1
00:00:02,520 --> 00:00:13,240
Speaker 1: Bloomberg Audio Studios, Podcasts, Radio News.

2
00:00:17,880 --> 00:00:21,040
Speaker 2: Hello and welcome to another episode of the Odd Lots podcast.

3
00:00:21,079 --> 00:00:22,480
Speaker 2: I'm Joe Wisenthal and.

4
00:00:22,440 --> 00:00:23,440
Speaker 3: I'm Tracy Alloway.

5
00:00:23,800 --> 00:00:27,120
Speaker 2: So, Tracy, a thing that keeps coming up is that

6
00:00:27,360 --> 00:00:29,200
Speaker 2: all these AI data centers are going to use a

7
00:00:29,200 --> 00:00:31,600
Speaker 2: lot of electricity. I keep hearing that.

8
00:00:32,040 --> 00:00:35,680
Speaker 3: Yes. Also, I just realized every time you use chat

9
00:00:35,720 --> 00:00:39,960
Speaker 3: GPT to write like a satirical song, you're diverting energy

10
00:00:40,000 --> 00:00:42,680
Speaker 3: away from someone turning on a light bulb or something

11
00:00:42,720 --> 00:00:45,280
Speaker 3: like that, potentially something like zero sum game.

12
00:00:45,400 --> 00:00:46,000
Speaker 4: Yeah, so be.

13
00:00:46,040 --> 00:00:50,560
Speaker 2: Careful about your your random chat GPT queries. Although I

14
00:00:50,600 --> 00:00:53,360
Speaker 2: think the training is like the more Yeah, like I think,

15
00:00:53,560 --> 00:00:56,240
Speaker 2: but you know, like so maybe your song there, maybe

16
00:00:56,280 --> 00:00:57,880
Speaker 2: it's okay. I don't think it's that bad.

17
00:00:58,280 --> 00:01:00,760
Speaker 3: In the grand scheme of things, probably not. But there

18
00:01:00,840 --> 00:01:05,200
Speaker 3: is this overarching conversation about AI's energy use. So what

19
00:01:05,319 --> 00:01:07,960
Speaker 3: exactly it is? This is a big question I have.

20
00:01:08,120 --> 00:01:12,280
Speaker 3: How do you disaggregate AI servers from your run of

21
00:01:12,319 --> 00:01:15,720
Speaker 3: the mill software servers, How much it's going to consume,

22
00:01:16,240 --> 00:01:20,120
Speaker 3: how that capacity is going to get allocated and built out?

23
00:01:20,520 --> 00:01:23,199
Speaker 3: And I think there is this sense that we could

24
00:01:23,319 --> 00:01:28,240
Speaker 3: end up going into very different, very extreme directions here.

25
00:01:28,319 --> 00:01:32,679
Speaker 3: So you could have this great situation that because AI

26
00:01:33,040 --> 00:01:36,800
Speaker 3: is a desirable activity, because it's profitable in many respects,

27
00:01:36,800 --> 00:01:41,280
Speaker 3: that big tech ends up accelerating the energy capacity build out,

28
00:01:41,400 --> 00:01:46,000
Speaker 3: maybe they even start building more green technology capabilities in

29
00:01:46,040 --> 00:01:49,040
Speaker 3: an ideal world. But then you have the polar opposite

30
00:01:49,080 --> 00:01:52,680
Speaker 3: scenario where you need all this power to develop this

31
00:01:52,800 --> 00:01:56,120
Speaker 3: technology there isn't enough and it's sort of a race

32
00:01:56,160 --> 00:01:58,559
Speaker 3: to the bottom where you have tech companies just trying

33
00:01:58,560 --> 00:02:01,640
Speaker 3: to get energy wherever they can, maybe they even start

34
00:02:01,720 --> 00:02:04,960
Speaker 3: using coal and things like that. So it feels like

35
00:02:05,880 --> 00:02:08,519
Speaker 3: there's two very different paths that we could be going

36
00:02:08,560 --> 00:02:08,959
Speaker 3: down here.

37
00:02:09,040 --> 00:02:10,639
Speaker 2: Yeah, there's a lot here for us.

38
00:02:10,880 --> 00:02:11,040
Speaker 4: You know.

39
00:02:11,040 --> 00:02:13,480
Speaker 2: I remember Jigger shaw at Or when we interviewed him

40
00:02:13,480 --> 00:02:15,839
Speaker 2: at the Texas Tribune conference like over a year ago,

41
00:02:16,560 --> 00:02:19,000
Speaker 2: he brought this up. It's getting more and more attention.

42
00:02:19,120 --> 00:02:21,840
Speaker 2: It keeps coming up. On the side of episode. Steve

43
00:02:21,880 --> 00:02:25,240
Speaker 2: Eisman obviously recently talked about it. But I feel like

44
00:02:25,840 --> 00:02:28,120
Speaker 2: it's time to like sort of make it a central

45
00:02:28,160 --> 00:02:31,240
Speaker 2: part of the conversation and actually learn about numbers and

46
00:02:31,280 --> 00:02:33,480
Speaker 2: where this power is generated from and like, yeah, like

47
00:02:33,520 --> 00:02:35,600
Speaker 2: how much are we really talking about here. We know

48
00:02:35,680 --> 00:02:37,560
Speaker 2: the you know, the tech company is highly aware of it.

49
00:02:37,560 --> 00:02:40,080
Speaker 2: There was a headline recently about Microsoft maybe wanting to

50
00:02:40,120 --> 00:02:43,960
Speaker 2: do something with on site nuclear development. They also, you know,

51
00:02:44,080 --> 00:02:45,320
Speaker 2: they did do something.

52
00:02:45,480 --> 00:02:48,400
Speaker 3: So I think they headline, well, didn't they buy a

53
00:02:48,480 --> 00:02:52,480
Speaker 3: data center maybe next to a nuclear power plant? Susqahanna thing?

54
00:02:52,520 --> 00:02:53,120
Speaker 3: I thought they did.

55
00:02:53,320 --> 00:02:55,000
Speaker 2: Yeah, I think you're right about that. But then the

56
00:02:55,000 --> 00:02:58,639
Speaker 2: other element too, is you mentioned that one solution here

57
00:02:58,760 --> 00:03:02,000
Speaker 2: is just fossil fuels and dirty energy, except that all

58
00:03:02,000 --> 00:03:05,280
Speaker 2: these tech companies are very like progressive minded and they

59
00:03:05,320 --> 00:03:07,600
Speaker 2: all have these net zero commitments by you know, we're

60
00:03:07,600 --> 00:03:11,040
Speaker 2: going to get it all from you know, windmills or sorry,

61
00:03:11,160 --> 00:03:13,640
Speaker 2: wind turbines and solar and batteries and stores.

62
00:03:13,720 --> 00:03:17,200
Speaker 3: Windmills next to AI servers would be an interesting one.

63
00:03:17,280 --> 00:03:19,480
Speaker 2: But like you know, at some point, the rubber's got

64
00:03:19,480 --> 00:03:21,840
Speaker 2: to hit the road with like how realistic are there

65
00:03:21,919 --> 00:03:24,680
Speaker 2: net zero commitments or how can they achieve them if

66
00:03:24,680 --> 00:03:27,960
Speaker 2: they're engaging in this investment activity that is highly energy intensive.

67
00:03:28,040 --> 00:03:28,079
Speaker 4: No.

68
00:03:28,200 --> 00:03:30,960
Speaker 3: Absolutely, And you are seeing a lot of this discussion

69
00:03:31,160 --> 00:03:34,920
Speaker 3: reflected in the conversation around AI investment at this point.

70
00:03:34,960 --> 00:03:37,040
Speaker 3: So I think a lot of people feel like they

71
00:03:37,080 --> 00:03:40,800
Speaker 3: missed that first wave of chips around in video, so

72
00:03:40,840 --> 00:03:44,600
Speaker 3: everyone's looking for the sort of second order investment play

73
00:03:44,680 --> 00:03:47,080
Speaker 3: and a lot of people now are talking about energy

74
00:03:47,200 --> 00:03:50,280
Speaker 3: or cooling and HVAC, so we need to talk about it.

75
00:03:50,560 --> 00:03:50,760
Speaker 4: Well.

76
00:03:50,800 --> 00:03:53,520
Speaker 2: I am really excited because I do believe we have

77
00:03:54,000 --> 00:03:58,120
Speaker 2: the perfect guest for this topic, someone who spent twelve

78
00:03:58,240 --> 00:04:02,040
Speaker 2: years at Microsoft, way before it was hot to talk

79
00:04:02,080 --> 00:04:06,560
Speaker 2: about how tech companies needed all this electricity and energy.

80
00:04:06,800 --> 00:04:10,320
Speaker 2: He was the first energy hire at Microsoft, and he

81
00:04:10,400 --> 00:04:13,119
Speaker 2: recently left last year. He left to start his own

82
00:04:13,160 --> 00:04:15,800
Speaker 2: firm to work on this problem specifically. So we were

83
00:04:15,840 --> 00:04:17,880
Speaker 2: going to be speaking with Brian Janis. He is the

84
00:04:17,920 --> 00:04:22,120
Speaker 2: co founder and chief strategy officer at Cloverleaf Infrastructure, a

85
00:04:22,120 --> 00:04:26,120
Speaker 2: power development company that works closely with utilities on solving

86
00:04:26,200 --> 00:04:28,760
Speaker 2: this problem. So, Brian, thank you so much for coming

87
00:04:28,760 --> 00:04:29,840
Speaker 2: on out lots.

88
00:04:29,720 --> 00:04:31,400
Speaker 4: Thank you for having me. Really excited to be here.

89
00:04:31,760 --> 00:04:35,080
Speaker 2: So you got hired for Microsoft twelve years ago to

90
00:04:35,120 --> 00:04:38,520
Speaker 2: do energy, and at the time, I don't think anyone

91
00:04:38,720 --> 00:04:42,400
Speaker 2: was talking about energy as being like a particularly important

92
00:04:42,440 --> 00:04:47,719
Speaker 2: aspect of these software companies or these big tech companies strategies.

93
00:04:48,000 --> 00:04:50,080
Speaker 2: What was going on back then, or like, what did

94
00:04:50,120 --> 00:04:52,320
Speaker 2: they see when they felt like, hey, we need to

95
00:04:52,400 --> 00:04:53,640
Speaker 2: hire a VP of energy here.

96
00:04:54,040 --> 00:04:56,359
Speaker 4: Yeah. It was actually funny because I had spent my

97
00:04:56,440 --> 00:04:59,679
Speaker 4: career prior to that working with mainly large energy consumers

98
00:04:59,720 --> 00:05:01,400
Speaker 4: who were the big who you'd expect it to be,

99
00:05:01,480 --> 00:05:05,400
Speaker 4: the big industrial companies, And so when Microsoft came calling

100
00:05:05,480 --> 00:05:07,480
Speaker 4: and said, hey, we need to get a full time

101
00:05:07,920 --> 00:05:10,640
Speaker 4: energy person, I told them it sounded like a dead

102
00:05:10,720 --> 00:05:13,599
Speaker 4: end job to be the energy person at a tech company,

103
00:05:13,600 --> 00:05:16,920
Speaker 4: because why would they ever actually care about this issue.

104
00:05:17,200 --> 00:05:19,000
Speaker 4: And the person that was recruiting me said, Hey, I

105
00:05:19,040 --> 00:05:21,560
Speaker 4: think there is something to this, this whole cloud thing,

106
00:05:21,680 --> 00:05:23,479
Speaker 4: and I think energy is going to start to be

107
00:05:24,000 --> 00:05:27,680
Speaker 4: pretty central to what we're doing as a company. And

108
00:05:27,960 --> 00:05:31,640
Speaker 4: you'll fast forward a decade and I remember having a

109
00:05:31,640 --> 00:05:33,719
Speaker 4: conversation before I left the company. I was talking to

110
00:05:33,720 --> 00:05:36,160
Speaker 4: the head of corporate strategy and he said to me,

111
00:05:36,240 --> 00:05:39,279
Speaker 4: He's like, I don't think people quite realize the degree

112
00:05:39,279 --> 00:05:42,800
Speaker 4: to which Microsoft is really just an energy company. We

113
00:05:42,839 --> 00:05:46,320
Speaker 4: need power and we need silicon. We need chips. That's it,

114
00:05:46,400 --> 00:05:48,200
Speaker 4: that's the business. If we don't have one of those

115
00:05:48,240 --> 00:05:50,120
Speaker 4: two things, we're in a lot of trouble and so

116
00:05:50,200 --> 00:05:52,960
Speaker 4: it was it was remarkable to see the shift over

117
00:05:53,000 --> 00:05:56,840
Speaker 4: that decade plus of maybe one or two people at

118
00:05:56,839 --> 00:05:59,440
Speaker 4: the company starting to think that energy might be important

119
00:05:59,440 --> 00:06:03,800
Speaker 4: to us, when day to energy is actually absolutely central

120
00:06:04,279 --> 00:06:05,800
Speaker 4: to everything the business does.

