1
00:00:11,039 --> 00:00:15,120
Speaker 1: Hello, and welcome to another episode of the Odd Lots podcast.

2
00:00:15,200 --> 00:00:19,520
Speaker 1: I'm Joe Wisn'tal And I'm Tracy all Away. Tracy, remember

3
00:00:19,520 --> 00:00:21,640
Speaker 1: when I said a few weeks ago that I thought

4
00:00:21,680 --> 00:00:26,599
Speaker 1: we should do a semiconductor Odd Lots spinoff series. No whole,

5
00:00:26,680 --> 00:00:28,560
Speaker 1: not a whole, not a series, like a whole new show.

6
00:00:29,120 --> 00:00:31,280
Speaker 1: The idea, a whole new show on the subject. No,

7
00:00:31,480 --> 00:00:33,040
Speaker 1: that's not what you said. You said you wanted to

8
00:00:33,080 --> 00:00:37,640
Speaker 1: transform All Thoughts into a semiconductor podcast. So this is

9
00:00:37,680 --> 00:00:42,280
Speaker 1: the latest step in your plan to do that, I assume. Yeah,

10
00:00:42,440 --> 00:00:45,000
Speaker 1: so we're not going to do that, And realistically, I

11
00:00:45,040 --> 00:00:47,760
Speaker 1: don't think we're going to have a new spinoff Odd

12
00:00:47,800 --> 00:00:52,640
Speaker 1: Lots branded uh semi conductor show, but I do think

13
00:00:53,000 --> 00:00:54,880
Speaker 1: but we are going to do some more episodes on

14
00:00:54,920 --> 00:00:58,480
Speaker 1: the topic because it really does just intersect with a

15
00:00:58,560 --> 00:01:02,280
Speaker 1: lot of different themes. There's sort of the pure tech

16
00:01:02,400 --> 00:01:05,920
Speaker 1: question of their role in the economy, why different companies

17
00:01:05,959 --> 00:01:09,360
Speaker 1: are rising and falling, but there's it also intersects with

18
00:01:09,720 --> 00:01:14,399
Speaker 1: um things like national security and sort of economic strength

19
00:01:14,560 --> 00:01:17,560
Speaker 1: and some of the questions we've seen during the COVID

20
00:01:17,600 --> 00:01:23,120
Speaker 1: crisis about say domestic manufacturing and reliance on manufacturing in

21
00:01:23,160 --> 00:01:27,680
Speaker 1: other countries, So it's a it's a rich load to mine. Yeah, well,

22
00:01:27,680 --> 00:01:29,800
Speaker 1: I mean you could go even further. So it kind

23
00:01:29,840 --> 00:01:32,760
Speaker 1: of gets to, I guess, the heart of a large

24
00:01:32,760 --> 00:01:36,160
Speaker 1: body of economic thought, right, how do you balance domestic

25
00:01:36,280 --> 00:01:40,000
Speaker 1: interests with market forces? So when the entire market is

26
00:01:40,000 --> 00:01:43,920
Speaker 1: telling you to outsource manufacturing to a cheaper place and

27
00:01:43,959 --> 00:01:46,960
Speaker 1: they can do it more efficiently, is that the way

28
00:01:46,959 --> 00:01:50,320
Speaker 1: to go for something as important as a sensitive technology

29
00:01:50,880 --> 00:01:55,200
Speaker 1: like a semiconductor. And I think we're seeing those tensions

30
00:01:55,280 --> 00:01:57,280
Speaker 1: over and over. I think we see them reflected in

31
00:01:57,480 --> 00:02:00,320
Speaker 1: a lot of US politics and the way people feel

32
00:02:00,360 --> 00:02:02,560
Speaker 1: at the moment. But of course we also see that

33
00:02:02,640 --> 00:02:07,800
Speaker 1: reflected in the U. S. China trade war, and technology

34
00:02:07,880 --> 00:02:11,880
Speaker 1: is really ground zero of that. And within technology, semiconductors

35
00:02:11,919 --> 00:02:16,280
Speaker 1: are the epicenter of the ground zero. I'm mixing my metaphors,

36
00:02:16,320 --> 00:02:18,519
Speaker 1: but I think you get it. The epicenter of the

37
00:02:18,560 --> 00:02:20,840
Speaker 1: ground zero. No, I get it, And I think you're

38
00:02:20,840 --> 00:02:23,640
Speaker 1: absolutely right. Like it's kind of a it's a business topic,

39
00:02:24,160 --> 00:02:27,320
Speaker 1: but it intersects with a lot of our favorite macro

40
00:02:27,720 --> 00:02:30,160
Speaker 1: themes that we discuss all the time. And I think

41
00:02:30,160 --> 00:02:33,680
Speaker 1: another sort of macro theme is you know, we talked

42
00:02:33,680 --> 00:02:37,840
Speaker 1: about it a lot in but the sort of rethinking

43
00:02:37,880 --> 00:02:40,520
Speaker 1: the role of the public sector and rethinking the role

44
00:02:40,520 --> 00:02:44,720
Speaker 1: of government spending, rethinking the role of government investment, and

45
00:02:44,760 --> 00:02:46,519
Speaker 1: I think that's also I mean, that's a that's a

46
00:02:46,560 --> 00:02:49,480
Speaker 1: big topic for us, it's a big topic nationally. I mean,

47
00:02:49,520 --> 00:02:51,640
Speaker 1: I think you know, if you look at one of

48
00:02:51,680 --> 00:02:55,320
Speaker 1: the successes that we've had in this crisis, most people

49
00:02:55,320 --> 00:02:58,519
Speaker 1: would say the speed with which we were able to

50
00:02:59,120 --> 00:03:02,560
Speaker 1: develop a X seeing for covid um was in part

51
00:03:02,680 --> 00:03:06,320
Speaker 1: due to the government investment and operation warp speed, and

52
00:03:06,360 --> 00:03:10,079
Speaker 1: of course the longstanding investment in pharma. So there's all

53
00:03:10,160 --> 00:03:13,840
Speaker 1: kinds of reasons to sort of use this moment as

54
00:03:14,000 --> 00:03:18,160
Speaker 1: a chance to reconceptualize what is the role of the

55
00:03:18,160 --> 00:03:22,079
Speaker 1: public sector. How can the public sector spur more innovation,

56
00:03:22,720 --> 00:03:26,280
Speaker 1: manufacturing production. So it's a it's a great topic because

57
00:03:26,320 --> 00:03:29,640
Speaker 1: it hits a lot of odd themes. I would say, yeah,

58
00:03:29,639 --> 00:03:32,600
Speaker 1: that's right, And I think we have seen a proposal

59
00:03:32,680 --> 00:03:38,720
Speaker 1: in Washington about how to boost domestic semiconductor capabilities, So yes,

60
00:03:38,840 --> 00:03:41,560
Speaker 1: it hits all the sweet spots. It's something that's happening

61
00:03:41,760 --> 00:03:43,880
Speaker 1: right now as well. And I think it's going to

62
00:03:43,920 --> 00:03:47,160
Speaker 1: be an interesting conversation. Yeah, I'm really excited about this.

63
00:03:47,400 --> 00:03:51,000
Speaker 1: Um our guest today is um Professor of Management Practice

64
00:03:51,080 --> 00:03:54,720
Speaker 1: at Harvard Business School, Willie She. He writes a lot

65
00:03:54,760 --> 00:03:59,840
Speaker 1: about the topic about semiconductors, about manufacturing, US manufacturing, and

66
00:04:00,000 --> 00:04:02,960
Speaker 1: particular what it takes to make it work. So we're

67
00:04:02,960 --> 00:04:05,400
Speaker 1: going to dive into all of these questions with him.

68
00:04:05,920 --> 00:04:09,400
Speaker 1: Professor She, thank you so much for joining us. Well,

69
00:04:09,440 --> 00:04:12,040
Speaker 1: thank you for having me, Joe. So I'm trying to

70
00:04:12,080 --> 00:04:14,640
Speaker 1: think like where the best place to start is. But

71
00:04:15,080 --> 00:04:18,440
Speaker 1: you know, obviously, uh, we're talking about Intel, and for

72
00:04:18,480 --> 00:04:21,840
Speaker 1: a long time, Uh, Intel is not just the leader

73
00:04:22,560 --> 00:04:26,520
Speaker 1: in semiconductors, but the sort of like the vanguard of

74
00:04:26,839 --> 00:04:32,080
Speaker 1: US tech and tech manufacturing capability. And of course the

75
00:04:32,160 --> 00:04:34,920
Speaker 1: story over the last couple of years has been the

76
00:04:35,000 --> 00:04:39,960
Speaker 1: problems they've run into in manufacturing. Also business model questions.

77
00:04:40,000 --> 00:04:44,520
Speaker 1: But to start, why is it important for a country

78
00:04:44,600 --> 00:04:47,279
Speaker 1: like the US or a country like China for that matter,

79
00:04:47,320 --> 00:04:49,960
Speaker 1: But why is it an important for country like the

80
00:04:50,080 --> 00:04:55,400
Speaker 1: US to have this capability at scale to produce high

81
00:04:55,400 --> 00:04:58,599
Speaker 1: tech goods such as semi conductors. Well, I think the

82
00:04:58,680 --> 00:05:03,200
Speaker 1: reason semiconductors have attracted so much focus in this regard

83
00:05:03,800 --> 00:05:07,440
Speaker 1: is because they are a building block for so much

84
00:05:07,560 --> 00:05:11,159
Speaker 1: modern technology. In fact, if you look at and it's

85
00:05:11,160 --> 00:05:14,280
Speaker 1: been really highlighted by the pandemic kind of the role

86
00:05:14,520 --> 00:05:20,279
Speaker 1: of electronics and technology communications technology, whether it is conducting

87
00:05:20,279 --> 00:05:25,120
Speaker 1: classes and meetings and business on Zoom or it's ordering online,

88
00:05:25,279 --> 00:05:28,920
Speaker 1: underpinning all this sort of the the core building block

89
00:05:29,240 --> 00:05:34,000
Speaker 1: are semiconductors. When you start to look at the numbers

90
00:05:34,880 --> 00:05:38,880
Speaker 1: on what percentage of the manufacturing happens in the US

91
00:05:38,960 --> 00:05:42,440
Speaker 1: and how that has declined over the years, it's only

92
00:05:42,440 --> 00:05:47,320
Speaker 1: about twelve for the microelectronics that we're talking about, for

93
00:05:47,440 --> 00:05:51,360
Speaker 1: things like Intel microprocessors or the chips that go into smartphones.

94
00:05:52,080 --> 00:05:54,000
Speaker 1: And you look at how much of the dependence is

95
00:05:54,040 --> 00:05:59,039
Speaker 1: in Asia, and combine that with the rhetoric around you know,

96
00:05:59,160 --> 00:06:04,040
Speaker 1: dependency on China or areas that China claims like Taiwan.

97
00:06:04,440 --> 00:06:08,480
Speaker 1: I think that has really brought attention to a concern

98
00:06:08,600 --> 00:06:11,560
Speaker 1: that I would say has been around in some circles

99
00:06:11,600 --> 00:06:16,360
Speaker 1: for the past twenty years. Mhm. So, I think there

100
00:06:16,440 --> 00:06:21,720
Speaker 1: is a recognition that semiconductors are strategically important for the

101
00:06:21,760 --> 00:06:25,359
Speaker 1: reasons that you just laid out. But we've we've discussed

102
00:06:25,360 --> 00:06:29,440
Speaker 1: this in some previous episodes. It's an incredibly competitive industry

103
00:06:29,839 --> 00:06:34,560
Speaker 1: and it has huge startup costs. How difficult is it

104
00:06:34,600 --> 00:06:42,560
Speaker 1: to actually um set up and create a semiconductor manufacturing capability?

