WEBVTT
1
00:00:07.269 --> 00:00:10.830
With Laurent Seguelin from London and Gerard Reid from Berlin,
2
00:00:11.310 --> 00:00:13.130
this is Redefining Energy.
3
00:00:13.509 --> 00:00:16.699
Today on Redefining Energy, we're going to talk about data centers.
4
00:00:16.730 --> 00:00:18.539
And in particular, we're going to look at the U.S.
5
00:00:18.570 --> 00:00:20.089
market and what's going on there.
6
00:00:20.710 --> 00:00:23.859
Gerard, if you look at the global numbers, it's insane.
7
00:00:24.699 --> 00:00:29.079
The worldwide data center CapEx is now expected to surpass
8
00:00:29.079 --> 00:00:32.880
$ 3 trillion by 2030, doubling what we have right now.
9
00:00:33.359 --> 00:00:36.619
It's the biggest GDP bet in America since the railroad.
10
00:00:37.229 --> 00:00:38.630
But first of all, from a partner.
11
00:00:39.130 --> 00:00:41.810
Right now, it feels like the energy world is one
12
00:00:41.950 --> 00:00:42.869
wild ride.
13
00:00:43.149 --> 00:00:44.590
Changing at dizzying speed.
14
00:00:45.210 --> 00:00:50.310
Twists, turns, steep climbs, sudden drops. But volatility doesn't have
15
00:00:50.369 --> 00:00:54.049
to slow you down. With the industry's largest proprietary data set,
16
00:00:54.670 --> 00:00:59.590
expert analysis, interconnected insight, and synoptic AI, Wood Mackenzie helps
17
00:00:59.609 --> 00:01:02.189
you quickly react to any sudden change.
18
00:01:02.570 --> 00:01:03.479
So you can be sure.
19
00:01:04.140 --> 00:01:04.859
Wood Mack short.
20
00:01:05.379 --> 00:01:07.819
Wood Mackenzie, Intelligence Connected.
21
00:01:08.519 --> 00:01:12.219
Back to the show. And it's burning cash. It's insane.
22
00:01:12.299 --> 00:01:16.609
So I've got a few numbers. KKR Credit estimate that
23
00:01:17.150 --> 00:01:22.349
AI-linked debt for data center, 600 billion. AI infrastructure company,
24
00:01:22.349 --> 00:01:24.469
40% of the S & P 500. AI-related companies, 49%
25
00:01:24.469 --> 00:01:31.849
of investment-grade bond insurance. And if you look at the
26
00:01:31.909 --> 00:01:36.250
private market, 87% of VC funding goes to AI. Of course,
27
00:01:36.290 --> 00:01:40.200
we've seen the entropic filing for IPO. I mean, it's wonderful.
28
00:01:40.200 --> 00:01:46.760
600 billion of commitment to Amazon, Google, and Microsoft XAI.
29
00:01:47.000 --> 00:01:48.799
The amount of cash is crazy.
30
00:01:49.420 --> 00:01:51.719
And I think it's really important to dig into this
31
00:01:51.780 --> 00:01:53.859
team because it's highly complex.
32
00:01:54.329 --> 00:01:54.890
Yeah, I know.
33
00:01:54.989 --> 00:01:57.510
And Laurent, as a result, it's really complex. important to
34
00:01:57.609 --> 00:02:00.109
dig into this theme. And that's why it was really
35
00:02:00.150 --> 00:02:02.659
great to get Eugene McGrain to come on the podcast
36
00:02:02.810 --> 00:02:04.780
to have a discussion around this topic.
37
00:02:05.379 --> 00:02:09.659
Eugene is executive managing director at Cushman Wakefield, which is
38
00:02:09.759 --> 00:02:13.300
one of the world's largest real estate broker. They used
39
00:02:13.319 --> 00:02:16.259
to do warehouses and they used to do office and
40
00:02:16.919 --> 00:02:21.740
somehow they acquired the expertise to do data centers. Now,
41
00:02:22.000 --> 00:02:24.819
when somebody say, okay, that data centers is worth a
42
00:02:24.879 --> 00:02:28.250
hundred billion. In fact, you get 50% chips, and then
43
00:02:28.289 --> 00:02:30.960
you get 20% power, and then the rest is cooling
44
00:02:31.000 --> 00:02:34.199
and everything. So it doesn't enter in the chips per se,
45
00:02:34.740 --> 00:02:38.599
but you organize everything that's around the world, the permitting,
46
00:02:38.639 --> 00:02:40.580
the connection, and so on. So I think it's very
47
00:02:40.620 --> 00:02:44.620
important to see that data center wave from the inside.
48
00:02:45.300 --> 00:02:45.800
Absolutely.
49
00:02:45.819 --> 00:02:48.860
So let's bring him on the show. Eugene, great to
50
00:02:48.879 --> 00:02:49.620
have you on the show.
51
00:02:50.319 --> 00:02:51.659
Thanks for having me.
52
00:02:51.680 --> 00:02:55.069
Eugene, a very simple question to start with. You are
53
00:02:55.330 --> 00:02:59.909
a real estate expert. How did you become a data
54
00:02:59.969 --> 00:03:01.289
center specialist?
55
00:03:01.990 --> 00:03:04.490
Well, like all things in life, a series of coincidence.
56
00:03:04.650 --> 00:03:07.400
I've been a commercial real estate broker for 20 plus
57
00:03:07.409 --> 00:03:10.840
years now. I started in California and I was at
58
00:03:10.879 --> 00:03:13.039
the beginning just an office broker, you know, working for
59
00:03:13.139 --> 00:03:16.199
corporations and dealing with their office leases, which are kind
60
00:03:16.219 --> 00:03:20.099
of just complex contract negotiations and market knowledge. Through that time,
61
00:03:20.180 --> 00:03:22.710
I worked for companies like PG & E and people
62
00:03:22.719 --> 00:03:26.169
like Chevron dealing with their real estate issues and assets.
63
00:03:26.889 --> 00:03:28.909
And I was very fortunate. I had a dear friend
64
00:03:28.930 --> 00:03:32.250
at Chevron, and he recommended that I, you know, I
65
00:03:32.289 --> 00:03:34.330
wanted to learn more about their business. He told me
66
00:03:34.370 --> 00:03:37.110
to read Daniel Juergen's book. So three books later, I
67
00:03:37.259 --> 00:03:39.599
felt like I knew enough about the energy business to
68
00:03:39.639 --> 00:03:43.639
not know much. And in 2021, my wife, who's from
69
00:03:43.699 --> 00:03:46.460
Rhode Island, needed to get home. And so we moved
70
00:03:46.520 --> 00:03:50.719
our family across the country. in that very unsettled time
71
00:03:50.740 --> 00:03:52.840
in the world where, you know, people were not signing
72
00:03:52.879 --> 00:03:57.500
new office leases or, you know, warehouse leases with much certainty.
73
00:03:57.520 --> 00:04:01.319
The political climate of the time made a giant push
74
00:04:01.419 --> 00:04:04.639
for offshore wind. And there was a commitment to do
75
00:04:04.639 --> 00:04:08.900
35 gigawatts of power and deliverables and contracts were done
76
00:04:08.939 --> 00:04:13.280
and investors were attracted. And, you know, with my background,
77
00:04:13.599 --> 00:04:16.240
I could see two things. One, that this was real,
78
00:04:16.519 --> 00:04:19.410
that the investments were going to be made. And two,
79
00:04:19.810 --> 00:04:23.089
that the companies, especially the Danes, were coming over with
80
00:04:23.129 --> 00:04:26.629
a completely different understanding of the power grid and how
81
00:04:26.709 --> 00:04:30.149
development worked here as to how it worked in Europe.
82
00:04:30.560 --> 00:04:34.019
And so I started advising those companies on what the
83
00:04:34.079 --> 00:04:36.300
kind of different obligations were going to be in order
84
00:04:36.319 --> 00:04:39.040
to reach the development. But they had interconnections going through
85
00:04:39.060 --> 00:04:41.839
towns that they didn't really understand the permitting process and
86
00:04:41.899 --> 00:04:45.000
what the obstacles were going to be. And so as
87
00:04:45.100 --> 00:04:47.000
I was doing that, I was doing a lot of
88
00:04:47.079 --> 00:04:50.860
infrastructure level work. power because that's what those developments were.
89
00:04:51.560 --> 00:04:54.800
And then when ChatGPT hit the market and our kind
90
00:04:54.839 --> 00:04:57.180
of worlds changed, I just happened to be one of
91
00:04:57.220 --> 00:05:00.310
those people that was working with companies that were doing 200,
92
00:05:00.310 --> 00:05:05.060
300 megawatt projects and kind of understood the basic stuff
93
00:05:05.079 --> 00:05:08.500
that those people did, which was not typical in the
94
00:05:08.540 --> 00:05:10.660
data center world prior to that. You know, a large
95
00:05:10.699 --> 00:05:14.720
requirement was five megawatts, 20 megawatts. all of a sudden
96
00:05:14.740 --> 00:05:18.319
we were looking at 200 megawatts was, oh, maybe that's
97
00:05:18.360 --> 00:05:21.560
not enough. So it was 200 to 500 to gigawatt
98
00:05:21.579 --> 00:05:25.639
to now what it is. And so the expertise that
99
00:05:25.660 --> 00:05:28.740
I've been able to grow and build is one that's
100
00:05:28.920 --> 00:05:32.879
based in real estate and complexities of developing property, the
101
00:05:32.939 --> 00:05:36.589
financing aspects of developing property, what has to happen, how
102
00:05:36.610 --> 00:05:39.009
the steps go, what you have to get through to
103
00:05:40.290 --> 00:05:43.600
somebody that kind of understood the PPAs and process that
104
00:05:43.720 --> 00:05:46.180
power companies go through. So I just happen to have
105
00:05:46.199 --> 00:05:48.079
kind of a unique set of where I am. And
106
00:05:48.120 --> 00:05:51.170
my partners, who are very, very good at the data
107
00:05:51.189 --> 00:05:53.730
center end of this, just partner with them, and I
108
00:05:53.769 --> 00:05:55.529
sit in between, and I think about this stuff all
109
00:05:55.569 --> 00:05:57.490
the time. So that's who I am. That's how I
110
00:05:57.509 --> 00:05:57.829
got here.
111
00:05:58.529 --> 00:06:01.009
Yeah, it's a great story. So in a way, you
112
00:06:01.029 --> 00:06:03.449
are the right man at the right place. Because I
113
00:06:03.490 --> 00:06:05.480
think all the data center guys, they're good in the
114
00:06:05.509 --> 00:06:08.240
soft thing, but when it comes to hard, the hard part,
115
00:06:08.319 --> 00:06:12.329
which is the walls, the terrain, the the relationship with
116
00:06:12.370 --> 00:06:17.170
the authorities and the connections, you have done that before.
117
00:06:17.470 --> 00:06:21.199
How fast has your practice grown in the past three years?
118
00:06:22.060 --> 00:06:25.620
The amount of time I spend on this has grown significantly.
119
00:06:26.319 --> 00:06:31.220
These projects are so complex and so large that from
120
00:06:31.259 --> 00:06:34.610
a percentage-wise, they're like 100% of my practice, but they
121
00:06:34.709 --> 00:06:37.750
take so long to get there that if I do
122
00:06:37.850 --> 00:06:39.670
two or three of these or four or five of these,
123
00:06:40.430 --> 00:06:42.730
That's a lot of them. These are sort of kind
124
00:06:42.769 --> 00:06:47.089
of career-sized projects to work on. It's scaled significantly.
125
00:06:47.810 --> 00:06:49.610
Eugene, can you just talk a little bit about the
126
00:06:50.129 --> 00:06:54.310
hype in this whole area and how you navigate through
127
00:06:54.329 --> 00:06:55.990
that and what your thoughts are on it?