121
00:06:06,600 --> 00:06:08,440
Speaker 3: So talk to us a little bit more about that

122
00:06:08,520 --> 00:06:12,799
Speaker 3: cultural shift, because Joe and I heard from someone else recently,

123
00:06:12,839 --> 00:06:15,159
Speaker 3: but they were saying that a bunch of the big

124
00:06:15,200 --> 00:06:18,960
Speaker 3: tech companies that you would be very familiar with had

125
00:06:19,000 --> 00:06:23,359
Speaker 3: representatives down in Houston for zero Week, so the annual

126
00:06:23,480 --> 00:06:26,039
Speaker 3: energy conference, which they were describing as kind of a

127
00:06:26,080 --> 00:06:30,599
Speaker 3: new development. But how familiar are tech companies nowadays with

128
00:06:30,880 --> 00:06:35,240
Speaker 3: energy usage or needs and how much expertise have they

129
00:06:35,279 --> 00:06:37,640
Speaker 3: actually built out in that capacity.

130
00:06:39,000 --> 00:06:41,680
Speaker 4: Yeah, it's been a tremendous shift. And I mean if

131
00:06:41,680 --> 00:06:43,800
Speaker 4: you would have gone to Seria Week even five years ago,

132
00:06:44,520 --> 00:06:47,120
Speaker 4: you would not have seen a whole lot of engagement

133
00:06:47,160 --> 00:06:52,040
Speaker 4: from the tech industry. But as their businesses have shifted

134
00:06:52,040 --> 00:06:56,240
Speaker 4: to the cloud, and as the business opportunities that sit

135
00:06:56,279 --> 00:06:58,599
Speaker 4: in front of them, particularly when it comes down to AI,

136
00:07:00,040 --> 00:07:03,760
Speaker 4: as those have arisen, there's been a recognition that power

137
00:07:03,800 --> 00:07:06,760
Speaker 4: really is central to what they're doing. And it was

138
00:07:06,800 --> 00:07:09,679
Speaker 4: a slow shift. If you go back to the advent

139
00:07:09,800 --> 00:07:14,200
Speaker 4: of the first cloud data centers. It really was about

140
00:07:14,400 --> 00:07:17,080
Speaker 4: being close to the network, and so that was the

141
00:07:17,200 --> 00:07:20,440
Speaker 4: driver of strategically, where do you put data centers, Well,

142
00:07:20,440 --> 00:07:22,320
Speaker 4: you put them where the biggest network cubs are. So

143
00:07:22,600 --> 00:07:24,840
Speaker 4: that's why we have lots of data centers in Northern Virginia.

144
00:07:24,960 --> 00:07:27,360
Speaker 4: That's why we have lots of data centers in Amsterdam.

145
00:07:27,640 --> 00:07:31,720
Speaker 4: Everyone was chasing network. Probably middle of last decade, there

146
00:07:31,760 --> 00:07:34,320
Speaker 4: was a shift and it started to go, actually, we

147
00:07:34,320 --> 00:07:38,040
Speaker 4: want to be close to eyeballs. So this started a

148
00:07:38,240 --> 00:07:41,440
Speaker 4: sort of a land grab of all the cloud data

149
00:07:41,440 --> 00:07:44,160
Speaker 4: centers starting to build lots of new data SISIs new

150
00:07:44,160 --> 00:07:46,480
Speaker 4: countries because they wanted to be close to where the

151
00:07:46,520 --> 00:07:50,960
Speaker 4: customers were. And so from about fall of twenty nineteen

152
00:07:51,200 --> 00:07:54,000
Speaker 4: through probably spring of twenty twenty two, I think Microsoft

153
00:07:54,040 --> 00:07:56,720
Speaker 4: added was adding close to a region a month in

154
00:07:56,800 --> 00:07:59,640
Speaker 4: terms of new data center regions they were establishing around

155
00:07:59,640 --> 00:08:04,320
Speaker 4: the world. And then in mid twenty twenty two that's

156
00:08:04,360 --> 00:08:08,560
Speaker 4: when the realization started to sink in that wait a minute,

157
00:08:08,640 --> 00:08:11,360
Speaker 4: this whole game is about power, because that's when we

158
00:08:11,360 --> 00:08:16,040
Speaker 4: were first starting to hear rumblings of what open aye

159
00:08:16,160 --> 00:08:18,880
Speaker 4: was working on and the scale of what chat gp

160
00:08:19,360 --> 00:08:21,600
Speaker 4: T three was going to be, which was sort of

161
00:08:21,640 --> 00:08:23,920
Speaker 4: the first big release where everyone was like, wait a minute,

162
00:08:23,960 --> 00:08:26,240
Speaker 4: this is this is kind of a big deal what

163
00:08:26,280 --> 00:08:29,120
Speaker 4: AI is doing and how fast it's moving. And then

164
00:08:29,240 --> 00:08:32,720
Speaker 4: when we had that release of chat gpt T three

165
00:08:32,840 --> 00:08:35,600
Speaker 4: in the fall of twenty two, and then shortly thereafter,

166
00:08:37,600 --> 00:08:39,960
Speaker 4: three point five was released, and there was a massive

167
00:08:40,520 --> 00:08:44,040
Speaker 4: increase in capability in that release if you recall, you know,

168
00:08:44,080 --> 00:08:46,360
Speaker 4: in terms of what it could do on you know,

169
00:08:46,480 --> 00:08:51,000
Speaker 4: getting scores on various tests and things. And it was

170
00:08:51,040 --> 00:08:55,720
Speaker 4: that moment that I realized, this technology is moving way

171
00:08:55,800 --> 00:09:00,559
Speaker 4: faster than the utility industry is moving. If we can

172
00:09:00,640 --> 00:09:04,880
Speaker 4: make this much improvement in this technology in a six

173
00:09:04,920 --> 00:09:09,320
Speaker 4: month time horizon, we're in a lot of trouble because

174
00:09:09,400 --> 00:09:11,600
Speaker 4: the power industry does not move that fast.

175
00:09:12,160 --> 00:09:15,320
Speaker 2: So I'm really fascinated by this idea that the release,

176
00:09:15,520 --> 00:09:17,840
Speaker 2: like you know, like you were at Microsoft and so

177
00:09:17,960 --> 00:09:20,200
Speaker 2: you had a front row seat to what open ai

178
00:09:20,400 --> 00:09:23,800
Speaker 2: was doing with GPT one and GPD two, and there

179
00:09:23,800 --> 00:09:25,839
Speaker 2: were a lot of people aware of this, and I'm

180
00:09:25,880 --> 00:09:28,760
Speaker 2: sort of fascinated by this idea that it was that

181
00:09:28,880 --> 00:09:32,600
Speaker 2: commercialization or sort of making easy in public CHATJEP. When

182
00:09:32,640 --> 00:09:35,800
Speaker 2: it became chat GPT, they're like, oh, this is serious,

183
00:09:35,840 --> 00:09:38,400
Speaker 2: and then we sell everyone rushing to buy in video

184
00:09:38,480 --> 00:09:41,559
Speaker 2: chips and all these vcs pivoting to AI, et cetera.

185
00:09:41,720 --> 00:09:45,320
Speaker 2: So talk to us about like the math there. It

186
00:09:45,400 --> 00:09:48,360
Speaker 2: feels like there has been this sort of level shift

187
00:09:48,520 --> 00:09:52,760
Speaker 2: up in sort of expectations of data center demand growth

188
00:09:53,160 --> 00:09:56,439
Speaker 2: basically as a function of all of the excitement for AI.

189
00:09:57,960 --> 00:09:59,960
Speaker 4: Yeah. I think then you're right. I mean, it's not

190
00:10:00,080 --> 00:10:03,520
Speaker 4: like we didn't know that Microsoft had a partnership with

191
00:10:03,559 --> 00:10:06,120
Speaker 4: open Ai and that that you know, AI was going

192
00:10:06,200 --> 00:10:10,320
Speaker 4: to consume energy. I think everyone though, was a bit

193
00:10:10,400 --> 00:10:18,240
Speaker 4: surprised at just how quickly what chat GPT could do

194
00:10:18,280 --> 00:10:21,200
Speaker 4: just captured the collective consciousness. Yeah, and I think, I

195
00:10:21,200 --> 00:10:24,000
Speaker 4: mean you probably remember when when that was released. I

196
00:10:24,040 --> 00:10:27,200
Speaker 4: mean it it really sort of surprised everyone, and it

197
00:10:27,280 --> 00:10:30,360
Speaker 4: became this thing where suddenly, even though we sort of

198
00:10:30,440 --> 00:10:32,520
Speaker 4: knew what we were working on, it wasn't until you

199
00:10:32,600 --> 00:10:35,320
Speaker 4: sort of put it out into the world that you

200
00:10:35,400 --> 00:10:39,360
Speaker 4: realize maybe what you've created. And I mean that's where

201
00:10:39,520 --> 00:10:42,000
Speaker 4: we realized we are we are running up this curve

202
00:10:42,160 --> 00:10:46,040
Speaker 4: of capability a lot faster than we thought, uh, and

203
00:10:46,320 --> 00:10:48,720
Speaker 4: the a number of applications that are getting built on this,

204
00:10:48,800 --> 00:10:51,000
Speaker 4: in the number of different ways that it's being used,

205
00:10:51,000 --> 00:10:54,199
Speaker 4: and how it's just become sort of common parlance. I mean,

206
00:10:54,200 --> 00:10:56,880
Speaker 4: everyone knows what chat GPG three is, and no one

207
00:10:56,920 --> 00:10:59,720
Speaker 4: what it was the month before that, right, So there

208
00:10:59,720 --> 00:11:01,480
Speaker 4: there as a bit of a I think of a

209
00:11:01,960 --> 00:11:04,960
Speaker 4: surprise in terms of just how quickly it was going

210
00:11:05,040 --> 00:11:09,040
Speaker 4: to capture, you know, the collective consciousness and then you know,

211
00:11:09,360 --> 00:11:12,240
Speaker 4: obviously lead to everything that's sort of being created as

212
00:11:12,240 --> 00:11:13,920
Speaker 4: a result. And so we just we just moved up

213
00:11:13,920 --> 00:11:16,440
Speaker 4: that curve so quickly, and I think that's that's where

214
00:11:16,480 --> 00:11:20,360
Speaker 4: the industry maybe got you know, certainly the utilities were

215
00:11:20,400 --> 00:11:23,560
Speaker 4: behind because as you may have seen there, a lot

216
00:11:23,600 --> 00:11:26,559
Speaker 4: of them are starting to restate their load growth expectations

217
00:11:26,920 --> 00:11:30,480
Speaker 4: and that was something that was not happening right before that.

218
00:11:30,520 --> 00:11:33,080
Speaker 4: And so we've had massive changes just in the last

219
00:11:33,080 --> 00:11:35,560
Speaker 4: two years of how utilities are one of the number

220
00:11:35,559 --> 00:11:39,240
Speaker 4: cap So you know, if you take a look at

221
00:11:39,320 --> 00:11:43,959
Speaker 4: a utility like Dominion in Virginia, so that's the largest

222
00:11:44,000 --> 00:11:46,120
Speaker 4: concentration of data centers in the United States, so they're

223
00:11:46,160 --> 00:11:49,600
Speaker 4: pretty good representative of what's happening. If you go back

224
00:11:49,600 --> 00:11:53,120
Speaker 4: to twenty twenty one, they were forecasting load growth over

225
00:11:53,160 --> 00:11:56,080
Speaker 4: a period of fifteen years of just a few percent.

226
00:11:56,240 --> 00:11:59,120
Speaker 4: I mean it was about it was single digit growth

227
00:11:59,160 --> 00:12:03,120
Speaker 4: over that entire period, so not yearly growth, but over

228
00:12:03,200 --> 00:12:07,000
Speaker 4: fifteen years, single digit growth. By twenty twenty three, they

229
00:12:07,000 --> 00:12:10,880
Speaker 4: were forecasting to grow two x over fifteen years. Wow.

230
00:12:10,960 --> 00:12:14,000
Speaker 4: Now keep in mind this is electric utility. They do

231
00:12:14,080 --> 00:12:17,959
Speaker 4: ten year planning cycles. So because they have very long

232
00:12:18,040 --> 00:12:21,680
Speaker 4: lead times for equipment, for getting rights of way for

233
00:12:21,800 --> 00:12:27,920
Speaker 4: transmission lines, they aren't companies that easily respond to a

234
00:12:28,080 --> 00:12:32,120
Speaker 4: two x order of magnitude, you know, growth change over

235
00:12:32,120 --> 00:12:33,559
Speaker 4: a period of fifteen years. I mean that is a

236
00:12:34,200 --> 00:12:37,679
Speaker 4: that is a massive change for electric utility, particularly given

237
00:12:37,720 --> 00:12:41,200
Speaker 4: the fact that the growth rate over the last fifteen

238
00:12:41,240 --> 00:12:44,520
Speaker 4: to twenty years has been close to zero, so there's

239
00:12:44,679 --> 00:12:47,720
Speaker 4: been relatively no load growth in fifteen to twenty years.

240
00:12:48,200 --> 00:12:51,400
Speaker 4: Now suddenly you have utilities having to pivot to doubling

241
00:12:51,440 --> 00:12:54,280
Speaker 4: the size of their system in that same horizon.