105
00:06:42,880 --> 00:06:45,880
Speaker 1: And what are the business decisions that kind of go

106
00:06:46,000 --> 00:06:50,720
Speaker 1: into that. Well, there are a couple of things that

107
00:06:51,320 --> 00:06:54,440
Speaker 1: I think have evolved as you get to the more

108
00:06:54,480 --> 00:06:58,640
Speaker 1: advanced technologies. The cost of setting up kind of a

109
00:06:58,720 --> 00:07:04,080
Speaker 1: minimum efficient ale manufacturing facility. It used to be, you know,

110
00:07:05,160 --> 00:07:07,279
Speaker 1: back thirty years ago, you could do this for a

111
00:07:07,279 --> 00:07:10,480
Speaker 1: few hundred million dollars, okay, or at the beginning it

112
00:07:10,520 --> 00:07:14,280
Speaker 1: was a few million dollars. Now to set up the

113
00:07:14,360 --> 00:07:18,240
Speaker 1: most advanced, leading edge fab will probably cost about twenty

114
00:07:18,280 --> 00:07:23,440
Speaker 1: billion dollars, right, And that's a substantial capital cost That

115
00:07:24,320 --> 00:07:29,360
Speaker 1: really strikes fear in the hearts of many uh CFOs

116
00:07:30,040 --> 00:07:33,240
Speaker 1: as they look at the level investment. And and when

117
00:07:33,280 --> 00:07:36,080
Speaker 1: you make that level investment, you have to make it

118
00:07:36,120 --> 00:07:39,320
Speaker 1: successful in order to earn a return. That means you

119
00:07:39,520 --> 00:07:44,040
Speaker 1: have to be able to learn all those processed technologies.

120
00:07:44,560 --> 00:07:49,400
Speaker 1: You have to be able to manufacturer to high yield, okay,

121
00:07:49,440 --> 00:07:53,840
Speaker 1: because if you don't have high yield, then you're throwing

122
00:07:54,120 --> 00:07:56,240
Speaker 1: out half of your output and then your cost just

123
00:07:56,920 --> 00:07:58,880
Speaker 1: you know, you don't have anything to absorb your costs.

124
00:07:58,960 --> 00:08:03,240
Speaker 1: So it's a technological challenge in this the financial challenge

125
00:08:03,240 --> 00:08:08,600
Speaker 1: as well. So, uh, you know we're talking about Intel.

126
00:08:08,640 --> 00:08:11,880
Speaker 1: You're the co author of a book Producing Prosperity, why

127
00:08:11,920 --> 00:08:17,119
Speaker 1: American needs a manufacturing renaissance. So when when thinking about

128
00:08:17,200 --> 00:08:21,000
Speaker 1: this question, it seems like there's kind of two parts

129
00:08:21,000 --> 00:08:24,680
Speaker 1: to it. One is, Okay, something is going on at Intel,

130
00:08:25,280 --> 00:08:29,360
Speaker 1: they're having tech problems, they're having business model problems as well,

131
00:08:29,760 --> 00:08:35,520
Speaker 1: But there's also a sort of broader US manufacturing malaise

132
00:08:35,760 --> 00:08:38,120
Speaker 1: or you know, a reason that there's less and less

133
00:08:38,160 --> 00:08:42,000
Speaker 1: of it happening here. How do you sort of disentangle

134
00:08:42,120 --> 00:08:43,960
Speaker 1: the two things? How much of it is sort of

135
00:08:44,000 --> 00:08:47,360
Speaker 1: macro factors about something that's going on in the US

136
00:08:47,440 --> 00:08:51,920
Speaker 1: that's causing this decline in high tech manufacturing capability and

137
00:08:51,960 --> 00:08:56,920
Speaker 1: know how versus specific micro things going on at a

138
00:08:57,000 --> 00:09:00,600
Speaker 1: company that happens to be stumbling. Well, so let me

139
00:09:00,840 --> 00:09:04,959
Speaker 1: start with the macro thing, because I think in essence,

140
00:09:05,000 --> 00:09:10,199
Speaker 1: the financial risk and the technological challenge a are linked.

141
00:09:10,480 --> 00:09:15,760
Speaker 1: Because what we saw, we've seen this type of crisis before.

142
00:09:16,120 --> 00:09:18,280
Speaker 1: It was actually in the nineteen eighties, and at the

143
00:09:18,360 --> 00:09:22,960
Speaker 1: time it wasn't China. It was a Japanese who were

144
00:09:23,080 --> 00:09:28,720
Speaker 1: manufacturing memory chips so called d rans dynamic random access memories.

145
00:09:29,280 --> 00:09:32,680
Speaker 1: And in the nineteen eighties, the Japanese had invested in

146
00:09:32,760 --> 00:09:37,840
Speaker 1: manufacturing and they were actually much better at it than

147
00:09:38,000 --> 00:09:42,800
Speaker 1: American companies, even like Intel. Right, Intel was founded as

148
00:09:42,800 --> 00:09:47,120
Speaker 1: a memory company. Many people forget that, right, but Intel

149
00:09:47,760 --> 00:09:51,080
Speaker 1: Texas Instruments was a big player in memories, moss tech

150
00:09:51,800 --> 00:09:56,920
Speaker 1: companies like that, and the Japanese out executed American manufacturers

151
00:09:56,960 --> 00:09:59,520
Speaker 1: because they were delivering higher yields. Right, So if you

152
00:09:59,559 --> 00:10:02,040
Speaker 1: think out, I'm going to sink all this money into

153
00:10:02,080 --> 00:10:05,280
Speaker 1: a factory. If I have higher yields, that means my

154
00:10:05,400 --> 00:10:11,320
Speaker 1: cost is lower. And uh, Intel really was pushed the

155
00:10:11,360 --> 00:10:14,840
Speaker 1: wall on memories and they decided to get out of

156
00:10:14,920 --> 00:10:19,320
Speaker 1: memory manufacturing and focused on microprocessor manufacturing. At the time,

157
00:10:19,559 --> 00:10:23,240
Speaker 1: you know, IBM had picked UH their microprocessor for the

158
00:10:23,240 --> 00:10:27,840
Speaker 1: IBM PC. There were early comments in Intel where they

159
00:10:27,840 --> 00:10:30,760
Speaker 1: weren't sure if you know, this PC thing was going

160
00:10:30,800 --> 00:10:34,360
Speaker 1: to be big, Okay, But in some sense they were

161
00:10:34,400 --> 00:10:39,760
Speaker 1: smart and they were lucky in moving two microprocessors, okay,

162
00:10:39,800 --> 00:10:43,120
Speaker 1: and then they were able to escape this problem of

163
00:10:43,200 --> 00:10:47,000
Speaker 1: being the best manufacturer. Okay. Now, at the same time,

164
00:10:47,520 --> 00:10:51,679
Speaker 1: what we saw in American industry was this separation of

165
00:10:51,720 --> 00:10:56,439
Speaker 1: design from manufacturing, which was really enabled by a lot

166
00:10:56,480 --> 00:10:59,800
Speaker 1: of these computer aided design tools, right, which you know,

167
00:11:00,120 --> 00:11:04,120
Speaker 1: separated out the manufacturing process. And then we saw the

168
00:11:04,240 --> 00:11:07,840
Speaker 1: rise of some Asian companies, in particular in Taiwan, but

169
00:11:07,880 --> 00:11:10,880
Speaker 1: also Korea and in Japan at the time, who said, well,

170
00:11:10,920 --> 00:11:14,160
Speaker 1: we're very happy to do the manufacturing because we're not

171
00:11:14,240 --> 00:11:16,640
Speaker 1: the experts of the design, right. So we saw that

172
00:11:16,760 --> 00:11:21,640
Speaker 1: bifurcation okay. And now as that has developed, what we've

173
00:11:21,679 --> 00:11:26,640
Speaker 1: seen is as you go to the more advanced manufacturing processes,

174
00:11:27,240 --> 00:11:32,040
Speaker 1: there were very few companies who could kind of keep

175
00:11:32,080 --> 00:11:35,560
Speaker 1: in that game. Intel has historically been one of them, right,

176
00:11:35,559 --> 00:11:39,080
Speaker 1: because after they left the memory business, they invested a

177
00:11:39,120 --> 00:11:44,959
Speaker 1: lot in manufacturing to give their microprocessor business a lead, right,

178
00:11:45,000 --> 00:11:50,120
Speaker 1: And historically Intel lead with manufacturing in that regard, okay,

179
00:11:50,200 --> 00:11:53,800
Speaker 1: and they were able to beat competitors like a m

180
00:11:53,880 --> 00:12:00,200
Speaker 1: D because their manufacturing processes were a generation or two ahead, okay,

181
00:12:00,200 --> 00:12:02,720
Speaker 1: So it didn't matter if you had a design that

182
00:12:03,040 --> 00:12:06,319
Speaker 1: might have not been quite as efficient. Okay. Now a

183
00:12:06,559 --> 00:12:10,200
Speaker 1: m D also had their own manufacturing, but then they

184
00:12:10,280 --> 00:12:14,320
Speaker 1: spun it off a decade ago into Global Foundaries. And

185
00:12:14,440 --> 00:12:18,880
Speaker 1: once a m D separated their manufacturing from their design

186
00:12:19,720 --> 00:12:22,640
Speaker 1: and uh, you know, they left Global Foundaries and they

187
00:12:22,640 --> 00:12:27,720
Speaker 1: shifted over to Taiwan Semiconductor. Okay. Once am D now

188
00:12:27,880 --> 00:12:33,280
Speaker 1: had access to a more advanced process than Intel did,

189
00:12:33,320 --> 00:12:35,439
Speaker 1: they have been able to take a lot of market

190
00:12:35,440 --> 00:12:39,640
Speaker 1: share from Intel. Right. So you know this this problem

191
00:12:39,679 --> 00:12:42,400
Speaker 1: has been a long time in developing. But the roots

192
00:12:42,440 --> 00:12:46,679
Speaker 1: of it are the manufacturing part is capital intensive and

193
00:12:46,760 --> 00:12:49,520
Speaker 1: it's hard, it takes a lot of discipline, Okay, and

194
00:12:49,559 --> 00:12:52,439
Speaker 1: you can make more money on the design side, right.

195
00:12:52,480 --> 00:12:55,320
Speaker 1: So hey, if somebody else wants to do the manufacturing,

196
00:12:55,520 --> 00:12:57,480
Speaker 1: let them do it. I'll be more profitable. And a

197
00:12:57,480 --> 00:12:59,800
Speaker 1: lot of people have said, well, the U S. Semi

198
00:12:59,800 --> 00:13:04,040
Speaker 1: can actor industry had a renaissance in the nineteen nineties

199
00:13:04,080 --> 00:13:07,640
Speaker 1: because a lot of those companies let go of the

200
00:13:07,679 --> 00:13:10,959
Speaker 1: manufacturing part. Not everybody did, but a lot of them did,

201
00:13:11,080 --> 00:13:14,040
Speaker 1: and they've been able to be very successful. Companies like

202
00:13:14,080 --> 00:13:19,400
Speaker 1: Qualcom and in Video in particular. Uh, and now am

203
00:13:19,480 --> 00:13:25,000
Speaker 1: d by letting go of manufacturing. So this is the

204
00:13:25,840 --> 00:13:29,560
Speaker 1: I think it's called idem model, where you design and

205
00:13:29,600 --> 00:13:33,319
Speaker 1: manufacture chips like what Intel has been doing, versus the

206
00:13:33,400 --> 00:13:37,480
Speaker 1: fabulous foundry model like in Video and Qualcom. We got

207
00:13:37,520 --> 00:13:40,040
Speaker 1: into this a little bit with um Well in the

208
00:13:40,080 --> 00:13:44,800
Speaker 1: episode with Stacy raskin the semiconductor analysts over at Bernstein.

