialized parts solely for 0:24:20.080 --> 0:24:23.199 their machines, their suppliers are producing thirty one or so 0:24:23.560 --> 0:24:26.439 of the components needed each year, and so there's just 0:24:26.480 --> 0:24:28.880 no way to ramp How many three hundred billion dollar 0:24:29.000 --> 0:24:33.679 companies in the world make produce thirty one make thirty 0:24:33.680 --> 0:24:35.720 one units a year, so we're talking like each one 0:24:35.800 --> 0:24:38.119 is like half a goon or something. There's thirty one 0:24:38.160 --> 0:24:41.199 of the of EUV machines. They will also sell some 0:24:41.240 --> 0:24:45.120 of the older equipment, but they sold units in total 0:24:45.200 --> 0:24:47.119 last year, so it's still a tiny number of units. 0:24:47.119 --> 0:24:49.240 It's still not very much. How much is it if 0:24:49.280 --> 0:24:51.439 you or I wanted to pull together and by a 0:24:51.600 --> 0:24:53.919 EUV machine? Like what are they? What are they retail for? 0:24:54.320 --> 0:24:57.480 Average average revenue per UV machine last year was around 0:24:57.480 --> 0:25:01.880 a hundred forty million euros. Got it? Okay, So this 0:25:01.920 --> 0:25:03.879 is something that comes up a lot in our supply 0:25:04.040 --> 0:25:08.320 chain discussions. But like how does ordering actually work and 0:25:08.480 --> 0:25:12.080 is there preference given to certain customers over others? Like, 0:25:12.200 --> 0:25:14.720 you know, if one company wants to buy an EUV machine, 0:25:14.760 --> 0:25:16.760 I imagine there there are plenty of other companies that 0:25:16.800 --> 0:25:18.639 want to do the same thing. For a limited supply 0:25:19.400 --> 0:25:22.320 How does a SML actually make the decisions about who 0:25:22.359 --> 0:25:25.400 gets allocated what? Also, how long does it take? Like 0:25:25.800 --> 0:25:30.159 what's the waiting time to actually get one of these things? Yeah, 0:25:30.440 --> 0:25:33.480 we don't really know the details as to how SML 0:25:33.560 --> 0:25:36.560 decides to allocate the number of potential customers for a 0:25:36.640 --> 0:25:39.800 hundred fifty million dollar machine. Is is limited? It's a 0:25:39.840 --> 0:25:43.160 half dozen potential customers in the world, you be realistically 0:25:43.160 --> 0:25:45.080 looking to buy one. But if you look at the 0:25:45.119 --> 0:25:48.600 main customers, it's t SMC, which has around half of 0:25:48.600 --> 0:25:52.680 operating the UV machines in its own fabs, Samsung, Intel, 0:25:52.760 --> 0:25:56.800 a handful of others, and there's almost certainly a premium 0:25:56.800 --> 0:26:00.520 that t SMC is paid for getting so many machines valuable. 0:26:00.600 --> 0:26:02.320 If if if a new company came online and wanted to 0:26:02.359 --> 0:26:04.800 buy machines, they'd face a weight of at least a 0:26:04.800 --> 0:26:08.679 couple of years because capacity has been purchased and advanced. 0:26:09.160 --> 0:26:11.600 Intel has said it's going to be the first customer 0:26:11.680 --> 0:26:15.000 of the next generation EUV machine, which would be online 0:26:15.040 --> 0:26:18.960 around Presumably it's paid something for the right to get 0:26:18.960 --> 0:26:21.760 the first generation, but we don't know any of the details. Yeah, 0:26:21.880 --> 0:26:23.840 speaking of Intel, and I want to back up to 0:26:23.880 --> 0:26:26.640 something we talked about earlier. Why was this never part 0:26:26.680 --> 0:26:30.800 of Intel's own ambitions? Because you know, over the years, 0:26:31.080 --> 0:26:34.760 I guess the degree to which Intel has wanted to 0:26:34.840 --> 0:26:37.679 be an i P first company or a manufacturing company 0:26:37.680 --> 0:26:39.920 at Waxes and wind, so at one point it was 0:26:39.960 --> 0:26:43.560 a manufacturing powerhouse, then has sort of scaled. Back then 0:26:43.680 --> 0:26:45.400 it was more of an I P company, and that's 0:26:45.400 --> 0:26:49.080 sally the anxiety these days now they seem to want 0:26:49.200 --> 0:26:52.680 to get back into being uh manufacturing, and the new 0:26:52.760 --> 0:26:55.320 CEO has made a point of like, we're not going 0:26:55.400 --> 0:27:00.879 to give up on being a manufacturing powerhouse. Why was 0:27:01.000 --> 0:27:07.359 lithography or advanced lithography never part of the Intel strategy? Well, 0:27:07.359 --> 0:27:09.920 when INTIL was founded, it was. It was founded at 0:27:09.920 --> 0:27:12.560 a time where you could already buy lithography equipment on 0:27:12.600 --> 0:27:15.040 the open market, so they always decided they were going 0:27:15.119 --> 0:27:19.480 to produce ships, but by lithography machines from suppliers. Over 0:27:19.520 --> 0:27:22.040 their fifty years, they've been one of the biggest buyers 0:27:22.040 --> 0:27:25.119 of lithography equipment in the industry, and the development of 0:27:25.160 --> 0:27:28.200 e V actually wouldn't have happened without Intel. When Andy 0:27:28.240 --> 0:27:31.080 Grove was still