242
00:13:10,600 --> 00:13:13,240
Speaker 3: I want to ask a very basic question, but I

243
00:13:13,280 --> 00:13:16,520
Speaker 3: think it will probably inform the rest of this conversation.

244
00:13:16,960 --> 00:13:20,760
Speaker 3: But when we say that AI consumes a lot of energy.

245
00:13:21,559 --> 00:13:25,120
Speaker 3: Where is that consumption actually coming from? And Joe touched

246
00:13:25,160 --> 00:13:27,040
Speaker 3: on this in the intro, but is it you know,

247
00:13:27,120 --> 00:13:30,839
Speaker 3: the sheer scale of users on these platforms, is it

248
00:13:31,000 --> 00:13:34,440
Speaker 3: I imagine the training that you need in order to

249
00:13:34,480 --> 00:13:39,120
Speaker 3: develop these models, and then does that energy usage differ

250
00:13:39,280 --> 00:13:42,960
Speaker 3: in any way from more traditional technologies.

251
00:13:43,320 --> 00:13:47,920
Speaker 4: Yeah. So whenever I think about the consumption of electricity

252
00:13:47,960 --> 00:13:50,840
Speaker 4: for AI or really any other application, I think you

253
00:13:50,840 --> 00:13:53,800
Speaker 4: have to start at sort of the core of what

254
00:13:53,840 --> 00:13:56,559
Speaker 4: we're talking about, which is really the human capacity for data.

255
00:13:57,120 --> 00:14:02,160
Speaker 4: Like whether it's AI or cloud. Humans have a massive

256
00:14:02,280 --> 00:14:06,200
Speaker 4: capacity to consume data. And if you think about where

257
00:14:06,240 --> 00:14:08,920
Speaker 4: we are in this curve, I mean we're on some

258
00:14:09,080 --> 00:14:13,920
Speaker 4: form of S curve right of human data consumption, which

259
00:14:14,040 --> 00:14:20,640
Speaker 4: then directly ties to data centers, devices, energy consumption ultimately,

260
00:14:20,640 --> 00:14:22,840
Speaker 4: because what we're doing is we're turning energy into data.

261
00:14:22,880 --> 00:14:26,200
Speaker 4: We take electrons, we convert them to light, we move

262
00:14:26,240 --> 00:14:29,360
Speaker 4: them around to your TV screens and your phones and

263
00:14:29,400 --> 00:14:34,800
Speaker 4: your laptops, etc. So that's the uber trend that we're

264
00:14:34,880 --> 00:14:37,520
Speaker 4: riding up right now. And so we're climbing this S curve.

265
00:14:38,000 --> 00:14:40,240
Speaker 4: I don't know that anyone has a good sense of

266
00:14:41,320 --> 00:14:45,000
Speaker 4: how steep or how long this curve will go. If

267
00:14:45,040 --> 00:14:47,480
Speaker 4: you go back to look at something like electricity, it

268
00:14:47,520 --> 00:14:51,200
Speaker 4: was roughly about one hundred year. S curve started in

269
00:14:51,240 --> 00:14:54,480
Speaker 4: the beginning of last century, and it really started to flatline,

270
00:14:54,720 --> 00:14:57,560
Speaker 4: as I mentioned before, towards the beginning of this century.

271
00:14:57,880 --> 00:15:00,440
Speaker 4: Now we have this new trajectory that we're ringing, this

272
00:15:00,480 --> 00:15:02,320
Speaker 4: new S curve that we're entering, that's going to sort

273
00:15:02,360 --> 00:15:04,760
Speaker 4: of change that narrative. But you know that S curve

274
00:15:04,760 --> 00:15:07,200
Speaker 4: for electricity took about one hundred years. No one knows

275
00:15:07,240 --> 00:15:10,880
Speaker 4: where we are on that data curve today. So when

276
00:15:10,880 --> 00:15:14,760
Speaker 4: you inject something like AI, you create a whole new

277
00:15:14,960 --> 00:15:18,320
Speaker 4: opportunity for humans to consume data, to do new things

278
00:15:18,360 --> 00:15:21,000
Speaker 4: with data that we couldn't do before, and so you

279
00:15:21,160 --> 00:15:24,040
Speaker 4: accelerate us up this curve. Right, So, we were sitting

280
00:15:24,400 --> 00:15:27,280
Speaker 4: somewhere along this curve. AI comes along, and now we're

281
00:15:27,320 --> 00:15:29,800
Speaker 4: just moving up even further, And of course that means

282
00:15:30,240 --> 00:15:35,160
Speaker 4: more energy consumption because the energy intensity of running an

283
00:15:35,160 --> 00:15:39,000
Speaker 4: AI query versus a traditional search is much higher. Now

284
00:15:39,280 --> 00:15:42,480
Speaker 4: what you can do with AI obviously is also much

285
00:15:42,480 --> 00:15:45,320
Speaker 4: greater than what you can do with a traditional search,

286
00:15:45,360 --> 00:15:50,600
Speaker 4: So there is a positive return on that invested energy.

287
00:15:51,080 --> 00:15:54,160
Speaker 4: So that's you know when when oftentimes when this conversation

288
00:15:54,280 --> 00:15:58,000
Speaker 4: comes up, there's a lot of consternation and panic over well,

289
00:15:58,000 --> 00:15:59,840
Speaker 4: what are we going to do? You know, we're going

290
00:15:59,920 --> 00:16:02,400
Speaker 4: to we're going to run out of energy. The nice

291
00:16:02,400 --> 00:16:04,960
Speaker 4: thing about electricity is we can always make more. We're

292
00:16:05,160 --> 00:16:07,360
Speaker 4: we're never going to run out of electricity. Not to

293
00:16:07,400 --> 00:16:10,000
Speaker 4: say that there's not times where the grid is under

294
00:16:10,000 --> 00:16:12,760
Speaker 4: constraint and there you know, you have risks of brownouts

295
00:16:12,760 --> 00:16:16,480
Speaker 4: and blackouts. That's that's the reality. But we can invest

296
00:16:16,600 --> 00:16:19,200
Speaker 4: we can invest more in transmission lines, we can invest

297
00:16:19,240 --> 00:16:23,239
Speaker 4: more in power plants, and we can create enough electricity

298
00:16:23,680 --> 00:16:26,120
Speaker 4: to to match that demand.

299
00:16:26,560 --> 00:16:29,400
Speaker 2: Just to sort of clarify a point in adding on

300
00:16:29,440 --> 00:16:33,760
Speaker 2: to Tracy's question, you mentioned that doing an AI query

301
00:16:34,160 --> 00:16:36,400
Speaker 2: is more energy intensive than say, if I had just

302
00:16:36,480 --> 00:16:38,760
Speaker 2: done a Google search, or if I had done a

303
00:16:38,800 --> 00:16:41,400
Speaker 2: Being search or something like that. Like, what is it

304
00:16:41,560 --> 00:16:47,640
Speaker 2: about the process of delivering these capabilities that makes it

305
00:16:47,720 --> 00:16:53,000
Speaker 2: more computationally intensive or energy intensive then the previous generation

306
00:16:53,360 --> 00:16:56,360
Speaker 2: of data usage or data querying online.

307
00:16:57,560 --> 00:17:00,840
Speaker 4: There's two aspects to it, and we sort of alluded

308
00:17:00,880 --> 00:17:04,239
Speaker 4: to it earlier. But the first is the training. So

309
00:17:04,240 --> 00:17:06,680
Speaker 4: The first is the building of the large language model

310
00:17:07,000 --> 00:17:12,280
Speaker 4: that itself is very energy intensive. These are extraordinarily large

311
00:17:12,400 --> 00:17:18,520
Speaker 4: machines collections of machines that use very dense chips to

312
00:17:18,600 --> 00:17:22,440
Speaker 4: create these language models that ultimately then get queried when

313
00:17:22,440 --> 00:17:26,040
Speaker 4: you do an inference. So then you go to CHATGBT

314
00:17:26,200 --> 00:17:29,119
Speaker 4: and you ask it to give you a menu or

315
00:17:29,840 --> 00:17:32,680
Speaker 4: dinner party you want to have this weekend. It's then

316
00:17:33,280 --> 00:17:36,399
Speaker 4: referencing that large language model and creating this response. And

317
00:17:36,440 --> 00:17:40,479
Speaker 4: of course that process is more computationally intensive because it's

318
00:17:40,920 --> 00:17:43,160
Speaker 4: it's doing a lot more things than a traditional searchces

319
00:17:43,400 --> 00:17:47,159
Speaker 4: personal search just matched the words you put into a

320
00:17:48,160 --> 00:17:51,640
Speaker 4: database of knowledge that it put together. But these large

321
00:17:51,680 --> 00:17:54,359
Speaker 4: language models are much more complex, and then the therefore

322
00:17:54,480 --> 00:17:56,320
Speaker 4: the things you're asking to do is more complex. So

323
00:17:56,720 --> 00:18:00,439
Speaker 4: it will will almost by definition, be a more energy

324
00:18:00,480 --> 00:18:03,640
Speaker 4: intensive process. Now, it's not to say that it can't

325
00:18:03,640 --> 00:18:08,360
Speaker 4: get more efficient, and it will. And Nvidio just last

326
00:18:08,400 --> 00:18:11,720
Speaker 4: week was releasing, you know, some data on some of

327
00:18:11,760 --> 00:18:14,840
Speaker 4: its next generation chips that are going to be significantly

328
00:18:14,880 --> 00:18:18,159
Speaker 4: more efficient than the prior generation. But one of the

329
00:18:18,160 --> 00:18:20,240
Speaker 4: things that we need to be careful of is to

330
00:18:21,200 --> 00:18:24,680
Speaker 4: think that because something becomes more efficient, then therefore we're

331
00:18:24,720 --> 00:18:28,080
Speaker 4: going to use less of the input resource, in this

332
00:18:28,160 --> 00:18:32,560
Speaker 4: case electricity. That's that's not how it works, because going

333
00:18:32,600 --> 00:18:37,760
Speaker 4: back to the concept of human capacity for consuming data,

334
00:18:38,119 --> 00:18:40,480
Speaker 4: all we do is we find more things to compute.

335
00:18:40,840 --> 00:18:43,760
Speaker 4: And this is you've probably heard of Jabon's paradox, and

336
00:18:43,760 --> 00:18:46,520
Speaker 4: this is the idea that, well, if we make more

337
00:18:46,640 --> 00:18:49,360
Speaker 4: efficient steam engines. He was an economists in the eighteen

338
00:18:49,440 --> 00:18:50,959
Speaker 4: hundreds and he said, well, if you make more efficient

339
00:18:51,000 --> 00:18:53,480
Speaker 4: steam engines, then we'll use us coal. And he's like, no,

340
00:18:53,520 --> 00:18:54,959
Speaker 4: that's not what's going to happen. We're going to use

341
00:18:54,960 --> 00:18:57,199
Speaker 4: more coal because we're going to mechanize more things. And

342
00:18:57,240 --> 00:18:59,919
Speaker 4: that's exactly what we do with data just because we

343
00:19:00,320 --> 00:19:02,439
Speaker 4: because we've had more law for years and so chips

344
00:19:02,440 --> 00:19:06,640
Speaker 4: have become incredibly more efficient than they were decades ago,

345
00:19:06,720 --> 00:19:10,120
Speaker 4: but we didn't use less energy. We used much more

346
00:19:10,240 --> 00:19:14,480
Speaker 4: energy because we could put chips in everything. So that's

347
00:19:14,640 --> 00:19:17,040
Speaker 4: the trend line that we're on. It's still climbing that

348
00:19:17,160 --> 00:19:21,000
Speaker 4: curve of consumption, and so no amount of efficiency is

349
00:19:21,040 --> 00:19:24,200
Speaker 4: going to take us at this point at least, because

350
00:19:24,440 --> 00:19:27,119
Speaker 4: I don't believe we're anywhere close to the bend in

351
00:19:27,160 --> 00:19:30,600
Speaker 4: that s curve. No amount of efficiency is going to

352
00:19:30,640 --> 00:19:34,760
Speaker 4: take us off of continuing to consume more electricity, and

353
00:19:34,760 --> 00:19:35,760
Speaker 4: at least in the near term.

354
00:19:36,400 --> 00:19:39,800
Speaker 3: So I have another basic building block kind of question.

355
00:19:40,000 --> 00:19:43,560
Speaker 3: But when we say that technology companies are aware of

356
00:19:43,640 --> 00:19:48,679
Speaker 3: the importance of energy usage or availability, and that this

357
00:19:48,760 --> 00:19:52,040
Speaker 3: is something they happened working on, what exactly is the

358
00:19:52,119 --> 00:19:56,720
Speaker 3: process by which a tech company gets its energy. So

359
00:19:56,840 --> 00:19:59,720
Speaker 3: you know you have a big data center, I imagine

360
00:19:59,720 --> 00:20:03,760
Speaker 3: you have some sort of agreement with whatever utility is

361
00:20:03,800 --> 00:20:06,879
Speaker 3: in that area. But I also imagine that that agreement

362
00:20:07,000 --> 00:20:12,639
Speaker 3: looks very very different to like my household energy bill

363
00:20:12,840 --> 00:20:13,720
Speaker 3: or something like that.