128
00:06:56.930 --> 00:06:59.930
Well, it's interesting because it's evolved. When this first happened,
129
00:07:00.149 --> 00:07:04.750
every farmer and every brownfield owner of property in the
130
00:07:04.829 --> 00:07:07.449
country all of a sudden told everybody that they had
131
00:07:07.449 --> 00:07:11.459
200 megawatts or 500 megawatts or some level of capacity,
132
00:07:11.500 --> 00:07:14.350
or they could get interconnection because they knew a guy
133
00:07:14.389 --> 00:07:18.949
at the PUC or whatever. And because of the size
134
00:07:18.970 --> 00:07:21.949
and complexity of these assets, I mean, we're almost at
135
00:07:21.970 --> 00:07:24.709
the other end of that now, in my opinion, where
136
00:07:24.829 --> 00:07:28.889
the hype is more narrative-driven than it is on the
137
00:07:28.970 --> 00:07:31.730
actual ground of people that are making decisions and people
138
00:07:31.769 --> 00:07:35.089
that are writing checks. This is now the part where
139
00:07:35.769 --> 00:07:40.470
the really sophisticated developers have taken it to a place
140
00:07:40.550 --> 00:07:43.269
where they have to walk through the individual steps and
141
00:07:43.629 --> 00:07:47.569
the press is against them because the story is against them.
142
00:07:48.250 --> 00:07:52.699
So the good comparison is essentially where the fracking business
143
00:07:52.740 --> 00:07:56.259
came from. First, it was this energy revolution to an
144
00:07:56.459 --> 00:08:00.560
environmental disaster to what it is now, which is really
145
00:08:01.000 --> 00:08:04.560
kind of energy independence and changing the map of the world.
146
00:08:05.110 --> 00:08:09.230
The press and the public sentiment became so strong against
147
00:08:09.310 --> 00:08:11.930
these projects that they really had to fine tune what
148
00:08:11.949 --> 00:08:14.189
the technology was and they had to fine tune what
149
00:08:14.209 --> 00:08:18.730
the message was. So the height now is much more
150
00:08:19.490 --> 00:08:22.370
in the financial markets than it is on the ground.
151
00:08:23.500 --> 00:08:26.600
That makes sense to me. And maybe I'd just follow
152
00:08:26.620 --> 00:08:28.339
it on from that then. When you look at it,
153
00:08:28.759 --> 00:08:30.060
I suppose I'm trying to get my head around.
154
00:08:30.339 --> 00:08:30.860
Do you see the.
155
00:08:30.819 --> 00:08:33.740
Glass as half full or half empty when you look
156
00:08:33.759 --> 00:08:34.759
at the potential, right?
157
00:08:35.789 --> 00:08:38.370
I'm a pessimistic optimist. So one of the questions that
158
00:08:38.389 --> 00:08:40.750
we always get and that I've answered in a bunch
159
00:08:40.789 --> 00:08:43.710
of places and I think is an ongoing question is,
160
00:08:44.269 --> 00:08:47.580
are we building data centers and infrastructure? Or is this
161
00:08:47.620 --> 00:08:51.879
the 1870s railroad boom? Or is this some sort of
162
00:08:52.559 --> 00:08:55.139
critical resource that we're building and this is going to
163
00:08:55.179 --> 00:08:58.620
define the next 50 years? We're somewhere in the middle.
164
00:08:58.840 --> 00:09:00.919
There is hype in the sector. We talked a bit
165
00:09:00.940 --> 00:09:03.980
about some of the players that have emerged as people
166
00:09:04.019 --> 00:09:07.529
that can deliver these products. But the vacancy rate, which is,
167
00:09:07.769 --> 00:09:09.990
you know, as a real estate guy, the first thing
168
00:09:10.009 --> 00:09:12.289
you look at is the vacancy rate. That tells you
169
00:09:12.330 --> 00:09:15.139
what kind of delivery there is. The vacancy rate in
170
00:09:15.179 --> 00:09:17.879
these large scale, let's call them 200 megawatts and above
171
00:09:18.139 --> 00:09:21.919
projects is less than 1%. The overall market of the
172
00:09:21.980 --> 00:09:26.269
Americas is 3.4%. And that includes everything, which is a
173
00:09:26.330 --> 00:09:29.570
lot of legacy stuff that can't really be repositioned to
174
00:09:29.649 --> 00:09:33.830
deal with the kind of higher compute requirements. So if
175
00:09:33.870 --> 00:09:37.679
there's a collapse in demand for compute, that would be
176
00:09:38.179 --> 00:09:41.000
pretty fatal for the industry. I don't believe we're going
177
00:09:41.019 --> 00:09:43.039
to see a collapse in demand for compute. We may
178
00:09:43.080 --> 00:09:47.500
see a collapse in paying for compute, and that may
179
00:09:47.879 --> 00:09:51.610
rattle the people that have done the infrastructure investments. But
180
00:09:52.190 --> 00:09:54.809
this is stuff that, no matter what happens, is going
181
00:09:54.830 --> 00:09:59.230
to be used. It's equivalent to the fiber build-out of
182
00:09:59.269 --> 00:10:02.759
the late 90s. 90s, where fiber got built everywhere because
183
00:10:03.059 --> 00:10:06.980
there was a supposedly endless demand for fiber. Well, it
184
00:10:07.019 --> 00:10:08.539
sat in the ground in a lot of places for
185
00:10:08.580 --> 00:10:11.990
a long time, and now it's being used. So glass
186
00:10:12.009 --> 00:10:15.129
half full, glass half empty. I think we have a
187
00:10:15.590 --> 00:10:19.809
near insatiable demand for compute. We have a very limited
188
00:10:19.850 --> 00:10:23.169
ability to deliver it at the scale that the demand
189
00:10:23.230 --> 00:10:26.889
is asking for. And this is essentially the infrastructure build
190
00:10:26.970 --> 00:10:30.519
out to meet that demand. Will there be carnage in
191
00:10:30.740 --> 00:10:34.259
the financials? Certainly. Somebody's going to lose something somewhere, but
192
00:10:34.279 --> 00:10:35.559
this is not pets.com.
193
00:10:36.539 --> 00:10:38.440
Eugene, I like the way you also started there because
194
00:10:39.019 --> 00:10:42.120
You also put it in the context of the railroad
195
00:10:42.179 --> 00:10:45.100
build-out and all that type of stuff. That was a revolution,
196
00:10:45.139 --> 00:10:47.779
a transport revolution. I think that we're going through the
197
00:10:47.820 --> 00:10:51.090
exact same thing. There are going to be winners and losers,
198
00:10:51.210 --> 00:10:53.350
but the bottom line is you just need to have
199
00:10:53.370 --> 00:10:53.919
to be careful.
200
00:10:54.990 --> 00:10:55.769
Do sensible things.
201
00:10:55.830 --> 00:10:56.850
Don't do stupid things.
202
00:10:57.370 --> 00:11:01.149
It's always easier to do stupid things. But every day
203
00:11:01.210 --> 00:11:03.389
I read the headlines, and then every day I answer
204
00:11:03.429 --> 00:11:06.500
my phone. And the headlines on the phone don't match up.
205
00:11:06.950 --> 00:11:09.129
in a lot of ways, which is the demand continues
206
00:11:09.169 --> 00:11:12.299
to be very strong for the needs for this. And
207
00:11:12.320 --> 00:11:16.580
then everybody you're dealing with is using AI. It's incredibly useful.
208
00:11:16.639 --> 00:11:21.139
We've been generating data for 30 plus years that allows
209
00:11:21.179 --> 00:11:24.370
for real productivity. You know, everybody wants better information, and
210
00:11:24.419 --> 00:11:25.669
this is the tool to get you there.
211
00:11:26.330 --> 00:11:29.970
If I go back to power and real estate, I remember,
212
00:11:30.049 --> 00:11:33.409
because I'm an old man, 27 years ago, Mark Mills...
213
00:11:35.399 --> 00:11:40.710
Mark Mills, who is worshipped every time he goes to Sarawak, said,
214
00:11:41.769 --> 00:11:44.990
so end of the 90s, by 2010, half of the
215
00:11:45.009 --> 00:11:49.009
electricity will be used by the internet. Okay. And basically
216
00:11:49.029 --> 00:11:52.840
that triggered a lot of construction of gas and so on,
217
00:11:52.879 --> 00:11:57.200
which then led to the collapse of Calpine and all
218
00:11:57.220 --> 00:11:59.799
the bad. He was the big evangelist at the time.
219
00:11:59.840 --> 00:12:02.379
He's still around. So now, of course, he's saying we
220
00:12:02.399 --> 00:12:05.210
need to multiply power by a factor of 10 and Now,
221
00:12:05.230 --> 00:12:08.440
I'm interested, if I look from a basic point of view,
222
00:12:08.960 --> 00:12:13.200
NVIDIA chips. So you had the Hopper, then the Blackwell,
223
00:12:13.639 --> 00:12:17.340
now the Rubin, and we're going next generation to the Fineman.
224
00:12:17.759 --> 00:12:21.059
The energy efficiency of the Fineman is going to be
225
00:12:21.059 --> 00:12:25.399
500 times more than the Hopper, which goes to two questions.
226
00:12:26.169 --> 00:12:28.710
The number one is, what's going to be the value,
227
00:12:28.929 --> 00:12:31.909
especially if they have those 20-year contract of a Hopper
228
00:12:32.090 --> 00:12:36.610
data center in 20 years? considering the speed of the improvement.
229
00:12:37.309 --> 00:12:40.149
And the second one is the guy everybody loves to hate,
230
00:12:40.299 --> 00:12:42.720
but is an engineering genius, is Elon Musk. If I
231
00:12:42.759 --> 00:12:45.100
look at the development of Colossus, now we've got this
232
00:12:45.179 --> 00:12:49.340
new building called Mini Hard. And Mini Hard is about
233
00:12:49.740 --> 00:12:52.090
a quarter of the size of the one he built
234
00:12:52.110 --> 00:12:54.929
two years ago, and it's even more powerful. It's very
235
00:12:54.970 --> 00:12:58.450
difficult to assess, but that's something that the hyperscaler industry
236
00:12:58.470 --> 00:13:02.429
has done. The energy efficiency gains that people are going
237
00:13:02.450 --> 00:13:06.200
to do which for me is a bigger risk than
238
00:13:06.240 --> 00:13:08.860
the demand for compute. The question is, do you have
239
00:13:08.899 --> 00:13:12.100
this type of thinking or no, as long as the
240
00:13:12.139 --> 00:13:14.919
telephone rings, everything's okay? No, no.
241
00:13:15.100 --> 00:13:19.440
The answer is the projects themselves are the limiting factor
242
00:13:19.480 --> 00:13:23.919
on the growth. And because of the size of the projects,
243
00:13:24.639 --> 00:13:27.940
because of the complexity, because of the software guys have
244
00:13:28.059 --> 00:13:31.039
never had the plan for power before. They just iterate
245
00:13:31.120 --> 00:13:34.350
the next level of software. And it's just been easy.
246
00:13:34.669 --> 00:13:38.190
So they've never had to do this hard infrastructure level work.
247
00:13:38.789 --> 00:13:42.210
Meta at one point wanted to build its own Bayside
248
00:13:42.409 --> 00:13:45.870
mass transit rail station to run from Palo Alto up
249
00:13:45.909 --> 00:13:48.720
to San Francisco. And they hired a whole team. They
250
00:13:48.740 --> 00:13:51.940
did all the things that would build an actual rail system.