the Intel CEO in the early nine nineties, 0:27:31.560 --> 0:27:33.920 he made the first big bet on the development of 0:27:33.960 --> 0:27:36.760 EV lithography, putting up two or million dollars in nearly 0:27:37.160 --> 0:27:39.959 nines to begin to develop this on the grounds not 0:27:39.960 --> 0:27:42.760 that Intel was ever going to produce lithography equipment, but 0:27:42.800 --> 0:27:46.200 that it would eventually need EUV to produce the most 0:27:46.240 --> 0:27:50.320 advanced ships. And even as as recently as when a 0:27:50.440 --> 0:27:53.560 SML needed to raise more capital to to fund its 0:27:53.560 --> 0:27:57.200 development of UV. It went to Intel, t SMC, and Samsung, 0:27:57.480 --> 0:27:59.640 and Intel was the biggest investor in a SML at 0:27:59.640 --> 0:28:02.480 the time, putting in a several billion dollars to help 0:28:02.520 --> 0:28:06.080 fund SML's development. So until quite recently, it seemed like 0:28:06.119 --> 0:28:09.960 Intel would be the biggest user of e V lithography machines. 0:28:10.440 --> 0:28:13.560 It's only in the past couple of years that Intel decided, 0:28:14.080 --> 0:28:16.560 in what looks to be a mistake in hindsight, that 0:28:16.920 --> 0:28:20.160 EV wouldn't be ready by around now, where t SMC 0:28:20.359 --> 0:28:23.080 maybe opposite that that EUV would be ready for high 0:28:23.160 --> 0:28:26.240 value manufacturing. And t SMC was proved right, which is 0:28:26.240 --> 0:28:28.159 why it's done so well the past couple of years, 0:28:28.200 --> 0:28:31.480 and Intel has proven wrong, which I think most observers 0:28:31.480 --> 0:28:34.119 can explain some of the manufacturing problems it's had in 0:28:34.160 --> 0:28:36.960 recent years by not using EUV and trying to use 0:28:37.000 --> 0:28:40.000 older versions of lithography to produce its most advanced ships. 0:28:40.480 --> 0:28:43.920 Now Intel is changing it's it's it's plans, it's it 0:28:43.960 --> 0:28:46.440 says it's investing very heavily in UV, but it's gonna 0:28:46.480 --> 0:28:49.160 take a couple of years for them to learn how 0:28:49.200 --> 0:28:52.720 to actually use the UV and high volume manufacturing. Wait, 0:28:52.760 --> 0:28:55.520 so I have a slightly related question, although maybe it's 0:28:55.560 --> 0:28:59.800 sort of reversed, I guess, but like, given a SML 0:29:00.840 --> 0:29:05.880 competitive edge in producing a key technological component for semiconductors, 0:29:07.000 --> 0:29:11.120 could they ever have just gone into making semiconductors themselves? 0:29:11.240 --> 0:29:13.480 Like if they have a monopoly on this technology, no 0:29:13.520 --> 0:29:15.840 one can do it as well as them, Like why 0:29:15.880 --> 0:29:20.440 not just start making the finished product yourself. To make 0:29:20.440 --> 0:29:22.880 a chip, you not only need lithography, which is one 0:29:22.880 --> 0:29:25.360 of the key steps, but there are other steps as well. 0:29:25.440 --> 0:29:28.880 You need to be able to deposit films of materials 0:29:28.960 --> 0:29:32.360 um with with atomic level precision. There's there are different 0:29:32.360 --> 0:29:35.720 companies Applied Materials, for example, in California that make that 0:29:35.800 --> 0:29:38.959 type of equipment. You need measurement equipment that can measure 0:29:39.360 --> 0:29:42.840 individual atomic level errors in your final chips to make 0:29:42.840 --> 0:29:46.360 sure you understand what errors you have. That's a different 0:29:46.360 --> 0:29:48.800 set of companies that makes that equipment. UM. So there's 0:29:48.920 --> 0:29:51.720 a lot of different specialized equipment that you need to 0:29:52.160 --> 0:29:55.080 make chips. The cost of a new fab that for example, 0:29:55.120 --> 0:29:57.480 t SMC will build more than half of that cost 0:29:57.560 --> 0:29:59.440 is the equipment that goes in it and a SML 0:29:59.720 --> 0:30:02.040 and as ptography machines are a critical part of that, 0:30:02.120 --> 0:30:05.320 but that's far from enough to make chips, and SML 0:30:05.440 --> 0:30:08.719 specialty is really solely in lithography UH and not at 0:30:08.720 --> 0:30:11.000 all in deposition or matching or the other types of 0:30:11.240 --> 0:30:14.360 equipment that you need to actually make finished chips. This 0:30:14.440 --> 0:30:16.320 at this point is so fascinating to me, like to 0:30:16.400 --> 0:30:19.280 think about, like, okay, a s mL among the many 0:30:19.360 --> 0:30:22.840 extraordinary things. They also lay claim to having the flattest 0:30:22.880 --> 0:30:26.400 service in the world, and presumably in order to get 0:30:26.400 --> 0:30:28.959 the flattest service of the world, there's some technology, as 0:30:29.000 --> 0:30:31.040 you sort of just mentioned, that has to be able 0:30:31.040 --> 0:30:34.680 to measure flatness and actually measure if the services not 0:30:34.800 --> 0:30:37.320 flat and it sort of speaks to like you know, 0:30:37.400 --> 0:30:39.640 we think about like in the US these days, and 0:30:39.760 --> 0:30:42.800 there is UH. There's a bill in d C that's 0:30:42.880 --> 0:30:48.440 designed to invest in UH, designed to bolster US capacity. 