364
00:20:14,560 --> 00:20:19,680
Speaker 4: I'm certain it does hopefully significant orders of magnitude. Yes,

365
00:20:20,640 --> 00:20:24,560
Speaker 4: So there's two components. I mean, one is, if you're

366
00:20:24,920 --> 00:20:26,840
Speaker 4: building a data center, you have to plug it in somewhere.

367
00:20:26,880 --> 00:20:28,399
Speaker 4: You've got to plug it into the grid, right, So

368
00:20:29,040 --> 00:20:33,400
Speaker 4: there you're working with your local electric utility or transmission

369
00:20:33,400 --> 00:20:38,440
Speaker 4: company and doing planning for how big is the facility

370
00:20:38,440 --> 00:20:40,080
Speaker 4: to be, how much power is it going to pull

371
00:20:40,119 --> 00:20:42,080
Speaker 4: off the grid at any given time, and then over

372
00:20:42,119 --> 00:20:44,440
Speaker 4: a period of time, because these facilities just tend to

373
00:20:44,480 --> 00:20:48,679
Speaker 4: grow forever, and so that's the physical nuts and bolts

374
00:20:48,680 --> 00:20:52,160
Speaker 4: of connecting to the grid. Now, the second piece, of course,

375
00:20:52,200 --> 00:20:54,600
Speaker 4: is there needs to be some generation source as well,

376
00:20:55,040 --> 00:20:57,679
Speaker 4: like where's the power going to come from? And so

377
00:20:57,760 --> 00:21:01,359
Speaker 4: those two things are related, but they could be somewhat disconnected.

378
00:21:01,400 --> 00:21:04,720
Speaker 4: And so this is where you see you know, these

379
00:21:04,880 --> 00:21:07,160
Speaker 4: especially the tech companies who've really been leaders in this space,

380
00:21:07,720 --> 00:21:11,960
Speaker 4: entering into all these power purchase agreements for wind energy

381
00:21:12,000 --> 00:21:15,359
Speaker 4: and for solar energy and in some cases nuclear. You

382
00:21:15,640 --> 00:21:19,600
Speaker 4: mentioned the project earlier that's actually an AWS project where

383
00:21:19,600 --> 00:21:23,080
Speaker 4: they cite it next to the Susquehanna Nuclear plant. Right,

384
00:21:23,160 --> 00:21:26,919
Speaker 4: So all of that is around where are the electrons

385
00:21:26,960 --> 00:21:29,760
Speaker 4: going to come from? And how can with that purchasing

386
00:21:29,840 --> 00:21:32,080
Speaker 4: power of being some of the largest energy consumers in

387
00:21:32,119 --> 00:21:36,560
Speaker 4: the planet, how can they begin to influence the mix

388
00:21:36,640 --> 00:21:40,000
Speaker 4: of generation on the grid? Right? And that's the critical issue,

389
00:21:40,040 --> 00:21:43,159
Speaker 4: is that you're trying to influence where that power is

390
00:21:43,200 --> 00:21:47,040
Speaker 4: being generated from. It's not. And one thing just to

391
00:21:47,760 --> 00:21:50,600
Speaker 4: keep in mind is that you know the electrons you get,

392
00:21:50,840 --> 00:21:53,320
Speaker 4: you know, whether it's at your house or the data

393
00:21:53,320 --> 00:21:55,560
Speaker 4: center down the street, they're all the same electrons. You're

394
00:21:55,560 --> 00:21:57,840
Speaker 4: all pulling from the same grid. But what you're trying

395
00:21:57,880 --> 00:22:02,040
Speaker 4: to do is influence how that that generation is being created,

396
00:22:02,440 --> 00:22:05,920
Speaker 4: and that's where these purchase agreements come in for all

397
00:22:05,920 --> 00:22:07,200
Speaker 4: these different sources of energy.

398
00:22:07,840 --> 00:22:10,440
Speaker 2: All right, now, let's bring the question back to say

399
00:22:10,560 --> 00:22:11,919
Speaker 2: the utility side or save.

400
00:22:11,840 --> 00:22:12,680
Speaker 4: The dominion side.

401
00:22:12,720 --> 00:22:16,760
Speaker 2: So the dominion executives for decades have basically seen no growth,

402
00:22:17,160 --> 00:22:19,760
Speaker 2: and then suddenly in the span of the year, they're like, oh,

403
00:22:19,840 --> 00:22:22,000
Speaker 2: actually we're going to double. What do they do? What

404
00:22:22,040 --> 00:22:25,720
Speaker 2: are they doing right now today on we're recording this

405
00:22:25,760 --> 00:22:28,600
Speaker 2: April tenth, twenty twenty four. What are they doing right

406
00:22:28,640 --> 00:22:32,480
Speaker 2: now to expand generation or expand the grid or whatever

407
00:22:32,520 --> 00:22:34,359
Speaker 2: it is to meet that doubling of demand.

408
00:22:35,160 --> 00:22:37,920
Speaker 4: Well, this is where this is where it gets a

409
00:22:37,960 --> 00:22:41,280
Speaker 4: little concerning, is that you have these tech companies that

410
00:22:41,320 --> 00:22:46,480
Speaker 4: have these really ambitious commitments to being carbon neutral, carbon negative,

411
00:22:47,040 --> 00:22:49,400
Speaker 4: having one hundred percent zero carbon energy one hundred percent

412
00:22:49,440 --> 00:22:51,720
Speaker 4: of the time, and you have to give them credit

413
00:22:51,720 --> 00:22:53,680
Speaker 4: for the work they've done. I mean, that industry has

414
00:22:53,720 --> 00:22:58,919
Speaker 4: done amazing work over the last decade to build absolutely

415
00:22:59,000 --> 00:23:03,639
Speaker 4: just gigawatts upon gigawatts of new renewable energy projects in

416
00:23:03,680 --> 00:23:06,520
Speaker 4: the United States, all over the world. They've been some

417
00:23:06,560 --> 00:23:11,840
Speaker 4: of the biggest drivers in the corporate focus on decarbonization,

418
00:23:12,400 --> 00:23:15,320
Speaker 4: and so you really have to give that industry credit

419
00:23:15,320 --> 00:23:18,520
Speaker 4: for all it's done, and all the big tech companies

420
00:23:18,520 --> 00:23:22,080
Speaker 4: have done some amazing work there. The challenge though, that

421
00:23:22,119 --> 00:23:24,199
Speaker 4: we have is the environment that they did that in.

422
00:23:25,119 --> 00:23:28,639
Speaker 4: Was that no growth environment we were talking about. They were

423
00:23:28,680 --> 00:23:32,160
Speaker 4: all growing, but they were starting from a relatively small

424
00:23:32,240 --> 00:23:36,520
Speaker 4: denominator ten or fifteen years ago. So there and there

425
00:23:36,520 --> 00:23:39,439
Speaker 4: was a lot of overhang in the utility system at

426
00:23:39,480 --> 00:23:42,880
Speaker 4: that time because the utilities had overbuilt ahead of that

427
00:23:42,920 --> 00:23:46,480
Speaker 4: sort of flatlining, so there was excess capacity on the system.

428
00:23:46,960 --> 00:23:50,879
Speaker 4: They were growing inside of a system that wasn't itself

429
00:23:50,920 --> 00:23:54,800
Speaker 4: growing on a net basis. Yeah, So everything they did,

430
00:23:55,000 --> 00:23:58,440
Speaker 4: every new wind project you brought on, every new solar

431
00:23:58,440 --> 00:24:02,359
Speaker 4: project you brought on, those we're all incrementally reducing the

432
00:24:02,359 --> 00:24:05,760
Speaker 4: amount of carbon in the system. It was all net positive.

433
00:24:06,359 --> 00:24:09,000
Speaker 4: Now we get into this new world where their growth

434
00:24:09,080 --> 00:24:13,600
Speaker 4: rates are exceeding what the utilities have ever imagine in

435
00:24:13,680 --> 00:24:17,960
Speaker 4: terms of the absolute impact on the system. The utilities

436
00:24:18,040 --> 00:24:21,800
Speaker 4: response is the only thing we can do in the

437
00:24:21,840 --> 00:24:25,480
Speaker 4: time horizon that we have is basically build more gas

438
00:24:25,480 --> 00:24:28,679
Speaker 4: plants or keep online gas plants or coal plants that

439
00:24:28,680 --> 00:24:32,679
Speaker 4: we were planning on shuttering. And so now the commitments

440
00:24:32,680 --> 00:24:37,320
Speaker 4: that they have to zero carbon energy, to be carbon negative,

441
00:24:37,440 --> 00:24:41,760
Speaker 4: et cetera, are coming into contrast with the response that

442
00:24:41,800 --> 00:24:45,760
Speaker 4: the utilities are laying out in their what's called integrated

443
00:24:45,800 --> 00:24:49,159
Speaker 4: resource plans or IRPs. And we've seen this recently, just

444
00:24:49,240 --> 00:24:52,760
Speaker 4: last week in Georgia. We've seen it in Duke in

445
00:24:52,800 --> 00:24:56,719
Speaker 4: North Carolina, Dominion and Virginia. Every single one of those

446
00:24:56,840 --> 00:24:59,720
Speaker 4: utilities is saying, with all the demand that we're seeing

447
00:24:59,760 --> 00:25:03,439
Speaker 4: come into our system, we have to put more fossil

448
00:25:03,480 --> 00:25:05,439
Speaker 4: fuel resources on the grid. It's the only way that

449
00:25:05,440 --> 00:25:07,399
Speaker 4: we can manage it in the time rise that we have.

450
00:25:07,520 --> 00:25:09,920
Speaker 4: Now there's a lot of debate about whether that is true,

451
00:25:10,200 --> 00:25:11,240
Speaker 4: but it is what's happening.

452
00:25:11,920 --> 00:25:15,280
Speaker 3: So when push comes to shove, it seems like some

453
00:25:15,359 --> 00:25:21,520
Speaker 3: of the green priorities are getting superseded by existential pressures

454
00:25:21,560 --> 00:25:24,880
Speaker 3: on the business model. Perhaps, and we could debate how

455
00:25:25,080 --> 00:25:28,800
Speaker 3: transferable AI actually is at this point and how big

456
00:25:28,840 --> 00:25:32,040
Speaker 3: a moat you have over something like chat, GPT or

457
00:25:32,119 --> 00:25:34,320
Speaker 3: Claude or something like that, but there does seem to

458
00:25:34,359 --> 00:25:38,760
Speaker 3: be the sense of urgency among tech companies where if

459
00:25:38,800 --> 00:25:41,439
Speaker 3: you're not building something out right now and trying to

460
00:25:41,560 --> 00:25:45,080
Speaker 3: dominate the market and really produce the best thing possible,

461
00:25:45,359 --> 00:25:48,880
Speaker 3: well you're either losing, you know, billions of dollars or

462
00:25:49,200 --> 00:25:52,399
Speaker 3: you're going to be superseded by someone who does manage

463
00:25:52,400 --> 00:25:53,600
Speaker 3: to do that successfully.

464
00:25:55,520 --> 00:25:58,240
Speaker 4: That's exactly right, and it's probably not billions of dollars.

465
00:25:58,280 --> 00:26:00,640
Speaker 4: There's probably trillions of time. Yes, yes, And that's where

466
00:26:00,640 --> 00:26:03,399
Speaker 4: the competitive pressure is coming in. And this is why

467
00:26:03,440 --> 00:26:06,640
Speaker 4: there's there's such a focus right now in this industry

468
00:26:06,680 --> 00:26:09,240
Speaker 4: on where is the power going to come from? Because

469
00:26:09,320 --> 00:26:15,640
Speaker 4: the ability to at least envision and on paper design

470
00:26:16,720 --> 00:26:21,440
Speaker 4: training models that are absolutely enormous, just orders of magnitude

471
00:26:21,480 --> 00:26:24,240
Speaker 4: bigger than anything that we've ever built in terms of

472
00:26:24,240 --> 00:26:26,560
Speaker 4: a data center are coming in to start contrast with

473
00:26:26,680 --> 00:26:29,719
Speaker 4: the reality of the power system of one, is that

474
00:26:29,760 --> 00:26:33,200
Speaker 4: power even available? And two, if it could be available,

475
00:26:33,359 --> 00:26:36,600
Speaker 4: is there a way to do it with a zero

476
00:26:36,640 --> 00:26:39,560
Speaker 4: carbon approach, which is again what these companies are committed to.

