1
00:00:00,440 --> 00:00:03,160
Speaker 1: This is Kamala Shelling, and you're listening to Switched on,

2
00:00:03,560 --> 00:00:06,840
Speaker 1: the podcast brought to you by Bloomberg and EF. Today,

3
00:00:06,840 --> 00:00:10,399
Speaker 1: we're bringing you another analyst reaction podcast. It's a mini

4
00:00:10,440 --> 00:00:13,560
Speaker 1: conversation about one specific event that we think says a

5
00:00:13,560 --> 00:00:16,439
Speaker 1: lot about the energy transition writ large, and it's based

6
00:00:16,520 --> 00:00:18,880
Speaker 1: on a type of story that BNF publishes on its

7
00:00:18,920 --> 00:00:21,400
Speaker 1: website and on the terminal in the hours after an

8
00:00:21,440 --> 00:00:25,160
Speaker 1: event occurs. On this episode, I'm joined by Aten Jane,

9
00:00:25,200 --> 00:00:28,000
Speaker 1: a specialist on BNF's WIN team, to talk about what

10
00:00:28,080 --> 00:00:31,200
Speaker 1: is maybe the marquee story of twenty twenty five AI

11
00:00:31,480 --> 00:00:34,960
Speaker 1: and data centers. Specifically, we're delving into his story next

12
00:00:35,000 --> 00:00:38,159
Speaker 1: era's AI leed gas pivot leaves space for renewables, and

13
00:00:38,200 --> 00:00:41,120
Speaker 1: we're talking about one utilities decision to start building new

14
00:00:41,159 --> 00:00:43,760
Speaker 1: gas capacity and what this shift can tell us about

15
00:00:43,760 --> 00:00:46,720
Speaker 1: the future balance between gas renewables and twenty four to

16
00:00:46,760 --> 00:00:49,839
Speaker 1: seven clean power as utilities race to serve data center's

17
00:00:50,200 --> 00:00:54,160
Speaker 1: massive electricity needs. BNF clients can find Athen's note and

18
00:00:54,200 --> 00:00:57,200
Speaker 1: other analyst reactions by heading to BNF go on the

19
00:00:57,240 --> 00:01:00,640
Speaker 1: Bloomberg Terminal, or at BNF dot com if you'd like

20
00:01:00,720 --> 00:01:03,520
Speaker 1: to learn more about how BNF approaches strategy research on

21
00:01:03,560 --> 00:01:07,720
Speaker 1: the energy transition, including developments in commodity markets, trends across

22
00:01:07,760 --> 00:01:11,080
Speaker 1: different sectors, and the cross cutting technologies shaping the future.

23
00:01:11,240 --> 00:01:14,240
Speaker 1: You can find more information on BNF dot com, and

24
00:01:14,280 --> 00:01:16,000
Speaker 1: if you'd like to speak with a member of our

25
00:01:16,040 --> 00:01:19,560
Speaker 1: team about becoming a client, email US at Sales dot

26
00:01:19,640 --> 00:01:23,240
Speaker 1: BNF at Bloomberg dot net. All right, let's get into

27
00:01:23,280 --> 00:01:26,640
Speaker 1: the conversation and hear how AI driven data center demand

28
00:01:26,840 --> 00:01:39,560
Speaker 1: is reshaping the way utilities plan their power mix. Aughtson,

29
00:01:39,760 --> 00:01:40,720
Speaker 1: welcome to switch down.

30
00:01:41,319 --> 00:01:42,080
Speaker 2: Thanks for having me.

31
00:01:42,560 --> 00:01:45,120
Speaker 1: Let's start with the headline of your react What is

32
00:01:45,200 --> 00:01:48,320
Speaker 1: driving next? There is renewed push toward gas generation and

33
00:01:48,360 --> 00:01:51,440
Speaker 1: what does that say about the broader energy landscape these days?