0:30:48.840 --> 0:30:51.680 And again that's part of why you keep having these discussions, 0:30:51.720 --> 0:30:54.880 because there's this effort on the way. It's kind of bipartisan. 0:30:54.960 --> 0:30:57.440 The bill might pass, it might not pass, but there's 0:30:57.440 --> 0:31:01.440 this sort of bipartisan interest to bolster domestic capacity. But 0:31:01.520 --> 0:31:04.440 I don't even know what that means sometimes because obviously, 0:31:04.440 --> 0:31:08.280 as you describe, the internationalization and complexity of the chip 0:31:08.480 --> 0:31:13.120 supply chain is so extreme, and as we've talked about 0:31:13.120 --> 0:31:16.240 with other guests, chips across borders a million times before 0:31:16.240 --> 0:31:20.520 they arrive in your xbox or your iPhone or whatever. 0:31:20.560 --> 0:31:23.160 It is, Like, what does it even mean in your 0:31:23.240 --> 0:31:25.160 view just to think about like this question of like 0:31:25.440 --> 0:31:30.840 expanding domestic capacity in an industry that but that just 0:31:31.080 --> 0:31:36.240 is so extremely fragmented and international. Yeah, I think the 0:31:36.280 --> 0:31:38.000 first thing is you've got to be specific as to 0:31:38.080 --> 0:31:41.520 what type of capacity you wanted to spand expand domestically. 0:31:41.720 --> 0:31:44.760 Certainly the US could expand production of chips domestically if 0:31:44.800 --> 0:31:47.800 we wanted to, but that wouldn't have any effect on 0:31:47.880 --> 0:31:51.080 the reality that there's no way to buy lithography equipment, 0:31:51.120 --> 0:31:55.560 for example, except from foreign suppliers, either Nicon or a SML. 0:31:56.160 --> 0:31:58.480 I think domestic production is a is a great thing 0:31:58.600 --> 0:32:01.160 to support UM, but thesis that we're going to have 0:32:01.360 --> 0:32:05.160 a entirely domestically produced supply chain is a fantasy. The 0:32:05.200 --> 0:32:08.160 only reason that we're able to produce ships with such 0:32:08.160 --> 0:32:10.680 small features is because we're able to take advantage of 0:32:11.000 --> 0:32:14.400 expertise from companies in many different countries around the world, 0:32:14.800 --> 0:32:17.000 and there's no one in the industry who thinks there's 0:32:17.040 --> 0:32:19.520 any conceivable future even if you're to spend a trillion 0:32:19.520 --> 0:32:22.400 dollars over a decade where you'd get a domestically produced 0:32:22.400 --> 0:32:24.760 supply chain as anywhere near as efficient. Is what we've 0:32:24.760 --> 0:32:27.840 got now, you know. I think the supply chain risk 0:32:28.240 --> 0:32:31.400 discussion is often takes place at a thirty foot level, 0:32:31.440 --> 0:32:33.040 when what you really need to look at is what 0:32:33.120 --> 0:32:36.200 of the specific components you're worried about, Are there specific 0:32:36.280 --> 0:32:39.959 suppliers you're worried about, and how can you mitigate those 0:32:40.000 --> 0:32:43.440 specific risks? But just talking about domestic versus foreign production 0:32:43.480 --> 0:32:45.800 is not nearly specific enough to have any sort of 0:32:46.000 --> 0:32:49.160 real meaning. So we kind of mentioned this in the intro. 0:32:49.400 --> 0:32:52.600 But again, one of the themes that comes up repeatedly 0:32:52.760 --> 0:32:56.440 on these episodes is the idea of supply chains all 0:32:56.440 --> 0:32:59.600 the way down. So if there's a bottleneck in one thing, 0:32:59.840 --> 0:33:03.760 there's probably a bottleneck um and a component and even 0:33:03.840 --> 0:33:06.560 smaller component that leads into that one thing. So if 0:33:06.560 --> 0:33:11.720 there's a bottleneck in lithography equipment that's impacting semiconductors, I 0:33:11.720 --> 0:33:14.640 guess my question is is there a bottleneck in something 0:33:14.680 --> 0:33:17.840 that goes into the lithography machines or the eu V 0:33:18.080 --> 0:33:22.360 machines that is preventing a s m L from expanding production. 0:33:23.800 --> 0:33:26.800 It certainly could be there's not enough public information about 0:33:26.920 --> 0:33:30.320 SML supply chain to know, and it's very plausible that 0:33:30.400 --> 0:33:33.560 a SML, despite being real experts at management supply chain, 0:33:33.600 --> 0:33:38.040 doesn't always uh No. They report having around four thousands suppliers, 0:33:38.520 --> 0:33:40.640 and all of their suppliers who you speak to will 0:33:40.680 --> 0:33:44.160 say they ask lots of questions about their suppliers suppliers, 0:33:44.960 --> 0:33:48.320 But the reality is that there are you know, multiple 0:33:48.400 --> 0:33:51.160 orders of magnitude more suppliers of their suppliers and so 0:33:51.240 --> 0:33:54.320 tracing them all down the chain is is basically impossible. 