251
00:13:52.700 --> 00:13:55.840
And they got nowhere because it involved things that are,
252
00:13:56.440 --> 00:13:59.240
you've got to get approvals, you've got to get Materials,
253
00:13:59.299 --> 00:14:03.220
like all of the hard structural costs that are involved
254
00:14:03.240 --> 00:14:07.240
with infrastructure, are in themselves a limiting factor in the
255
00:14:07.360 --> 00:14:11.820
speed of this sort of build. Now, this feels like
256
00:14:11.879 --> 00:14:15.029
it's moving at some level of warp speed, but I
257
00:14:15.070 --> 00:14:18.669
would make the argument that it is basically rate limited
258
00:14:19.649 --> 00:14:23.490
or expansion limited by the funnel of, you can't get
259
00:14:23.509 --> 00:14:26.899
a turbine from GE Vranova or Hitachi for three years.
260
00:14:27.000 --> 00:14:29.440
You can't get a transformer from those same people for
261
00:14:29.720 --> 00:14:32.320
two or three years. There's a queue. So the demand
262
00:14:32.379 --> 00:14:35.519
is so strong and the inputs have so much trouble
263
00:14:35.570 --> 00:14:39.149
flexing up that I don't look at it as, oh,
264
00:14:39.230 --> 00:14:43.149
we're 50 gigawatts ahead of where compute demand is going
265
00:14:43.169 --> 00:14:46.529
to be in 10 years. I think where we are
266
00:14:46.730 --> 00:14:49.850
is we're probably par or less than par of where
267
00:14:49.870 --> 00:14:53.049
we're going to be in five to 10 years especially
268
00:14:53.210 --> 00:14:56.049
if the chips are able to get more efficient, because
269
00:14:56.080 --> 00:14:58.200
I don't think you're going to see compute demand drop.
270
00:14:58.580 --> 00:15:01.659
What we've had is it's worth looking at the first
271
00:15:01.700 --> 00:15:04.799
internet boom, because that is kind of what's informing a
272
00:15:04.860 --> 00:15:07.720
lot of people's prejudices on this. It certainly informs mine.
273
00:15:07.960 --> 00:15:10.419
I grew up in San Francisco and watched the first
274
00:15:10.500 --> 00:15:13.720
dot-com boom. That was based so much on software.
275
00:15:13.759 --> 00:15:14.379
It was free.
276
00:15:14.460 --> 00:15:17.200
Everybody could figure out how to write a new website.
277
00:15:18.320 --> 00:15:20.820
And the money that was being thrown at those companies
278
00:15:20.860 --> 00:15:23.460
was being thrown at companies that were just basically ideas.
279
00:15:23.620 --> 00:15:26.419
They didn't have to go out and really do anything.
280
00:15:27.450 --> 00:15:31.210
In order to get this level of compute that's demanded today,
281
00:15:32.009 --> 00:15:35.289
you've got to build infrastructure that will be valuable in
282
00:15:35.289 --> 00:15:38.169
20 years. And if it becomes a factor of the
283
00:15:38.320 --> 00:15:41.360
chips continue to get more energy efficient, it just allows
284
00:15:41.399 --> 00:15:43.679
them to meet. Now, there's going to be some overlap.
285
00:15:43.740 --> 00:15:45.340
There's going to be a tipping point where we get
286
00:15:45.379 --> 00:15:47.940
beyond that. I don't think we're there. And I think
287
00:15:48.000 --> 00:15:50.980
that what we'll see is projects get abandoned as opposed
288
00:15:51.019 --> 00:15:53.519
to projects being built and then not being used.
289
00:15:54.779 --> 00:15:56.399
What I just like to add to that, to put
290
00:15:56.450 --> 00:16:01.190
it into context, is that Google, just Google alone, Google
291
00:16:01.289 --> 00:16:05.889
has had a sevenfold increase in demand for processing power
292
00:16:05.909 --> 00:16:10.149
in the last 12 months. 700%. Now, of course, what
293
00:16:10.190 --> 00:16:11.710
you don't have is 700% increase in energy demand, but
294
00:16:14.289 --> 00:16:16.779
But let's say you have a doubling of energy demand
295
00:16:16.820 --> 00:16:19.100
because of energy efficiency and new chips and this, that,
296
00:16:19.120 --> 00:16:21.320
and the other thing. The point I'm making is that's
297
00:16:21.360 --> 00:16:23.840
going to continue. It mightn't be seven-fold next year, but
298
00:16:23.860 --> 00:16:28.250
let's say it doubles. That's what's driving demand. And that
299
00:16:28.309 --> 00:16:30.789
is all of us sitting there. And I don't use
300
00:16:30.970 --> 00:16:34.629
Google anymore. I use ChatGBT to do my searches. I
301
00:16:34.669 --> 00:16:37.610
want them to do an invitation for something.
302
00:16:37.669 --> 00:16:38.149
Guess what?
303
00:16:38.269 --> 00:16:38.850
I'm using it.
304
00:16:39.230 --> 00:16:42.559
That's the bottom line. We are driving as consumers demand.
305
00:16:43.429 --> 00:16:48.080
massive demand for data because it's useful. And I think
306
00:16:48.139 --> 00:16:49.980
it's just going to become more useful. Yeah.
307
00:16:50.820 --> 00:16:54.860
Eugene, as you said, there is this crazy thirst for
308
00:16:55.440 --> 00:16:58.470
more compute, more buildings and more chips, more everything. And
309
00:16:58.500 --> 00:17:01.210
of course the turbines, I mean, we know that, but
310
00:17:01.429 --> 00:17:05.910
you as a professional, when your telephone rings, And it's
311
00:17:06.089 --> 00:17:10.009
a new client or new developer. How do you assess
312
00:17:10.150 --> 00:17:15.039
how serious that this deal will go through rather than
313
00:17:15.099 --> 00:17:17.920
just pie in the sky? You have a secret sauce here.
314
00:17:18.480 --> 00:17:20.700
My first boss always answered the phone. You never know
315
00:17:20.720 --> 00:17:22.160
when somebody is going to want to buy the Empire
316
00:17:22.200 --> 00:17:24.880
State Building. I've been trained with a certain level of
317
00:17:25.000 --> 00:17:28.720
optimism in answering my phone. These projects, one, they're so
318
00:17:28.759 --> 00:17:31.490
much more interesting for me to work on than a
319
00:17:31.529 --> 00:17:35.440
million square foot warehouse for Amazon. It does a great project.
320
00:17:35.460 --> 00:17:38.690
They're very lucrative. But even at the most technical, it's
321
00:17:38.789 --> 00:17:41.829
just high bay warehouse. An office building is just a
322
00:17:42.029 --> 00:17:45.410
lease with a subordination, non-disturbance, and atonement act. It's nothing
323
00:17:45.950 --> 00:17:49.019
real great shakes. This is a real technical issue. So
324
00:17:49.359 --> 00:17:52.660
the way you qualify these things is the level of
325
00:17:52.700 --> 00:17:55.200
professionalism on the other side. One of the groups we're
326
00:17:55.240 --> 00:17:57.880
working for, the way that I was introduced to them
327
00:17:57.960 --> 00:18:01.730
was the former head of BP's trading desk went to
328
00:18:01.769 --> 00:18:03.829
go work for them. That, to me, spoke of a
329
00:18:03.869 --> 00:18:06.430
level of professionalism and a level of understanding of the
330
00:18:06.470 --> 00:18:11.349
market and the economies of scale. That was a real person.
331
00:18:11.410 --> 00:18:14.849
And I think that's something that we should understand. My
332
00:18:14.910 --> 00:18:16.619
friend who I was on the phone with yesterday was
333
00:18:16.650 --> 00:18:20.799
a nuclear engineer for Dominion before he went over and
334
00:18:20.839 --> 00:18:23.019
worked for Meta as one of their heads of energy.
335
00:18:23.319 --> 00:18:26.240
They're not hiring guys like me to run these things.
336
00:18:26.279 --> 00:18:28.240
The guys like me are figuring out how to get
337
00:18:28.279 --> 00:18:30.690
them delivered. But the quality of the people that are
338
00:18:30.910 --> 00:18:33.950
involved in this are really world-class. And if you don't
339
00:18:34.049 --> 00:18:36.549
have that group of people, if you don't have the
340
00:18:36.589 --> 00:18:39.829
market expert to get into, we have a team. I
341
00:18:40.509 --> 00:18:43.710
have a broker in Dallas who has been doing data
342
00:18:43.750 --> 00:18:46.720
centers in Dallas for 30 years and worked for HP
343
00:18:46.799 --> 00:18:50.200
when HP was building out their first infrastructure. You have
344
00:18:50.240 --> 00:18:52.720
to be willing to assemble that kind of group. That's
345
00:18:52.779 --> 00:18:56.680
the qualification. In the United States, we're broken up into ISOs.
346
00:18:56.920 --> 00:18:59.799
And the ISOs we're getting I wouldn't call it artificial demand,
347
00:18:59.819 --> 00:19:01.660
but they were getting a lot of, hey, maybe this
348
00:19:01.680 --> 00:19:04.460
will happen in their queues because people wanted to just
349
00:19:04.500 --> 00:19:07.309
try to get it placed in line. They've all basically
350
00:19:07.369 --> 00:19:09.910
come back and said, you need to pay to play
351
00:19:10.150 --> 00:19:12.789
and you need to pay real money. So ERCOT's charging
352
00:19:12.970 --> 00:19:16.269
quite a bit of money that's non-refundable if you want
353
00:19:16.289 --> 00:19:19.329
to place on their interconnection queue. And you start talking
354
00:19:19.369 --> 00:19:23.369
about 300 or 500 or 1,000 megawatts, you start talking
355
00:19:23.410 --> 00:19:26.609
about having to put real money into the utility system.
356
00:19:26.670 --> 00:19:29.640
So That's a lot of what drives how serious I
357
00:19:29.660 --> 00:19:32.539
take somebody that picks up the phone. These are large projects.
358
00:19:32.579 --> 00:19:33.400
You should be real people.
359
00:19:34.200 --> 00:19:38.359
Eugene, recently we've seen the political landscape, which used to
360
00:19:38.420 --> 00:19:43.009
be pretty favorable. And now you've got governors all of
361
00:19:43.049 --> 00:19:46.049
a sudden. So the light, they have an epiphany and
362
00:19:46.130 --> 00:19:49.950
they think that politically it's easier to go with the grumbling.
363
00:19:50.200 --> 00:19:53.359
of the people rather than the developers. So do you
364
00:19:53.380 --> 00:19:56.740
think it's short term or there's really something in it
365
00:19:56.920 --> 00:19:59.259
that's going to hamper the possible growth?
366
00:19:59.980 --> 00:20:01.980
I mean, this is one of my kind of favorite observations.
367
00:20:02.670 --> 00:20:05.490
So I was in California when Jerry Brown, whose father
368
00:20:05.529 --> 00:20:07.670
had been governor, who had been governor in the 70s
369
00:20:08.009 --> 00:20:11.670
and then got reelected as governor after Arnold Schwarzenegger was governor,
370
00:20:12.029 --> 00:20:14.589
and really pulled California out of the recession. And it
371
00:20:14.630 --> 00:20:18.769
was a very, very, very savvy operator and was somebody
372
00:20:18.809 --> 00:20:21.529
who had been very pro-development. He turned the city of
373
00:20:21.710 --> 00:20:26.809
Oakland around. He banned fracking in California. Now, most people
374
00:20:26.869 --> 00:20:31.730
don't think of California as one of the largest oil-producing states.