209
00:13:45,840 --> 00:13:48,880
Speaker 1: I'm curious, like the consensus in the industry right now

210
00:13:48,960 --> 00:13:51,880
Speaker 1: seems to be if you move to the fabulous foundry model,

211
00:13:51,880 --> 00:13:56,000
Speaker 1: where you sort of outsource your manufacturing to a foundry

212
00:13:56,080 --> 00:13:59,000
Speaker 1: like t SMC, that that's the way to go. But

213
00:13:59,200 --> 00:14:03,280
Speaker 1: your argue that there are some hidden costs or some

214
00:14:03,400 --> 00:14:07,720
Speaker 1: synergies that a company might lose in that model, Is

215
00:14:07,720 --> 00:14:11,800
Speaker 1: that right? Well, yes, I think so. And here I

216
00:14:11,800 --> 00:14:15,960
Speaker 1: would argue that more broadly speaking, there's a set of

217
00:14:16,040 --> 00:14:23,000
Speaker 1: capabilities that comes with your manufacturing which can feedback into

218
00:14:23,040 --> 00:14:28,119
Speaker 1: design as well. Now, I think in the chip industry

219
00:14:28,480 --> 00:14:33,200
Speaker 1: what we have seen is manufacturers have continued to work

220
00:14:33,520 --> 00:14:37,320
Speaker 1: very closely with the foundaries. Uh. And you know a

221
00:14:37,400 --> 00:14:42,200
Speaker 1: lot of that capability is transmitted through like software, design

222
00:14:42,280 --> 00:14:45,880
Speaker 1: tools and so on. Right, So that chip industry has

223
00:14:45,920 --> 00:14:48,760
Speaker 1: been able to the fabulous companies have been able to

224
00:14:50,120 --> 00:14:56,120
Speaker 1: leverage those manufacturing processes by working closely with their foundry partners. Okay.

225
00:14:55,920 --> 00:15:01,920
Speaker 1: I think the question in Semiconnectors is really what would

226
00:15:01,920 --> 00:15:05,320
Speaker 1: it be like for the US if it was unable

227
00:15:05,360 --> 00:15:11,040
Speaker 1: to manufacture anything domestically okay. And I think here the

228
00:15:11,120 --> 00:15:16,160
Speaker 1: COVID pandemic has really shined a light on that because

229
00:15:17,080 --> 00:15:22,520
Speaker 1: on a lot simpler commodities, things like personal protective equipment okay,

230
00:15:22,680 --> 00:15:26,520
Speaker 1: or ninety masks and things like that, you know, we

231
00:15:26,640 --> 00:15:30,880
Speaker 1: didn't have the manufacturing capability. Uh. And the question would

232
00:15:30,920 --> 00:15:38,040
Speaker 1: be what if Asian manufacturing capability was suddenly turned off?

233
00:15:38,480 --> 00:15:40,120
Speaker 1: You know. I'll point out that an awful lot of

234
00:15:40,120 --> 00:15:45,000
Speaker 1: the semiconductor manufacturing capability is in three science parks in Taiwan,

235
00:15:45,720 --> 00:15:49,880
Speaker 1: which is a seismically active region okay, And there have

236
00:15:50,040 --> 00:15:54,160
Speaker 1: been minor disruptions over the last few decades in terms

237
00:15:54,200 --> 00:15:57,920
Speaker 1: of ta typhoons and earthquakes and so on. But you know,

238
00:15:58,600 --> 00:16:03,040
Speaker 1: there's also the risk of geopolitical disruptions. Okay. So I

239
00:16:03,080 --> 00:16:09,000
Speaker 1: think that has heightened the sensitivity about having manufacturing concentrated

240
00:16:09,000 --> 00:16:13,280
Speaker 1: in such a small geographic area. It's a little different,

241
00:16:13,360 --> 00:16:17,560
Speaker 1: I would say than the the know how from manufacturing

242
00:16:17,600 --> 00:16:23,360
Speaker 1: in a lot of other processes, but you do see

243
00:16:23,360 --> 00:16:27,680
Speaker 1: it in semiconductors. In terms of the tool manufacturers okay,

244
00:16:27,680 --> 00:16:31,960
Speaker 1: and here American companies have a large presence and tool

245
00:16:32,000 --> 00:16:36,880
Speaker 1: manufacturing okay, and being close to the actual manufacturing facility

246
00:16:37,000 --> 00:16:39,360
Speaker 1: is very important, and you see a lot of those

247
00:16:39,360 --> 00:16:45,480
Speaker 1: toolmakers moving their development activities and manufacturing activities closer to

248
00:16:45,520 --> 00:16:49,400
Speaker 1: the market, be at Singapore or Taiwan, or Korea or

249
00:16:49,720 --> 00:17:08,560
Speaker 1: or even China. Can you explain in a little more

250
00:17:08,840 --> 00:17:15,320
Speaker 1: detail the mechanism by which learning by doing actual know

251
00:17:15,440 --> 00:17:20,040
Speaker 1: how in the manufacturing feeds through back to innovation, because

252
00:17:20,280 --> 00:17:22,600
Speaker 1: you know, it sounds good this idea that like, okay,

253
00:17:22,640 --> 00:17:25,560
Speaker 1: if you're actually in the process of sort of building chips,

254
00:17:25,600 --> 00:17:29,960
Speaker 1: if you're not outsourcing it, that also helps the technological innovation.

255
00:17:30,119 --> 00:17:32,800
Speaker 1: And that this idea, this dream of okay, you can

256
00:17:32,880 --> 00:17:36,040
Speaker 1: just separate the design and the manufacturing and still be

257
00:17:36,240 --> 00:17:39,840
Speaker 1: a world leader. That it's problematic. How does that work

258
00:17:39,960 --> 00:17:42,439
Speaker 1: in practice? Walk us through a little bit why that

259
00:17:42,520 --> 00:17:47,120
Speaker 1: link is important? So I think the key idea here

260
00:17:47,320 --> 00:17:53,879
Speaker 1: is tacit knowledge okay, And tacit knowledge is that which

261
00:17:54,359 --> 00:17:58,320
Speaker 1: you know but you maybe haven't written down or haven't

262
00:17:58,359 --> 00:18:02,280
Speaker 1: really explained to people. Right, So there is a lot

263
00:18:02,320 --> 00:18:06,760
Speaker 1: of know how on the shop floor that comes from

264
00:18:06,840 --> 00:18:10,679
Speaker 1: kind of learning as I produce more product, then the

265
00:18:10,760 --> 00:18:13,679
Speaker 1: question is how is that knowledge captured and how is

266
00:18:13,760 --> 00:18:19,639
Speaker 1: it uh communicated? Okay? Oftentimes you will see there's this

267
00:18:19,720 --> 00:18:24,520
Speaker 1: concept called design for manufacture. In other words, I was

268
00:18:24,560 --> 00:18:28,960
Speaker 1: in one factory once where the people assembling this piece

269
00:18:29,000 --> 00:18:31,800
Speaker 1: of equipment and said, you know, the guys who design this,

270
00:18:32,240 --> 00:18:35,920
Speaker 1: they haven't thought about you know, assembly sequence for example.

271
00:18:36,040 --> 00:18:38,480
Speaker 1: It's like, you know, you can't put that part there

272
00:18:38,520 --> 00:18:42,479
Speaker 1: because I can't I can't install it. But there's a

273
00:18:42,520 --> 00:18:46,000
Speaker 1: lot of those kind of tricks of the trade that

274
00:18:46,119 --> 00:18:50,080
Speaker 1: come from practice and manufacturing, which then feeds back into

275
00:18:50,160 --> 00:18:55,320
Speaker 1: design that says, okay, when you design it, recognize that

276
00:18:55,359 --> 00:18:59,800
Speaker 1: we have to do these things in manufacturing, and how

277
00:19:00,240 --> 00:19:04,600
Speaker 1: that change your design? Right? So tacit knowledge is really

278
00:19:05,240 --> 00:19:08,679
Speaker 1: the main concept you know that which I know, which

279
00:19:08,760 --> 00:19:16,480
Speaker 1: I cannot have not or cannot explained necessarily. So we've

280
00:19:16,520 --> 00:19:20,400
Speaker 1: sort of laid out the scene here and laid out

281
00:19:20,600 --> 00:19:24,480
Speaker 1: the differences between the two business models and why semiconductors

282
00:19:24,480 --> 00:19:28,199
Speaker 1: are important, and also why the industry is different to

283
00:19:28,680 --> 00:19:32,800
Speaker 1: a lot of other types of manufacturing and design. What

284
00:19:32,920 --> 00:19:37,240
Speaker 1: can actually be done about this? If if we agree

285
00:19:37,280 --> 00:19:40,639
Speaker 1: that Intel is falling behind on semiconductors, that the U

286
00:19:40,720 --> 00:19:45,159
Speaker 1: S should do more to boost its domestic capacity in

287
00:19:45,320 --> 00:19:48,639
Speaker 1: order to have a more resilient supply chain and you know,

288
00:19:49,400 --> 00:19:53,920
Speaker 1: be able to withstand unexpected disruptions like the pandemic, what's

289
00:19:53,960 --> 00:19:59,159
Speaker 1: the first step towards actually achieving that? Well, Tracy, I

290
00:19:59,160 --> 00:20:02,080
Speaker 1: think this is a tough question, uh, And a lot

291
00:20:02,119 --> 00:20:06,680
Speaker 1: of people I've talked to agree on the problem, but

292
00:20:06,760 --> 00:20:09,360
Speaker 1: you know, I think there's really a searching for what

293
00:20:09,440 --> 00:20:13,800
Speaker 1: kinds of steps would help. Now, what some people say

294
00:20:14,240 --> 00:20:19,080
Speaker 1: is Intel needs to establish a foundry operation. Okay, I

295
00:20:19,440 --> 00:20:25,000
Speaker 1: think I agree with that because in manufacturing, and especially

296
00:20:25,080 --> 00:20:29,119
Speaker 1: in manufacturing semiconductors, you need to have value, right because

297
00:20:29,960 --> 00:20:33,760
Speaker 1: through manufacturing lots of volumes, that's where you get your

298
00:20:33,840 --> 00:20:37,320
Speaker 1: process learning, that's where you get your process improvement, that's

299
00:20:37,320 --> 00:20:40,520
Speaker 1: where you improve your yields, okay, and find out what

300
00:20:40,640 --> 00:20:46,000
Speaker 1: works and what doesn't work. Okay. As Intel's microprocessor volumes

301
00:20:46,320 --> 00:20:51,800
Speaker 1: come under competitive assault from other designs, where people go

302
00:20:51,960 --> 00:20:55,240
Speaker 1: to say t SMC or a foundry to manufacture it,

303
00:20:55,480 --> 00:20:57,760
Speaker 1: Intel needs to keep its volumes up. One way it

304
00:20:57,840 --> 00:21:00,520
Speaker 1: could do that is it could establish a foundry operation

305
00:21:01,280 --> 00:21:04,720
Speaker 1: that would make both its own chips and chips for

306
00:21:04,800 --> 00:21:10,920
Speaker 1: other customers. That has its challenges, though, because Intel has

307
00:21:10,960 --> 00:21:13,840
Speaker 1: its own way of doing things which have served its

308
00:21:13,920 --> 00:21:16,800
Speaker 1: own needs historically, and so what it needs to do

309
00:21:16,880 --> 00:21:21,240
Speaker 1: then it needs to learn how to be a service

310
00:21:21,280 --> 00:21:27,080
Speaker 1: business that will service the needs of other customers, including

311
00:21:27,119 --> 00:21:31,520
Speaker 1: its competitors. Okay. And and that's that's kind of a

312
00:21:31,680 --> 00:21:35,879
Speaker 1: tough transition because that means it has to give priority

313
00:21:36,000 --> 00:21:42,040
Speaker 1: to its uh customer slash competitor needs okay, and it's

314
00:21:42,080 --> 00:21:48,920
Speaker 1: competitors who might use it as a foundry then will

315
00:21:49,240 --> 00:21:52,480
Speaker 1: ask the question of why do I want to help

316
00:21:52,520 --> 00:21:56,760
Speaker 1: Intel because they've been a tough competitor over time, and

317
00:21:56,880 --> 00:22:01,320
Speaker 1: if I give business to this foundry and my helping Intel, okay.