0:33:54.440 --> 0:33:56.720 So what a SML tries to do is understand what 0:33:56.760 --> 0:33:59.600 other biggest risks. They've even gone so far as to 0:33:59.600 --> 0:34:02.920 purchase some of their suppliers to gain more detailed control 0:34:03.080 --> 0:34:08.200 over managing those risks. But they simply can't know every 0:34:08.280 --> 0:34:11.840 ultimate component that goes into all of all their suppliers systems, 0:34:12.160 --> 0:34:14.239 and so that's that's where the supply chain management just 0:34:14.280 --> 0:34:17.000 becomes extraordinarily difficult. Now what they've been really good at 0:34:17.320 --> 0:34:20.399 I think better than their competitors over the past couple 0:34:20.440 --> 0:34:23.840 of decades is managing that so it hasn't generally caused 0:34:24.760 --> 0:34:27.560 any sort of serious delays. And one of the things 0:34:27.560 --> 0:34:30.319 that they're known for with their customers is being able 0:34:30.360 --> 0:34:33.240 to deliver mostly on time when they promise a machine 0:34:33.360 --> 0:34:35.399 and managing this, which is something that no one else 0:34:35.440 --> 0:34:37.879 has been able to manage. I think. The other thing 0:34:38.200 --> 0:34:40.200 to note is that you know, when you've got this 0:34:40.280 --> 0:34:45.279 equipment that is is manipulating individual atoms or trying to 0:34:46.120 --> 0:34:50.480 trying to control the movement of light with extraordinary precision. 0:34:51.000 --> 0:34:52.960 It's it's one thing to have a machine that will 0:34:53.000 --> 0:34:55.640 do this once or twice or sporadically. It's another thing 0:34:55.680 --> 0:34:57.279 to have a machine that will do this day and 0:34:57.360 --> 0:35:00.880 day out, operating twenty four hours a day. And that's 0:35:00.920 --> 0:35:03.440 that's the other thing that a SML has done very 0:35:03.480 --> 0:35:06.319 successfully over the past couple of years. It was clear 0:35:06.360 --> 0:35:08.480 as early as the nine nineties that it was possible 0:35:08.520 --> 0:35:13.040 to make a chip with evy lithography. What's been difficult 0:35:13.160 --> 0:35:15.880 is making millions of ships with the euvy lithography and 0:35:15.920 --> 0:35:18.719 doing in a cost effective way. And that's what a 0:35:18.840 --> 0:35:22.360 smlls really stood out is that their machines are rarely broken, 0:35:22.840 --> 0:35:26.879 always functioning um They need less um, less tuning, less 0:35:26.920 --> 0:35:29.839 cleaning than their competitors. That's what a sml has done 0:35:30.200 --> 0:35:32.680 quite well. So it's not simply managing the physics, which 0:35:32.719 --> 0:35:35.120 is very hard, but it's also making sure that you've 0:35:35.160 --> 0:35:38.319 got this extraordinary precise physics that's always operating in these 0:35:38.360 --> 0:35:41.120 act way you expected to operate. Yeah, I'm thinking back 0:35:41.160 --> 0:35:45.000 to one of our discussions with HBS professor. Really, she 0:35:45.680 --> 0:35:48.279 and I don't remember the math exactly, but if like 0:35:48.400 --> 0:35:52.160 chip making is like a seven thousand step process, then 0:35:52.239 --> 0:35:56.480 even you know, ninety nine point nine nine percent execution 0:35:56.600 --> 0:36:00.759 in each step is insufficient in many cases because by 0:36:00.800 --> 0:36:04.000 the time you're down to the seventh seven thousand, you've 0:36:04.080 --> 0:36:06.480 like basically lost all your chips. I don't remember the 0:36:06.520 --> 0:36:09.640 exact math, but it is interesting to think about like 0:36:09.960 --> 0:36:14.239 building up that comp that competent now just in can 0:36:14.280 --> 0:36:16.640 the machine make the chip, but can make it over 0:36:16.760 --> 0:36:20.920 and over and over again without without many errors. If 0:36:20.920 --> 0:36:24.520 you look at a s mls revenue statements, what you'll 0:36:24.520 --> 0:36:26.760 find is they've got a growing share of their revenue 0:36:26.800 --> 0:36:29.600 coming from services which is servicing the machines that they operate. 0:36:29.600 --> 0:36:34.160 They've got staff in t SMCS, facilities in in Samsung, etcetera, 0:36:34.280 --> 0:36:35.960 making sure that not only the machines are operating, but 0:36:36.000 --> 0:36:39.720 they're operating exactly according to plan, they're not breaking down. 0:36:40.160 --> 0:36:42.640 The other thing that SML is is doing more of 0:36:42.800 --> 0:36:46.360 is managing the software for their machines. And the way 0:36:46.400 --> 0:36:49.280 that lithography works at at the scale that we're talking 0:36:49.320 --> 0:36:52.520 about is that if you want to print a certain pictures, 0:36:52.520 --> 0:36:55.360 so you want to print an AX, you don't reflect 0:36:55.480 --> 0:36:58.080 the shape of the ax on your silicon because the 0:36:58.160 --> 0:37:00.680 way that light reflects if you print and actually get 0:37:00.719 --> 0:37:03.600 something different, so you actually learn over time all of 