375
00:20:33.069 --> 00:20:35.849
But it certainly was. And the reason that Chevron was
376
00:20:35.910 --> 00:20:39.450
headquartered there is that was Standard Oil's West Coast branch
377
00:20:39.509 --> 00:20:44.170
of their oil production. And the history of oil and
378
00:20:44.329 --> 00:20:48.109
energy production in California runs so deep and the expertise
379
00:20:48.150 --> 00:20:51.329
is so, so serious and the geologists are so good
380
00:20:51.829 --> 00:20:55.410
and the technology is so good. We did not understand
381
00:20:55.750 --> 00:20:58.289
what hydraulic fracking was going to end up being at
382
00:20:58.450 --> 00:21:02.700
all in maturations. And that governor, who should have probably
383
00:21:02.740 --> 00:21:06.940
known better, led an effort to ban fracking in California
384
00:21:07.000 --> 00:21:11.569
because hydrocarbons were such a dirty word. And I grew
385
00:21:11.630 --> 00:21:15.990
up in California in the 70s, 80s, 90s. And I
386
00:21:16.009 --> 00:21:18.049
remember going to Southern California and you could eat the
387
00:21:18.150 --> 00:21:20.549
air because it was so bad. So there was certainly
388
00:21:20.710 --> 00:21:24.880
reason to be anti-development of new energy resources.
389
00:21:25.440 --> 00:21:25.980
However.
390
00:21:26.690 --> 00:21:29.630
Here we are now. We still need that energy in
391
00:21:29.690 --> 00:21:33.839
order to make the transition into a cleaner energy environment.
392
00:21:34.190 --> 00:21:36.539
I look at data centers. I look at this growth
393
00:21:36.579 --> 00:21:41.920
of hyperscaler mass data collection center as the major tool
394
00:21:42.019 --> 00:21:45.119
to modernize our grid because we haven't had an economic
395
00:21:45.160 --> 00:21:48.960
driver that has done this as a business case in
396
00:21:49.140 --> 00:21:52.839
my lifetime. There hasn't been anything that could put this
397
00:21:52.960 --> 00:21:56.410
kind of capital to work. to upgrade what is a
398
00:21:56.430 --> 00:22:00.970
100-year-old grid. And so I think there's a deep level
399
00:22:01.019 --> 00:22:05.440
of short-sightedness for the governors that are reacting to this
400
00:22:05.759 --> 00:22:09.440
with what will be long-term bans. I think that there's
401
00:22:09.519 --> 00:22:13.109
room for both the technical improvement that the data center
402
00:22:13.130 --> 00:22:17.329
developers are doing to address things like noise, water, and
403
00:22:17.410 --> 00:22:20.930
energy consumption. But the fact that these guys are a
404
00:22:20.950 --> 00:22:27.099
24-7 baseload, standard load operator allows so much work to
405
00:22:27.119 --> 00:22:30.240
be done on the grids that need the work. I
406
00:22:30.279 --> 00:22:31.660
go to California all the time, but I'm in the
407
00:22:31.720 --> 00:22:34.599
Northeast now. The grid is 100 years old. Natural gas
408
00:22:34.640 --> 00:22:37.960
pipeline just south of San Francisco exploded, took out a
409
00:22:38.000 --> 00:22:41.460
whole neighborhood because it was a 100-year-old pipe. And so
410
00:22:41.500 --> 00:22:44.019
the only thing that can drive this level of investment
411
00:22:44.140 --> 00:22:47.700
is private industry with a real demand for the power.
412
00:22:47.940 --> 00:22:49.660
And that's what I think is coming. And that's what
413
00:22:49.700 --> 00:22:53.890
I think will be the product of this as the
414
00:22:53.970 --> 00:22:56.990
noise of the opposition calms down. And we're also in
415
00:22:57.009 --> 00:22:58.170
an election year.
416
00:22:58.190 --> 00:23:00.670
So Eugene, I like your vision and I agree with
417
00:23:00.690 --> 00:23:02.289
it because I think we actually do need to see
418
00:23:03.289 --> 00:23:07.410
AI as part of the solution. There's a huge amount
419
00:23:07.450 --> 00:23:08.380
of flexibility there.
420
00:23:08.410 --> 00:23:09.700
They all have backup power.
421
00:23:09.740 --> 00:23:12.180
They've got batteries in place. They've storage everywhere.
422
00:23:12.539 --> 00:23:13.539
You can move.
423
00:23:13.380 --> 00:23:18.000
Demand loads around the place. These guys have got great technology. Google,
424
00:23:18.019 --> 00:23:21.160
for example, has tapestry. I'm totally with you on this
425
00:23:21.240 --> 00:23:23.420
is that I think it's, and by the way, as
426
00:23:23.460 --> 00:23:27.009
you just said, they're willing to invest themselves and they're
427
00:23:27.109 --> 00:23:29.710
willing to sign offtake agreements. So use it to see
428
00:23:29.730 --> 00:23:30.589
it as an opportunity.
429
00:23:31.210 --> 00:23:33.230
I can't get past that the grid's 100 years old
430
00:23:33.329 --> 00:23:35.069
and we need we need to invest in it and
431
00:23:35.109 --> 00:23:36.789
nothing's going to drive it you're not going to go
432
00:23:36.809 --> 00:23:40.339
to the pucs and demand repairs pay more money because
433
00:23:40.880 --> 00:23:41.819
they can't afford.
434
00:23:41.579 --> 00:23:47.099
It yeah eugene i don't know if the glass is
435
00:23:47.140 --> 00:23:50.579
half full or half empty but there's certainly liquid into
436
00:23:50.640 --> 00:23:53.819
it so thank you very much for coming on the
437
00:23:53.839 --> 00:23:57.910
show and You know, having this kind of freewheeling conversation because,
438
00:23:57.930 --> 00:24:00.390
you know, there's a lot of data, but what's important
439
00:24:00.410 --> 00:24:03.349
for us are the people. And you're one of them.
440
00:24:03.430 --> 00:24:04.309
So thank you very much.
441
00:24:04.990 --> 00:24:07.599
It's my pleasure. And I wouldn't be a real estate
442
00:24:07.660 --> 00:24:10.559
broker at heart if I didn't give you no answer
443
00:24:10.579 --> 00:24:13.140
to whether or not the glass was half full or not. Yes,
444
00:24:13.160 --> 00:24:17.079
there is water. So I appreciate you having me on.
445
00:24:17.220 --> 00:24:18.819
Yeah, thanks, Eugene. Great having you.
446
00:24:19.920 --> 00:24:24.170
So Gerard, I know you are always a bull, always
447
00:24:24.250 --> 00:24:25.660
a bull. And you know why?
448
00:24:25.680 --> 00:24:26.880
Tell me.
449
00:24:27.700 --> 00:24:34.269
Because you're an equity guy. You love stories. And me,
450
00:24:34.349 --> 00:24:37.210
I'm a fixed income guy. So I look at credit.
451
00:24:37.789 --> 00:24:42.829
I'm quoting Apollo here. The credit story in hyperscaler rests
452
00:24:42.930 --> 00:24:47.390
on the single consensus assumption that the operating cash flow
453
00:24:47.509 --> 00:24:51.230
triples from $ 600 billion to $ 2 trillion. If it does
454
00:24:51.269 --> 00:24:54.230
not happen, everything's going to go to junk bonds. And
455
00:24:54.279 --> 00:24:57.880
it started. SoftBank, when they borrow now, it's junk bonds.
456
00:24:58.619 --> 00:25:02.769
We've seen the default, sorry, the force measure from Oracle
457
00:25:02.930 --> 00:25:06.549
on the Jupiter project in New Mexico. What that means
458
00:25:06.650 --> 00:25:09.349
now is if you look at the weakest part of
459
00:25:09.390 --> 00:25:13.309
the market, and I talk about it almost a year ago,
460
00:25:13.670 --> 00:25:17.849
the Blue Owl, the CoreWeaves, the Neoclouds, now they borrow
461
00:25:18.069 --> 00:25:22.490
at 8%, 9%, 10%, 11%. What the market says, not
462
00:25:22.529 --> 00:25:27.309
the equity jar, the credit market says, It's less risky
463
00:25:27.589 --> 00:25:31.730
to lend to Zambia and Angola than those new clouds.
464
00:25:32.069 --> 00:25:32.829
That's not going to work.
465
00:25:33.430 --> 00:25:35.730
Well, actually, I think it's probably worth talking about the
466
00:25:35.789 --> 00:25:38.019
equity side as well, because when you look at some
467
00:25:38.039 --> 00:25:43.740
of the valuations, obviously, Anthropic and OpenAI have not IPO'd
468
00:25:43.759 --> 00:25:46.839
at this point in time, but the numbers floated around
469
00:25:47.119 --> 00:25:50.049
are just... off the charts and you sort of wonder,
470
00:25:50.069 --> 00:25:51.589
are you ever going to really be able to get
471
00:25:51.609 --> 00:25:55.150
a return on an investment as an equity investor in
472
00:25:55.390 --> 00:25:56.450
those type of companies?
473
00:25:56.630 --> 00:25:58.930
I can tell you that all the early investors of
474
00:25:59.150 --> 00:26:03.880
their best PR companies going non-stop, non-stop, non-stop on media
475
00:26:03.900 --> 00:26:06.960
and pushing it, pushing it, pushing it. There are very,
476
00:26:07.019 --> 00:26:10.720
very few people in the equity who's going to say anything. Okay,
477
00:26:10.740 --> 00:26:13.140
so you always have the big short, Michael Burry and
478
00:26:13.160 --> 00:26:16.180
so on. But the debt, it's the debt. You need
479
00:26:16.240 --> 00:26:19.759
to check the debt. If your business model, and we've
480
00:26:19.779 --> 00:26:22.200
seen the new cloud, is to borrow at 9% and
481
00:26:22.240 --> 00:26:26.609
then lend at 1%, look, you've got the competition from
482
00:26:26.670 --> 00:26:29.910
the open weights. You see the price of compute going down.
483
00:26:29.990 --> 00:26:33.619
So I'm not saying AI is not a revolution. But
484
00:26:33.640 --> 00:26:36.900
what I'm saying is that the financial assumptions around the
485
00:26:37.119 --> 00:26:42.180
investment wave don't make any sense at the end of 2026.
486
00:26:42.180 --> 00:26:44.660
Listen, I think the point you made there is a revolution.
487
00:26:44.680 --> 00:26:47.740
In a revolution, what actually happens is there's a huge
488
00:26:47.759 --> 00:26:50.099
amount of money made and there's also a huge amount
489
00:26:50.119 --> 00:26:53.140
of money lost. And I think this revolution is going
490
00:26:53.160 --> 00:26:53.779
to be no different.
491
00:26:54.220 --> 00:26:57.880
But nevertheless, Gerard, the one who managed to build a
492
00:26:57.900 --> 00:27:02.059
data center, that's where Eugene was great. So whether it's
493
00:27:02.119 --> 00:27:05.359
the whole one or those type of things, when you
494
00:27:05.460 --> 00:27:07.660
manage to build a data center, a data center is
495
00:27:07.680 --> 00:27:08.829
very profitable right now.
496
00:27:09.380 --> 00:27:09.900
Oh, absolutely.
497
00:27:10.140 --> 00:27:10.740
Totally with Gerard.
498
00:27:11.180 --> 00:27:16.460
But that's all development pipeline was just insane. The reality
499
00:27:16.480 --> 00:27:19.690
is great, but the promises, just don't believe them.
500
00:27:20.329 --> 00:27:23.750
Good, my friend. Well, listen, as ever, good chatting. So
501
00:27:24.029 --> 00:27:25.089
I'm looking forward to next week.
502
00:27:25.390 --> 00:27:26.710
Okay, John. Cheers.
503
00:27:28.009 --> 00:27:31.730
Thank you for listening to Redefining Energy. Don't forget to
504
00:27:31.799 --> 00:27:36.500
rate the show and subscribe on Apple Podcasts, Spotify, or
505
00:27:36.519 --> 00:27:37.920
the platform of your choice.