318
00:22:01,320 --> 00:22:05,600
Speaker 1: And the other question as a potential US foundary customer

319
00:22:05,600 --> 00:22:07,960
Speaker 1: I'm gonna ask is like, are you going to give

320
00:22:08,000 --> 00:22:13,199
Speaker 1: me the same level of technology and process capability that

321
00:22:13,320 --> 00:22:19,400
Speaker 1: I can reliably get in Taiwan today? So there there

322
00:22:19,400 --> 00:22:23,280
Speaker 1: are a lot of barriers to making a foundry business

323
00:22:23,320 --> 00:22:27,960
Speaker 1: like that work. But I would argue that Intel needs

324
00:22:28,000 --> 00:22:30,760
Speaker 1: to do something like that in order to keep its

325
00:22:30,920 --> 00:22:36,200
Speaker 1: manufacturing volumes up. You know why Intel wow? Because Intel

326
00:22:36,480 --> 00:22:40,240
Speaker 1: is probably the last best chance for the US to

327
00:22:40,320 --> 00:22:44,760
Speaker 1: have a leader into this technology. Let me just actually

328
00:22:44,920 --> 00:22:47,600
Speaker 1: back up here because I'm still sort of struggling with

329
00:22:48,040 --> 00:22:52,600
Speaker 1: a key question, which is when we talk about the

330
00:22:52,640 --> 00:22:55,439
Speaker 1: stumbles at Intel or the decline. I mean, here is

331
00:22:55,480 --> 00:23:00,200
Speaker 1: a company that has been integrated, it's uh connected, you

332
00:23:00,280 --> 00:23:03,600
Speaker 1: know sort of, it has this lots of tacit knowledge

333
00:23:03,640 --> 00:23:06,719
Speaker 1: and having manufactured its own chips for so long, unlike

334
00:23:06,720 --> 00:23:10,119
Speaker 1: some of its competitors. What is your core diagnosis for

335
00:23:10,200 --> 00:23:13,879
Speaker 1: what's wrong with the company? And also, um, you know, recently,

336
00:23:13,920 --> 00:23:16,320
Speaker 1: I think it was the hedge fund Third Point came

337
00:23:16,359 --> 00:23:20,159
Speaker 1: out with some proposals the activist firm to try to

338
00:23:20,200 --> 00:23:24,040
Speaker 1: revive the company. But a, um, how would you sort

339
00:23:24,040 --> 00:23:27,359
Speaker 1: of summarize your diagnosis of what went wrong there? And

340
00:23:27,400 --> 00:23:31,280
Speaker 1: then what do you make of some of the suggestions

341
00:23:31,320 --> 00:23:36,600
Speaker 1: by them? Well, I think uh, Intel needs to invest

342
00:23:36,760 --> 00:23:42,720
Speaker 1: heavily on process leadership. Ten years ago, they were two

343
00:23:42,720 --> 00:23:47,280
Speaker 1: generations ahead of the SMC. Now ten years later there

344
00:23:47,320 --> 00:23:50,720
Speaker 1: behind t SMC. Right, so they they need to invest

345
00:23:50,960 --> 00:23:56,840
Speaker 1: heavily on that. I think Intel historically has tried to

346
00:23:56,880 --> 00:24:00,600
Speaker 1: be very self sufficient in terms of using a lot

347
00:24:00,640 --> 00:24:04,600
Speaker 1: of their own tools and a lot of their own methods. Okay,

348
00:24:04,640 --> 00:24:07,560
Speaker 1: I mean I think one thing that would be helpful

349
00:24:08,440 --> 00:24:14,639
Speaker 1: would be to work more with others in the industry

350
00:24:14,680 --> 00:24:20,679
Speaker 1: in broadening the funnel of ideas that come to them. Okay,

351
00:24:20,720 --> 00:24:26,520
Speaker 1: because in some sense, trying to do it all yourself

352
00:24:27,160 --> 00:24:34,280
Speaker 1: in this very challenging technological era may not be the

353
00:24:34,320 --> 00:24:37,359
Speaker 1: most competitive path. If you look at companies like t SMC,

354
00:24:38,640 --> 00:24:44,600
Speaker 1: they have heavily leveraged partnerships with a SML who's the

355
00:24:44,680 --> 00:24:52,320
Speaker 1: lithography leader for DPUV. They have heavily leveraged organizations like

356
00:24:53,320 --> 00:24:57,879
Speaker 1: im EC in Europe, which is kind of a shared

357
00:24:57,960 --> 00:25:03,160
Speaker 1: research collaborative. Okay. And t SMC is very good about

358
00:25:04,000 --> 00:25:07,720
Speaker 1: working with a broad range of parties, and I think

359
00:25:07,880 --> 00:25:12,440
Speaker 1: that would help Intel, right, And that that means they

360
00:25:12,480 --> 00:25:15,920
Speaker 1: have to kind of set aside some of their established

361
00:25:15,920 --> 00:25:20,439
Speaker 1: ways of doing things and open themselves up to maybe

362
00:25:20,720 --> 00:25:26,080
Speaker 1: broader approaches to the competitiveness. Okay. Now on the third

363
00:25:26,119 --> 00:25:31,280
Speaker 1: point question, Okay, I think I agree with the need

364
00:25:31,920 --> 00:25:38,520
Speaker 1: for Intel to really establish a foundry business because, as

365
00:25:38,600 --> 00:25:42,960
Speaker 1: I said earlier, I think volume is very important for them.

366
00:25:43,119 --> 00:25:50,080
Speaker 1: But just the simplistic idea of separating their design business,

367
00:25:50,160 --> 00:25:56,160
Speaker 1: which is designing microprocessors, and manufacturing, which is manufacturing semiconductors,

368
00:25:57,359 --> 00:26:00,920
Speaker 1: taking them down the path that am Be went down

369
00:26:01,040 --> 00:26:05,760
Speaker 1: before and many other companies went down before, that will

370
00:26:05,840 --> 00:26:10,600
Speaker 1: create value for shareholders. Okay, but then you have to

371
00:26:10,640 --> 00:26:15,520
Speaker 1: ask how likely is that foundry going to be absent

372
00:26:16,240 --> 00:26:19,639
Speaker 1: some other moves to ensure that it has customers and

373
00:26:19,720 --> 00:26:23,399
Speaker 1: ensure that it is competitive. Right, So at a simple level,

374
00:26:24,720 --> 00:26:26,480
Speaker 1: I could see how it would create a lot of

375
00:26:26,520 --> 00:26:30,760
Speaker 1: shareholder value, but is that the right answer for the

376
00:26:30,840 --> 00:26:36,359
Speaker 1: United States? So one thing I'm wondering is, you know,

377
00:26:36,440 --> 00:26:40,200
Speaker 1: we're having this conversation now, and uh, there is a

378
00:26:40,200 --> 00:26:44,040
Speaker 1: proposal in Washington, as I mentioned, to do some of this,

379
00:26:44,320 --> 00:26:46,560
Speaker 1: and we can talk about that in some more detail

380
00:26:46,640 --> 00:26:51,760
Speaker 1: in just a bit. But why hasn't this been done before,

381
00:26:52,119 --> 00:26:55,800
Speaker 1: or why hasn't there been more of a concerted effort

382
00:26:56,200 --> 00:27:00,240
Speaker 1: to boost America's you know, strategic capability when it comes

383
00:27:00,280 --> 00:27:07,119
Speaker 1: to a technology that people agree is increasingly important. I

384
00:27:07,359 --> 00:27:14,320
Speaker 1: think we have h in some sense been looking at

385
00:27:14,359 --> 00:27:18,919
Speaker 1: a rosy picture of the semiconductor industry, reflected by the

386
00:27:19,040 --> 00:27:25,879
Speaker 1: success of fabulous companies like Nvidia, like Broadcom, like Qualcom,

387
00:27:25,920 --> 00:27:30,359
Speaker 1: like Apple by the way, uh, and we're seeing it

388
00:27:30,440 --> 00:27:34,280
Speaker 1: as well in companies like Amazon and Google who say

389
00:27:34,359 --> 00:27:37,760
Speaker 1: I don't need the manufacturing, but you see the valuations

390
00:27:37,800 --> 00:27:41,560
Speaker 1: of these companies and you see the financial success. As

391
00:27:41,640 --> 00:27:46,680
Speaker 1: long as you know, Asia represents a secure manufacturing supply

392
00:27:46,800 --> 00:27:52,040
Speaker 1: based one could argue if they want to subsidize that,

393
00:27:52,280 --> 00:27:57,439
Speaker 1: or they want to do lower value work, then we

394
00:27:57,480 --> 00:28:00,000
Speaker 1: should let them. And there are people in Silicon Valley

395
00:28:00,040 --> 00:28:02,119
Speaker 1: I've talked to who said, you know, there's a strong

396
00:28:02,320 --> 00:28:07,000
Speaker 1: argument for Hey, if Asian countries want to subsidize manufacturing,

397
00:28:07,000 --> 00:28:09,160
Speaker 1: we should let them. We should take advantage of that.