34
00:01:51,880 --> 00:01:55,280
Speaker 2: The short answer is really the speed, scale and reliability

35
00:01:55,320 --> 00:01:59,040
Speaker 2: needs for data centers. Data centers are the fastest doing

36
00:01:59,040 --> 00:02:01,600
Speaker 2: segment of POD demand in the US, and we'll account

37
00:02:01,680 --> 00:02:04,720
Speaker 2: for about twelve percent of the country's electricity usage by

38
00:02:04,800 --> 00:02:07,680
Speaker 2: twenty thirty five. That's a tripling from about focus cen

39
00:02:07,960 --> 00:02:11,560
Speaker 2: currently according to BNAF estimates. That is an extremely steep

40
00:02:11,800 --> 00:02:14,440
Speaker 2: ramp in a very short time window. NETCA has been

41
00:02:14,520 --> 00:02:17,680
Speaker 2: very explicit that we are entering what they are calling

42
00:02:17,880 --> 00:02:20,760
Speaker 2: a golden age for par demand with AI data centers

43
00:02:20,960 --> 00:02:23,560
Speaker 2: a major driver of future growth. And when really the

44
00:02:23,600 --> 00:02:28,200
Speaker 2: customers of netce are wanting multi gigabaut campuses built and

45
00:02:28,240 --> 00:02:31,160
Speaker 2: they warn them online fast, they're really running into a

46
00:02:31,160 --> 00:02:34,919
Speaker 2: practical constraint. Clean energy may be the cheapest, but sometimes

47
00:02:34,960 --> 00:02:37,280
Speaker 2: may take a little bit time to arrive at the

48
00:02:37,320 --> 00:02:42,120
Speaker 2: point of interconnection. Interconnection QS, transmission build timelines, and equipment

49
00:02:42,160 --> 00:02:45,920
Speaker 2: supply chains are all a binding constraint. So what's really

50
00:02:45,960 --> 00:02:49,600
Speaker 2: happening is not up pivot that renewals are over. It

51
00:02:49,680 --> 00:02:52,679
Speaker 2: is more like grid timeline is a problem. Gas is

52
00:02:52,680 --> 00:02:55,560
Speaker 2: one of the few tools that really help move at

53
00:02:55,560 --> 00:02:57,280
Speaker 2: the AI ready's speed.

54
00:02:57,639 --> 00:02:59,919
Speaker 1: You mentioned that attaching to the grid and the tie

55
00:03:00,120 --> 00:03:02,160
Speaker 1: line of the grid connection cues has long been a

56
00:03:02,200 --> 00:03:05,200
Speaker 1: problem for renewables that can stretch for years. Is it

57
00:03:05,240 --> 00:03:07,880
Speaker 1: actually faster to connect a new gas project to the

58
00:03:07,919 --> 00:03:11,839
Speaker 1: grid or it's nextstair and other independent power producers like them?

59
00:03:11,960 --> 00:03:14,280
Speaker 1: Are they using this as an opportunity to develop new

60
00:03:14,320 --> 00:03:17,600
Speaker 1: techniques of bringing generation to demand centers.

61
00:03:18,120 --> 00:03:22,000
Speaker 2: Well, it really depends on the location of the projects.

62
00:03:22,080 --> 00:03:24,880
Speaker 2: There are vast parts of the country where renewables are

63
00:03:24,919 --> 00:03:27,800
Speaker 2: still faster to connect then gas. But then there are

64
00:03:27,840 --> 00:03:31,560
Speaker 2: so many other locations, particularly the ones where people who

65
00:03:31,600 --> 00:03:34,880
Speaker 2: want to build these data centers where it's really challenging

66
00:03:34,920 --> 00:03:38,640
Speaker 2: to bring renewables faster, and maybe some gas capacity can

67
00:03:38,680 --> 00:03:40,960
Speaker 2: come faster with little grid upade. So it's a little

68
00:03:40,960 --> 00:03:43,920
Speaker 2: bit more nuanced in that sense. But when I kind

69
00:03:43,920 --> 00:03:46,360
Speaker 2: of come back to what NEXA is trying to do here,

70
00:03:46,480 --> 00:03:49,160
Speaker 2: essentially they're building some kind of like a data center

71
00:03:49,200 --> 00:03:52,480
Speaker 2: power product and not just trying to sell electrons to

72
00:03:52,560 --> 00:03:55,480
Speaker 2: the grid. One phase I keep hearing is bring your

73
00:03:55,520 --> 00:04:00,880
Speaker 2: own generation or BYOG, and that's more about hyperskis trying

74
00:04:00,920 --> 00:04:05,640
Speaker 2: to build these dedicated generation facilities that can be done

75
00:04:05,680 --> 00:04:08,520
Speaker 2: faster and closer to the load. Next. HERA has talked

76
00:04:08,560 --> 00:04:12,000
Speaker 2: about doing about fifteen gigawards of new generation capacity for

77
00:04:12,520 --> 00:04:15,760
Speaker 2: data center hubs by twenty thirty five, and if all

78
00:04:15,800 --> 00:04:19,039
Speaker 2: those the plan maybe increase that to about thirty gig awards.