477
00:26:39,800 --> 00:26:42,520
Speaker 4: And that's that's the tension that we're in right now

478
00:26:42,920 --> 00:26:49,919
Speaker 4: of how do we quickly accelerate the delivery and growth

479
00:26:49,960 --> 00:26:53,800
Speaker 4: of the electric grid? Which is which and I think

480
00:26:53,920 --> 00:26:57,480
Speaker 4: I just want to the quick aside on this consuming

481
00:26:57,520 --> 00:27:00,639
Speaker 4: electricity is in the context where it's talking about is

482
00:27:00,680 --> 00:27:02,680
Speaker 4: a really great thing. I mean, this is something that

483
00:27:03,280 --> 00:27:07,200
Speaker 4: leads to economic growth, it leads to job creation. All

484
00:27:07,280 --> 00:27:10,040
Speaker 4: of this, I mean, this whole problem that we have

485
00:27:10,200 --> 00:27:13,239
Speaker 4: right now of electric utilities having to think about this

486
00:27:13,280 --> 00:27:16,520
Speaker 4: whole new era of growth. It's all because we're on

487
00:27:16,640 --> 00:27:20,040
Speaker 4: shoring manufacturing in the United States. We're building these data

488
00:27:20,040 --> 00:27:23,879
Speaker 4: centers and creating all sorts of amazing tools and creating

489
00:27:23,920 --> 00:27:27,720
Speaker 4: efficiency across all sorts of sectors. And we're also in

490
00:27:27,760 --> 00:27:32,119
Speaker 4: the same vein we're also electrifying transportation and heating. Like,

491
00:27:32,560 --> 00:27:35,320
Speaker 4: all of this is good. It's all goodness. And we

492
00:27:35,320 --> 00:27:38,200
Speaker 4: didn't even get to things like hydrogen production and other

493
00:27:38,200 --> 00:27:40,600
Speaker 4: ways that we're going to use electricity. The real rub

494
00:27:40,600 --> 00:27:43,560
Speaker 4: of this, though, is that we're in this situation right

495
00:27:43,600 --> 00:27:49,560
Speaker 4: now where again the electric electricity industry was somewhat surprised

496
00:27:49,560 --> 00:27:53,240
Speaker 4: by this. They weren't prepared for over a period of

497
00:27:53,240 --> 00:27:55,960
Speaker 4: a couple of years. Again going back to the case

498
00:27:55,960 --> 00:28:00,600
Speaker 4: of Dominion having to double their load forecast flexively, They're

499
00:28:00,640 --> 00:28:01,840
Speaker 4: going to go to the one thing they know how

500
00:28:01,880 --> 00:28:05,040
Speaker 4: to do, which is build gas plants because they know

501
00:28:05,119 --> 00:28:08,000
Speaker 4: that works. That's the easy way out. There are other

502
00:28:08,080 --> 00:28:10,720
Speaker 4: things we can do, though. There are ways we can

503
00:28:10,800 --> 00:28:14,600
Speaker 4: leverage the existing system more effectively. We can use things

504
00:28:14,600 --> 00:28:19,480
Speaker 4: called grid enhancing technologies, where we through sensoring, through better

505
00:28:19,720 --> 00:28:23,160
Speaker 4: dynamic rating of power lines, we can actually get more

506
00:28:23,240 --> 00:28:25,840
Speaker 4: out of the existing system we have. There's ways we

507
00:28:25,880 --> 00:28:30,120
Speaker 4: can use storage more effectively, because really what we're trying

508
00:28:30,119 --> 00:28:32,240
Speaker 4: to manage is just these system peaks. Most of the

509
00:28:32,280 --> 00:28:35,200
Speaker 4: time there's plenty of power. It's really just during the

510
00:28:35,240 --> 00:28:39,520
Speaker 4: modest summer hours or the coldest winter hours that the

511
00:28:39,560 --> 00:28:41,880
Speaker 4: system gets constrained, and that's what's driving a lot of

512
00:28:41,920 --> 00:28:45,200
Speaker 4: the need for utilities to want to build this new capacity.

513
00:28:45,200 --> 00:28:47,600
Speaker 4: But we can manage it in other ways. And it's

514
00:28:47,640 --> 00:28:51,760
Speaker 4: really incumbent upon the data center industry to lean in

515
00:28:51,800 --> 00:28:55,320
Speaker 4: on this, to think through how can we be more

516
00:28:55,360 --> 00:28:58,240
Speaker 4: of a party to solving this problem. Because data centers

517
00:28:58,280 --> 00:29:01,440
Speaker 4: have lots of opportunities to be more flexible. They have

518
00:29:01,760 --> 00:29:04,680
Speaker 4: behind the meter generation, they have behind the meter storage.

519
00:29:04,800 --> 00:29:07,240
Speaker 4: They can actually be part of the solution, not just

520
00:29:07,280 --> 00:29:08,040
Speaker 4: part of the problem.

521
00:29:08,600 --> 00:29:10,840
Speaker 3: I just want to press you on this point because

522
00:29:10,880 --> 00:29:14,240
Speaker 3: I know people will have questions about this, and I

523
00:29:14,320 --> 00:29:17,080
Speaker 3: take the point about in many respects we're talking about

524
00:29:17,160 --> 00:29:21,160
Speaker 3: increased energy usage as a result of new things that

525
00:29:21,200 --> 00:29:25,280
Speaker 3: are leading to you know, new jobs and new productive industry,

526
00:29:25,760 --> 00:29:28,760
Speaker 3: and also the idea that, well, we can produce more

527
00:29:28,800 --> 00:29:31,960
Speaker 3: electricity in different ways, or we can make the delivery

528
00:29:31,960 --> 00:29:35,240
Speaker 3: of electricity more efficient, and all those types of things.

529
00:29:35,840 --> 00:29:39,120
Speaker 3: But I think one of the reservations people might have

530
00:29:39,200 --> 00:29:43,560
Speaker 3: about this is the idea of you know, competition with

531
00:29:43,840 --> 00:29:47,320
Speaker 3: large tech companies that have a lot of money and

532
00:29:47,440 --> 00:29:51,160
Speaker 3: that potentially have a lot of influence over the utility companies,

533
00:29:51,440 --> 00:29:54,080
Speaker 3: and the idea that maybe you could get a situation

534
00:29:54,560 --> 00:29:58,240
Speaker 3: where I don't know, Amazon gets like one hundred percent

535
00:29:58,320 --> 00:30:03,160
Speaker 3: off take from some power plant in whatever state, and

536
00:30:03,320 --> 00:30:06,600
Speaker 3: maybe other people are left with either you know, not

537
00:30:06,720 --> 00:30:11,800
Speaker 3: enough electricity or more likely much more expensive electricity. Can

538
00:30:11,840 --> 00:30:16,400
Speaker 3: you talk about that zero? I was being somewhat facetious

539
00:30:16,200 --> 00:30:19,760
Speaker 3: in the intro talking about zero sum game. But there

540
00:30:19,920 --> 00:30:23,240
Speaker 3: is this idea of like competition and there might not

541
00:30:23,360 --> 00:30:26,760
Speaker 3: be enough to go around, at least at the precise

542
00:30:26,840 --> 00:30:28,320
Speaker 3: times that everyone might want it.

543
00:30:29,920 --> 00:30:32,560
Speaker 4: That's right, And that's the big challenge that good planners

544
00:30:32,760 --> 00:30:36,400
Speaker 4: have today is what loads do you say yes to

545
00:30:37,200 --> 00:30:41,360
Speaker 4: and what are the long term implications of that and this?

546
00:30:41,520 --> 00:30:46,080
Speaker 4: And we've seen this play out over the rest of

547
00:30:46,120 --> 00:30:49,000
Speaker 4: the globe where you've had these concentrations of data centers.

548
00:30:49,480 --> 00:30:52,400
Speaker 4: This is a story that we saw in Dublin, We've

549
00:30:52,440 --> 00:30:54,800
Speaker 4: seen it in Singapore, we've seen it in Amsterdam, and

550
00:30:55,120 --> 00:30:57,440
Speaker 4: these governments start to get really worried of wait a minute,

551
00:30:57,520 --> 00:30:59,920
Speaker 4: we have too many data centers as a sort of

552
00:31:00,000 --> 00:31:05,840
Speaker 4: percentage of overall energy consumption. And what inevitably happens is

553
00:31:06,440 --> 00:31:09,760
Speaker 4: a move towards putting either moratoriums on data center build

554
00:31:09,760 --> 00:31:13,960
Speaker 4: out or putting very tight restrictions on what they can

555
00:31:14,000 --> 00:31:15,479
Speaker 4: do and the scale at which they can do it.

556
00:31:15,880 --> 00:31:19,320
Speaker 4: And so, you know, we haven't yet seen that to

557
00:31:19,400 --> 00:31:22,080
Speaker 4: any material degree in the United States, but I do

558
00:31:22,120 --> 00:31:24,040
Speaker 4: think that's a real risk, and it's a risk that

559
00:31:24,160 --> 00:31:27,440
Speaker 4: the data center industry faces, I think somewhat uniquely in that,

560
00:31:28,080 --> 00:31:30,440
Speaker 4: you know, if you're the governor of state and you

561
00:31:30,600 --> 00:31:34,480
Speaker 4: have a choice to give power to a say new

562
00:31:35,360 --> 00:31:39,880
Speaker 4: you know, ev car factory that's going to produce fifteen

563
00:31:39,960 --> 00:31:42,680
Speaker 4: hundred and two thousand jobs versus a data center that's

564
00:31:42,720 --> 00:31:45,880
Speaker 4: going to produce significantly less than that you're going to

565
00:31:45,880 --> 00:31:48,840
Speaker 4: give it to the factory. Right, the data centers are

566
00:31:48,880 --> 00:31:51,160
Speaker 4: actually the ones that are going to face likely the

567
00:31:51,240 --> 00:31:57,840
Speaker 4: most constraints as governments, utilities, regulators start wrestling with this

568
00:31:57,960 --> 00:31:59,760
Speaker 4: trade off of Oh, we're going to have to say

569
00:31:59,760 --> 00:32:02,520
Speaker 4: no to somebody. And that's the real risk that I

570
00:32:02,560 --> 00:32:06,800
Speaker 4: think the AI and data center industry faces today is

571
00:32:06,800 --> 00:32:11,520
Speaker 4: that they are the easiest target because everyone loves what

572
00:32:11,640 --> 00:32:15,200
Speaker 4: data centers do, but no one particularly loves just having

573
00:32:15,240 --> 00:32:17,840
Speaker 4: a data center next door to their house. And so

574
00:32:17,920 --> 00:32:22,040
Speaker 4: that's a real challenge for the industry is that they

575
00:32:22,080 --> 00:32:27,080
Speaker 4: will start to get in the crosshairs of these regulators, leaders,

576
00:32:27,120 --> 00:32:31,080
Speaker 4: whomever who's pulling the strings as these decisions start to get.

577
00:32:47,160 --> 00:32:49,720
Speaker 2: So I just want to make two random thoughts that

578
00:32:49,760 --> 00:32:51,760
Speaker 2: were in my head. I walked by a film set

579
00:32:51,880 --> 00:32:53,680
Speaker 2: in the East Villaers the other day. They were filming

580
00:32:53,720 --> 00:32:55,680
Speaker 2: this movie, and they are all these like big, like

581
00:32:55,760 --> 00:32:59,200
Speaker 2: thick electrical cables like you know, like that are powering

582
00:32:59,240 --> 00:33:01,440
Speaker 2: the lights and and all that stuff. And I thought

583
00:33:01,480 --> 00:33:03,280
Speaker 2: to myself, Oh, it would be so great when they

584
00:33:03,320 --> 00:33:06,720
Speaker 2: can just make all the movies on AI with Sora

585
00:33:06,960 --> 00:33:09,240
Speaker 2: or something like that, and then you know, we'll also

586
00:33:09,280 --> 00:33:11,520
Speaker 2: get electricity savings because we won't have to have human

587
00:33:11,560 --> 00:33:14,160
Speaker 2: actors with actual lights and stuff like that, so that'll

588
00:33:14,200 --> 00:33:17,120
Speaker 2: be exciting and then being a little facetious about the

589
00:33:17,160 --> 00:33:19,440
Speaker 2: end of human actors, but you know, in theory that

590
00:33:19,480 --> 00:33:21,280
Speaker 2: could be exciting, and then you could you know, you

591
00:33:21,600 --> 00:33:24,640
Speaker 2: mentioned it was like, well, the utilities got surprised by

592
00:33:24,800 --> 00:33:27,880
Speaker 2: the low you know, this the spike in demand. But

593
00:33:27,960 --> 00:33:29,440
Speaker 2: it sounds to me like we can't really blame the

594
00:33:29,520 --> 00:33:33,160
Speaker 2: utilities too much because if even the people inside Microsoft

595
00:33:33,360 --> 00:33:37,240
Speaker 2: got a bit caught unsurprised by the explosion of AI

596
00:33:37,440 --> 00:33:40,320
Speaker 2: interest in the fall of twenty twenty two, then I

597
00:33:40,320 --> 00:33:42,800
Speaker 2: guess like we can't really blame Dominion if they hit

598
00:33:42,840 --> 00:33:46,360
Speaker 2: they were private. Further away from the issue, you mentioned

599
00:33:46,560 --> 00:33:49,920
Speaker 2: peak demand, and this gets to like power, the type

600
00:33:49,960 --> 00:33:52,560
Speaker 2: of power, because people talk about like this sort of need.

601
00:33:52,640 --> 00:33:55,360
Speaker 2: The problem with renewables as well, at least when we're

602
00:33:55,360 --> 00:33:58,480
Speaker 2: talking about solar and wind, there's this intermittency problem. It's

603
00:33:58,520 --> 00:34:01,160
Speaker 2: not always sunny, even if it's hot. When it's hot,

604
00:34:01,320 --> 00:34:04,840
Speaker 2: it's not always windy, there's nighttime, et cetera. How much

605
00:34:04,920 --> 00:34:10,120
Speaker 2: does that constrain the ability of more renewables to be

606
00:34:10,280 --> 00:34:13,560
Speaker 2: sort of the solution to the utilities problem.

