1
00:00:00,000 --> 00:00:01,800
Brad, welcome back to Austin.

2
00:00:02,120 --> 00:00:03,520
Parker, good to be here, man.

3
00:00:05,860 --> 00:00:12,180
Last episode with Pierre, we went real deep on the difficulty target and the difficulty adjustment.

4
00:00:13,100 --> 00:00:24,160
Today, we're going to talk about mining and ERCOT, ERCOT generally, grid mining, but also talking about the broader market and how Bitcoin mining fits into that.

5
00:00:24,160 --> 00:00:28,560
So we'll inevitably get deep, but appreciate you being here and recording.

6
00:00:28,560 --> 00:00:34,960
got brad cuddy from cholla inc what does the hat say never stop exploring never stopped exploring

7
00:00:34,960 --> 00:00:40,080
yep we've got to figure out uh if that was a gideon powell original or if it was passed down

8
00:00:40,080 --> 00:00:47,420
but love that i shot him a text no answer yet so not yet tbd all right we're gonna dive right in

9
00:00:47,420 --> 00:01:01,020
to an actual example of how bitcoin mining is interacting with the ercott grid in terms of

10
00:01:01,020 --> 00:01:07,380
winter storm and around scarcity events and then we'll kind of come back up for air to talk about

11
00:01:10,120 --> 00:01:16,980
how different levers why things are happening but there's a chart that ercott had put out during

12
00:01:16,980 --> 00:01:27,840
winter storm heather in january of 2024 that the the storm spanned two or three days and

13
00:01:27,840 --> 00:01:39,280
the the price of power effectively increased from you know 15 dollars a megawatt hour to

14
00:01:39,280 --> 00:01:49,900
on a log scale looks like 750 dollars to thousand dollars a megawatt hour and approximately 1.1

15
00:01:49,900 --> 00:01:58,660
gigawatts of power specifically large flexible load came offline the power price came back down

16
00:01:58,660 --> 00:02:07,140
basically peaked came back down another scarcity event this time almost 1.6 gigawatts of large

17
00:02:07,140 --> 00:02:15,520
flexible load came offline and then there was a third peak and similarly so talk about

18
00:02:15,520 --> 00:02:23,540
the actual incentives that are driving that and what might be happening to create a scarcity event

19
00:02:23,540 --> 00:02:31,620
in ERCOT during the winter and then how ERCOT helps manage that and how Bitcoin miners play into it

20
00:02:31,620 --> 00:02:51,060
Yeah, I think this is really hitting on the beauty of Bitcoin mining in ERCOT. It's a completely economically rational responder to a price signal. You can see there that they have a line drawn at, what is it, 120 megawatt hours? It's like an assumed average of breakeven price.

21
00:02:51,060 --> 00:03:09,860
Right, 122 megawatt hours, which is what ERCOT had calculated as what they viewed Bitcoin mining's break-even being based on a S19J Pro, which is just interesting that someone at ERCOT is doing this math, presenting this to ERCOT.

22
00:03:09,860 --> 00:03:28,320
Yeah, when Evan Neal was at ERCOT, he was in the weeds. He's a Bitcoiner. So we had one of us in there. He's since left, but I digress. So what you can see is that the pink highlights on the graph, those are ERCOT issued conservation appeals.

23
00:03:28,320 --> 00:03:36,540
Those are alerts that are coming out via your retailer or your queasy or even like to a residential user on your ERCOT app.

24
00:03:36,620 --> 00:03:39,100
And like, hey, looks tight.

25
00:03:39,460 --> 00:03:42,360
We don't know if we have enough generation to serve the projected load.

26
00:03:42,480 --> 00:03:46,760
So if you could, you know, not use as much load, that would be appreciated.

27
00:03:47,120 --> 00:03:51,860
And this is when in the winter, everyone's running their heaters.

28
00:03:51,880 --> 00:03:52,180
Yes.

29
00:03:52,780 --> 00:03:55,920
Freezing temperatures, maxing it out.

30
00:03:55,980 --> 00:03:56,160
Right.