79
00:04:19,120 --> 00:04:22,960
Speaker 2: They've also said that they've identified about twenty potential hubs

80
00:04:23,080 --> 00:04:27,160
Speaker 2: to do this kind of solution with these data center operators.

81
00:04:27,480 --> 00:04:31,120
Speaker 2: And importantly, gas has now become a major part of

82
00:04:31,160 --> 00:04:33,840
Speaker 2: how they make that fast part product career. They're targeting

83
00:04:33,880 --> 00:04:37,320
Speaker 2: about four to eight gigaward or new gas projects through

84
00:04:37,360 --> 00:04:39,600
Speaker 2: twenty thirty two as part of that build out plan.

85
00:04:39,960 --> 00:04:42,440
Speaker 1: It's one thing to have these major buildout plans in

86
00:04:42,480 --> 00:04:46,440
Speaker 1: the abstract, but does next stare have any concrete agreements

87
00:04:46,560 --> 00:04:48,920
Speaker 1: with hyperscalers or big tech companies who are looking to

88
00:04:48,920 --> 00:04:49,880
Speaker 1: build data centers.

89
00:04:50,200 --> 00:04:53,239
Speaker 2: Yeah, so next Heerra has very recently expanded some of

90
00:04:53,279 --> 00:04:56,880
Speaker 2: its partnerships with companies like Google and Meta to develop

91
00:04:56,920 --> 00:05:00,920
Speaker 2: these multi gigaward scale data center campuses, and these would

92
00:05:00,960 --> 00:05:04,159
Speaker 2: be paired with new generation capacity, plus a plan to

93
00:05:04,240 --> 00:05:07,520
Speaker 2: launch what they're describing as an AI driven product by

94
00:05:08,000 --> 00:05:11,640
Speaker 2: midnext year, which is aimed to improve the grid operations.

95
00:05:11,839 --> 00:05:15,160
Speaker 2: It recently signed about eleven PPS and DO energy storage

96
00:05:15,160 --> 00:05:18,240
Speaker 2: agreements tied to Meta. This is about two and a

97
00:05:18,279 --> 00:05:22,839
Speaker 2: half gigaward of clean energy contracts with projects expected to

98
00:05:22,880 --> 00:05:25,680
Speaker 2: come online between twenty twenty six and twenty twenty eight.

99
00:05:25,880 --> 00:05:29,120
Speaker 2: In short, next Serra's messages will wrap renewables and storage

100
00:05:29,120 --> 00:05:31,440
Speaker 2: around the load, but we will also deliver firm mega

101
00:05:31,480 --> 00:05:34,320
Speaker 2: awards that gets really the shovels for these data centers.

102
00:05:34,320 --> 00:05:35,240
Speaker 2: In the ground quickly.

103
00:05:35,560 --> 00:05:37,720
Speaker 1: You have a great line in your react which is

104
00:05:37,920 --> 00:05:40,479
Speaker 1: a renewables to gas tilt, and I like it because

105
00:05:40,520 --> 00:05:42,920
Speaker 1: it's not an about face, it's not a U turn. Instead,

106
00:05:42,920 --> 00:05:46,040
Speaker 1: it's a sense that things are shifting slightly while maintaining

107
00:05:46,080 --> 00:05:49,080
Speaker 1: the same course. So tell me more about this gas

108
00:05:49,160 --> 00:05:52,440
Speaker 1: tilt that is emerging as US utilities respond to AI

109
00:05:52,520 --> 00:05:53,279
Speaker 1: data center demand.