607
00:34:14,160 --> 00:34:18,600
Speaker 4: It's a real challenge because again, as you noted, we're

608
00:34:18,600 --> 00:34:21,200
Speaker 4: trying to manage peak demand. That's what all this growth

609
00:34:21,239 --> 00:34:24,400
Speaker 4: is about. So peak demand is about the certainty that

610
00:34:24,440 --> 00:34:28,000
Speaker 4: you're going to have power during those highest system piece,

611
00:34:28,160 --> 00:34:31,640
Speaker 4: the hottest days, the coldest winter nights, and you can't

612
00:34:31,680 --> 00:34:36,920
Speaker 4: always guarantee that renewable generation will be online during those times.

613
00:34:36,920 --> 00:34:38,759
Speaker 4: And this is the role of the system planner is

614
00:34:38,800 --> 00:34:42,440
Speaker 4: to look at all these different resources and figure out

615
00:34:42,520 --> 00:34:46,120
Speaker 4: how can we assure that we have the sufficient reserve

616
00:34:46,200 --> 00:34:48,560
Speaker 4: margin to ensure that we're not going to have things

617
00:34:48,600 --> 00:34:52,160
Speaker 4: like rolling brownouts or black eyes. Now there's a lot

618
00:34:52,160 --> 00:34:56,319
Speaker 4: of tools though, that we have to help manage that uncertainty,

619
00:34:56,600 --> 00:35:00,520
Speaker 4: and we have increasingly, you know, month after a month,

620
00:35:00,560 --> 00:35:03,799
Speaker 4: it seems like lower cost battery options which give us

621
00:35:03,920 --> 00:35:06,840
Speaker 4: more duration that we can deploy to solve some of

622
00:35:06,560 --> 00:35:10,240
Speaker 4: these issues. We have the ability of even the loads

623
00:35:10,280 --> 00:35:15,160
Speaker 4: through like virtual power plants to be more responsive during

624
00:35:15,200 --> 00:35:18,520
Speaker 4: these times of system peaks, right, So we have tools

625
00:35:18,600 --> 00:35:23,000
Speaker 4: that we can use to manage that uncertainty. The problem

626
00:35:23,080 --> 00:35:26,000
Speaker 4: is that it is a very complex problem. I mean,

627
00:35:26,040 --> 00:35:29,960
Speaker 4: you're talking about, you know, millions of different data points

628
00:35:29,960 --> 00:35:32,520
Speaker 4: that you're trying to manage, and the way that utilities

629
00:35:32,520 --> 00:35:36,239
Speaker 4: have historically managed these things has been fairly rudimentary in

630
00:35:36,320 --> 00:35:40,080
Speaker 4: terms of their sophistication, and so they're having to go

631
00:35:40,120 --> 00:35:44,120
Speaker 4: through this learning curve of how do we ensure that

632
00:35:44,160 --> 00:35:47,320
Speaker 4: we can achieve the load growth that all these industries

633
00:35:47,840 --> 00:35:54,200
Speaker 4: you know, are expecting and meet the reliability, cost availability

634
00:35:54,200 --> 00:35:57,480
Speaker 4: expectations of our customers. And that's where that's where the

635
00:35:57,560 --> 00:35:59,839
Speaker 4: challenge comes in. And this is where the whole problem,

636
00:36:00,080 --> 00:36:03,279
Speaker 4: it's frankly really interesting, is that there are lots of

637
00:36:03,360 --> 00:36:05,880
Speaker 4: levers that we have and we don't just have to

638
00:36:05,880 --> 00:36:09,680
Speaker 4: throw more fossil fuel plants at this problem. Does that

639
00:36:09,719 --> 00:36:11,920
Speaker 4: mean we're not going to build any new gas plants

640
00:36:11,920 --> 00:36:15,839
Speaker 4: in this country? I certainly will. I don't think there's

641
00:36:15,880 --> 00:36:19,040
Speaker 4: a way around this problem, at least in the short run,

642
00:36:19,080 --> 00:36:24,000
Speaker 4: without having some incremental addition of fossil based resources. But

643
00:36:24,560 --> 00:36:26,560
Speaker 4: there's also a lot of other things we could be

644
00:36:26,640 --> 00:36:32,280
Speaker 4: doing that would significantly reduce dependence on fossil based resources

645
00:36:32,320 --> 00:36:35,560
Speaker 4: to achieve the growth objectives that we have as a country.

646
00:36:36,440 --> 00:36:40,600
Speaker 3: What are the levers specifically on the tech company or

647
00:36:40,640 --> 00:36:44,200
Speaker 3: the data center side, because I again, so much of

648
00:36:44,239 --> 00:36:46,799
Speaker 3: the focus of this conversation is on what can the

649
00:36:46,920 --> 00:36:50,200
Speaker 3: utilities do, what can we do in terms of enhancing

650
00:36:50,200 --> 00:36:55,120
Speaker 3: the grid managing supply more efficiently? But are there novel

651
00:36:55,320 --> 00:36:59,000
Speaker 3: or interesting things that the data centers themselves can do

652
00:36:59,120 --> 00:37:01,600
Speaker 3: here in terms of managing their own energy usage.

653
00:37:02,800 --> 00:37:05,640
Speaker 4: Yes, I there's there's a few things. I mean, one is,

654
00:37:06,600 --> 00:37:11,400
Speaker 4: data centers have substantial ability to be more flexible in

655
00:37:11,520 --> 00:37:14,839
Speaker 4: terms of the power that they're taking from the grid

656
00:37:14,880 --> 00:37:17,680
Speaker 4: at any given time. As I mentioned before, every data

657
00:37:17,680 --> 00:37:20,000
Speaker 4: center or nearly every data data center has some form

658
00:37:20,000 --> 00:37:23,720
Speaker 4: of backup generation. They have some form of energy storage

659
00:37:23,960 --> 00:37:26,359
Speaker 4: built into this. So there the way a data center

660
00:37:26,400 --> 00:37:29,080
Speaker 4: is designed. Its designed like a power plant with an

661
00:37:29,160 --> 00:37:31,799
Speaker 4: energy storage plant that just happens to be sitting next

662
00:37:31,800 --> 00:37:33,920
Speaker 4: to a room full of servers, right, And so when

663
00:37:33,960 --> 00:37:36,560
Speaker 4: you when you break it down to those components, you say, okay, well,

664
00:37:36,600 --> 00:37:39,640
Speaker 4: how can we better optimize this power plant to be

665
00:37:40,120 --> 00:37:42,560
Speaker 4: more of a grid resource? How can we have to

666
00:37:42,640 --> 00:37:45,120
Speaker 4: optimize the storage plant to be more of a grid resource?

667
00:37:45,360 --> 00:37:47,960
Speaker 4: And then in terms of even the servers themselves, how

668
00:37:48,000 --> 00:37:51,840
Speaker 4: can we optimize the way the software actually operates and

669
00:37:51,960 --> 00:37:55,200
Speaker 4: is architected to be more of a grid resource. And

670
00:37:55,680 --> 00:37:59,000
Speaker 4: that is that sort of thinking is what is being

671
00:37:59,040 --> 00:38:02,279
Speaker 4: forced on the industry. Frankly, we've always had this capability.

672
00:38:02,560 --> 00:38:04,520
Speaker 4: I mean, we were doing I mean we did a

673
00:38:04,560 --> 00:38:07,600
Speaker 4: project like twenty sixteen with a utility where we put

674
00:38:07,640 --> 00:38:11,480
Speaker 4: in flexible gas generators behind our meter because the utility

675
00:38:11,520 --> 00:38:14,640
Speaker 4: was going to have to build a new power plant

676
00:38:15,200 --> 00:38:17,400
Speaker 4: if we didn't have a way to be more flexible.

677
00:38:17,680 --> 00:38:20,680
Speaker 4: So we've always known that we can do this, but

678
00:38:21,040 --> 00:38:25,319
Speaker 4: the industry has never been pressurized to really think innovatively

679
00:38:25,440 --> 00:38:28,759
Speaker 4: about how can we utilize all these assets that we

680
00:38:28,840 --> 00:38:33,279
Speaker 4: have inside of the data center plant itself to be

681
00:38:33,520 --> 00:38:35,960
Speaker 4: more part of the grid. Right, So that's I think

682
00:38:36,000 --> 00:38:39,080
Speaker 4: the most important thing is really thinking about how data

683
00:38:39,120 --> 00:38:42,080
Speaker 4: centers become more flexible. There's a whole nother line of thinking,

684
00:38:42,120 --> 00:38:45,440
Speaker 4: which is this idea of well, utility is not going

685
00:38:45,480 --> 00:38:48,000
Speaker 4: to be fast enough, so data centers just need to

686
00:38:48,000 --> 00:38:49,640
Speaker 4: build all their own power plants. And this is where

687
00:38:49,680 --> 00:38:53,320
Speaker 4: you start hearing about nuclear and SMRs in confusion, which

688
00:38:54,080 --> 00:38:58,520
Speaker 4: is interesting except it doesn't solve the problem this decade.

689
00:38:59,239 --> 00:39:01,760
Speaker 4: It doesn't solve the problem that we're facing right now

690
00:39:01,800 --> 00:39:05,439
Speaker 4: because none of that stuff is actually ready for prime time.

691
00:39:05,719 --> 00:39:10,239
Speaker 4: We don't have an SMR that we can build today predictably,

692
00:39:10,400 --> 00:39:15,239
Speaker 4: on time, on budget, So we are dependent on the

693
00:39:15,280 --> 00:39:18,200
Speaker 4: tools that we have today, which are things like batteries,

694
00:39:18,800 --> 00:39:25,439
Speaker 4: great enhancing technologies, flexible load reconductoring, transmission lines to get

695
00:39:25,480 --> 00:39:29,120
Speaker 4: more power over existing rights of ways. So there's a

696
00:39:29,239 --> 00:39:31,600
Speaker 4: number of things we can do with technologies we have

697
00:39:32,040 --> 00:39:35,960
Speaker 4: today that are going to be very meaningful this decade,

698
00:39:36,520 --> 00:39:38,799
Speaker 4: and we should keep investing in things that are going

699
00:39:38,840 --> 00:39:41,719
Speaker 4: to be really meaningful next decade. I'm very bullish on

700
00:39:42,000 --> 00:39:45,680
Speaker 4: what we can do with new forms of nuclear technology.

701
00:39:46,719 --> 00:39:49,600
Speaker 4: They're just not relevant in the time horise and the

702
00:39:49,640 --> 00:39:50,680
Speaker 4: problem we're talking about.

703
00:39:50,880 --> 00:39:53,799
Speaker 2: At some point, at some point, we're going to do

704
00:39:53,880 --> 00:39:58,040
Speaker 2: an Odd Lots episode specifically on the promise of small

705
00:39:58,320 --> 00:40:01,280
Speaker 2: modular reactors and why we still don't have them despite

706
00:40:01,280 --> 00:40:03,160
Speaker 2: the seeming benefits. But do you have like a sort

707
00:40:03,160 --> 00:40:06,920
Speaker 2: of succinct answer for why this sort of seeming solution

708
00:40:07,239 --> 00:40:10,800
Speaker 2: of manufacturing them faster, et cetera like has not translated

709
00:40:11,000 --> 00:40:13,080
Speaker 2: into anything in production.

710
00:40:14,600 --> 00:40:17,200
Speaker 4: Well, quite simply, we just forgot how to do it.

711
00:40:17,719 --> 00:40:19,720
Speaker 4: We used to be able to build nuclear in this country.

712
00:40:19,800 --> 00:40:21,720
Speaker 4: We did in the seventies, we did in the eighties,

713
00:40:22,400 --> 00:40:25,120
Speaker 4: but every person that was involved in any one of

714
00:40:25,120 --> 00:40:29,360
Speaker 4: those projects is either not alive or certainly not still

715
00:40:29,360 --> 00:40:31,760
Speaker 4: a project manager at a company that would be building

716
00:40:31,840 --> 00:40:36,880
Speaker 4: nuclear plants. Right. We I think we underestimate human capacity

717
00:40:36,920 --> 00:40:39,560
Speaker 4: to forget things right. Just because we've done something in

718
00:40:39,560 --> 00:40:42,920
Speaker 4: the past doesn't mean that we necessarily can do it again.

719
00:40:43,000 --> 00:40:46,000
Speaker 4: We have to relearn these things. And as a country

720
00:40:46,360 --> 00:40:48,640
Speaker 4: like we do not have a supply chain, we don't

721
00:40:48,680 --> 00:40:51,680
Speaker 4: have a labor force, we don't have people that manage

722
00:40:51,680 --> 00:40:54,640
Speaker 4: construction projects that know how to do any of these things.

723
00:40:55,000 --> 00:40:57,879
Speaker 4: And so when you look at what South Korea is doing,

724
00:40:58,000 --> 00:41:00,640
Speaker 4: you look at what China's doing, you know, they are

725
00:41:00,640 --> 00:41:04,960
Speaker 4: building nuclear plants with regularity, they're doing it at at

726
00:41:05,000 --> 00:41:07,920
Speaker 4: a very attractive costs, they're doing it on a predictable

727
00:41:08,040 --> 00:41:11,319
Speaker 4: time horizon. But they have actually built all of those

728
00:41:11,360 --> 00:41:13,960
Speaker 4: resources that we just simply don't have in this country

729
00:41:14,239 --> 00:41:17,520
Speaker 4: that we need and we need to rebuild that capability.