31
00:03:56,160 --> 00:04:03,060
and temperatures are going down so it's likely at some point early in like overnight early in the

32
00:04:03,060 --> 00:04:09,020
morning right it's it's likely yeah overnight early in the morning uh we have that in the in

33
00:04:09,020 --> 00:04:13,940
the winter time it'll peak in the morning time sometimes you'll get an evening peak too but the

34
00:04:13,940 --> 00:04:19,940
load curves look completely different on uh on a winter versus summer but here what you will what

35
00:04:19,940 --> 00:04:28,440
you'll see is the blue, which is the total large flexible load, start to step down as that price

36
00:04:28,440 --> 00:04:34,900
line, the black line starts to go up. So the miners are doing this purely off of like an

37
00:04:34,900 --> 00:04:39,900
economically rational decision. Like if you're making 120 bucks a megawatt hour and power costs

38
00:04:39,900 --> 00:04:45,200
$200 a megawatt hour, you're losing money. It's a quick way to go out of business. Like you do not

39
00:04:45,200 --> 00:04:51,080
want to be doing this. There's a potential to have some layered on hedges or be committed to

40
00:04:51,080 --> 00:04:55,760
an ancillary service so you have to keep your load on. Don't want to get too nuanced here,

41
00:04:56,400 --> 00:05:05,120
but the Bitcoin miners are responding to the price. The price is a signal of generation scarcity.

42
00:05:05,940 --> 00:05:12,540
As generation becomes more and more scarce, it goes up and up the generation offer curve,

43
00:05:12,540 --> 00:05:19,940
and it settles at the lowest price generator that clears the market. And so in this situation,

44
00:05:19,940 --> 00:05:27,680
there is less and less generation. They're likely hitting the EORDC operating reserve demand curve.

45
00:05:27,940 --> 00:05:37,960
That's a long string of EORDC. Sorry, sorry. EORDC? Operating reserve demand curve. And so

46
00:05:37,960 --> 00:05:44,080
that basically when ERCOT gets to, I think it's two or 3000 megawatts worth of reserve capacity

47
00:05:44,080 --> 00:05:49,620
left, meaning late in gen that can turn on, like when it's, when load in gen is within $2,000,

48
00:05:50,160 --> 00:05:56,200
they start increasing the prices to send the right price signals to have every generator turn on,

49
00:05:56,260 --> 00:06:00,320
whether it's like this old steam generator, whether it's a marginally economic coal unit,

50
00:06:00,680 --> 00:06:07,280
like they're sending the price sky high and then hoping that the market responds and, you know,

51
00:06:07,280 --> 00:06:11,000
acts rationally in an economic basis and Bitcoin money does here.

52
00:06:11,620 --> 00:06:12,580
And so on the other side,

53
00:06:12,660 --> 00:06:13,240
on the other side,

54
00:06:13,260 --> 00:06:13,620
exactly.

55
00:06:13,880 --> 00:06:14,220
Yeah.

56
00:06:14,600 --> 00:06:19,080
Cause the price is going up and nobody wants to eat a thousand or $2,000 a

57
00:06:19,080 --> 00:06:19,660
megawatt hour.

58
00:06:19,760 --> 00:06:19,960
It's a,

59
00:06:19,960 --> 00:06:21,700
it's a very quick way to,

60
00:06:21,740 --> 00:06:22,360
to lose money.

61
00:06:23,360 --> 00:06:26,240
So if someone is looking at the chart,

62
00:06:26,240 --> 00:06:29,700
when this winter storm starts,

63
00:06:29,820 --> 00:06:35,440
there's about 2.2 gigawatts of large flexible load,

64
00:06:35,560 --> 00:06:36,680
which is,

65
00:06:37,280 --> 00:06:41,040
In your mind, is that virtually all Bitcoin mining?

66
00:06:41,800 --> 00:06:42,140
Yes.

67
00:06:43,060 --> 00:06:43,760
Practically speaking.

68
00:06:43,920 --> 00:06:44,900
Yeah, there is.

69
00:06:45,220 --> 00:06:47,400
And we'll talk about why that's the case,

70
00:06:47,400 --> 00:06:51,300
but just wanted to touch on this now,

71
00:06:51,820 --> 00:06:54,860
that as the market price started moving up,

72
00:06:54,980 --> 00:06:58,860
before the first conservation notice,

73
00:06:59,340 --> 00:07:04,320
about 500 megawatts had already come offline.