398
00:28:09,800 --> 00:28:12,960
Speaker 1: And then the only question is what happens if you

399
00:28:13,000 --> 00:28:16,960
Speaker 1: get cut off from that supply. Now, I think we

400
00:28:17,119 --> 00:28:20,080
Speaker 1: tend to believe in the US that people wouldn't cut

401
00:28:20,160 --> 00:28:24,520
Speaker 1: us off from supply. Okay. In the last couple of years,

402
00:28:24,560 --> 00:28:29,000
Speaker 1: the US has really weaponized supply chains against other countries,

403
00:28:29,760 --> 00:28:34,439
Speaker 1: notably against companies like Huawei or z t E or

404
00:28:34,600 --> 00:28:37,919
Speaker 1: you know a lot more recently. Okay, So then the

405
00:28:38,000 --> 00:28:41,760
Speaker 1: question is do we open ourselves to that kind of

406
00:28:41,840 --> 00:28:47,120
Speaker 1: quid quid pro quo behavior. So let's talk a little

407
00:28:47,120 --> 00:28:50,760
Speaker 1: bit more about national policy and you talk about the

408
00:28:50,920 --> 00:28:55,280
Speaker 1: case for why it should be a DC priority. Now,

409
00:28:55,400 --> 00:29:00,560
Speaker 1: just two for my recollection or knowledge. In the later Uh,

410
00:29:00,800 --> 00:29:04,040
Speaker 1: there's the Foundry Act, which I sought to encourage more

411
00:29:04,120 --> 00:29:07,560
Speaker 1: domestic investment. Has that been passed? Now? Well, there was

412
00:29:07,600 --> 00:29:12,320
Speaker 1: the American Foundries Act, Okay, and then uh there was

413
00:29:12,920 --> 00:29:18,560
Speaker 1: another act. Those got merged into the National Defense Authorization Act,

414
00:29:18,640 --> 00:29:22,440
Speaker 1: which was passed, but the funding still has to be

415
00:29:22,480 --> 00:29:24,920
Speaker 1: worked out. Okay. And if you and if you read

416
00:29:25,560 --> 00:29:28,280
Speaker 1: the language in the n d A A uh, I

417
00:29:28,400 --> 00:29:31,440
Speaker 1: forgot exactly how many pages it was, but it was

418
00:29:32,440 --> 00:29:36,959
Speaker 1: around sev hundred something. And uh, you know, so I

419
00:29:37,000 --> 00:29:40,920
Speaker 1: went to and found all the language on the semi

420
00:29:40,960 --> 00:29:45,720
Speaker 1: conductor stuff, and I would describe it as aspirational, right

421
00:29:46,120 --> 00:29:51,000
Speaker 1: understanding that we're gonna we're gonna direct Commerce Department to

422
00:29:51,040 --> 00:29:53,280
Speaker 1: do these things. We're going to direct all these people

423
00:29:53,320 --> 00:29:57,080
Speaker 1: to report and all that. The challenge right now, I mean,

424
00:29:57,160 --> 00:30:02,000
Speaker 1: I think Congressional staff have good intense here, right but

425
00:30:02,080 --> 00:30:07,560
Speaker 1: the challenge right now is this micro electronics leadership in

426
00:30:07,600 --> 00:30:12,400
Speaker 1: this segment. You know that intel is in. There is

427
00:30:12,480 --> 00:30:17,880
Speaker 1: a dearth of what I would describe as good ideas,

428
00:30:19,080 --> 00:30:23,640
Speaker 1: good strategic moves on how to handle this. I've been

429
00:30:23,720 --> 00:30:28,560
Speaker 1: racking my brain on this recently, but I think it's

430
00:30:28,560 --> 00:30:32,600
Speaker 1: a very tough problem that has been decades in the making.

431
00:30:33,480 --> 00:30:37,800
Speaker 1: I hope people, you know, don't kind of get dissolution

432
00:30:37,840 --> 00:30:41,360
Speaker 1: of these things. It's like my mother always told me,

433
00:30:41,640 --> 00:30:43,600
Speaker 1: problems that we're a long time in the making are

434
00:30:43,640 --> 00:30:46,200
Speaker 1: a long time in solving. Right. This is not a

435
00:30:46,240 --> 00:30:49,800
Speaker 1: problem that gets solved in three years or five years. Right.

436
00:30:49,840 --> 00:30:53,320
Speaker 1: If you look at t SMC, which was found in

437
00:30:53,320 --> 00:30:58,360
Speaker 1: the nineteen eighties, Okay, they've been working very, very hard

438
00:30:58,840 --> 00:31:01,959
Speaker 1: for decades to get to where they are, all right,

439
00:31:02,080 --> 00:31:07,800
Speaker 1: and they're not about to lose this either. So what

440
00:31:07,920 --> 00:31:11,200
Speaker 1: is I mean you say there's this dearth of ideas.

441
00:31:11,240 --> 00:31:14,560
Speaker 1: But obviously, and we talked about this on the previous episode,

442
00:31:14,560 --> 00:31:16,239
Speaker 1: there's sort of a limit to how much you can

443
00:31:16,280 --> 00:31:20,640
Speaker 1: achieve just by throwing money at the problem. But of

444
00:31:20,680 --> 00:31:23,080
Speaker 1: course money it has to be part of it. Um

445
00:31:23,120 --> 00:31:26,520
Speaker 1: So in your view, what would be the sort of

446
00:31:26,600 --> 00:31:32,680
Speaker 1: basics of the beginning of rebuilding that capability domestically, both

447
00:31:32,720 --> 00:31:35,960
Speaker 1: in terms of how much investment would need from d C,

448
00:31:36,160 --> 00:31:39,160
Speaker 1: but also investment trained in the right way such that

449
00:31:39,200 --> 00:31:42,200
Speaker 1: it really fostered a sort of self sustaining and more,

450
00:31:42,760 --> 00:31:47,440
Speaker 1: you know, robust capabilities. So one of the ideas, Joe,

451
00:31:47,520 --> 00:31:52,000
Speaker 1: that I have been floating is they're an awful lot

452
00:31:52,040 --> 00:31:56,240
Speaker 1: of proposals that focus on the supply side, like let's

453
00:31:56,280 --> 00:31:59,160
Speaker 1: get people to build fabs in the US, you know,

454
00:31:59,360 --> 00:32:01,720
Speaker 1: let's get bowl to do more R and D in

455
00:32:01,760 --> 00:32:04,680
Speaker 1: the U s Okay, And that's all well and good.

456
00:32:05,560 --> 00:32:09,840
Speaker 1: But you know, if you build a new foundry in

457
00:32:09,920 --> 00:32:15,520
Speaker 1: the US and nobody shops there, it's not sustainable, right.

458
00:32:15,560 --> 00:32:20,000
Speaker 1: So what I've been saying is instead of only focusing

459
00:32:20,040 --> 00:32:22,160
Speaker 1: on the supply side, we should also think about the

460
00:32:22,200 --> 00:32:26,040
Speaker 1: demand side, okay. And the reason I highlight that is

461
00:32:26,640 --> 00:32:29,360
Speaker 1: if you go back to the nineteen sixties and I'm

462
00:32:29,400 --> 00:32:32,120
Speaker 1: I'm dating myself now, but you know, if you go

463
00:32:32,160 --> 00:32:37,080
Speaker 1: back to the nineteen sixties, at the birth of the

464
00:32:37,120 --> 00:32:41,040
Speaker 1: integrated circuit and the birth of the semiconductor industry in

465
00:32:41,080 --> 00:32:45,959
Speaker 1: the United States, d D and NASA bought of all

466
00:32:46,000 --> 00:32:49,600
Speaker 1: the integrated circuits that were made in the world, Okay,

467
00:32:49,640 --> 00:32:54,160
Speaker 1: And that was something that helped companies like Texas Instruments,

468
00:32:54,160 --> 00:32:58,480
Speaker 1: who was one of the pioneers okay, and and Fairchild

469
00:32:59,720 --> 00:33:03,240
Speaker 1: which gave birth to Intel and others. Okay, But it

470
00:33:03,360 --> 00:33:08,080
Speaker 1: was that demand that really helped foster the growth of

471
00:33:08,080 --> 00:33:11,760
Speaker 1: the industry. It was a customer for kind of those

472
00:33:11,800 --> 00:33:18,000
Speaker 1: early products who worked with the suppliers to improve. Okay.

473
00:33:18,000 --> 00:33:20,959
Speaker 1: So I've been saying, it's like, let's focus on demand

474
00:33:21,000 --> 00:33:25,120
Speaker 1: as well. Now, one of the things that we do

475
00:33:25,320 --> 00:33:28,160
Speaker 1: have in the United States is we do have demand

476
00:33:28,200 --> 00:33:29,840
Speaker 1: for a lot of high and semi conductors. If you

477
00:33:29,840 --> 00:33:33,400
Speaker 1: look at all these hyper scaled data centers along the

478
00:33:33,440 --> 00:33:37,360
Speaker 1: Columbia River or around the DC area, or spaced across

479
00:33:37,440 --> 00:33:41,120
Speaker 1: the country. Let's take some of these big hyper scale

480
00:33:41,200 --> 00:33:44,360
Speaker 1: data centers from you know, Google or Microsoft or Amazon

481
00:33:45,000 --> 00:33:49,040
Speaker 1: along the Columbia River. Well, where did those zon microprocessors

482
00:33:49,640 --> 00:33:53,280
Speaker 1: come from? Well, probably Hillsboro, Oregon, where Intel has a fab.

483
00:33:53,400 --> 00:33:56,360
Speaker 1: You know, that's one of their leading fabs in the US.

484
00:33:56,480 --> 00:34:00,320
Speaker 1: And then what happened as well, Well, those chips then

485
00:34:00,480 --> 00:34:04,280
Speaker 1: went to the finished wafers with the chips on them

486
00:34:05,120 --> 00:34:08,800
Speaker 1: that were then sent to either end China or Vietnam

487
00:34:09,000 --> 00:34:13,560
Speaker 1: or Malaysia where they were tested and cut apart into

488
00:34:13,680 --> 00:34:17,680
Speaker 1: individual die and then they were put into packages, and

489
00:34:17,719 --> 00:34:21,359
Speaker 1: then they would go to a distribution center somewhere in

490
00:34:21,400 --> 00:34:26,400
Speaker 1: Asia and then go to a serverboard manufacturer in China probably,

491
00:34:27,440 --> 00:34:29,880
Speaker 1: and then they ship them back to the Columbia River Valley,

492
00:34:29,920 --> 00:34:33,120
Speaker 1: you know, a hundred miles away from where the chips originated.

493
00:34:33,120 --> 00:34:36,279
Speaker 1: And it's like, now, why do they take that round trip? Well,

494
00:34:36,320 --> 00:34:38,359
Speaker 1: they take that round trip because in the early days

495
00:34:38,400 --> 00:34:43,800
Speaker 1: of the chip industry, when you had to cut these

496
00:34:44,600 --> 00:34:46,960
Speaker 1: chips apart with the diamond saw and then put them

497
00:34:47,000 --> 00:34:49,759
Speaker 1: into little packages, you would have rows and rows of

498
00:34:49,800 --> 00:34:57,120
Speaker 1: workers looking through microscopes. Why are bonding those die into packages? Okay?

499
00:34:57,120 --> 00:35:00,760
Speaker 1: And it was very labor intensive. And semi conector chips,

500
00:35:00,880 --> 00:35:02,680
Speaker 1: you know, they have a very high value density. So

501
00:35:02,719 --> 00:35:06,080
Speaker 1: it's no problem to air air fraight them over to Asia.