1
00:00:07.269 --> 00:00:10.830
With Laurent Seguelin from London and Gerard Reid from Berlin,
2
00:00:11.310 --> 00:00:13.130
this is Redefining Energy.
3
00:00:13.509 --> 00:00:16.699
Today on Redefining Energy, we're going to talk about data centers.
4
00:00:16.730 --> 00:00:18.539
And in particular, we're going to look at the U.S.
5
00:00:18.570 --> 00:00:20.089
market and what's going on there.
6
00:00:20.710 --> 00:00:23.859
Gerard, if you look at the global numbers, it's insane.
7
00:00:24.699 --> 00:00:29.079
The worldwide data center CapEx is now expected to surpass
8
00:00:29.079 --> 00:00:32.880
$ 3 trillion by 2030, doubling what we have right now.
9
00:00:33.359 --> 00:00:36.619
It's the biggest GDP bet in America since the railroad.
10
00:00:37.229 --> 00:00:38.630
But first of all, from a partner.
11
00:00:39.130 --> 00:00:41.810
Right now, it feels like the energy world is one
12
00:00:41.950 --> 00:00:42.869
wild ride.
13
00:00:43.149 --> 00:00:44.590
Changing at dizzying speed.
14
00:00:45.210 --> 00:00:50.310
Twists, turns, steep climbs, sudden drops. But volatility doesn't have
15
00:00:50.369 --> 00:00:54.049
to slow you down. With the industry's largest proprietary data set,
16
00:00:54.670 --> 00:00:59.590
expert analysis, interconnected insight, and synoptic AI, Wood Mackenzie helps
17
00:00:59.609 --> 00:01:02.189
you quickly react to any sudden change.
18
00:01:02.570 --> 00:01:03.479
So you can be sure.
19
00:01:04.140 --> 00:01:04.859
Wood Mack short.
20
00:01:05.379 --> 00:01:07.819
Wood Mackenzie, Intelligence Connected.
21
00:01:08.519 --> 00:01:12.219
Back to the show. And it's burning cash. It's insane.
22
00:01:12.299 --> 00:01:16.609
So I've got a few numbers. KKR Credit estimate that
23
00:01:17.150 --> 00:01:22.349
AI-linked debt for data center, 600 billion. AI infrastructure company,
24
00:01:22.349 --> 00:01:24.469
40% of the S & P 500. AI-related companies, 49%
25
00:01:24.469 --> 00:01:31.849
of investment-grade bond insurance. And if you look at the
26
00:01:31.909 --> 00:01:36.250
private market, 87% of VC funding goes to AI. Of course,
27
00:01:36.290 --> 00:01:40.200
we've seen the entropic filing for IPO. I mean, it's wonderful.
28
00:01:40.200 --> 00:01:46.760
600 billion of commitment to Amazon, Google, and Microsoft XAI.
29
00:01:47.000 --> 00:01:48.799
The amount of cash is crazy.
30
00:01:49.420 --> 00:01:51.719
And I think it's really important to dig into this
31
00:01:51.780 --> 00:01:53.859
team because it's highly complex.
32
00:01:54.329 --> 00:01:54.890
Yeah, I know.
33
00:01:54.989 --> 00:01:57.510
And Laurent, as a result, it's really complex. important to
34
00:01:57.609 --> 00:02:00.109
dig into this theme. And that's why it was really
35
00:02:00.150 --> 00:02:02.659
great to get Eugene McGrain to come on the podcast
36
00:02:02.810 --> 00:02:04.780
to have a discussion around this topic.
37
00:02:05.379 --> 00:02:09.659
Eugene is executive managing director at Cushman Wakefield, which is
38
00:02:09.759 --> 00:02:13.300
one of the world's largest real estate broker. They used
39
00:02:13.319 --> 00:02:16.259
to do warehouses and they used to do office and
40
00:02:16.919 --> 00:02:21.740
somehow they acquired the expertise to do data centers. Now,
41
00:02:22.000 --> 00:02:24.819
when somebody say, okay, that data centers is worth a
42
00:02:24.879 --> 00:02:28.250
hundred billion. In fact, you get 50% chips, and then
43
00:02:28.289 --> 00:02:30.960
you get 20% power, and then the rest is cooling
44
00:02:31.000 --> 00:02:34.199
and everything. So it doesn't enter in the chips per se,
45
00:02:34.740 --> 00:02:38.599
but you organize everything that's around the world, the permitting,
46
00:02:38.639 --> 00:02:40.580
the connection, and so on. So I think it's very
47
00:02:40.620 --> 00:02:44.620
important to see that data center wave from the inside.
48
00:02:45.300 --> 00:02:45.800
Absolutely.
49
00:02:45.819 --> 00:02:48.860
So let's bring him on the show. Eugene, great to
50
00:02:48.879 --> 00:02:49.620
have you on the show.
51
00:02:50.319 --> 00:02:51.659
Thanks for having me.
52
00:02:51.680 --> 00:02:55.069
Eugene, a very simple question to start with. You are
53
00:02:55.330 --> 00:02:59.909
a real estate expert. How did you become a data
54
00:02:59.969 --> 00:03:01.289
center specialist?
55
00:03:01.990 --> 00:03:04.490
Well, like all things in life, a series of coincidence.
56
00:03:04.650 --> 00:03:07.400
I've been a commercial real estate broker for 20 plus
57
00:03:07.409 --> 00:03:10.840
years now. I started in California and I was at
58
00:03:10.879 --> 00:03:13.039
the beginning just an office broker, you know, working for
59
00:03:13.139 --> 00:03:16.199
corporations and dealing with their office leases, which are kind
60
00:03:16.219 --> 00:03:20.099
of just complex contract negotiations and market knowledge. Through that time,
61
00:03:20.180 --> 00:03:22.710
I worked for companies like PG & E and people
62
00:03:22.719 --> 00:03:26.169
like Chevron dealing with their real estate issues and assets.
63
00:03:26.889 --> 00:03:28.909
And I was very fortunate. I had a dear friend
64
00:03:28.930 --> 00:03:32.250
at Chevron, and he recommended that I, you know, I
65
00:03:32.289 --> 00:03:34.330
wanted to learn more about their business. He told me
66
00:03:34.370 --> 00:03:37.110
to read Daniel Juergen's book. So three books later, I
67
00:03:37.259 --> 00:03:39.599
felt like I knew enough about the energy business to
68
00:03:39.639 --> 00:03:43.639
not know much. And in 2021, my wife, who's from
69
00:03:43.699 --> 00:03:46.460
Rhode Island, needed to get home. And so we moved
70
00:03:46.520 --> 00:03:50.719
our family across the country. in that very unsettled time
71
00:03:50.740 --> 00:03:52.840
in the world where, you know, people were not signing
72
00:03:52.879 --> 00:03:57.500
new office leases or, you know, warehouse leases with much certainty.
73
00:03:57.520 --> 00:04:01.319
The political climate of the time made a giant push
74
00:04:01.419 --> 00:04:04.639
for offshore wind. And there was a commitment to do
75
00:04:04.639 --> 00:04:08.900
35 gigawatts of power and deliverables and contracts were done
76
00:04:08.939 --> 00:04:13.280
and investors were attracted. And, you know, with my background,
77
00:04:13.599 --> 00:04:16.240
I could see two things. One, that this was real,
78
00:04:16.519 --> 00:04:19.410
that the investments were going to be made. And two,
79
00:04:19.810 --> 00:04:23.089
that the companies, especially the Danes, were coming over with
80
00:04:23.129 --> 00:04:26.629
a completely different understanding of the power grid and how
81
00:04:26.709 --> 00:04:30.149
development worked here as to how it worked in Europe.
82
00:04:30.560 --> 00:04:34.019
And so I started advising those companies on what the
83
00:04:34.079 --> 00:04:36.300
kind of different obligations were going to be in order
84
00:04:36.319 --> 00:04:39.040
to reach the development. But they had interconnections going through
85
00:04:39.060 --> 00:04:41.839
towns that they didn't really understand the permitting process and
86
00:04:41.899 --> 00:04:45.000
what the obstacles were going to be. And so as
87
00:04:45.100 --> 00:04:47.000
I was doing that, I was doing a lot of
88
00:04:47.079 --> 00:04:50.860
infrastructure level work. power because that's what those developments were.
89
00:04:51.560 --> 00:04:54.800
And then when ChatGPT hit the market and our kind
90
00:04:54.839 --> 00:04:57.180
of worlds changed, I just happened to be one of
91
00:04:57.220 --> 00:05:00.310
those people that was working with companies that were doing 200,
92
00:05:00.310 --> 00:05:05.060
300 megawatt projects and kind of understood the basic stuff
93
00:05:05.079 --> 00:05:08.500
that those people did, which was not typical in the
94
00:05:08.540 --> 00:05:10.660
data center world prior to that. You know, a large
95
00:05:10.699 --> 00:05:14.720
requirement was five megawatts, 20 megawatts. all of a sudden
96
00:05:14.740 --> 00:05:18.319
we were looking at 200 megawatts was, oh, maybe that's
97
00:05:18.360 --> 00:05:21.560
not enough. So it was 200 to 500 to gigawatt
98
00:05:21.579 --> 00:05:25.639
to now what it is. And so the expertise that
99
00:05:25.660 --> 00:05:28.740
I've been able to grow and build is one that's
100
00:05:28.920 --> 00:05:32.879
based in real estate and complexities of developing property, the
101
00:05:32.939 --> 00:05:36.589
financing aspects of developing property, what has to happen, how
102
00:05:36.610 --> 00:05:39.009
the steps go, what you have to get through to
103
00:05:40.290 --> 00:05:43.600
somebody that kind of understood the PPAs and process that
104
00:05:43.720 --> 00:05:46.180
power companies go through. So I just happen to have
105
00:05:46.199 --> 00:05:48.079
kind of a unique set of where I am. And
106
00:05:48.120 --> 00:05:51.170
my partners, who are very, very good at the data
107
00:05:51.189 --> 00:05:53.730
center end of this, just partner with them, and I
108
00:05:53.769 --> 00:05:55.529
sit in between, and I think about this stuff all
109
00:05:55.569 --> 00:05:57.490
the time. So that's who I am. That's how I
110
00:05:57.509 --> 00:05:57.829
got here.
111
00:05:58.529 --> 00:06:01.009
Yeah, it's a great story. So in a way, you
112
00:06:01.029 --> 00:06:03.449
are the right man at the right place. Because I
113
00:06:03.490 --> 00:06:05.480
think all the data center guys, they're good in the
114
00:06:05.509 --> 00:06:08.240
soft thing, but when it comes to hard, the hard part,
115
00:06:08.319 --> 00:06:12.329
which is the walls, the terrain, the the relationship with
116
00:06:12.370 --> 00:06:17.170
the authorities and the connections, you have done that before.
117
00:06:17.470 --> 00:06:21.199
How fast has your practice grown in the past three years?
118
00:06:22.060 --> 00:06:25.620
The amount of time I spend on this has grown significantly.
119
00:06:26.319 --> 00:06:31.220
These projects are so complex and so large that from
120
00:06:31.259 --> 00:06:34.610
a percentage-wise, they're like 100% of my practice, but they
121
00:06:34.709 --> 00:06:37.750
take so long to get there that if I do
122
00:06:37.850 --> 00:06:39.670
two or three of these or four or five of these,
123
00:06:40.430 --> 00:06:42.730
That's a lot of them. These are sort of kind
124
00:06:42.769 --> 00:06:47.089
of career-sized projects to work on. It's scaled significantly.
125
00:06:47.810 --> 00:06:49.610
Eugene, can you just talk a little bit about the
126
00:06:50.129 --> 00:06:54.310
hype in this whole area and how you navigate through
127
00:06:54.329 --> 00:06:55.990
that and what your thoughts are on it?