74
00:07:04,320 --> 00:07:09,680
and then as the price ramped further

75
00:07:09,680 --> 00:07:15,560
another one gigawatt of power came offline

76
00:07:15,560 --> 00:07:21,480
and a bitcoin a bitcoiner would look at that and say that's that's bitcoin

77
00:07:21,480 --> 00:07:29,940
mining helping the grid but this interesting note from from ercott which is basically saying not all

78
00:07:29,940 --> 00:07:34,020
large flux not all not 100 came offline but making making that point right

79
00:07:34,020 --> 00:07:41,720
Um, but then basically the market price comes down, the large flexible load comes back up to

80
00:07:41,720 --> 00:07:48,960
like 1.9 gigawatts. And then the next time there's a conservation notice, 1.6 comes offline.

81
00:07:48,960 --> 00:07:56,920
Yeah. Um, but just talk about the, um, the relationship there. Like it's not a,

82
00:07:57,000 --> 00:08:02,500
it's a request a hundred percent of the market, right? Even like grandma and grandpa at their

83
00:08:02,500 --> 00:08:08,040
house this is a this isn't a notice that's going out just to bitcoin miners correct yeah this is is

84
00:08:08,040 --> 00:08:14,220
uh this is an you know ercott wide notice conservation appeal and you can even see on

85
00:08:14,220 --> 00:08:19,960
that third section there wasn't even an ercott conservation appeal which is indicated by the pink

86
00:08:19,960 --> 00:08:27,640
but the bitcoin miners responded to the price anyways and it's interesting because the the price

87
00:08:27,640 --> 00:08:33,100
if you if you look at the price on that third peak goes higher goes higher but there's not

88
00:08:33,100 --> 00:08:38,660
the conservation notice so what would is that because they have more generation available

89
00:08:38,660 --> 00:08:49,240
you know i'm not sure on this specific um that specific peak it could been um they were forecasted

90
00:08:49,240 --> 00:08:53,600
a warmer temperature and it actually was colder so they didn't have the time to

91
00:08:53,600 --> 00:08:57,600
put out a conservation nobody's going to read it at three o'clock in the morning if this is

92
00:08:57,600 --> 00:09:04,440
like a morning hour peak um or it could be that there's there's more to system reliability than

93
00:09:04,440 --> 00:09:12,440
just price um so they might have felt that they had enough operating reserves uh to to satisfy

94
00:09:12,440 --> 00:09:19,080
enough ancillaries or whatever it might have been um but there is there is more to reliability than

95
00:09:19,080 --> 00:09:23,860
just price and then the last thing because i have a few of these charts we won't bring the other ones

96
00:09:23,860 --> 00:09:34,680
up but so to lead into the broader discussion it is relevant that ERCOT is looking at this

97
00:09:34,680 --> 00:09:41,640
and tracking it and paying attention to not just what's supposed to happen in theory but what's

98
00:09:41,640 --> 00:09:48,860
happening in practice but that if we looked at a similar analysis that ERCOT had done in

99
00:09:48,860 --> 00:09:57,400
2022 there was approximately 1.5 gigawatts of large flexible load around a scarcity event on

100
00:09:57,400 --> 00:10:03,620
the grid to start and then in 2023 1.9 gigawatts and then this example that we were talking about

101
00:10:03,620 --> 00:10:13,220
from september 2023 to january 2024 2.2 gigawatts so this amount of large flexible load continuing to

102
00:10:13,220 --> 00:10:22,800
increase. Approximately how much Bitcoin mining is on grid in ERCOT today, would you estimate?

103
00:10:23,480 --> 00:10:30,800
I think the LFL, large flexible load estimation, is right around three and a half gigawatts.

104
00:10:31,480 --> 00:10:39,460
LFL's classification is greater than 75 megawatts. There's also a lot of miners sub-75 megawatts,

105
00:10:39,460 --> 00:10:45,960
whether it's transmission level, but just on a 50 megawatt interconnect or distribution level like that we are too.

106
00:10:46,140 --> 00:10:52,240
So if you add all of that up, I think we're probably in the high three gigawatts, maybe low four gigawatts.