502
00:35:06,200 --> 00:35:10,359
Speaker 1: Haven't packaging and haven't come back. Okay. These days, all

503
00:35:10,440 --> 00:35:13,279
Speaker 1: that stuff is automated, so there's no reason for it

504
00:35:13,320 --> 00:35:16,560
Speaker 1: to go to Asia except over the last four or

505
00:35:16,560 --> 00:35:20,000
Speaker 1: five decades, Asia has developed the capability for chip packaging

506
00:35:20,040 --> 00:35:23,200
Speaker 1: and automation and so on. So one of the things

507
00:35:23,239 --> 00:35:25,720
Speaker 1: I would focus on is I would focus on let's

508
00:35:26,360 --> 00:35:29,759
Speaker 1: bring kind of chip packaging back to the US. We

509
00:35:29,800 --> 00:35:33,120
Speaker 1: still have companies in the US like am Corps who

510
00:35:33,160 --> 00:35:37,759
Speaker 1: are leaders in that technology, you know, and let's help

511
00:35:37,880 --> 00:35:43,680
Speaker 1: them to do packaging in the US and just kind

512
00:35:43,680 --> 00:35:47,640
Speaker 1: of move a key part of that value chain where

513
00:35:47,680 --> 00:35:50,480
Speaker 1: we have some of that demand. You know. The other

514
00:35:50,560 --> 00:35:54,239
Speaker 1: reason I think packaging is important is we're seeing the

515
00:35:54,360 --> 00:35:57,920
Speaker 1: end of Moore's law where it is becoming harder and

516
00:35:58,040 --> 00:36:03,320
Speaker 1: harder to push the frontier and these advanced processes okay,

517
00:36:03,320 --> 00:36:06,440
Speaker 1: but an awful lot of the horsepower you need in

518
00:36:06,480 --> 00:36:10,240
Speaker 1: these chips come from lots of different processes and maybe

519
00:36:10,239 --> 00:36:12,160
Speaker 1: some not so advanced, and you can get a lot

520
00:36:12,200 --> 00:36:16,120
Speaker 1: of the benefit by going to what people call three

521
00:36:16,200 --> 00:36:21,000
Speaker 1: D packaging, where I'm going to put multiple chips into

522
00:36:21,080 --> 00:36:25,000
Speaker 1: a single package and then kind of bond them together

523
00:36:25,160 --> 00:36:28,440
Speaker 1: with what are called vas or you know. Basically, I'm

524
00:36:28,440 --> 00:36:31,480
Speaker 1: going to use multiple chips in one package because if

525
00:36:31,480 --> 00:36:34,719
Speaker 1: i can cut the distance between them and pack them

526
00:36:34,719 --> 00:36:36,200
Speaker 1: more densely, I'm going to get a lot of the

527
00:36:36,200 --> 00:36:39,759
Speaker 1: benefits from Moore's law without having to go to the

528
00:36:39,760 --> 00:36:45,719
Speaker 1: most advanced processes. Right. So, by focusing on parts of

529
00:36:45,760 --> 00:36:50,600
Speaker 1: the value chain that we can win over the shorter

530
00:36:50,840 --> 00:36:55,400
Speaker 1: term or get some some of that manufacturing capability back

531
00:36:56,320 --> 00:37:02,600
Speaker 1: while you are working to restore process leadership, particularly looking

532
00:37:02,680 --> 00:37:08,440
Speaker 1: at next generation, like what comes after all the technologies

533
00:37:08,880 --> 00:37:12,200
Speaker 1: that we know will take us to three nimeter today

534
00:37:12,239 --> 00:37:16,120
Speaker 1: with Moore's Law that has gotten very expensive. Investing in

535
00:37:17,280 --> 00:37:23,520
Speaker 1: new process technologies, new technologies that will take US and

536
00:37:23,600 --> 00:37:28,920
Speaker 1: global semiconductor production to kind of the next era that

537
00:37:28,960 --> 00:37:31,480
Speaker 1: would be a good investment as well. Right, So, I mean,

538
00:37:31,800 --> 00:37:36,000
Speaker 1: I think there are things that we can do that

539
00:37:36,080 --> 00:37:39,480
Speaker 1: will take investment, okay, but those also play to some

540
00:37:39,560 --> 00:37:58,080
Speaker 1: of the strengths of this country. Has I want to

541
00:37:58,080 --> 00:38:01,759
Speaker 1: go back to something said earlier about the idea that

542
00:38:01,960 --> 00:38:06,840
Speaker 1: the US has been weaponizing supply chains, particularly in China

543
00:38:07,080 --> 00:38:11,920
Speaker 1: and with its actions against Huawei UM. We've seen Huawei

544
00:38:12,080 --> 00:38:16,239
Speaker 1: not necessarily struggle to source semiconductors, but they've had to

545
00:38:16,440 --> 00:38:18,719
Speaker 1: stockpile a bunch of them UM and they're sort of

546
00:38:18,760 --> 00:38:21,680
Speaker 1: starting to run through their supply because they can't get

547
00:38:21,760 --> 00:38:24,880
Speaker 1: them in the way that they used to. There's a

548
00:38:24,960 --> 00:38:29,319
Speaker 1: perception within China, or at least there's one strand of

549
00:38:29,360 --> 00:38:33,600
Speaker 1: thinking within China that this isn't necessarily the worst thing

550
00:38:33,760 --> 00:38:38,120
Speaker 1: for the country because it's going to end up accelerating

551
00:38:38,160 --> 00:38:43,279
Speaker 1: their own domestic manufacturing ability. So they're going to sort

552
00:38:43,280 --> 00:38:46,440
Speaker 1: of rush through development UM and the US has in

553
00:38:46,480 --> 00:38:50,080
Speaker 1: some ways done them a favor by sort of sparking

554
00:38:50,520 --> 00:38:53,360
Speaker 1: a big fire under them to get this done. Is

555
00:38:53,880 --> 00:38:58,240
Speaker 1: that a credible argument in your view? Or I guess

556
00:38:58,239 --> 00:39:00,640
Speaker 1: in your view, what are the dawn sites and the

557
00:39:00,680 --> 00:39:07,120
Speaker 1: benefits to targeting um external manufacturing or targeting China in

558
00:39:07,160 --> 00:39:11,640
Speaker 1: this way? Well, in in some sense what China is

559
00:39:11,680 --> 00:39:16,440
Speaker 1: doing is uh there you know, on an imports substitution

560
00:39:16,560 --> 00:39:21,919
Speaker 1: drive now in response to this weaponization. Okay, A lot

561
00:39:21,960 --> 00:39:26,879
Speaker 1: of economies have used import substitution very successfully to kind

562
00:39:26,880 --> 00:39:31,440
Speaker 1: of bootstrap themselves up. Taiwan notably one of them, right, Korea, Japan,

563
00:39:32,120 --> 00:39:36,280
Speaker 1: a lot of Asian nations have used that. Okay, will

564
00:39:36,920 --> 00:39:40,960
Speaker 1: China be able to get there? Right? You will hear

565
00:39:41,080 --> 00:39:44,239
Speaker 1: a lot of skeptics who say, well, China won't be

566
00:39:44,360 --> 00:39:48,560
Speaker 1: able to get there fast because they still need technology

567
00:39:48,800 --> 00:39:52,640
Speaker 1: from the West. Okay, and it may take them a while.

568
00:39:53,160 --> 00:39:56,120
Speaker 1: But you know, I have friends in China who said,

569
00:39:56,160 --> 00:40:00,000
Speaker 1: you know, you guys in the US, you're a young country,

570
00:40:00,040 --> 00:40:03,920
Speaker 1: free your less than years old. That's like, you know,

571
00:40:04,000 --> 00:40:08,120
Speaker 1: one dynasty, right, and it's like they have a very

572
00:40:08,200 --> 00:40:13,000
Speaker 1: different scale on time. Okay. And I would point out

573
00:40:13,760 --> 00:40:19,080
Speaker 1: that the Europeans in the late sixties and early seventies,

574
00:40:19,960 --> 00:40:24,920
Speaker 1: we're looking at the commercial hygenemy in commercial aircraft in

575
00:40:24,960 --> 00:40:28,719
Speaker 1: the US, and they were concerned about it. They put

576
00:40:28,719 --> 00:40:33,560
Speaker 1: together air Bus as a consortium. Right Now, air Bus

577
00:40:33,600 --> 00:40:39,520
Speaker 1: had lots of political issues early on, okay, But here

578
00:40:39,560 --> 00:40:45,560
Speaker 1: fifty years later, air Bus arguably is the global leader. Right.

579
00:40:45,880 --> 00:40:49,279
Speaker 1: Is fifty years a long time for China? Well, it

580
00:40:49,440 --> 00:40:52,280
Speaker 1: is in the context of their five year plans, okay,

581
00:40:52,320 --> 00:40:58,600
Speaker 1: but I wouldn't underestimate, you know, their ability to get there. Now.

582
00:40:59,160 --> 00:41:02,160
Speaker 1: I personally don't think that's a good idea, right because

583
00:41:03,080 --> 00:41:07,319
Speaker 1: I still believe that the world is better off if

584
00:41:07,360 --> 00:41:12,000
Speaker 1: we trade with each other. But that also requires, uh,

585
00:41:12,440 --> 00:41:15,919
Speaker 1: some notion of balance, right. I mean what you're hearing

586
00:41:15,920 --> 00:41:20,600
Speaker 1: out of China analysis so called dual circulation strategy, right,

587
00:41:20,680 --> 00:41:25,600
Speaker 1: which is they're going to be self sufficient as much

588
00:41:25,640 --> 00:41:29,840
Speaker 1: as possible, but they still want to be a source

589
00:41:30,440 --> 00:41:35,200
Speaker 1: of manufacturers for the world. Right. That's something I would

590
00:41:35,320 --> 00:41:38,280
Speaker 1: question because if you don't want to buy from anybody

591
00:41:38,280 --> 00:41:41,279
Speaker 1: else and you want them to buy from you. I

592
00:41:41,320 --> 00:41:43,279
Speaker 1: don't see how that's going to play very well over

593
00:41:43,320 --> 00:41:46,000
Speaker 1: the long term, right, So I don't think you can

594
00:41:46,280 --> 00:41:52,400
Speaker 1: have it both ways. But I wouldn't underestimate China's ability

595
00:41:52,600 --> 00:41:55,399
Speaker 1: to get there. There are still some people, I think,

596
00:41:55,400 --> 00:41:58,239
Speaker 1: who believe all China is just about copycats and so on.

597
00:41:58,880 --> 00:42:01,920
Speaker 1: There are some very innovative people in China. They're very

598
00:42:02,120 --> 00:42:05,640
Speaker 1: some very innovative companies. So I would not underestimate there

599
00:42:05,680 --> 00:42:11,120
Speaker 1: are capabilities over the long term. Might they spend much

600
00:42:11,160 --> 00:42:17,319
Speaker 1: more money getting to you know, parody and semiconductors than

601
00:42:17,400 --> 00:42:22,240
Speaker 1: they would have if they took a more global approach.

602
00:42:23,400 --> 00:42:28,359
Speaker 1: Probably does that matter to them? Maybe not? I want

603
00:42:28,360 --> 00:42:32,319
Speaker 1: no asking a slightly shift the conversation just slightly to

604
00:42:32,520 --> 00:42:35,440
Speaker 1: another sort of timely topic, but it's related to this.