128
00:06:56.930 --> 00:06:59.930
Well, it's interesting because it's evolved. When this first happened,
129
00:07:00.149 --> 00:07:04.750
every farmer and every brownfield owner of property in the
130
00:07:04.829 --> 00:07:07.449
country all of a sudden told everybody that they had
131
00:07:07.449 --> 00:07:11.459
200 megawatts or 500 megawatts or some level of capacity,
132
00:07:11.500 --> 00:07:14.350
or they could get interconnection because they knew a guy
133
00:07:14.389 --> 00:07:18.949
at the PUC or whatever. And because of the size
134
00:07:18.970 --> 00:07:21.949
and complexity of these assets, I mean, we're almost at
135
00:07:21.970 --> 00:07:24.709
the other end of that now, in my opinion, where
136
00:07:24.829 --> 00:07:28.889
the hype is more narrative-driven than it is on the
137
00:07:28.970 --> 00:07:31.730
actual ground of people that are making decisions and people
138
00:07:31.769 --> 00:07:35.089
that are writing checks. This is now the part where
139
00:07:35.769 --> 00:07:40.470
the really sophisticated developers have taken it to a place
140
00:07:40.550 --> 00:07:43.269
where they have to walk through the individual steps and
141
00:07:43.629 --> 00:07:47.569
the press is against them because the story is against them.
142
00:07:48.250 --> 00:07:52.699
So the good comparison is essentially where the fracking business
143
00:07:52.740 --> 00:07:56.259
came from. First, it was this energy revolution to an
144
00:07:56.459 --> 00:08:00.560
environmental disaster to what it is now, which is really
145
00:08:01.000 --> 00:08:04.560
kind of energy independence and changing the map of the world.
146
00:08:05.110 --> 00:08:09.230
The press and the public sentiment became so strong against
147
00:08:09.310 --> 00:08:11.930
these projects that they really had to fine tune what
148
00:08:11.949 --> 00:08:14.189
the technology was and they had to fine tune what
149
00:08:14.209 --> 00:08:18.730
the message was. So the height now is much more
150
00:08:19.490 --> 00:08:22.370
in the financial markets than it is on the ground.
151
00:08:23.500 --> 00:08:26.600
That makes sense to me. And maybe I'd just follow
152
00:08:26.620 --> 00:08:28.339
it on from that then. When you look at it,
153
00:08:28.759 --> 00:08:30.060
I suppose I'm trying to get my head around.
154
00:08:30.339 --> 00:08:30.860
Do you see the.
155
00:08:30.819 --> 00:08:33.740
Glass as half full or half empty when you look
156
00:08:33.759 --> 00:08:34.759
at the potential, right?
157
00:08:35.789 --> 00:08:38.370
I'm a pessimistic optimist. So one of the questions that
158
00:08:38.389 --> 00:08:40.750
we always get and that I've answered in a bunch
159
00:08:40.789 --> 00:08:43.710
of places and I think is an ongoing question is,
160
00:08:44.269 --> 00:08:47.580
are we building data centers and infrastructure? Or is this
161
00:08:47.620 --> 00:08:51.879
the 1870s railroad boom? Or is this some sort of
162
00:08:52.559 --> 00:08:55.139
critical resource that we're building and this is going to
163
00:08:55.179 --> 00:08:58.620
define the next 50 years? We're somewhere in the middle.
164
00:08:58.840 --> 00:09:00.919
There is hype in the sector. We talked a bit
165
00:09:00.940 --> 00:09:03.980
about some of the players that have emerged as people
166
00:09:04.019 --> 00:09:07.529
that can deliver these products. But the vacancy rate, which is,
167
00:09:07.769 --> 00:09:09.990
you know, as a real estate guy, the first thing
168
00:09:10.009 --> 00:09:12.289
you look at is the vacancy rate. That tells you
169
00:09:12.330 --> 00:09:15.139
what kind of delivery there is. The vacancy rate in
170
00:09:15.179 --> 00:09:17.879
these large scale, let's call them 200 megawatts and above
171
00:09:18.139 --> 00:09:21.919
projects is less than 1%. The overall market of the
172
00:09:21.980 --> 00:09:26.269
Americas is 3.4%. And that includes everything, which is a
173
00:09:26.330 --> 00:09:29.570
lot of legacy stuff that can't really be repositioned to
174
00:09:29.649 --> 00:09:33.830
deal with the kind of higher compute requirements. So if
175
00:09:33.870 --> 00:09:37.679
there's a collapse in demand for compute, that would be
176
00:09:38.179 --> 00:09:41.000
pretty fatal for the industry. I don't believe we're going
177
00:09:41.019 --> 00:09:43.039
to see a collapse in demand for compute. We may
178
00:09:43.080 --> 00:09:47.500
see a collapse in paying for compute, and that may
179
00:09:47.879 --> 00:09:51.610
rattle the people that have done the infrastructure investments. But
180
00:09:52.190 --> 00:09:54.809
this is stuff that, no matter what happens, is going
181
00:09:54.830 --> 00:09:59.230
to be used. It's equivalent to the fiber build-out of
182
00:09:59.269 --> 00:10:02.759
the late 90s. 90s, where fiber got built everywhere because
183
00:10:03.059 --> 00:10:06.980
there was a supposedly endless demand for fiber. Well, it
184
00:10:07.019 --> 00:10:08.539
sat in the ground in a lot of places for
185
00:10:08.580 --> 00:10:11.990
a long time, and now it's being used. So glass
186
00:10:12.009 --> 00:10:15.129
half full, glass half empty. I think we have a
187
00:10:15.590 --> 00:10:19.809
near insatiable demand for compute. We have a very limited
188
00:10:19.850 --> 00:10:23.169
ability to deliver it at the scale that the demand
189
00:10:23.230 --> 00:10:26.889
is asking for. And this is essentially the infrastructure build
190
00:10:26.970 --> 00:10:30.519
out to meet that demand. Will there be carnage in
191
00:10:30.740 --> 00:10:34.259
the financials? Certainly. Somebody's going to lose something somewhere, but
192
00:10:34.279 --> 00:10:35.559
this is not pets.com.
193
00:10:36.539 --> 00:10:38.440
Eugene, I like the way you also started there because
194
00:10:39.019 --> 00:10:42.120
You also put it in the context of the railroad
195
00:10:42.179 --> 00:10:45.100
build-out and all that type of stuff. That was a revolution,
196
00:10:45.139 --> 00:10:47.779
a transport revolution. I think that we're going through the
197
00:10:47.820 --> 00:10:51.090
exact same thing. There are going to be winners and losers,
198
00:10:51.210 --> 00:10:53.350
but the bottom line is you just need to have
199
00:10:53.370 --> 00:10:53.919
to be careful.
200
00:10:54.990 --> 00:10:55.769
Do sensible things.
201
00:10:55.830 --> 00:10:56.850
Don't do stupid things.
202
00:10:57.370 --> 00:11:01.149
It's always easier to do stupid things. But every day
203
00:11:01.210 --> 00:11:03.389
I read the headlines, and then every day I answer
204
00:11:03.429 --> 00:11:06.500
my phone. And the headlines on the phone don't match up.
205
00:11:06.950 --> 00:11:09.129
in a lot of ways, which is the demand continues
206
00:11:09.169 --> 00:11:12.299
to be very strong for the needs for this. And
207
00:11:12.320 --> 00:11:16.580
then everybody you're dealing with is using AI. It's incredibly useful.
208
00:11:16.639 --> 00:11:21.139
We've been generating data for 30 plus years that allows
209
00:11:21.179 --> 00:11:24.370
for real productivity. You know, everybody wants better information, and
210
00:11:24.419 --> 00:11:25.669
this is the tool to get you there.
211
00:11:26.330 --> 00:11:29.970
If I go back to power and real estate, I remember,
212
00:11:30.049 --> 00:11:33.409
because I'm an old man, 27 years ago, Mark Mills...
213
00:11:35.399 --> 00:11:40.710
Mark Mills, who is worshipped every time he goes to Sarawak, said,
214
00:11:41.769 --> 00:11:44.990
so end of the 90s, by 2010, half of the
215
00:11:45.009 --> 00:11:49.009
electricity will be used by the internet. Okay. And basically
216
00:11:49.029 --> 00:11:52.840
that triggered a lot of construction of gas and so on,
217
00:11:52.879 --> 00:11:57.200
which then led to the collapse of Calpine and all
218
00:11:57.220 --> 00:11:59.799
the bad. He was the big evangelist at the time.
219
00:11:59.840 --> 00:12:02.379
He's still around. So now, of course, he's saying we
220
00:12:02.399 --> 00:12:05.210
need to multiply power by a factor of 10 and Now,
221
00:12:05.230 --> 00:12:08.440
I'm interested, if I look from a basic point of view,
222
00:12:08.960 --> 00:12:13.200
NVIDIA chips. So you had the Hopper, then the Blackwell,
223
00:12:13.639 --> 00:12:17.340
now the Rubin, and we're going next generation to the Fineman.
224
00:12:17.759 --> 00:12:21.059
The energy efficiency of the Fineman is going to be
225
00:12:21.059 --> 00:12:25.399
500 times more than the Hopper, which goes to two questions.
226
00:12:26.169 --> 00:12:28.710
The number one is, what's going to be the value,
227
00:12:28.929 --> 00:12:31.909
especially if they have those 20-year contract of a Hopper
228
00:12:32.090 --> 00:12:36.610
data center in 20 years? considering the speed of the improvement.
229
00:12:37.309 --> 00:12:40.149
And the second one is the guy everybody loves to hate,
230
00:12:40.299 --> 00:12:42.720
but is an engineering genius, is Elon Musk. If I
231
00:12:42.759 --> 00:12:45.100
look at the development of Colossus, now we've got this
232
00:12:45.179 --> 00:12:49.340
new building called Mini Hard. And Mini Hard is about
233
00:12:49.740 --> 00:12:52.090
a quarter of the size of the one he built
234
00:12:52.110 --> 00:12:54.929
two years ago, and it's even more powerful. It's very
235
00:12:54.970 --> 00:12:58.450
difficult to assess, but that's something that the hyperscaler industry
236
00:12:58.470 --> 00:13:02.429
has done. The energy efficiency gains that people are going
237
00:13:02.450 --> 00:13:06.200
to do which for me is a bigger risk than
238
00:13:06.240 --> 00:13:08.860
the demand for compute. The question is, do you have
239
00:13:08.899 --> 00:13:12.100
this type of thinking or no, as long as the
240
00:13:12.139 --> 00:13:14.919
telephone rings, everything's okay? No, no.
241
00:13:15.100 --> 00:13:19.440
The answer is the projects themselves are the limiting factor
242
00:13:19.480 --> 00:13:23.919
on the growth. And because of the size of the projects,
243
00:13:24.639 --> 00:13:27.940
because of the complexity, because of the software guys have
244
00:13:28.059 --> 00:13:31.039
never had the plan for power before. They just iterate
245
00:13:31.120 --> 00:13:34.350
the next level of software. And it's just been easy.
246
00:13:34.669 --> 00:13:38.190
So they've never had to do this hard infrastructure level work.
247
00:13:38.789 --> 00:13:42.210
Meta at one point wanted to build its own Bayside
248
00:13:42.409 --> 00:13:45.870
mass transit rail station to run from Palo Alto up
249
00:13:45.909 --> 00:13:48.720
to San Francisco. And they hired a whole team. They
250
00:13:48.740 --> 00:13:51.940
did all the things that would build an actual rail system.