0:37:03.640 --> 0:37:06.800 the unexpected errors and light refraction and errors in the 0:37:06.800 --> 0:37:10.560 way the chemicals react, and you print the error version 0:37:10.640 --> 0:37:11.960 and then it will give you an act. And so 0:37:12.000 --> 0:37:15.880 there's extradinally complex software that now tries to understand how 0:37:15.920 --> 0:37:18.399 all these different effects work. And you can actually look 0:37:18.400 --> 0:37:21.399 at the images that smls producing to produce a straight 0:37:21.440 --> 0:37:23.239 line and it looks nothing like a straight line. And 0:37:23.239 --> 0:37:25.759 so that that software as well as something that SML 0:37:25.840 --> 0:37:44.319 has been focusing on. So one of the things that 0:37:44.600 --> 0:37:46.680 we've talked about. You know, it's like the sort of 0:37:46.719 --> 0:37:50.239 the nanometer wars and the obvious. You know, people talking 0:37:50.239 --> 0:37:54.680 about Moore's law and whether it's t SMC or Intel 0:37:54.760 --> 0:37:57.040 or anyone else, so they're always or a MG. Maybe 0:37:57.200 --> 0:38:01.120 they're always bragging about like making smaller and smaller chips. 0:38:01.160 --> 0:38:02.920 And one of the things that we learned is that 0:38:03.360 --> 0:38:06.359 actually the chip companies all defined these measures a little 0:38:06.360 --> 0:38:09.000 bit differently, so to some extent it's made up. But 0:38:09.120 --> 0:38:13.160 how much are the chip manufacturers themselves as they advertised, 0:38:13.160 --> 0:38:15.080 like Okay, we're gonna be able to make a seven 0:38:15.160 --> 0:38:18.120 NIM chip or maybe a five NIM chip or whatever. 0:38:18.719 --> 0:38:24.120 How much are there timelines dependent on a s m 0:38:24.320 --> 0:38:27.560 l's I guess I would say a smls own learning curve. 0:38:28.000 --> 0:38:31.640 And what are as a sort of monopoly provider, I 0:38:31.640 --> 0:38:34.080 don't want to say monopoly, but is the sole provider 0:38:34.160 --> 0:38:37.480 of the most cutting edge technology. What are the forces 0:38:37.920 --> 0:38:43.440 that drive technological gains for a sm L itself. If 0:38:43.480 --> 0:38:46.640 you look at how a SML has begun to roll 0:38:46.680 --> 0:38:49.719 out at CV machines in into high volume manufacturing, which 0:38:49.760 --> 0:38:53.080 has mostly been at t SMC, the learning has happened 0:38:53.120 --> 0:38:56.200 collectively with a sm L and t SMC, so there's 0:38:56.239 --> 0:38:58.879 been lots of SMA personnel that spend tons of time 0:38:58.880 --> 0:39:02.479 in Taiwan and vice versa. So you really wouldn't have 0:39:02.800 --> 0:39:04.440 the rollout of e V over the past couple of 0:39:04.480 --> 0:39:07.160 years had you not have this collective effort between t 0:39:07.360 --> 0:39:10.879 s MC and a SML. And that's that puts other 0:39:10.880 --> 0:39:14.239 companies at a disadvantage because t SMC knows better than 0:39:14.280 --> 0:39:18.360 anyone how a s MLS machines actually work in practice, 0:39:18.840 --> 0:39:22.000 and and the highlight manufacturing is really crucial to understanding 0:39:22.080 --> 0:39:24.920 how how these machines work, because you don't really know 0:39:25.040 --> 0:39:27.600 until you've got them in production. And once you've got 0:39:27.600 --> 0:39:30.600 them in production, you've got thousands and thousands and thousands 0:39:30.600 --> 0:39:34.560 of iterations every single day where you can understand what 0:39:34.680 --> 0:39:38.280 the errors are, what the idiosyncrasies are of a given machine, 0:39:38.360 --> 0:39:41.520 and begin to correct for them. We talked about technological 0:39:41.560 --> 0:39:44.000 progress and that that's important, but in some ways the 0:39:44.600 --> 0:39:48.000 real challenge here is actually manufacturing progress. Understanding what the 0:39:48.360 --> 0:39:51.000 what what the ADO syncrasies are at the manufacturing stage 0:39:51.040 --> 0:39:52.600 and then learning to correct for them. And that's that's 0:39:52.600 --> 0:39:54.640 what t SMC has done extraordinarly well ad and it's 0:39:54.640 --> 0:39:58.719 been hand in hand with with a SML. So I 0:39:58.760 --> 0:40:02.040 have a sort of market oriented question. But you know, 0:40:02.080 --> 0:40:05.960 we think of semiconductors as this highly cyclical industry that 0:40:06.120 --> 0:40:09.440 usually moves in line with whatever is going on with 0:40:09.480 --> 0:40:12.680 GDP and economic growth, and that hasn't really been the 0:40:12.719 --> 0:40:16.040 case since the pandemic because we've had, you know, this 0:40:16.160 --> 0:40:20.319 big boom in demand for electronic goods and it's been 0:40:20.320 --> 0:40:22.960 a struggle to keep up. But I imagine for a 0:40:22.960 --> 0:40:25.640 company like a s mL, it has also traditionally been 0:40:25.640 --> 0:40:28.680 considered cyclical and its fortunes are sort of tied to 0:40:28.760 --> 0:40:32.960 what's going on with the actual semiconductor manufacturers. But