107
00:10:52,240 --> 00:10:59,800
And to put it all into perspective, total system wide peak load, I think is 85 or 86 gigawatts.

108
00:10:59,820 --> 00:11:01,180
So it's a significant portion.

109
00:11:02,000 --> 00:11:05,080
Is that fairly consistent between the summer peak and the winter peak?

110
00:11:05,080 --> 00:11:11,220
Or is that 85, 86 gigawatts of power being demanded summer peak?

111
00:11:11,980 --> 00:11:18,880
That's the summer peak, but we are shifting towards most likely to being a winter peaking system.

112
00:11:19,380 --> 00:11:19,720
Interesting.

113
00:11:19,720 --> 00:11:48,060
And then, so to transition this into a more general conversation, what in your mind allows Bitcoin mining to be flexible when other sources of demand for power on a grid system are not, and how that benefits the grid?

114
00:11:48,060 --> 00:11:55,360
yes for the first part um you know every other industrial electrified process on the grid

115
00:11:55,360 --> 00:12:03,520
has electricity plus some process equals some widget or output steel aluminum whatever it is

116
00:12:03,520 --> 00:12:08,700
sold for dollars so there's like because of the added complexity the added workload you know

117
00:12:08,700 --> 00:12:15,160
people staffing all like raw materials um if you don't finish a smelt maybe it doesn't go all the

118
00:12:15,160 --> 00:12:21,680
way through and you have to throw something away. Bitcoin mining, we have no customer. We have no

119
00:12:21,680 --> 00:12:27,980
output besides money. We're directly transferring electrons into money. And because of that, we have

120
00:12:27,980 --> 00:12:35,980
the purest, most rational price behavior and response that there is. And how that benefits

121
00:12:35,980 --> 00:12:41,980
is what you saw there, right? Price goes up. Bitcoin miners, as an economically rational

122
00:12:41,980 --> 00:12:47,980
consumer of electrons will respond to the price as long as the incentives are aligned. And that's

123
00:12:47,980 --> 00:12:53,140
why I think ERCOT is so great because you can, in the deregulated energy-only market,

124
00:12:53,640 --> 00:12:56,720
participate in this. And that's why you saw so many miners flock here.

125
00:12:58,360 --> 00:13:05,160
And what does it mean for ERCOT to be deregulated? And how does that compare to

126
00:13:05,160 --> 00:13:07,240
other grids in the United States?

127
00:13:07,240 --> 00:13:31,720
So what's still regulated in ERCA is the transmission and distribution companies. What is deregulated is the generators and then the retail side. That's why Texans have retail choice. You can go shop around for your electricity. That's part of the deregulated movement that came in, I think, the late 90s, early 2000s.

128
00:13:31,720 --> 00:13:41,620
um so because of that like you have the um the access to this real-time market and the day ahead

129
00:13:41,620 --> 00:13:51,540
market that urcott has created in their energy only market and so is it the fact that there's a

130
00:13:51,540 --> 00:14:01,460
real-time price that that real-time price doesn't necessarily exist in other grids because i think

131
00:14:01,460 --> 00:14:06,800
if i'm understanding correctly the fact that the real-time price exists and it's being communicated

132
00:14:06,800 --> 00:14:17,840
to all of the the market at any one time that allows a purely economic actor like a bitcoin

133
00:14:17,840 --> 00:14:22,440
miner who again doesn't have a widget that they're producing or a manufacturing plant that's running

134
00:14:22,440 --> 00:14:30,100
or a hospital that has patients or a grandmother grandfather in freezing weather to to respond

135
00:14:30,100 --> 00:14:36,200
because they have the pricing but is there something about ERCOT that has you know because

136
00:14:36,200 --> 00:14:42,200
I consistently hear the term real-time pricing does that not exist in the same way in other grids

137
00:14:42,200 --> 00:14:50,520
that makes ERCOT somewhat unique and I'm not an expert on like Kaiso, PJM, all these other grids

138
00:14:50,520 --> 00:14:58,460
I don't know for certain on their real-time pricing but I also don't know if like as a

139
00:14:58,460 --> 00:15:03,020
consumer as a Bitcoin miner, consumer of electrons, if you could have access to that real-time