251
00:13:52.700 --> 00:13:55.840
And they got nowhere because it involved things that are,
252
00:13:56.440 --> 00:13:59.240
you've got to get approvals, you've got to get Materials,
253
00:13:59.299 --> 00:14:03.220
like all of the hard structural costs that are involved
254
00:14:03.240 --> 00:14:07.240
with infrastructure, are in themselves a limiting factor in the
255
00:14:07.360 --> 00:14:11.820
speed of this sort of build. Now, this feels like
256
00:14:11.879 --> 00:14:15.029
it's moving at some level of warp speed, but I
257
00:14:15.070 --> 00:14:18.669
would make the argument that it is basically rate limited
258
00:14:19.649 --> 00:14:23.490
or expansion limited by the funnel of, you can't get
259
00:14:23.509 --> 00:14:26.899
a turbine from GE Vranova or Hitachi for three years.
260
00:14:27.000 --> 00:14:29.440
You can't get a transformer from those same people for
261
00:14:29.720 --> 00:14:32.320
two or three years. There's a queue. So the demand
262
00:14:32.379 --> 00:14:35.519
is so strong and the inputs have so much trouble
263
00:14:35.570 --> 00:14:39.149
flexing up that I don't look at it as, oh,
264
00:14:39.230 --> 00:14:43.149
we're 50 gigawatts ahead of where compute demand is going
265
00:14:43.169 --> 00:14:46.529
to be in 10 years. I think where we are
266
00:14:46.730 --> 00:14:49.850
is we're probably par or less than par of where
267
00:14:49.870 --> 00:14:53.049
we're going to be in five to 10 years especially
268
00:14:53.210 --> 00:14:56.049
if the chips are able to get more efficient, because
269
00:14:56.080 --> 00:14:58.200
I don't think you're going to see compute demand drop.
270
00:14:58.580 --> 00:15:01.659
What we've had is it's worth looking at the first
271
00:15:01.700 --> 00:15:04.799
internet boom, because that is kind of what's informing a
272
00:15:04.860 --> 00:15:07.720
lot of people's prejudices on this. It certainly informs mine.
273
00:15:07.960 --> 00:15:10.419
I grew up in San Francisco and watched the first
274
00:15:10.500 --> 00:15:13.720
dot-com boom. That was based so much on software.
275
00:15:13.759 --> 00:15:14.379
It was free.
276
00:15:14.460 --> 00:15:17.200
Everybody could figure out how to write a new website.
277
00:15:18.320 --> 00:15:20.820
And the money that was being thrown at those companies
278
00:15:20.860 --> 00:15:23.460
was being thrown at companies that were just basically ideas.
279
00:15:23.620 --> 00:15:26.419
They didn't have to go out and really do anything.
280
00:15:27.450 --> 00:15:31.210
In order to get this level of compute that's demanded today,
281
00:15:32.009 --> 00:15:35.289
you've got to build infrastructure that will be valuable in
282
00:15:35.289 --> 00:15:38.169
20 years. And if it becomes a factor of the
283
00:15:38.320 --> 00:15:41.360
chips continue to get more energy efficient, it just allows
284
00:15:41.399 --> 00:15:43.679
them to meet. Now, there's going to be some overlap.
285
00:15:43.740 --> 00:15:45.340
There's going to be a tipping point where we get
286
00:15:45.379 --> 00:15:47.940
beyond that. I don't think we're there. And I think
287
00:15:48.000 --> 00:15:50.980
that what we'll see is projects get abandoned as opposed
288
00:15:51.019 --> 00:15:53.519
to projects being built and then not being used.
289
00:15:54.779 --> 00:15:56.399
What I just like to add to that, to put
290
00:15:56.450 --> 00:16:01.190
it into context, is that Google, just Google alone, Google
291
00:16:01.289 --> 00:16:05.889
has had a sevenfold increase in demand for processing power
292
00:16:05.909 --> 00:16:10.149
in the last 12 months. 700%. Now, of course, what
293
00:16:10.190 --> 00:16:11.710
you don't have is 700% increase in energy demand, but
294
00:16:14.289 --> 00:16:16.779
But let's say you have a doubling of energy demand
295
00:16:16.820 --> 00:16:19.100
because of energy efficiency and new chips and this, that,
296
00:16:19.120 --> 00:16:21.320
and the other thing. The point I'm making is that's
297
00:16:21.360 --> 00:16:23.840
going to continue. It mightn't be seven-fold next year, but
298
00:16:23.860 --> 00:16:28.250
let's say it doubles. That's what's driving demand. And that
299
00:16:28.309 --> 00:16:30.789
is all of us sitting there. And I don't use
300
00:16:30.970 --> 00:16:34.629
Google anymore. I use ChatGBT to do my searches. I
301
00:16:34.669 --> 00:16:37.610
want them to do an invitation for something.
302
00:16:37.669 --> 00:16:38.149
Guess what?
303
00:16:38.269 --> 00:16:38.850
I'm using it.
304
00:16:39.230 --> 00:16:42.559
That's the bottom line. We are driving as consumers demand.
305
00:16:43.429 --> 00:16:48.080
massive demand for data because it's useful. And I think
306
00:16:48.139 --> 00:16:49.980
it's just going to become more useful. Yeah.
307
00:16:50.820 --> 00:16:54.860
Eugene, as you said, there is this crazy thirst for
308
00:16:55.440 --> 00:16:58.470
more compute, more buildings and more chips, more everything. And
309
00:16:58.500 --> 00:17:01.210
of course the turbines, I mean, we know that, but
310
00:17:01.429 --> 00:17:05.910
you as a professional, when your telephone rings, And it's
311
00:17:06.089 --> 00:17:10.009
a new client or new developer. How do you assess
312
00:17:10.150 --> 00:17:15.039
how serious that this deal will go through rather than
313
00:17:15.099 --> 00:17:17.920
just pie in the sky? You have a secret sauce here.
314
00:17:18.480 --> 00:17:20.700
My first boss always answered the phone. You never know
315
00:17:20.720 --> 00:17:22.160
when somebody is going to want to buy the Empire
316
00:17:22.200 --> 00:17:24.880
State Building. I've been trained with a certain level of
317
00:17:25.000 --> 00:17:28.720
optimism in answering my phone. These projects, one, they're so
318
00:17:28.759 --> 00:17:31.490
much more interesting for me to work on than a
319
00:17:31.529 --> 00:17:35.440
million square foot warehouse for Amazon. It does a great project.
320
00:17:35.460 --> 00:17:38.690
They're very lucrative. But even at the most technical, it's
321
00:17:38.789 --> 00:17:41.829
just high bay warehouse. An office building is just a
322
00:17:42.029 --> 00:17:45.410
lease with a subordination, non-disturbance, and atonement act. It's nothing
323
00:17:45.950 --> 00:17:49.019
real great shakes. This is a real technical issue. So
324
00:17:49.359 --> 00:17:52.660
the way you qualify these things is the level of
325
00:17:52.700 --> 00:17:55.200
professionalism on the other side. One of the groups we're
326
00:17:55.240 --> 00:17:57.880
working for, the way that I was introduced to them
327
00:17:57.960 --> 00:18:01.730
was the former head of BP's trading desk went to
328
00:18:01.769 --> 00:18:03.829
go work for them. That, to me, spoke of a
329
00:18:03.869 --> 00:18:06.430
level of professionalism and a level of understanding of the
330
00:18:06.470 --> 00:18:11.349
market and the economies of scale. That was a real person.
331
00:18:11.410 --> 00:18:14.849
And I think that's something that we should understand. My
332
00:18:14.910 --> 00:18:16.619
friend who I was on the phone with yesterday was
333
00:18:16.650 --> 00:18:20.799
a nuclear engineer for Dominion before he went over and
334
00:18:20.839 --> 00:18:23.019
worked for Meta as one of their heads of energy.
335
00:18:23.319 --> 00:18:26.240
They're not hiring guys like me to run these things.
336
00:18:26.279 --> 00:18:28.240
The guys like me are figuring out how to get
337
00:18:28.279 --> 00:18:30.690
them delivered. But the quality of the people that are
338
00:18:30.910 --> 00:18:33.950
involved in this are really world-class. And if you don't
339
00:18:34.049 --> 00:18:36.549
have that group of people, if you don't have the
340
00:18:36.589 --> 00:18:39.829
market expert to get into, we have a team. I
341
00:18:40.509 --> 00:18:43.710
have a broker in Dallas who has been doing data
342
00:18:43.750 --> 00:18:46.720
centers in Dallas for 30 years and worked for HP
343
00:18:46.799 --> 00:18:50.200
when HP was building out their first infrastructure. You have
344
00:18:50.240 --> 00:18:52.720
to be willing to assemble that kind of group. That's
345
00:18:52.779 --> 00:18:56.680
the qualification. In the United States, we're broken up into ISOs.
346
00:18:56.920 --> 00:18:59.799
And the ISOs we're getting I wouldn't call it artificial demand,
347
00:18:59.819 --> 00:19:01.660
but they were getting a lot of, hey, maybe this
348
00:19:01.680 --> 00:19:04.460
will happen in their queues because people wanted to just
349
00:19:04.500 --> 00:19:07.309
try to get it placed in line. They've all basically
350
00:19:07.369 --> 00:19:09.910
come back and said, you need to pay to play
351
00:19:10.150 --> 00:19:12.789
and you need to pay real money. So ERCOT's charging
352
00:19:12.970 --> 00:19:16.269
quite a bit of money that's non-refundable if you want
353
00:19:16.289 --> 00:19:19.329
to place on their interconnection queue. And you start talking
354
00:19:19.369 --> 00:19:23.369
about 300 or 500 or 1,000 megawatts, you start talking
355
00:19:23.410 --> 00:19:26.609
about having to put real money into the utility system.
356
00:19:26.670 --> 00:19:29.640
So That's a lot of what drives how serious I
357
00:19:29.660 --> 00:19:32.539
take somebody that picks up the phone. These are large projects.
358
00:19:32.579 --> 00:19:33.400
You should be real people.
359
00:19:34.200 --> 00:19:38.359
Eugene, recently we've seen the political landscape, which used to
360
00:19:38.420 --> 00:19:43.009
be pretty favorable. And now you've got governors all of
361
00:19:43.049 --> 00:19:46.049
a sudden. So the light, they have an epiphany and
362
00:19:46.130 --> 00:19:49.950
they think that politically it's easier to go with the grumbling.
363
00:19:50.200 --> 00:19:53.359
of the people rather than the developers. So do you
364
00:19:53.380 --> 00:19:56.740
think it's short term or there's really something in it
365
00:19:56.920 --> 00:19:59.259
that's going to hamper the possible growth?
366
00:19:59.980 --> 00:20:01.980
I mean, this is one of my kind of favorite observations.
367
00:20:02.670 --> 00:20:05.490
So I was in California when Jerry Brown, whose father
368
00:20:05.529 --> 00:20:07.670
had been governor, who had been governor in the 70s
369
00:20:08.009 --> 00:20:11.670
and then got reelected as governor after Arnold Schwarzenegger was governor,
370
00:20:12.029 --> 00:20:14.589
and really pulled California out of the recession. And it
371
00:20:14.630 --> 00:20:18.769
was a very, very, very savvy operator and was somebody
372
00:20:18.809 --> 00:20:21.529
who had been very pro-development. He turned the city of
373
00:20:21.710 --> 00:20:26.809
Oakland around. He banned fracking in California. Now, most people
374
00:20:26.869 --> 00:20:31.730
don't think of California as one of the largest oil-producing states.