just 0:40:33.000 --> 0:40:37.440 looking at SML's most recent results, they're forecasting basically like 0:40:37.800 --> 0:40:41.680 a boom in revenue for the next decade, something they 0:40:41.719 --> 0:40:44.960 expect to last for ten years. Is there anything that 0:40:45.000 --> 0:40:48.560 could sort of knock that revenue cycle at this point 0:40:48.600 --> 0:40:51.120 in time or is this business like, is there such 0:40:51.120 --> 0:40:54.360 a steep moat around the eu V technology that is 0:40:54.360 --> 0:40:57.400 just going to be impossible for anything to um to 0:40:57.520 --> 0:41:02.120 hit it. There's there's definitely at mote around SMILS technology, 0:41:02.160 --> 0:41:04.360 and they won't be overtaken into UV in a decade. 0:41:04.520 --> 0:41:08.800 I should say I'm mixing my metaphors and steep and 0:41:09.040 --> 0:41:12.080 impossible to overcome mote. There's no real competition that a 0:41:12.200 --> 0:41:14.600 small faces when it comes to UV. That the question 0:41:14.680 --> 0:41:17.800 is what is going to be demand for UV machines, 0:41:17.920 --> 0:41:20.440 and that's a question ultimately of final demand for the 0:41:20.440 --> 0:41:25.040 most advanced chips. A SMILS projections are are based on 0:41:25.719 --> 0:41:28.799 the assumption that we've reached a point where there's a 0:41:28.880 --> 0:41:32.040 secular increase in demand for chips, as we have more 0:41:32.360 --> 0:41:35.200 demand for data center capacity, um as we have the 0:41:35.200 --> 0:41:38.080 five G rollout and new devices that are taking advantage 0:41:38.120 --> 0:41:40.000 of five G networks. Their bat is that we're going 0:41:40.080 --> 0:41:42.000 to have more chips per g d P and therefore 0:41:42.000 --> 0:41:44.640 more demand for a SML. That's a bat that you know, 0:41:44.719 --> 0:41:46.200 it's not clear what that's going to play out. What 0:41:46.320 --> 0:41:49.239 is clear is that anyone who's producing advanced ships will 0:41:49.239 --> 0:41:51.360 have no choice but to turn to sm L and 0:41:51.400 --> 0:41:53.920 so in some ways they're perfectly exposed to the fluctuations 0:41:53.920 --> 0:41:56.640 in the semiconductor industry. The more chips that are produced, 0:41:56.640 --> 0:41:59.160 the more machines you need. But the opposite is also true. 0:41:59.800 --> 0:42:03.839 You kind of hinted at this early on. But the 0:42:03.880 --> 0:42:08.040 idea that at uh seven nnimms in below at cutting edge, 0:42:08.160 --> 0:42:13.960 maybe EUV in theory isn't the only technology. So it's like, okay, 0:42:14.000 --> 0:42:17.720 if if if EUV is the only technology that can 0:42:17.760 --> 0:42:20.600 perform the test, then there's no one who can attack 0:42:20.640 --> 0:42:24.200 as well. Are there other theoretical approaches for accomplishing the 0:42:24.239 --> 0:42:26.120 same thing besides the UV, I feel like you hinted 0:42:26.160 --> 0:42:29.719 at in the beginning, there are It's really not a 0:42:29.800 --> 0:42:33.200 question of science but of of manufacturing efficiency. So if 0:42:33.239 --> 0:42:36.560 you if you take the previous generation of lithography machines, 0:42:36.560 --> 0:42:39.760 which rather than thirteen point five nanometer light, we're working 0:42:39.800 --> 0:42:43.200 with one. It became it was possible over time to 0:42:43.440 --> 0:42:47.839 produce ever smaller feature sizes on silicon wafers by using 0:42:47.880 --> 0:42:50.080 a number of tricks. So, for example, you can shoot 0:42:50.120 --> 0:42:52.520 the light through water, and if you think back to 0:42:52.560 --> 0:42:56.400 high school physics, when light refracts differently through water, that 0:42:56.520 --> 0:42:59.799 same principle lets you shoot lithography machines through water and 0:43:00.000 --> 0:43:03.800 and carve more specific shapes. You can also use multiple 0:43:03.960 --> 0:43:09.120 steps of lithography to carve specific shapes that are more detailed. 0:43:09.239 --> 0:43:12.239 The challenges just can you do this efficiently? Um So, 0:43:12.320 --> 0:43:15.560 every step of lithography you need adds to the time 0:43:15.560 --> 0:43:17.640 it takes to produce a way for ads to your costs. 0:43:18.120 --> 0:43:20.840 And so there's no doubt that if you wanted to 0:43:20.880 --> 0:43:24.000 produce an equivalent of one of Apple's new iPhone ships 0:43:24.080 --> 0:43:26.440 using older generation lithography, you could do it in a 0:43:26.560 --> 0:43:28.560 lab and do one of them. The question is can 0:43:28.600 --> 0:43:30.520 you do a million of them as skin? And that 0:43:30.560 --> 0:43:34.760 seems pretty implausible right now. There's no really credible pathway 0:43:34.800 --> 0:43:37.680 of of how you could do that efficiently today, and 0:43:37.760 --> 0:43:40.919 especially when you project forward five or ten years, we're 0:43:40.960 --> 0:43:45.080 expecting to be making ever smaller transistors with more complex 0:43:45.080 --> 0:43:48.400 shapes