110
00:05:53,600 --> 00:05:57,279
Speaker 2: Really happening is a convergence of three things. Time to

111
00:05:57,320 --> 00:06:01,040
Speaker 2: power really is beating lease cost power. In what I

112
00:06:01,080 --> 00:06:04,360
Speaker 2: should describe as an early innings of this AI build out,

113
00:06:04,560 --> 00:06:08,039
Speaker 2: data centers really value twenty four to seven and very

114
00:06:08,200 --> 00:06:12,760
Speaker 2: location specific reliability, and they are increasingly willing to pay

115
00:06:12,800 --> 00:06:15,600
Speaker 2: the utilities for it. The grid is already constrained, as

116
00:06:15,600 --> 00:06:19,760
Speaker 2: we discuss interconnection delays and transmission scarcity really mean that

117
00:06:19,839 --> 00:06:23,440
Speaker 2: you can't always really deliver these clean megawards to the

118
00:06:23,560 --> 00:06:27,159
Speaker 2: exact notes you need to and on the timeline that

119
00:06:27,240 --> 00:06:29,560
Speaker 2: you're interested in. So you can see this all of

120
00:06:29,600 --> 00:06:32,560
Speaker 2: the above thinking not just at the utilities maybe, but

121
00:06:32,760 --> 00:06:36,200
Speaker 2: add the tech buiers themselves. Big tech is shifting towards

122
00:06:36,200 --> 00:06:39,080
Speaker 2: an all of the above strategy that includes renewables but

123
00:06:39,200 --> 00:06:43,360
Speaker 2: also faster deploying gas, more nuclear power, and other forms.

124
00:06:43,600 --> 00:06:47,200
Speaker 2: Given the delays and uncertainty in connections, it's not that

125
00:06:47,279 --> 00:06:51,000
Speaker 2: clean energy has stopped real penciling for these players, it's

126
00:06:51,040 --> 00:06:54,760
Speaker 2: that delivery risk has become the variable of some sorts.

127
00:06:55,240 --> 00:06:57,320
Speaker 1: So if we're looking at next Era as an entry

128
00:06:57,320 --> 00:07:00,520
Speaker 1: point to talk about a broader trend in the energy transition,

129
00:07:00,839 --> 00:07:03,279
Speaker 1: how does their approach compare to what the major peers

130
00:07:03,279 --> 00:07:04,240
Speaker 1: are doing well.

131
00:07:04,320 --> 00:07:07,800
Speaker 2: Nextc is still one of the most renewables scaled players,

132
00:07:08,080 --> 00:07:11,000
Speaker 2: but it is kind of converging with its spears on

133
00:07:11,240 --> 00:07:13,840
Speaker 2: some type of a similar strategy at play. That is

134
00:07:13,880 --> 00:07:16,960
Speaker 2: to secure firm capacity near the load first and then

135
00:07:17,000 --> 00:07:20,280
Speaker 2: maybe later pair it with renewables storage and all kinds

136
00:07:20,280 --> 00:07:24,200
Speaker 2: of contracting structures that can provide a bigger bridge in

137
00:07:24,240 --> 00:07:26,480
Speaker 2: the future. And yeah, you're right, this is not just

138
00:07:26,880 --> 00:07:32,520
Speaker 2: Nextterra story. It's a story beyond nex Sera Dominion, which

139
00:07:32,680 --> 00:07:35,480
Speaker 2: is the main utility that operates in Virginia where we

140
00:07:35,560 --> 00:07:38,600
Speaker 2: have that the world's famous data center Ali. They have

141
00:07:38,680 --> 00:07:42,040
Speaker 2: reported about forty gigaward of data center capacity in various

142
00:07:42,040 --> 00:07:47,120
Speaker 2: stages of contracting, highlighting how really the concentrated and massive

143
00:07:47,160 --> 00:07:49,920
Speaker 2: load pipeline is in its territory. It crans to add

144
00:07:49,960 --> 00:07:52,640
Speaker 2: about eight and a half gigawatts of new gas, along

145
00:07:52,680 --> 00:07:55,920
Speaker 2: with about seventeen and a half gigawaut of solar, three

146
00:07:55,920 --> 00:07:58,240
Speaker 2: and a half gig out of WIN and some storage.