605
00:42:35,480 --> 00:42:41,520
Speaker 1: So we're talking about these sort of global hubs of technology,

606
00:42:41,640 --> 00:42:46,920
Speaker 1: know how. Obviously Taiwan is a q on. Also Singapore

607
00:42:47,480 --> 00:42:52,560
Speaker 1: Silicon Valley has silicon right in its name, the home

608
00:42:52,680 --> 00:42:56,520
Speaker 1: of the home of this industry and lately and this

609
00:42:56,640 --> 00:42:59,319
Speaker 1: is sort of maybe more software outside of technology, but

610
00:42:59,760 --> 00:43:03,000
Speaker 1: we here Heart especially like it really is accelerating the

611
00:43:03,080 --> 00:43:07,359
Speaker 1: last several months people talking about moving to new tech

612
00:43:07,440 --> 00:43:10,040
Speaker 1: hubs or creating tech hubs. So maybe people are annoyed

613
00:43:10,080 --> 00:43:13,319
Speaker 1: with politics or the government in California and they're like,

614
00:43:13,320 --> 00:43:15,760
Speaker 1: you know, we're all gonna let's move to Austin, Texas,

615
00:43:16,400 --> 00:43:19,480
Speaker 1: or let's move to Miami and we'll just sort of

616
00:43:19,640 --> 00:43:22,800
Speaker 1: set up a new a new tech hub in this area,

617
00:43:23,400 --> 00:43:27,200
Speaker 1: get away from California regulation or whatever it is. What

618
00:43:27,320 --> 00:43:32,319
Speaker 1: does history say about sort of what will make a

619
00:43:32,480 --> 00:43:36,399
Speaker 1: tech hub succeed or not? And is this the kind

620
00:43:36,440 --> 00:43:39,080
Speaker 1: of thing that can sort of be done if you know,

621
00:43:39,400 --> 00:43:41,600
Speaker 1: have a few venture capital leaders say we're all moving

622
00:43:41,640 --> 00:43:44,399
Speaker 1: to Miami, We're gonna set up the next Silicon Valley here.

623
00:43:44,719 --> 00:43:47,080
Speaker 1: Does it work that way or does it sort of

624
00:43:47,120 --> 00:43:51,520
Speaker 1: need to be more emergent and organic. Well, I think

625
00:43:51,640 --> 00:43:56,160
Speaker 1: one of the benefits of physical proximity has always been

626
00:43:56,320 --> 00:43:58,640
Speaker 1: the movement of people. And this goes back to what

627
00:43:58,680 --> 00:44:01,440
Speaker 1: we were talking about earlier in terms of testent knowledge.

628
00:44:01,520 --> 00:44:04,880
Speaker 1: If you think about how testent knowledge moves, it generally

629
00:44:04,920 --> 00:44:07,680
Speaker 1: moves into the head of heads of people. I worked

630
00:44:07,719 --> 00:44:13,000
Speaker 1: for IBM for fourteen years, okay, and every disk drive

631
00:44:13,160 --> 00:44:18,160
Speaker 1: startup in Silicon Valley was started by people who came

632
00:44:18,200 --> 00:44:22,680
Speaker 1: out of the IBM research operations in San Jose, California,

633
00:44:22,719 --> 00:44:26,040
Speaker 1: where they did descrive development every single one. Right, So

634
00:44:27,080 --> 00:44:29,439
Speaker 1: what happened is they had different ideas or they wanted

635
00:44:29,480 --> 00:44:33,720
Speaker 1: to pursue different ideas, you know, Phineas Connor, Allan Sugar,

636
00:44:33,800 --> 00:44:38,440
Speaker 1: all those guys. And so that's one of the things

637
00:44:38,480 --> 00:44:44,320
Speaker 1: that makes hubs productive, right, which is the spreading of ideas,

638
00:44:44,400 --> 00:44:48,800
Speaker 1: this movement of ideas, movement of people. Now the phenomena

639
00:44:48,840 --> 00:44:51,120
Speaker 1: you describe, I mean, I lived in Austin, Texas for

640
00:44:51,160 --> 00:44:54,480
Speaker 1: a while, and you know, I like the no income

641
00:44:54,480 --> 00:44:58,200
Speaker 1: tax and the much lower cost of living there compared

642
00:44:58,239 --> 00:45:00,560
Speaker 1: to California. I lived in California as for a while.

643
00:45:01,600 --> 00:45:06,080
Speaker 1: Those movements, I would argue, are really about being able

644
00:45:06,120 --> 00:45:11,200
Speaker 1: to attract talent and lifestyle when you're unconstrained by having

645
00:45:11,239 --> 00:45:16,080
Speaker 1: to be in a physical office. I don't know if

646
00:45:17,080 --> 00:45:25,200
Speaker 1: zoom and uh Microsoft teams or virtual whatever, what the

647
00:45:25,239 --> 00:45:30,880
Speaker 1: impact will be that movement of tested knowledge between organizations.

648
00:45:31,000 --> 00:45:35,200
Speaker 1: And if if what we have as a new era

649
00:45:35,440 --> 00:45:40,640
Speaker 1: of that kind of movement, maybe we'll see some more dispersion. Okay,

650
00:45:40,640 --> 00:45:44,440
Speaker 1: But I would argue one of the things that really

651
00:45:44,520 --> 00:45:48,160
Speaker 1: has help Silicon Valley over time has been the people

652
00:45:48,280 --> 00:45:51,520
Speaker 1: movement between jobs to the extent that they're willing to

653
00:45:52,040 --> 00:45:57,520
Speaker 1: geographically relocate, which is something that has historically been very

654
00:45:57,560 --> 00:46:00,759
Speaker 1: strong in the US but has diminished over the last decade.

655
00:46:01,560 --> 00:46:06,600
Speaker 1: It could work, right, but we'll see on that, Professor

656
00:46:06,640 --> 00:46:09,600
Speaker 1: she this is a great conversation before you go. I

657
00:46:09,640 --> 00:46:12,080
Speaker 1: wanna sort of I was reading one of your articles

658
00:46:12,120 --> 00:46:14,840
Speaker 1: and I thought this was super interesting and maybe super

659
00:46:14,840 --> 00:46:19,360
Speaker 1: helpful for listeners. And that is the brutal math the

660
00:46:19,360 --> 00:46:23,719
Speaker 1: brutal mathematics of UM yields, and we've talked about it

661
00:46:24,120 --> 00:46:27,160
Speaker 1: on a few a couple of previous episodes. But this

662
00:46:27,280 --> 00:46:30,960
Speaker 1: idea you sort of enter do this process. As you mentioned,

663
00:46:31,640 --> 00:46:33,880
Speaker 1: a fab could cost twenty billion dollars here, making this

664
00:46:34,000 --> 00:46:38,520
Speaker 1: huge investment extraordinarily complex. And then to hit your margins

665
00:46:38,560 --> 00:46:41,640
Speaker 1: and to actually make a return, you need to guarantee

666
00:46:41,760 --> 00:46:45,320
Speaker 1: that if you produce that on the way to creating

667
00:46:45,360 --> 00:46:48,319
Speaker 1: a finished product UM, that it actually works, or that

668
00:46:48,400 --> 00:46:50,560
Speaker 1: most of them actually work, so that you're not throwing

669
00:46:50,560 --> 00:46:53,239
Speaker 1: away a bunch or whatever, and that even if you

670
00:46:53,280 --> 00:46:57,839
Speaker 1: have say like effectiveness on various steps, that that might

671
00:46:57,920 --> 00:47:00,920
Speaker 1: not be enough. Can you talk us through how yields

672
00:47:00,920 --> 00:47:06,720
Speaker 1: are calculated and why the requirements for excellence are so demanding? Yes, okay,

673
00:47:06,719 --> 00:47:10,359
Speaker 1: so in semiconductors. Uh. The thing that makes this one

674
00:47:10,360 --> 00:47:13,640
Speaker 1: of the most challenging manufacturing processes is that it has

675
00:47:13,920 --> 00:47:16,640
Speaker 1: so many steps. Okay. And if you just do the

676
00:47:16,719 --> 00:47:22,400
Speaker 1: math of saying, let's say I had seven steps, okay,

677
00:47:22,560 --> 00:47:29,840
Speaker 1: and let's say I have yield for the first step. Okay.

678
00:47:29,880 --> 00:47:32,680
Speaker 1: If I have yield for the first step and then

679
00:47:33,320 --> 00:47:37,480
Speaker 1: for the second, for the third, after each step, I'm

680
00:47:37,520 --> 00:47:40,080
Speaker 1: getting fewer and fewer, Okay, By the time I get

681
00:47:40,120 --> 00:47:44,400
Speaker 1: to seven steps, I will get no good output. Okay. Therefore,

682
00:47:44,840 --> 00:47:47,279
Speaker 1: what that means is I actually have to have my

683
00:47:48,320 --> 00:47:50,680
Speaker 1: yields way up there. That means I have to have

684
00:47:52,400 --> 00:47:55,680
Speaker 1: point nine nine more than that. Right. If I have

685
00:47:58,320 --> 00:48:01,359
Speaker 1: yield at each step, and most of us think, oh,

686
00:48:03,000 --> 00:48:08,399
Speaker 1: that's pretty good, right, ninety nine times seven steps will

687
00:48:08,440 --> 00:48:11,080
Speaker 1: give me fifty yield out the other end. Okay. Now,

688
00:48:11,160 --> 00:48:15,840
Speaker 1: let's think about the economics of a twenty billion dollar

689
00:48:16,000 --> 00:48:22,920
Speaker 1: fab that might produce let's say wafers a month. Okay,

690
00:48:22,920 --> 00:48:26,360
Speaker 1: And on those twenty thou wafers, if I have to

691
00:48:26,400 --> 00:48:30,480
Speaker 1: throw out half of them. You can see the economic

692
00:48:30,520 --> 00:48:33,840
Speaker 1: cost of that is unbelievable. Right, So when you have

693
00:48:33,880 --> 00:48:37,520
Speaker 1: an expensive fab you need everything coming out of it

694
00:48:37,560 --> 00:48:39,560
Speaker 1: to be good, okay, and when you have a lot

695
00:48:39,600 --> 00:48:43,440
Speaker 1: of steps, that means you've got to be almost flawless

696
00:48:43,640 --> 00:48:46,160
Speaker 1: on every single step. And that's why this process, this

697
00:48:46,400 --> 00:48:51,319
Speaker 1: manufacturing process is so hard. I love that. That's a

698
00:48:51,360 --> 00:48:54,560
Speaker 1: great place to, uh to stop and I think I

699
00:48:54,680 --> 00:48:58,320
Speaker 1: sort of perfectly clear explanation of how this industry is

700
00:48:58,400 --> 00:49:02,200
Speaker 1: just so brutal. Professor she thank you so much for

701
00:49:02,239 --> 00:49:05,839
Speaker 1: coming on. Well, thank you for having you. Thank you.

702
00:49:20,040 --> 00:49:23,520
Speaker 1: I really like that conversation. You know, obviously, that last

703
00:49:23,600 --> 00:49:27,600
Speaker 1: answer that he gave about yield math, to me, it's

704
00:49:27,640 --> 00:49:30,600
Speaker 1: sort of like what I really liked about his perspective overall.

705
00:49:30,600 --> 00:49:33,399
Speaker 1: I mean, obviously that was super interesting. Is this sort

706
00:49:33,400 --> 00:49:38,440
Speaker 1: of way business and technological demands sort of intersect, And

707
00:49:38,480 --> 00:49:40,839
Speaker 1: I feel like you just had like super clear and

708
00:49:40,880 --> 00:49:43,759
Speaker 1: also like concrete examples of how that really works. Because

709
00:49:43,800 --> 00:49:46,120
Speaker 1: we could talk to sort of like vague ideas about oh,

710
00:49:46,120 --> 00:49:49,200
Speaker 1: it's important to have technological know how or it's important

711
00:49:49,239 --> 00:49:53,080
Speaker 1: to uh, you know, be able to trade with advanced partners,

712
00:49:53,120 --> 00:49:56,719
Speaker 1: but professor she really had a very concrete way of

713
00:49:56,800 --> 00:50:01,640
Speaker 1: describing what's actually going on. Yeah, was a good example. Um.