140
00:15:03,020 --> 00:15:19,617
pricing So I think that really what ERCOT has different is like the ability as a Bitcoin miner to access that real pricing and respond economically whether it to the load zone pricing or whether it to the node pricing

141
00:15:19,617 --> 00:15:30,317
and then as we've seen the large flexible load increase so from 2022 at 1.5 gigawatts

142
00:15:30,317 --> 00:15:38,857
september 2023 1.9 gigawatts january 2024 2.2 and now estimating between 3.5 before

143
00:15:38,857 --> 00:16:03,337
Or how does scale factor in to not the Bitcoin mining operation itself, but the ability to have more power come offline at points where there is scarcity relative to generation coming online?

144
00:16:03,997 --> 00:16:04,558
Right.

145
00:16:04,558 --> 00:16:13,178
I mean, naturally, as the amount of Bitcoin mining load increases on the system, it gives you more flexibility.

146
00:16:13,538 --> 00:16:21,777
Four gigawatts of load that is able to turn off on a price signal is a massive boon to the system operator.

147
00:16:22,438 --> 00:16:24,977
What they don't want is it all turning off at once.

148
00:16:24,977 --> 00:16:39,438
And we have a responsibility as Bitcoin miners, as stakeholders in ERCOT, to be good stewards, to follow ramp rates, to act responsibly and economically rationally.

149
00:16:40,197 --> 00:16:51,178
But it's only good because of how much intermittent generation is coming to ERCOT, whether it's wind, because the wind belt intersects North Dakota to West Texas.

150
00:16:51,178 --> 00:16:54,317
and then the solar belt goes from California to West Texas.

151
00:16:54,837 --> 00:16:56,717
We have a ton of intermittent generation.

152
00:16:57,397 --> 00:16:58,737
For the most part, it was stranded.

153
00:16:58,997 --> 00:17:00,357
For a long time, it was stranded.

154
00:17:01,258 --> 00:17:03,138
I think over the last eight years,

155
00:17:03,938 --> 00:17:07,918
so pretty much since the inception of Hottel Ranch

156
00:17:07,918 --> 00:17:10,078
when Gideon went out to West Texas,

157
00:17:10,858 --> 00:17:13,138
and this isn't just Bitcoin mining,

158
00:17:13,217 --> 00:17:14,418
it's some oil and gas load,

159
00:17:14,938 --> 00:17:18,658
but the load in West Texas has increased.

160
00:17:18,658 --> 00:17:26,737
i think it's at least threefold load is is demand for power load is demand for power instead of

161
00:17:26,737 --> 00:17:30,638
generation the generation was already going to west texas because that's where the wind and

162
00:17:30,638 --> 00:17:37,017
that's where the solar is where bitcoin bitcoin's a location agnostic consumer of that load it

163
00:17:37,017 --> 00:17:42,757
doesn't need to be close to a city center it goes where like it goes where it can it goes where it

164
00:17:42,757 --> 00:17:44,517
where power is Jesus.

165
00:17:45,037 --> 00:17:45,457
Power, exactly.

166
00:17:46,277 --> 00:17:47,857
And so if you have all this generation

167
00:17:47,857 --> 00:17:48,777
siding out somewhere

168
00:17:48,777 --> 00:17:50,617
that doesn't really have that much load,

169
00:17:51,017 --> 00:17:52,277
it was a natural incentive

170
00:17:52,277 --> 00:17:53,537
created by the market

171
00:17:53,537 --> 00:17:56,457
to show that Bitcoin mining,

172
00:17:56,658 --> 00:17:57,737
if you can curtail,

173
00:17:58,257 --> 00:17:59,877
like if you can respond to price,

174
00:18:00,497 --> 00:18:01,797
West Texas is the place to be.