730
00:41:18,080 --> 00:41:19,480
Speaker 4: It just doesn't exist today. You know.

731
00:41:19,600 --> 00:41:21,799
Speaker 2: One of the things that in the when we're talking

732
00:41:21,800 --> 00:41:24,840
Speaker 2: about utilities, they're like weird companies because they're not like

733
00:41:24,920 --> 00:41:29,080
Speaker 2: normal businesses. They're sort of natural monopolies. They price set,

734
00:41:29,360 --> 00:41:32,560
Speaker 2: in my understanding, is based on how much they invest,

735
00:41:32,640 --> 00:41:35,480
Speaker 2: and so they have to then petition some local regulators say, look,

736
00:41:35,480 --> 00:41:37,360
Speaker 2: we had to invest this much, and that's why I

737
00:41:37,640 --> 00:41:40,520
Speaker 2: want to raise the prices this much, et cetera. Are

738
00:41:40,560 --> 00:41:45,319
Speaker 2: there regulatory hurdles or things about the regulatory system right

739
00:41:45,360 --> 00:41:48,000
Speaker 2: now that are going to make that doubling of demand

740
00:41:48,280 --> 00:41:49,759
Speaker 2: more challenging than they need to be?

741
00:41:50,960 --> 00:41:55,239
Speaker 4: Absolutely, And so you go back to the era that

742
00:41:55,239 --> 00:41:57,960
Speaker 4: we've been in of relative no load growth.

743
00:41:58,040 --> 00:41:58,840
Speaker 2: Yeah.

744
00:41:58,920 --> 00:42:01,480
Speaker 4: You know, if you're a utility regulator and utility comes

745
00:42:01,520 --> 00:42:04,760
Speaker 4: and asks you for a billion dollars for new investment,

746
00:42:05,239 --> 00:42:08,279
Speaker 4: and you're used to saying no, used to saying, well,

747
00:42:08,280 --> 00:42:11,400
Speaker 4: wait a minute, why why do you need this, what's what?

748
00:42:11,400 --> 00:42:13,560
Speaker 4: What is this for? How is this going to help

749
00:42:13,920 --> 00:42:19,680
Speaker 4: you know, manage again, reliability, cost, predictability, et cetera. Now

750
00:42:19,719 --> 00:42:21,560
Speaker 4: you're in this whole new world and going back to

751
00:42:21,600 --> 00:42:25,759
Speaker 4: this concept of like we easily forget things. No one

752
00:42:25,760 --> 00:42:28,040
Speaker 4: who's a regulator today or the head of a utility

753
00:42:28,040 --> 00:42:31,319
Speaker 4: today has ever lived through an environment where we've had

754
00:42:31,360 --> 00:42:36,160
Speaker 4: this massive expansion of the demand for electricity. So everyone

755
00:42:36,360 --> 00:42:40,520
Speaker 4: now including the regulators, are having to relearn, Okay, how

756
00:42:40,560 --> 00:42:45,520
Speaker 4: do we enable utility investment in a growth environment. It's

757
00:42:45,560 --> 00:42:48,719
Speaker 4: not something they've ever done before, and so they're having

758
00:42:48,760 --> 00:42:52,040
Speaker 4: to figure out, Okay, how do we create the sort

759
00:42:52,080 --> 00:42:55,680
Speaker 4: of the bandwidth for utilities to make these investments, because

760
00:42:56,640 --> 00:43:01,239
Speaker 4: one of the fundamental challenges that utilities have is that

761
00:43:01,600 --> 00:43:07,040
Speaker 4: they struggle to invest. If there's no customer sitting there

762
00:43:07,160 --> 00:43:10,040
Speaker 4: you asking for the request, right, so they can't sort

763
00:43:10,040 --> 00:43:12,160
Speaker 4: of invest. I mean, if I'm in Vidia and I'm

764
00:43:12,440 --> 00:43:15,560
Speaker 4: thinking about the world five years from now and think, wow,

765
00:43:15,360 --> 00:43:18,440
Speaker 4: how many chips do I want to sell in twenty thirty,

766
00:43:18,640 --> 00:43:20,880
Speaker 4: I can go out and build a new factory. I

767
00:43:20,920 --> 00:43:23,239
Speaker 4: can go out and invest capital, and I can go

768
00:43:23,280 --> 00:43:25,319
Speaker 4: do all this. I mean, I don't need to have

769
00:43:25,360 --> 00:43:28,440
Speaker 4: an order from a Microsoft or an Amazon or a

770
00:43:28,520 --> 00:43:33,200
Speaker 4: Meta to go do that. I can build speculatively. Utilities

771
00:43:33,600 --> 00:43:37,160
Speaker 4: can't really do that. They're basically waiting for the customer

772
00:43:37,160 --> 00:43:39,279
Speaker 4: to come ask for it. But when you have all

773
00:43:39,320 --> 00:43:42,040
Speaker 4: this demand show up at the same time, well what happens.

774
00:43:42,080 --> 00:43:44,719
Speaker 4: The lead times start to extend and so instead of

775
00:43:44,760 --> 00:43:47,000
Speaker 4: saying yeah, I'll give you that power in a year

776
00:43:47,080 --> 00:43:48,560
Speaker 4: or two years. It's now like we I'll give it

777
00:43:48,560 --> 00:43:51,000
Speaker 4: to you in five to seven years. And so that's

778
00:43:51,000 --> 00:43:55,680
Speaker 4: an unsustainable way to run the electric utility grid. So

779
00:43:55,760 --> 00:43:59,319
Speaker 4: we do need regulators to adapt and evolve to this

780
00:43:59,360 --> 00:44:00,440
Speaker 4: new era of growth.

781
00:44:00,960 --> 00:44:03,840
Speaker 3: This is actually exactly something that I wanted to ask you,

782
00:44:03,880 --> 00:44:06,560
Speaker 3: which is we're sort of used to at this point

783
00:44:06,600 --> 00:44:10,879
Speaker 3: when we talk about industrial policy, the importance of an

784
00:44:11,120 --> 00:44:16,520
Speaker 3: end buyer for whatever capacity that we're building out, and utilities,

785
00:44:16,560 --> 00:44:19,680
Speaker 3: you know, to some degree, have struggled with that in

786
00:44:19,760 --> 00:44:23,440
Speaker 3: recent decades, at least this idea that they have huge

787
00:44:23,840 --> 00:44:29,880
Speaker 3: investment requirements. And while there is clearly demand for electricity

788
00:44:29,920 --> 00:44:32,800
Speaker 3: and maybe new types of electricity, it's not always certain

789
00:44:32,880 --> 00:44:35,400
Speaker 3: and you're sort of managing these day to day cycles

790
00:44:35,440 --> 00:44:38,680
Speaker 3: and things like that. But if we know that AI

791
00:44:38,920 --> 00:44:41,200
Speaker 3: is booming, and we know this is a future area

792
00:44:41,239 --> 00:44:45,120
Speaker 3: of growth, and we see these headlines like AI servers

793
00:44:45,160 --> 00:44:48,799
Speaker 3: are going to require like one hundred tarawatt hours per

794
00:44:48,880 --> 00:44:52,920
Speaker 3: year and things like that, does that potentially give utilities

795
00:44:53,160 --> 00:44:57,880
Speaker 3: more certainty or more confidence in the future investment outlook.

796
00:44:58,520 --> 00:45:02,000
Speaker 4: I mean, I would say in some respects it does.

797
00:45:02,120 --> 00:45:04,799
Speaker 4: I mean, they're certainly. And I've been spending a lot

798
00:45:04,880 --> 00:45:07,840
Speaker 4: of time with utilities well for most of my career,

799
00:45:07,880 --> 00:45:10,319
Speaker 4: but even in the last several months having this conversation

800
00:45:10,480 --> 00:45:14,600
Speaker 4: about how are they thinking about this future growth? And

801
00:45:15,320 --> 00:45:17,960
Speaker 4: you know, they're they're struggling a little bit because like,

802
00:45:18,280 --> 00:45:21,120
Speaker 4: all they know is what the customers, you know, show

803
00:45:21,200 --> 00:45:22,759
Speaker 4: up at their door and say that they want. Right,

804
00:45:22,800 --> 00:45:24,719
Speaker 4: they say, well, okay, I talk to X y Z

805
00:45:25,640 --> 00:45:28,120
Speaker 4: data center and this is what they say they want.

806
00:45:28,600 --> 00:45:32,360
Speaker 4: But they don't necessarily have view to the long term,

807
00:45:32,520 --> 00:45:35,560
Speaker 4: like what really is the demand behind that? Like I'm

808
00:45:35,560 --> 00:45:38,800
Speaker 4: getting a request because one data center bought one parsonal

809
00:45:38,880 --> 00:45:41,680
Speaker 4: Land and they need five hundred mega loots of power,

810
00:45:42,400 --> 00:45:44,239
Speaker 4: and then they're trying to extrapolate from that, well, what

811
00:45:44,320 --> 00:45:49,439
Speaker 4: is that underlying demand for data? Right? How much more

812
00:45:49,480 --> 00:45:51,840
Speaker 4: growth should I expect after that? And that's where the

813
00:45:51,960 --> 00:45:54,240
Speaker 4: utilities I think are really struggling is that they don't

814
00:45:54,440 --> 00:45:57,040
Speaker 4: they can't see much beyond the requests that they have,

815
00:45:57,960 --> 00:46:00,799
Speaker 4: and so they're trying to then extra appolid Okay, what

816
00:46:00,880 --> 00:46:03,520
Speaker 4: is it? What are these trends? You know? And really

817
00:46:03,560 --> 00:46:06,240
Speaker 4: the only the only way to get a good sense

818
00:46:06,320 --> 00:46:10,120
Speaker 4: of the the real demand for data and the trends

819
00:46:10,200 --> 00:46:12,800
Speaker 4: is you have to actually go back to the probably

820
00:46:12,800 --> 00:46:15,200
Speaker 4: to the NVIDIAs and the intels of the world and

821
00:46:15,239 --> 00:46:18,200
Speaker 4: go what's the forecast for chip sales? Like, what's the

822
00:46:18,239 --> 00:46:21,080
Speaker 4: forecast for how many chips you're going to make? Not

823
00:46:21,239 --> 00:46:22,960
Speaker 4: I mean not even sales, but really how much they

824
00:46:23,040 --> 00:46:25,680
Speaker 4: produce because frankly, I think every chip they can produce,

825
00:46:26,280 --> 00:46:27,880
Speaker 4: it will get plugged in something. Someone will buy it

826
00:46:27,920 --> 00:46:31,400
Speaker 4: and it will get plugged in. So that's that's probably

827
00:46:31,480 --> 00:46:33,520
Speaker 4: the best estimates that you can come up with for

828
00:46:34,360 --> 00:46:36,680
Speaker 4: what utility load growth should look like, at least as

829
00:46:36,719 --> 00:46:39,840
Speaker 4: it relates to data center. Right. But you know, you

830
00:46:39,920 --> 00:46:42,880
Speaker 4: have thousands of utilities in the United States, so you

831
00:46:42,880 --> 00:46:44,880
Speaker 4: don't have you know, there's not even like a single

832
00:46:44,920 --> 00:46:46,560
Speaker 4: source you can go to to say, Okay, what's the

833
00:46:46,600 --> 00:46:51,480
Speaker 4: forecast next year for electricity loads? Like nobody has that.

834
00:46:51,600 --> 00:46:54,480
Speaker 4: I mean people, there's numbers out there, but they're not

835
00:46:54,560 --> 00:46:58,640
Speaker 4: really based on anything other than speculation. So this is

836
00:46:58,680 --> 00:47:02,319
Speaker 4: the challenge that utilities is that they don't have a

837
00:47:02,360 --> 00:47:06,000
Speaker 4: good view into what load growth really is going to

838
00:47:06,000 --> 00:47:08,560
Speaker 4: look like over the next five, seven, ten years.

839
00:47:09,120 --> 00:47:13,200
Speaker 2: Brian Jenna's fascinating conversation. There's probably like ten more follow

840
00:47:13,280 --> 00:47:15,520
Speaker 2: ups that we could do specifically with you, and maybe

841
00:47:15,600 --> 00:47:17,839
Speaker 2: one day we'll do them. But in the meantime, thank

842
00:47:17,880 --> 00:47:20,239
Speaker 2: you so much for coming on, Odd Ladds. This is

843
00:47:20,239 --> 00:47:23,839
Speaker 2: a great conversation that we definitely needed to get done,

844
00:47:23,920 --> 00:47:25,200
Speaker 2: so really appreciate you joining it.

845
00:47:26,080 --> 00:47:27,920
Speaker 4: Thank you, Joe and Tracy. You really appreciate it.