147
00:07:58,400 --> 00:08:02,440
Speaker 2: Duke Energy another big uty. They have publicly disclosed agreements

148
00:08:02,440 --> 00:08:05,000
Speaker 2: to connect about two gigout of new data center capacity

149
00:08:05,160 --> 00:08:08,560
Speaker 2: and really expect material growth in the load later part

150
00:08:08,560 --> 00:08:11,440
Speaker 2: of the decade driven by these AI loads. Georgia Power,

151
00:08:11,600 --> 00:08:14,400
Speaker 2: another big Southern utility, projected about eight and a half

152
00:08:14,440 --> 00:08:16,880
Speaker 2: gig out of load growth over the next six years

153
00:08:16,920 --> 00:08:20,840
Speaker 2: in reporting around its planning processes, again really underscoring the

154
00:08:20,880 --> 00:08:24,520
Speaker 2: skale utilities are planning for. And there are some other

155
00:08:24,720 --> 00:08:29,040
Speaker 2: interesting things also happening. Earlier this year, Constellation Energy announced

156
00:08:29,040 --> 00:08:32,880
Speaker 2: they are acquiring Calpine in really signal that even clean

157
00:08:32,960 --> 00:08:36,480
Speaker 2: firm narratives, which Constellation has really championed, they are pairing

158
00:08:36,520 --> 00:08:40,120
Speaker 2: with flexible gas and large fleets essentially to tackle this

159
00:08:40,440 --> 00:08:43,720
Speaker 2: massive demand growth. So next set isn't really an outlier,

160
00:08:43,800 --> 00:08:47,080
Speaker 2: It's an example of a sector wide shift. Utilities are

161
00:08:47,120 --> 00:08:50,040
Speaker 2: really being paid to reduce time to power and they're

162
00:08:50,080 --> 00:08:52,760
Speaker 2: acting on these demand signals really.

163
00:08:52,800 --> 00:08:56,959
Speaker 1: How does this remixing of the combination of gas, renewables,

164
00:08:56,960 --> 00:09:00,439
Speaker 1: and firm clean resources fit into a broader net zero trichectory.

165
00:09:00,800 --> 00:09:03,880
Speaker 2: Well, the US is not really thinking a lot about

166
00:09:04,240 --> 00:09:06,720
Speaker 2: net zero right now, but if we still have to

167
00:09:06,800 --> 00:09:09,160
Speaker 2: kind of like consider what kind of a zero sub

168
00:09:09,200 --> 00:09:12,319
Speaker 2: scenario may play out into the future, we have to

169
00:09:12,400 --> 00:09:15,280
Speaker 2: kind of think of it in two time frames. In

170
00:09:15,320 --> 00:09:18,400
Speaker 2: the near term, what's really happening over the next decade,

171
00:09:18,559 --> 00:09:21,040
Speaker 2: gas will play a bigger role because it can be

172
00:09:21,120 --> 00:09:25,560
Speaker 2: delivered quickly and provide firm capacity while the great infrastructure

173
00:09:25,600 --> 00:09:29,040
Speaker 2: catches up for reneables. Even next, Serra's own projects describe

174
00:09:29,080 --> 00:09:33,640
Speaker 2: gas as a workhouse behind these early hub buildouts. And

175
00:09:33,679 --> 00:09:36,520
Speaker 2: then in the medium term that's like mid twenty thirties

176
00:09:36,520 --> 00:09:40,080
Speaker 2: and beyond, we see some kind of two reinforcing pushes.

177
00:09:40,360 --> 00:09:43,120
Speaker 2: One would be like a massive renewables and storage build

178
00:09:43,120 --> 00:09:46,200
Speaker 2: out because economics are strong, and then the another one

179
00:09:46,240 --> 00:09:49,959
Speaker 2: would be more firm keen procurement. Nuclear is the obvious

180
00:09:49,960 --> 00:09:52,559
Speaker 2: example here next Hera and Google. Earlier this year, is

181
00:09:52,600 --> 00:09:56,560
Speaker 2: announced their plan to restart tunr LL nuclear facility, which

182
00:09:56,559 --> 00:09:59,520
Speaker 2: has been dominant for some time, and so this kind

183
00:09:59,559 --> 00:10:04,040
Speaker 2: of a shift, we'll just accelerate beyond the trainy thirties