on them, and it seems really implausible that you'll 0:43:48.440 --> 0:43:51.239 be able to do that using anything besides UV. Chris, 0:43:51.400 --> 0:43:54.040 is there anything any other sort of least key things 0:43:54.040 --> 0:43:55.600 do you think we've missed? And I mean, I'm sure 0:43:55.640 --> 0:43:58.480 there's a million things, but other key ideas that I 0:43:58.480 --> 0:44:02.160 think we need to get across. I think if you're 0:44:02.160 --> 0:44:05.680 interested in US China dynamics, obviously it's a big. One 0:44:05.719 --> 0:44:09.680 of the key reasons why t SMC cut off huaweih 0:44:09.719 --> 0:44:14.600 In was because the US could restrict t SMCS access 0:44:14.640 --> 0:44:19.160 to machinery, of which lithography machinery was a It was 0:44:19.200 --> 0:44:22.160 a key example. So when when the Chinese ship industry 0:44:22.200 --> 0:44:23.920 looks out and says we're we're gonna get the tools 0:44:23.960 --> 0:44:26.800 that we need, the impact of the US support controls 0:44:26.880 --> 0:44:31.080 on companies like a SML, even foreign companies that nevertheless 0:44:31.239 --> 0:44:34.640 use US technology in their systems, is a pretty fundamental 0:44:34.719 --> 0:44:37.520 roadblock that China faces, and that the big concern that 0:44:37.880 --> 0:44:40.080 SML has right now is that the US is going 0:44:40.120 --> 0:44:42.759 to expand its restrictions on what you can send to 0:44:42.880 --> 0:44:45.719 China in terms of lithography machinery, and China has been 0:44:45.719 --> 0:44:47.680 a big growth market the past couple of years for 0:44:47.719 --> 0:44:50.560 older generational lithography machines, and so SML does face a 0:44:50.640 --> 0:44:55.440 risk that the US it expands these restrictions. What inspired 0:44:55.480 --> 0:44:57.560 you to write a book all about a s m 0:44:57.640 --> 0:45:00.239 L because it is like it's not something that ms 0:45:00.320 --> 0:45:04.240 up necessarily in daily conversation. UM, so I'm just wondering 0:45:04.280 --> 0:45:06.160 how it sort of came on your radar and what 0:45:06.360 --> 0:45:09.359 is it that piqued your interest? We like the two 0:45:09.360 --> 0:45:11.719 of you. I spent the past couple of years realizing 0:45:11.760 --> 0:45:15.760 that semiconductors were vastly more important than the average person, 0:45:16.160 --> 0:45:20.680 including myself realized, and also far cooler. The technology needed 0:45:20.719 --> 0:45:23.319 to make them as extraordinary. The fact that we're able 0:45:23.400 --> 0:45:27.480 to manipulate individual atoms in some cases is extraordinary, and 0:45:27.480 --> 0:45:30.320 to do it at the scale of trillions and trillions 0:45:30.320 --> 0:45:33.360 and transistors, I thought was was was really just wild 0:45:33.400 --> 0:45:36.080 in terms of what what what was possible? Uh? And 0:45:36.120 --> 0:45:39.160 it seemed to me that I took my iPhone for granted, 0:45:39.200 --> 0:45:41.760 I took my computer for granted. I took the cloud, 0:45:41.800 --> 0:45:43.919 which is just a bunch of silicon in big data 0:45:43.920 --> 0:45:46.480 centers in IO. I took all that for granted without 0:45:46.480 --> 0:45:50.880 thinking through how complex it was too actually make these 0:45:50.920 --> 0:45:53.759 tools work. And I think for for a long time 0:45:53.760 --> 0:45:55.800 we thought of the Internet is something out there. We 0:45:55.880 --> 0:45:58.640 thought of data processing is something that happens somewhere else. 0:45:58.640 --> 0:46:01.520 But it's all actually very physical, all things being carved 0:46:02.000 --> 0:46:06.000 onto silicon by shooting light at them and depositing uh, 0:46:06.440 --> 0:46:09.560 layers of atoms and using different chemicals. And the reality 0:46:09.600 --> 0:46:12.440 that our entire digital world is in fact existing on 0:46:12.880 --> 0:46:15.000 millions and millions of silicon wafers is something I don't 0:46:15.000 --> 0:46:17.560 think we think enough about and we're just having to 0:46:17.600 --> 0:46:19.880 come directon with that with the some innunctory shortage right 0:46:19.920 --> 0:46:22.600 now that you can't just imagine an increase in computing 0:46:22.600 --> 0:46:24.880 power and increase in memory, You've actually got to carve 0:46:24.920 --> 0:46:28.440 it onto silicon in billions and billions of tiny transistors. 0:46:29.400 --> 0:46:32.560 You know, it's interesting, I mean, your assistant professor at 0:46:32.560 --> 0:46:37.680 the Fletchers School, which I associate with diplomacy and government, 0:46:37.800 --> 0:46:41.560 and that seems like another sort of like fascinating dynamic here, 0:46:41.600 --> 0:46:44.799 which is like maybe it's it feels again, or maybe 0:46:44.800 --> 0:46:47.719 it goes in cycles, but there's appreciation. And you sort 0:46:47.719 --> 0:46:49.799 of said it for your last point about yours trying 0:46:49.840 --> 0:46:53.400 to like that this particular industry is sort of inseparable 0:46:53.600 --> 0:46:57.239 from thinking about how governments relate to each other. That's right. 