846
00:47:41,280 --> 00:47:43,920
Speaker 2: Tracy, I thought that was great. I think actually the

847
00:47:43,960 --> 00:47:46,120
Speaker 2: first thing that sort of stands out from my mind

848
00:47:46,200 --> 00:47:49,799
Speaker 2: sort of like working backwards through the conversation, is just

849
00:47:50,000 --> 00:47:52,600
Speaker 2: sort of exactly what you talked about, which is that

850
00:47:52,640 --> 00:47:55,279
Speaker 2: there is this weird situation where you have this very

851
00:47:55,480 --> 00:47:59,000
Speaker 2: unpredictable demand. No one knows like what the steady state

852
00:47:59,360 --> 00:48:01,959
Speaker 2: demand is going to be for this stuff, and yet

853
00:48:02,000 --> 00:48:06,080
Speaker 2: the utilities are sort of legally constricted in the degree

854
00:48:06,080 --> 00:48:08,960
Speaker 2: to which they can say overbuild now or sort of

855
00:48:09,000 --> 00:48:11,520
Speaker 2: operate or plan for that demand.

856
00:48:11,600 --> 00:48:11,680
Speaker 4: No.

857
00:48:11,880 --> 00:48:15,160
Speaker 3: Absolutely, And also, well, going back to the beginning of

858
00:48:15,200 --> 00:48:18,480
Speaker 3: the conversation with Brian, the idea of a mismatch between

859
00:48:18,920 --> 00:48:22,480
Speaker 3: just how fast technology is going at the moment in

860
00:48:22,560 --> 00:48:26,680
Speaker 3: terms of developing AI versus utilities and their you know,

861
00:48:26,840 --> 00:48:30,839
Speaker 3: ten year investment programs that they need to get regulatory

862
00:48:30,920 --> 00:48:34,200
Speaker 3: approval for and all of that stuff. Now, there was

863
00:48:34,280 --> 00:48:37,160
Speaker 3: so much to pick out from that conversation. I also

864
00:48:37,200 --> 00:48:40,440
Speaker 3: thought it was interesting. So I think there is a

865
00:48:40,640 --> 00:48:43,640
Speaker 3: sense among a lot of commentators that there is going

866
00:48:43,680 --> 00:48:47,080
Speaker 3: to be competition for power at least at certain times.

867
00:48:47,360 --> 00:48:50,600
Speaker 3: But I thought Brian's point about how, in some respects

868
00:48:50,680 --> 00:48:55,040
Speaker 3: data centers might be the easy target for politicians to

869
00:48:55,160 --> 00:48:58,120
Speaker 3: kind of ignore, I thought that was really interesting. And

870
00:48:58,160 --> 00:49:01,200
Speaker 3: again his example of what, well, if you know, if

871
00:49:01,280 --> 00:49:04,360
Speaker 3: you're a governor or something, and there's a Tesla factory

872
00:49:04,360 --> 00:49:07,839
Speaker 3: that wants energy versus a data center that probably has

873
00:49:07,960 --> 00:49:10,600
Speaker 3: I don't know, like a handful of employees. Maybe that's

874
00:49:10,600 --> 00:49:13,680
Speaker 3: an exaggeration, then you're gonna go with the Tesla factory.

875
00:49:13,400 --> 00:49:15,520
Speaker 2: Right, totally, so right, You're not going to shut down

876
00:49:15,520 --> 00:49:18,839
Speaker 2: the factory that employs people. You're not gonna politically tell

877
00:49:18,880 --> 00:49:21,600
Speaker 2: people to go without air conditioning on a hot day.

878
00:49:21,960 --> 00:49:24,840
Speaker 2: The data center is going to be the first target.

879
00:49:25,120 --> 00:49:27,160
Speaker 2: I thought that was interesting. You know, again, I do

880
00:49:27,239 --> 00:49:30,160
Speaker 2: think it's like striking, and I think this is not

881
00:49:30,200 --> 00:49:32,759
Speaker 2: even just in the energy context, But I still a

882
00:49:33,000 --> 00:49:35,680
Speaker 2: sort of fascinated by this idea that like open Ai

883
00:49:36,040 --> 00:49:38,360
Speaker 2: was this company, I think it was founded in twenty sixteen,

884
00:49:38,400 --> 00:49:41,880
Speaker 2: and people saw GPT one and GPT two and then

885
00:49:41,960 --> 00:49:44,400
Speaker 2: GPT three, which came out before Chad GPT.

886
00:49:45,000 --> 00:49:45,480
Speaker 3: But it was.

887
00:49:45,480 --> 00:49:47,560
Speaker 2: Really like that day. I mean, it was like that

888
00:49:47,640 --> 00:49:51,200
Speaker 2: day that Chick GPT was announced. Even though like the

889
00:49:51,200 --> 00:49:53,680
Speaker 2: technology was in development, there are also theories and stuff.

890
00:49:53,719 --> 00:49:55,799
Speaker 2: It was like that day of the commercialization of the

891
00:49:55,800 --> 00:50:00,399
Speaker 2: productization of this technology where everyone woke up and all

892
00:50:00,400 --> 00:50:02,720
Speaker 2: these different companies like, we're in like a totally different

893
00:50:02,760 --> 00:50:05,040
Speaker 2: new world and we have to revisit all of these

894
00:50:05,080 --> 00:50:08,400
Speaker 2: investment decisions, whether it's on chips or energy that we

895
00:50:08,520 --> 00:50:09,879
Speaker 2: had made maybe just a year ago.

896
00:50:10,120 --> 00:50:10,319
Speaker 1: Yeah.

897
00:50:10,360 --> 00:50:14,480
Speaker 3: It's also it's almost like bullwhip effect isn't the right term,

898
00:50:14,560 --> 00:50:17,279
Speaker 3: but I'm just thinking the utilities in some respects are

899
00:50:17,440 --> 00:50:21,239
Speaker 3: at the very end of that sort of demand cycle, right,

900
00:50:21,320 --> 00:50:24,279
Speaker 3: So even the tech companies woke up to it very

901
00:50:24,360 --> 00:50:27,680
Speaker 3: very suddenly, the boom and AI and how fast this

902
00:50:27,880 --> 00:50:29,719
Speaker 3: was all going to come about and all of that,

903
00:50:29,920 --> 00:50:32,200
Speaker 3: and the utilities are sort of the last ones to

904
00:50:32,280 --> 00:50:35,760
Speaker 3: know in that respect, and we're expecting them to react

905
00:50:35,920 --> 00:50:37,840
Speaker 3: very quickly to it. It's kind of funny.

906
00:50:37,960 --> 00:50:39,960
Speaker 2: The other thing too, is like it'll be fastating to

907
00:50:40,000 --> 00:50:42,200
Speaker 2: see if some of these net zero commitments just have

908
00:50:42,280 --> 00:50:44,759
Speaker 2: to give, yeah, or something is going to happen there.

909
00:50:44,800 --> 00:50:46,640
Speaker 2: It sounds like it sounds like the rubber is going

910
00:50:46,680 --> 00:50:50,480
Speaker 2: to meet the road. But it does not sound like,

911
00:50:50,560 --> 00:50:53,080
Speaker 2: in the short term anyway, that there is a way

912
00:50:53,080 --> 00:50:58,400
Speaker 2: to accommodate this much increased demand with renewable energy. It

913
00:50:58,400 --> 00:51:02,000
Speaker 2: doesn't seem like it. And so like something is gonna

914
00:51:02,120 --> 00:51:03,560
Speaker 2: it seems like something's gonna have to give.

915
00:51:03,680 --> 00:51:06,880
Speaker 3: I think we're back to the very start of this conversation,

916
00:51:07,120 --> 00:51:09,040
Speaker 3: which is the idea of we have these two very

917
00:51:09,040 --> 00:51:13,360
Speaker 3: different paths where in an ideal world, if everything goes perfectly,

918
00:51:13,719 --> 00:51:16,839
Speaker 3: you have all this new commercial interest in technology that

919
00:51:16,880 --> 00:51:19,120
Speaker 3: requires a lot of energy usage, and so some of

920
00:51:19,160 --> 00:51:23,560
Speaker 3: those dollars get diverted into building out additional capacity in

921
00:51:23,640 --> 00:51:27,719
Speaker 3: terms of energy and maybe even additional green capacity. But

922
00:51:28,400 --> 00:51:31,040
Speaker 3: the other path is kind of depressing, where you have

923
00:51:31,080 --> 00:51:34,240
Speaker 3: a bunch of big tech companies that feel existential pressure

924
00:51:34,560 --> 00:51:36,920
Speaker 3: to do whatever it takes to win the AI race,

925
00:51:37,040 --> 00:51:41,400
Speaker 3: and maybe whatever it takes includes getting energy through coal

926
00:51:41,680 --> 00:51:42,719
Speaker 3: or something like that.

927
00:51:42,920 --> 00:51:45,200
Speaker 2: You know what I think is interesting and it's sort

928
00:51:45,200 --> 00:51:47,719
Speaker 2: of it hadn't really clicked to me, but Brian talked

929
00:51:47,760 --> 00:51:50,839
Speaker 2: about how, you know, after the early eighties, the US

930
00:51:50,960 --> 00:51:54,080
Speaker 2: basically stopped building nuclear and we're like, oh, you know,

931
00:51:54,120 --> 00:51:56,240
Speaker 2: it's like a big mistake. Why do we stop building nuclear?

932
00:51:56,440 --> 00:51:59,160
Speaker 2: But you could sort of understand it in the context

933
00:51:59,200 --> 00:52:01,960
Speaker 2: of very little growth, right, So why make these like

934
00:52:02,040 --> 00:52:06,399
Speaker 2: really big investments in anything when obviously at the time

935
00:52:06,480 --> 00:52:09,880
Speaker 2: there wasn't as much concern or awareness about climate change

936
00:52:10,080 --> 00:52:12,480
Speaker 2: and the effects of fossil fuels, and there just wasn't

937
00:52:12,560 --> 00:52:14,520
Speaker 2: much demand growth, So why make these big things? And

938
00:52:14,560 --> 00:52:17,399
Speaker 2: so you think about like South Korea and China having

939
00:52:17,440 --> 00:52:20,200
Speaker 2: never really slowed down on the nuclear construction, but they're

940
00:52:20,239 --> 00:52:23,680
Speaker 2: also because they're developing countries or poor countries becoming richer,

941
00:52:24,080 --> 00:52:28,880
Speaker 2: they never presumably had that sort of demand plateau just

942
00:52:28,920 --> 00:52:31,040
Speaker 2: by dint of having started from somewhere lower.

943
00:52:31,160 --> 00:52:33,760
Speaker 3: You know what we need what we need chat GPT

944
00:52:34,080 --> 00:52:37,799
Speaker 3: to design a small modular reactor, and then we need

945
00:52:37,800 --> 00:52:40,560
Speaker 3: a robot, then we need a robot yees to build it. Yeah,

946
00:52:40,640 --> 00:52:43,520
Speaker 3: all right, well it sounds like we're probably far away

947
00:52:43,520 --> 00:52:46,560
Speaker 3: from that. Maybe one day, Okay, shall we leave it there?

948
00:52:46,640 --> 00:52:47,359
Speaker 4: Let's leave it there.

949
00:52:47,520 --> 00:52:50,160
Speaker 3: This has been another episode of the All Blots podcast.

950
00:52:50,200 --> 00:52:53,279
Speaker 3: I'm Tracy Alloway. You can follow me at Tracy Alloway and.

951
00:52:53,200 --> 00:52:55,719
Speaker 2: I'm Joe Wisenthal. You can follow me at the Stalwart.

952
00:52:55,920 --> 00:52:59,200
Speaker 2: Follow our producers Carmen Rodriguez at Carman Erman dash Oll

953
00:52:59,200 --> 00:53:02,360
Speaker 2: Bennett at dash Spot, Killbrooks at Kilbrooks. Thank you to

954
00:53:02,400 --> 00:53:05,320
Speaker 2: our producer Moses onm. For more Oddlots content, go to

955
00:53:05,360 --> 00:53:08,000
Speaker 2: Bloomberg dot com slash odd Lots, where we have transcripts

956
00:53:08,000 --> 00:53:10,480
Speaker 2: of blog in the newsletter and you could chat about

957
00:53:10,520 --> 00:53:14,239
Speaker 2: all of these things, including AI, energy and climate in

958
00:53:14,320 --> 00:53:18,000
Speaker 2: our chatroom Discord Discord dot gg slash odd Lots twenty

959
00:53:18,040 --> 00:53:19,280
Speaker 2: four to seven with fellow listener.

960
00:53:19,640 --> 00:53:21,960
Speaker 3: And if you enjoy odd Lots, if you like it

961
00:53:22,000 --> 00:53:25,120
Speaker 3: when we dive into the energy usage of AI, then

962
00:53:25,200 --> 00:53:28,880
Speaker 3: please leave us a positive review on your favorite podcast platform.

963
00:53:29,160 --> 00:53:31,759
Speaker 3: And remember, if you are a Bloomberg subscriber, you can

964
00:53:31,800 --> 00:53:35,120
Speaker 3: listen to all of our episodes absolutely ad free. All

965
00:53:35,160 --> 00:53:37,640
Speaker 3: you need to do is connect your Bloomberg account with

966
00:53:37,840 --> 00:53:39,920
Speaker 3: Apple Podcasts. Thanks for listening,