184
00:10:04,440 --> 00:10:07,439
Speaker 2: in the netzerio pathway. The direction of travel is clear

185
00:10:07,480 --> 00:10:11,320
Speaker 2: though that is unabated. Fossil fuel use must decline. So

186
00:10:11,360 --> 00:10:13,400
Speaker 2: if you build a lot of new gas now, the

187
00:10:13,520 --> 00:10:16,120
Speaker 2: key question that you're faced with is what's going to

188
00:10:16,160 --> 00:10:19,240
Speaker 2: be the utilization profile, what's a contracting structure, what's the

189
00:10:19,280 --> 00:10:24,920
Speaker 2: exit plan? And is a really credible decarbonization pathway like

190
00:10:25,200 --> 00:10:27,760
Speaker 2: CCS to support this build out.

191
00:10:27,920 --> 00:10:30,679
Speaker 1: And arten As a final question, I assume that data

192
00:10:30,679 --> 00:10:33,720
Speaker 1: center demand is not going to be going anywhere anytime soon.

193
00:10:33,920 --> 00:10:36,680
Speaker 1: So as somebody who watches this space, what are three

194
00:10:36,840 --> 00:10:38,920
Speaker 1: key things you think the rest of us should be

195
00:10:38,920 --> 00:10:40,840
Speaker 1: looking at going into twenty twenty six.

196
00:10:41,600 --> 00:10:45,640
Speaker 2: Well, the three things would be interconnection and transmission reformed

197
00:10:45,640 --> 00:10:49,839
Speaker 2: really because that's the long run enabler for renewables LED

198
00:10:49,960 --> 00:10:52,840
Speaker 2: solutions in the long term. Number two would be turbine

199
00:10:52,840 --> 00:10:56,200
Speaker 2: supply chains and delivery slots because that really constrains how

200
00:10:56,280 --> 00:10:59,360
Speaker 2: fast firm power can scale two and number three would

201
00:10:59,400 --> 00:11:02,560
Speaker 2: be the structure of hyperscular deeds, whether they are signing

202
00:11:02,760 --> 00:11:06,760
Speaker 2: long duration contracts or funding dedicated generation, and how they

203
00:11:06,760 --> 00:11:09,280
Speaker 2: are really thinking about matching their twenty four to seven

204
00:11:09,360 --> 00:11:12,120
Speaker 2: clean energy goals against physical delivery or powel.

205
00:11:12,600 --> 00:11:14,520
Speaker 1: Thank you so much for coming in today, Aton, It's

206
00:11:14,520 --> 00:11:15,600
Speaker 1: been a pleasure to speak with you.

207
00:11:15,800 --> 00:11:16,520
Speaker 2: Thanks for having me.

208
00:11:25,440 --> 00:11:28,560
Speaker 3: Today's episode of Switched On was produced by Cam Gray

209
00:11:28,760 --> 00:11:32,480
Speaker 3: with production assistants from Kamala Shelling. Bloomberg NEIF is a

210
00:11:32,520 --> 00:11:35,640
Speaker 3: service provided by Bloomberg Finance LP and its affiliates. This

211
00:11:35,720 --> 00:11:38,439
Speaker 3: recording does not constitute, nor should it be construed, as

212
00:11:38,480 --> 00:11:42,400
Speaker 3: investment advice, investment recommendations, or a recommendation as to an

213
00:11:42,400 --> 00:11:45,600
Speaker 3: investment or other strategy. Bloomberg ANIFF should not be considered

214
00:11:45,640 --> 00:11:48,920
Speaker 3: as information sufficient upon which to base an investment decision.

215
00:11:49,040 --> 00:11:52,000
Speaker 3: Neither Bloomberg Finance LP nor any of its affiliates makes

216
00:11:52,040 --> 00:11:55,760
Speaker 3: any representation or warranty as to the accuracy or completeness

217
00:11:55,760 --> 00:11:58,800
Speaker 3: of the information contained in this recording, and any liability

218
00:11:58,800 --> 00:12:01,480
Speaker 3: as a result of this recording is expressly disclaimed.

219
00:12:10,160 --> 00:12:10,480
Speaker 2: MMM