375
00:20:33.069 --> 00:20:35.849
But it certainly was. And the reason that Chevron was
376
00:20:35.910 --> 00:20:39.450
headquartered there is that was Standard Oil's West Coast branch
377
00:20:39.509 --> 00:20:44.170
of their oil production. And the history of oil and
378
00:20:44.329 --> 00:20:48.109
energy production in California runs so deep and the expertise
379
00:20:48.150 --> 00:20:51.329
is so, so serious and the geologists are so good
380
00:20:51.829 --> 00:20:55.410
and the technology is so good. We did not understand
381
00:20:55.750 --> 00:20:58.289
what hydraulic fracking was going to end up being at
382
00:20:58.450 --> 00:21:02.700
all in maturations. And that governor, who should have probably
383
00:21:02.740 --> 00:21:06.940
known better, led an effort to ban fracking in California
384
00:21:07.000 --> 00:21:11.569
because hydrocarbons were such a dirty word. And I grew
385
00:21:11.630 --> 00:21:15.990
up in California in the 70s, 80s, 90s. And I
386
00:21:16.009 --> 00:21:18.049
remember going to Southern California and you could eat the
387
00:21:18.150 --> 00:21:20.549
air because it was so bad. So there was certainly
388
00:21:20.710 --> 00:21:24.880
reason to be anti-development of new energy resources.
389
00:21:25.440 --> 00:21:25.980
However.
390
00:21:26.690 --> 00:21:29.630
Here we are now. We still need that energy in
391
00:21:29.690 --> 00:21:33.839
order to make the transition into a cleaner energy environment.
392
00:21:34.190 --> 00:21:36.539
I look at data centers. I look at this growth
393
00:21:36.579 --> 00:21:41.920
of hyperscaler mass data collection center as the major tool
394
00:21:42.019 --> 00:21:45.119
to modernize our grid because we haven't had an economic
395
00:21:45.160 --> 00:21:48.960
driver that has done this as a business case in
396
00:21:49.140 --> 00:21:52.839
my lifetime. There hasn't been anything that could put this
397
00:21:52.960 --> 00:21:56.410
kind of capital to work. to upgrade what is a
398
00:21:56.430 --> 00:22:00.970
100-year-old grid. And so I think there's a deep level
399
00:22:01.019 --> 00:22:05.440
of short-sightedness for the governors that are reacting to this
400
00:22:05.759 --> 00:22:09.440
with what will be long-term bans. I think that there's
401
00:22:09.519 --> 00:22:13.109
room for both the technical improvement that the data center
402
00:22:13.130 --> 00:22:17.329
developers are doing to address things like noise, water, and
403
00:22:17.410 --> 00:22:20.930
energy consumption. But the fact that these guys are a
404
00:22:20.950 --> 00:22:27.099
24-7 baseload, standard load operator allows so much work to
405
00:22:27.119 --> 00:22:30.240
be done on the grids that need the work. I
406
00:22:30.279 --> 00:22:31.660
go to California all the time, but I'm in the
407
00:22:31.720 --> 00:22:34.599
Northeast now. The grid is 100 years old. Natural gas
408
00:22:34.640 --> 00:22:37.960
pipeline just south of San Francisco exploded, took out a
409
00:22:38.000 --> 00:22:41.460
whole neighborhood because it was a 100-year-old pipe. And so
410
00:22:41.500 --> 00:22:44.019
the only thing that can drive this level of investment
411
00:22:44.140 --> 00:22:47.700
is private industry with a real demand for the power.
412
00:22:47.940 --> 00:22:49.660
And that's what I think is coming. And that's what
413
00:22:49.700 --> 00:22:53.890
I think will be the product of this as the
414
00:22:53.970 --> 00:22:56.990
noise of the opposition calms down. And we're also in
415
00:22:57.009 --> 00:22:58.170
an election year.
416
00:22:58.190 --> 00:23:00.670
So Eugene, I like your vision and I agree with
417
00:23:00.690 --> 00:23:02.289
it because I think we actually do need to see
418
00:23:03.289 --> 00:23:07.410
AI as part of the solution. There's a huge amount
419
00:23:07.450 --> 00:23:08.380
of flexibility there.
420
00:23:08.410 --> 00:23:09.700
They all have backup power.
421
00:23:09.740 --> 00:23:12.180
They've got batteries in place. They've storage everywhere.
422
00:23:12.539 --> 00:23:13.539
You can move.
423
00:23:13.380 --> 00:23:18.000
Demand loads around the place. These guys have got great technology. Google,
424
00:23:18.019 --> 00:23:21.160
for example, has tapestry. I'm totally with you on this
425
00:23:21.240 --> 00:23:23.420
is that I think it's, and by the way, as
426
00:23:23.460 --> 00:23:27.009
you just said, they're willing to invest themselves and they're
427
00:23:27.109 --> 00:23:29.710
willing to sign offtake agreements. So use it to see
428
00:23:29.730 --> 00:23:30.589
it as an opportunity.
429
00:23:31.210 --> 00:23:33.230
I can't get past that the grid's 100 years old
430
00:23:33.329 --> 00:23:35.069
and we need we need to invest in it and
431
00:23:35.109 --> 00:23:36.789
nothing's going to drive it you're not going to go
432
00:23:36.809 --> 00:23:40.339
to the pucs and demand repairs pay more money because
433
00:23:40.880 --> 00:23:41.819
they can't afford.
434
00:23:41.579 --> 00:23:47.099
It yeah eugene i don't know if the glass is
435
00:23:47.140 --> 00:23:50.579
half full or half empty but there's certainly liquid into
436
00:23:50.640 --> 00:23:53.819
it so thank you very much for coming on the
437
00:23:53.839 --> 00:23:57.910
show and You know, having this kind of freewheeling conversation because,
438
00:23:57.930 --> 00:24:00.390
you know, there's a lot of data, but what's important
439
00:24:00.410 --> 00:24:03.349
for us are the people. And you're one of them.
440
00:24:03.430 --> 00:24:04.309
So thank you very much.
441
00:24:04.990 --> 00:24:07.599
It's my pleasure. And I wouldn't be a real estate
442
00:24:07.660 --> 00:24:10.559
broker at heart if I didn't give you no answer
443
00:24:10.579 --> 00:24:13.140
to whether or not the glass was half full or not. Yes,
444
00:24:13.160 --> 00:24:17.079
there is water. So I appreciate you having me on.
445
00:24:17.220 --> 00:24:18.819
Yeah, thanks, Eugene. Great having you.
446
00:24:19.920 --> 00:24:24.170
So Gerard, I know you are always a bull, always
447
00:24:24.250 --> 00:24:25.660
a bull. And you know why?
448
00:24:25.680 --> 00:24:26.880
Tell me.
449
00:24:27.700 --> 00:24:34.269
Because you're an equity guy. You love stories. And me,
450
00:24:34.349 --> 00:24:37.210
I'm a fixed income guy. So I look at credit.
451
00:24:37.789 --> 00:24:42.829
I'm quoting Apollo here. The credit story in hyperscaler rests
452
00:24:42.930 --> 00:24:47.390
on the single consensus assumption that the operating cash flow
453
00:24:47.509 --> 00:24:51.230
triples from $ 600 billion to $ 2 trillion. If it does
454
00:24:51.269 --> 00:24:54.230
not happen, everything's going to go to junk bonds. And
455
00:24:54.279 --> 00:24:57.880
it started. SoftBank, when they borrow now, it's junk bonds.
456
00:24:58.619 --> 00:25:02.769
We've seen the default, sorry, the force measure from Oracle
457
00:25:02.930 --> 00:25:06.549
on the Jupiter project in New Mexico. What that means
458
00:25:06.650 --> 00:25:09.349
now is if you look at the weakest part of
459
00:25:09.390 --> 00:25:13.309
the market, and I talk about it almost a year ago,
460
00:25:13.670 --> 00:25:17.849
the Blue Owl, the CoreWeaves, the Neoclouds, now they borrow
461
00:25:18.069 --> 00:25:22.490
at 8%, 9%, 10%, 11%. What the market says, not
462
00:25:22.529 --> 00:25:27.309
the equity jar, the credit market says, It's less risky
463
00:25:27.589 --> 00:25:31.730
to lend to Zambia and Angola than those new clouds.
464
00:25:32.069 --> 00:25:32.829
That's not going to work.
465
00:25:33.430 --> 00:25:35.730
Well, actually, I think it's probably worth talking about the
466
00:25:35.789 --> 00:25:38.019
equity side as well, because when you look at some
467
00:25:38.039 --> 00:25:43.740
of the valuations, obviously, Anthropic and OpenAI have not IPO'd
468
00:25:43.759 --> 00:25:46.839
at this point in time, but the numbers floated around
469
00:25:47.119 --> 00:25:50.049
are just... off the charts and you sort of wonder,
470
00:25:50.069 --> 00:25:51.589
are you ever going to really be able to get
471
00:25:51.609 --> 00:25:55.150
a return on an investment as an equity investor in
472
00:25:55.390 --> 00:25:56.450
those type of companies?
473
00:25:56.630 --> 00:25:58.930
I can tell you that all the early investors of
474
00:25:59.150 --> 00:26:03.880
their best PR companies going non-stop, non-stop, non-stop on media
475
00:26:03.900 --> 00:26:06.960
and pushing it, pushing it, pushing it. There are very,
476
00:26:07.019 --> 00:26:10.720
very few people in the equity who's going to say anything. Okay,
477
00:26:10.740 --> 00:26:13.140
so you always have the big short, Michael Burry and
478
00:26:13.160 --> 00:26:16.180
so on. But the debt, it's the debt. You need
479
00:26:16.240 --> 00:26:19.759
to check the debt. If your business model, and we've
480
00:26:19.779 --> 00:26:22.200
seen the new cloud, is to borrow at 9% and
481
00:26:22.240 --> 00:26:26.609
then lend at 1%, look, you've got the competition from
482
00:26:26.670 --> 00:26:29.910
the open weights. You see the price of compute going down.
483
00:26:29.990 --> 00:26:33.619
So I'm not saying AI is not a revolution. But
484
00:26:33.640 --> 00:26:36.900
what I'm saying is that the financial assumptions around the
485
00:26:37.119 --> 00:26:42.180
investment wave don't make any sense at the end of 2026.
486
00:26:42.180 --> 00:26:44.660
Listen, I think the point you made there is a revolution.
487
00:26:44.680 --> 00:26:47.740
In a revolution, what actually happens is there's a huge
488
00:26:47.759 --> 00:26:50.099
amount of money made and there's also a huge amount
489
00:26:50.119 --> 00:26:53.140
of money lost. And I think this revolution is going
490
00:26:53.160 --> 00:26:53.779
to be no different.
491
00:26:54.220 --> 00:26:57.880
But nevertheless, Gerard, the one who managed to build a
492
00:26:57.900 --> 00:27:02.059
data center, that's where Eugene was great. So whether it's
493
00:27:02.119 --> 00:27:05.359
the whole one or those type of things, when you
494
00:27:05.460 --> 00:27:07.660
manage to build a data center, a data center is
495
00:27:07.680 --> 00:27:08.829
very profitable right now.
496
00:27:09.380 --> 00:27:09.900
Oh, absolutely.
497
00:27:10.140 --> 00:27:10.740
Totally with Gerard.
498
00:27:11.180 --> 00:27:16.460
But that's all development pipeline was just insane. The reality
499
00:27:16.480 --> 00:27:19.690
is great, but the promises, just don't believe them.
500
00:27:20.329 --> 00:27:23.750
Good, my friend. Well, listen, as ever, good chatting. So
501
00:27:24.029 --> 00:27:25.089
I'm looking forward to next week.
502
00:27:25.390 --> 00:27:26.710
Okay, John. Cheers.
503
00:27:28.009 --> 00:27:31.730
Thank you for listening to Redefining Energy. Don't forget to
504
00:27:31.799 --> 00:27:36.500
rate the show and subscribe on Apple Podcasts, Spotify, or
505
00:27:36.519 --> 00:27:37.920
the platform of your choice.