714
00:50:01,680 --> 00:50:04,160
Speaker 1: The other thing that really comes through in those discussions,

715
00:50:04,239 --> 00:50:07,200
Speaker 1: I mean it's coming through in every episode that we're

716
00:50:07,200 --> 00:50:11,400
Speaker 1: doing on semiconductors, but it's sort of the uniqueness of

717
00:50:11,440 --> 00:50:14,360
Speaker 1: the industry, right, It's sort of like an exceptional industry

718
00:50:14,440 --> 00:50:18,320
Speaker 1: in many ways. And again that yields problem that professor

719
00:50:18,360 --> 00:50:22,040
Speaker 1: she just described, as well as the startup costs for

720
00:50:22,160 --> 00:50:24,719
Speaker 1: creating a foundry, the idea that you used to be

721
00:50:24,760 --> 00:50:27,480
Speaker 1: able to do it for a few million dollars like

722
00:50:27,800 --> 00:50:30,920
Speaker 1: back in the seventies or eighties maybe, and now it

723
00:50:31,000 --> 00:50:35,399
Speaker 1: costs upwards of twenty billion. Uh. That kind of goes

724
00:50:35,440 --> 00:50:39,600
Speaker 1: to show just how complex and advance the technology has become.

725
00:50:39,640 --> 00:50:42,920
Speaker 1: But it also like as it advances, of course, it

726
00:50:42,960 --> 00:50:45,480
Speaker 1: becomes harder and harder to get in that game. And

727
00:50:45,520 --> 00:50:48,680
Speaker 1: that's where the question of public support comes in. Yeah,

728
00:50:48,760 --> 00:50:53,160
Speaker 1: because it's interesting, Like so obviously, um, there are companies

729
00:50:53,200 --> 00:50:55,560
Speaker 1: in the US, and you mentioned some of them, you know,

730
00:50:55,600 --> 00:50:57,960
Speaker 1: like a M D or in video they've just done

731
00:50:57,960 --> 00:51:01,080
Speaker 1: like phenomenally well that don't have their own fabs, So

732
00:51:01,640 --> 00:51:04,640
Speaker 1: you don't have to have that integration to work, and

733
00:51:04,880 --> 00:51:09,000
Speaker 1: from a pure shareholder perspective, those companies have like shout

734
00:51:09,040 --> 00:51:11,759
Speaker 1: the lights out, which then raises this question like for

735
00:51:11,800 --> 00:51:14,400
Speaker 1: the government, it's like, okay, well, maybe you could just

736
00:51:14,520 --> 00:51:18,400
Speaker 1: have this sort of total separation and some people in

737
00:51:18,560 --> 00:51:23,000
Speaker 1: a design shop somewhere in the US or California, Austin, Texas,

738
00:51:23,080 --> 00:51:26,440
Speaker 1: and then you ship the design to somewhere in Taiwan

739
00:51:26,719 --> 00:51:29,080
Speaker 1: and then they produce it. And maybe that model is

740
00:51:29,120 --> 00:51:32,120
Speaker 1: really great for shareholders, but it does raise the question

741
00:51:32,160 --> 00:51:36,360
Speaker 1: of a how does that serve sort of strategic public needs?

742
00:51:36,800 --> 00:51:40,000
Speaker 1: But then be is that sustainable the sort of the

743
00:51:40,080 --> 00:51:44,239
Speaker 1: long term separation of designing and building, because big part

744
00:51:44,280 --> 00:51:46,839
Speaker 1: of his whole core thesis is that the two really

745
00:51:46,840 --> 00:51:50,520
Speaker 1: do go together, right, It's the short term goals of

746
00:51:50,520 --> 00:51:54,560
Speaker 1: of shareholders who are interested in returns versus the long

747
00:51:54,680 --> 00:51:59,480
Speaker 1: term value of having a strategic industry that's competitive globally.

748
00:51:59,520 --> 00:52:03,239
Speaker 1: I guess yeah, another big macro theme to add to

749
00:52:03,239 --> 00:52:06,640
Speaker 1: this conversation. You know another thing that um and again

750
00:52:06,680 --> 00:52:09,719
Speaker 1: it's it's sort of thinking about how this fits in

751
00:52:09,880 --> 00:52:13,360
Speaker 1: our other themes. His whole um the whole discussion of

752
00:52:13,440 --> 00:52:17,200
Speaker 1: like demand side policy for tech. I thought it's super

753
00:52:17,239 --> 00:52:20,120
Speaker 1: interesting because you know, I think the clear you're just like,

754
00:52:20,120 --> 00:52:22,000
Speaker 1: all right, what do we need to do to create

755
00:52:22,120 --> 00:52:26,600
Speaker 1: a more robust manufacturing sector here for high tech? And

756
00:52:26,600 --> 00:52:29,520
Speaker 1: I think a lot of people would say, well, invest

757
00:52:29,600 --> 00:52:32,520
Speaker 1: in R and D, create tax credits for building foundries

758
00:52:32,560 --> 00:52:34,920
Speaker 1: here and so forth. But this idea that there is

759
00:52:34,960 --> 00:52:37,799
Speaker 1: also an important demand side component, and of course go

760
00:52:37,840 --> 00:52:40,719
Speaker 1: back to the history of Silicon Valley. The idea that

761
00:52:41,360 --> 00:52:44,880
Speaker 1: the Department of Defense and NASA as being NASA is

762
00:52:44,920 --> 00:52:47,680
Speaker 1: being customers of last resort or first resort for a

763
00:52:47,719 --> 00:52:49,759
Speaker 1: lot of these things was a big part of it.

764
00:52:49,760 --> 00:52:52,680
Speaker 1: It's also interesting, you know, bringing into the vaccine discussion,

765
00:52:53,160 --> 00:52:57,600
Speaker 1: a lot of what has allowed the fast pace of

766
00:52:57,640 --> 00:53:01,279
Speaker 1: vaccine development has been this sort of guaranteed buyer at

767
00:53:01,280 --> 00:53:03,600
Speaker 1: the end that the government pre committed to buy all

768
00:53:03,640 --> 00:53:07,279
Speaker 1: these shots from visor Maderia Johnson and Johnson if they

769
00:53:07,280 --> 00:53:10,440
Speaker 1: were successful, and the ability of the government on the

770
00:53:10,480 --> 00:53:15,080
Speaker 1: government alone to de risk the process by promising to

771
00:53:15,200 --> 00:53:18,799
Speaker 1: buy the end output. That's uh. I thought a super

772
00:53:18,840 --> 00:53:22,600
Speaker 1: interesting point, right, And I mean we've talked about the

773
00:53:22,719 --> 00:53:25,439
Speaker 1: history of Silicon Valley before, and I think there's still

774
00:53:25,440 --> 00:53:29,280
Speaker 1: a tendency to think about tech is this very entrepreneurial,

775
00:53:30,120 --> 00:53:33,880
Speaker 1: uh industry that you know, people started in their garage

776
00:53:34,000 --> 00:53:36,239
Speaker 1: and they're smart, and they sort of get lucky and

777
00:53:36,239 --> 00:53:39,080
Speaker 1: they're able to do it. And one of the things

778
00:53:39,120 --> 00:53:41,680
Speaker 1: people don't really realize is the degree to which having

779
00:53:41,719 --> 00:53:44,839
Speaker 1: that government backstop has actually helped the tech industry. I

780
00:53:44,880 --> 00:53:48,680
Speaker 1: do wonder, based off the experience of the vaccine development,

781
00:53:49,040 --> 00:53:52,160
Speaker 1: whether that starts to change. And again, as we've been

782
00:53:52,200 --> 00:53:56,120
Speaker 1: discussing for most of whether there is going to be

783
00:53:56,640 --> 00:54:00,879
Speaker 1: a larger recognized role for the government in a sparking book,

784
00:54:00,920 --> 00:54:05,320
Speaker 1: Demand and Supply. Yeah, one of our previous guests, actually

785
00:54:05,320 --> 00:54:07,040
Speaker 1: I think we've had them on a couple of times,

786
00:54:07,040 --> 00:54:08,920
Speaker 1: although maybe you weren't there for one of them, but uh,

787
00:54:09,400 --> 00:54:14,240
Speaker 1: Bill Janeway, the venture capitalist. He talks a lot about

788
00:54:14,280 --> 00:54:18,200
Speaker 1: exactly this point and the idea of like government underwriting

789
00:54:18,200 --> 00:54:21,560
Speaker 1: guaranteed government demand and how important that is for the

790
00:54:21,640 --> 00:54:25,640
Speaker 1: commercialization of innovation, and also like you know, with an

791
00:54:25,680 --> 00:54:29,160
Speaker 1: eye towards like puncturing a little bit of Silicon Valley's

792
00:54:29,200 --> 00:54:32,680
Speaker 1: own myth making that if the government just left tech alone,

793
00:54:32,719 --> 00:54:35,520
Speaker 1: that they would do all these great things and so forth,

794
00:54:35,560 --> 00:54:38,040
Speaker 1: and that how much of that all all the industries

795
00:54:38,080 --> 00:54:42,160
Speaker 1: that really take off, I'd say green technology or clean

796
00:54:42,200 --> 00:54:46,880
Speaker 1: technology is another area in which the government has had

797
00:54:46,880 --> 00:54:49,520
Speaker 1: a big role or could play a big role, both

798
00:54:49,520 --> 00:54:52,000
Speaker 1: in terms of the supply side investing in R and D,

799
00:54:52,480 --> 00:54:55,719
Speaker 1: but also the demand side in being a customer for

800
00:54:56,520 --> 00:54:59,080
Speaker 1: ventures that take a lot of risk, take a lot

801
00:54:59,120 --> 00:55:01,920
Speaker 1: of money, and therefore high rid Yeah. Well, it seems

802
00:55:01,960 --> 00:55:04,000
Speaker 1: like we might have a new example in the form

803
00:55:04,040 --> 00:55:07,080
Speaker 1: of a u S Semiconductor is not not too far

804
00:55:07,200 --> 00:55:11,440
Speaker 1: from now. Maybe to add to that pod, we'll see, Okay,

805
00:55:11,480 --> 00:55:13,680
Speaker 1: shall we leave it there? Yeah, let's leave it there.

806
00:55:14,120 --> 00:55:17,000
Speaker 1: This has been another episode of the All Thoughts Podcast.

807
00:55:17,120 --> 00:55:19,879
Speaker 1: I'm Tracy Alloway. You can follow me on Twitter at

808
00:55:19,920 --> 00:55:22,759
Speaker 1: Tracy Alloway and I'm Joe Why Isn't Though. You can

809
00:55:22,800 --> 00:55:26,120
Speaker 1: follow me on Twitter at the Stalwart. Follow our guest

810
00:55:26,280 --> 00:55:30,680
Speaker 1: professor Willie she He's on Twitter at Willie she Underscore

811
00:55:31,360 --> 00:55:36,400
Speaker 1: at HBS. Follow our producer Laura Carlson. She's at Laura M. Carlson.

812
00:55:36,840 --> 00:55:40,680
Speaker 1: Follow the Bloomberg head of podcast, Francesco Levi at Francisco Today,

813
00:55:40,840 --> 00:55:43,960
Speaker 1: and check out all of our podcasts under the handle

814
00:55:44,280 --> 00:56:03,160
Speaker 1: at podcast. Thanks for listening to