0:46:57.280 --> 0:47:01.080 It's it's it's crucial for for military systems, for example, 0:47:01.160 --> 0:47:05.560 it's crucial for controlling computing power in the future. And 0:47:05.640 --> 0:47:09.400 it's been a prominent tool of of of geopolitics for 0:47:09.440 --> 0:47:12.000 the past three quarters of a century. And I think 0:47:12.320 --> 0:47:14.560 we're seeing that more to the four today. But in fact, 0:47:14.560 --> 0:47:16.600 when you look at the history of lithography and of 0:47:16.640 --> 0:47:19.239 semiconductor is more generally you find that it's constantly been 0:47:19.880 --> 0:47:22.399 something that governments have thought about in in political terms 0:47:22.400 --> 0:47:25.000 as well as in economic terms, and constantly been an 0:47:25.000 --> 0:47:27.360 area of dispute between different governments as they tried to 0:47:27.760 --> 0:47:33.840 vy for a bigger chunk of the semiconductor ecosystem. Chris Miller, 0:47:34.040 --> 0:47:37.680 thank you so much for coming on. That was the 0:47:37.680 --> 0:47:40.040 the a s m L episode We needed to do it, 0:47:40.080 --> 0:47:41.520 and you are the perfect guest for it, and I 0:47:41.640 --> 0:47:44.200 just learned a lot, So thank you so much for 0:47:44.280 --> 0:47:48.560 coming out on Oblock Bain's invitation. Thanks Chris, that was 0:47:48.600 --> 0:48:01.759 so interesting. Obviously, I love that conversation, and you know, 0:48:01.840 --> 0:48:05.239 I sortaty interrupted like seven minutes in because like this 0:48:05.320 --> 0:48:09.120 idea of like thinking of like a component as in 0:48:09.200 --> 0:48:13.560 itself a supply chain story. Like the idea that really 0:48:13.640 --> 0:48:18.400 the breakthrough is how do you coordinate four thousand different 0:48:18.440 --> 0:48:23.440 suppliers of them of highly specific raw materials and machines 0:48:23.840 --> 0:48:26.880 into one thing that forms a cohesive whole. The idea 0:48:26.920 --> 0:48:31.120 that that is what the thing is is pretty fascinating 0:48:31.160 --> 0:48:35.040 to me. Here's the important question, which is, what, like 0:48:35.200 --> 0:48:37.480 what idea did you get out of that conversation For 0:48:37.480 --> 0:48:40.920 the next Semiconductor episode, so I'm sure there is one. 0:48:41.520 --> 0:48:44.480 What is the next one? Now we we we got 0:48:44.520 --> 0:48:48.080 to the end. Now it would be like really interesting actually, okay, 0:48:48.120 --> 0:48:51.719 so like it all seriousness, like I would like to 0:48:52.120 --> 0:48:56.680 learn more about that process, like the actual like the 0:48:56.760 --> 0:48:58.920 coordination the heart. It's almost like you think of like 0:48:58.960 --> 0:49:01.319 a conductor of an orchestra. Is sort of like the 0:49:01.320 --> 0:49:04.000 mental model I use for a company that has to 0:49:04.080 --> 0:49:07.439 like have four thousand parts all coming together to form 0:49:07.560 --> 0:49:11.120 thirty one units or units or whatever. It is like 0:49:11.400 --> 0:49:13.840 thinking about like how do you do that from like 0:49:13.880 --> 0:49:17.560 a management perspective even beyond this sort of tech perspective 0:49:17.719 --> 0:49:21.080 is like a super fascinating thing to explore, especially at 0:49:21.080 --> 0:49:24.520 its especially right now. I mean this is almost verging 0:49:24.600 --> 0:49:27.279 on state secrets. But we've got to get you know, 0:49:27.320 --> 0:49:28.680 we have to try to get the a s m 0:49:28.800 --> 0:49:31.360 L supply chain manager on all thoughts, so you know, 0:49:31.480 --> 0:49:34.120 a s m L hit us up. We're interested in 0:49:34.160 --> 0:49:36.799 how you're doing it. Um, and we have to keep 0:49:36.840 --> 0:49:42.280 the semiconductor series going. Yeah, No, that was fascinating, and um, 0:49:42.440 --> 0:49:46.360 Chris was the Chris was the perfect guest for that one. Yeah, definitely, 0:49:46.440 --> 0:49:49.960 shall we leave it there, Let's leave it there alright. 0:49:50.000 --> 0:49:52.760 This has been another episode of the All Thoughts podcast. 0:49:52.840 --> 0:49:55.640 I'm Tracy Alloway. You can follow me on Twitter at 0:49:55.640 --> 0:49:58.920 Tracy Alloway. And I'm Joe Wisenthal. You can follow me 0:49:59.080 --> 0:50:02.480 on Twitter at the Stalwart. Follow our guest Chris Miller. 0:50:02.520 --> 0:50:04.200 He has a book coming out next year on the 0:50:04.280 --> 0:50:07.360 chip industry. Assistant professor at the Fletcher School, he is 0:50:07.480 --> 0:50:12.279 at cr Miller One. Follow our producer Laura Carlson. She's 0:50:12.360 --> 0:50:15.600 at Laura M. Carlson. Follow the Bloomberg head of podcast, 0:50:15.640 --> 0:50:19.080 Francesca Levi at Francesca Today, and check out all of 0:50:19.080 --> 0:50:23.279 our podcasts at Bloomberg under the handle at podcasts. Thanks 0:50:23.320 --> 0:50:23.840 for listening.