175
00:18:02,078 --> 00:18:02,737
And we've seen that

176
00:18:02,737 --> 00:18:03,697
over the last eight years,

177
00:18:04,418 --> 00:18:05,037
three times,

178
00:18:05,318 --> 00:18:06,257
the total load

179
00:18:06,257 --> 00:18:08,398
in load zone West,

180
00:18:08,517 --> 00:18:09,578
which is a classification

181
00:18:09,578 --> 00:18:11,257
of all the way to

182
00:18:11,257 --> 00:18:20,138
like pecos piyo monahans midland basically the entire permian and delaware basin is in low zone

183
00:18:20,138 --> 00:18:28,797
west all the way down to big bend and then over to abilene i believe and so to have the frame of

184
00:18:28,797 --> 00:18:39,658
reference you mentioned that at peak like peak peak demand might be 85 gigawatts of power yeah

185
00:18:39,658 --> 00:18:50,217
and bitcoin mining might be four gigawatts that if it were only 100 megawatts

186
00:18:50,217 --> 00:18:57,578
the ability for the market to feel an impact of that to solve a problem is lower but what's

187
00:18:57,578 --> 00:19:05,898
functionally happening and again power has to get to specific points on the grid that it's

188
00:19:05,898 --> 00:19:31,277
Just to use a rough example, everything's more complicated than this, that as demand is going from 81 gigawatts to 85 gigawatts, 4 gigawatts coming offline versus the highest cost generators coming online is what helps mitigate maybe that price ceiling that the entire market would have to bear.

189
00:19:31,277 --> 00:19:41,777
And so as Bitcoin mining represents more, at least in peak times, it can help solve a larger problem for greater scarcity events.

190
00:19:42,318 --> 00:19:45,277
Correct. Because we're not on when the price is going high.

191
00:19:45,638 --> 00:19:51,237
Like $122 a megawatt hour is $0.12 a kilowatt hour.

192
00:19:51,398 --> 00:19:59,838
Most people's delivered power in the entire greater United States is greater than that price where pretty much all Bitcoin miners are off.

193
00:19:59,838 --> 00:20:05,158
right and this is in this was in january of 2024 hash rates increased significantly so that

194
00:20:05,158 --> 00:20:11,537
that perceived break even or marginal break even for bitcoin might actually be

195
00:20:11,537 --> 00:20:18,158
lower at six cents or seven cents a kilowatt hour today who knows what it is but um that it that it's

196
00:20:18,158 --> 00:20:25,057
not static two two concepts and we'll talk about one of them first you talked about ramp

197
00:20:25,057 --> 00:20:32,257
ramping down and ramping up somebody might generally understand bitcoin miners responding

198
00:20:32,257 --> 00:20:42,838
to price signals but ercott's the electricity reliability council of texas that talk about the

199
00:20:42,838 --> 00:20:52,178
actual challenge of why having to balance supply and demand is not just a market function but

200
00:20:52,178 --> 00:20:58,217
a reliability function of the grid itself and to go into a little bit of detail when you were

201
00:20:58,217 --> 00:21:04,898
talking about the importance of how say a large load might ram down or up to be a

202
00:21:04,898 --> 00:21:15,057
not a steward but a good market participant right ever since winter storm yuri which every texan

203
00:21:15,057 --> 00:21:21,057
remembers urkot's been in the in the limelight as much as they don't want to be but for those

204
00:21:21,057 --> 00:21:26,377
for those people who are not familiar with winter storm yuri that was it was actually before my time

205
00:21:26,377 --> 00:21:35,777
here it was 2021 i believe it was a cold snap that lasted something like nine days um and there was

206
00:21:35,777 --> 00:21:42,138
various reasons for uh for it but price price went to the cap but at that time it was nine

207
00:21:42,138 --> 00:21:49,517
thousand dollars a megawatt hour um generation was unable to black start pipelines were freezing

208
00:21:49,517 --> 00:21:55,297
wasn't sunny so the solar wasn't producing no wind it was like a black swan through and through

209
00:21:55,297 --> 00:22:01,377
but uh it ended up with you know a lot of rolling blackouts and when people lose power and it's

210
00:22:01,377 --> 00:22:06,318
super cold and you can't turn on the heat uh you know our young and our old are are susceptible to

211
00:22:06,318 --> 00:22:10,717
that and so there was there was dust that was associated with it it was it was top of mind for

212
00:22:10,717 --> 00:22:16,617
everybody um and that just highlights how important it is for urcott especially because it is

213
00:22:16,617 --> 00:22:20,457
an island and system. And I'll just explain that real quick. Yeah, if you could.

214
00:22:21,037 --> 00:22:26,557
So the entire Eastern part of the United States is under the Eastern interconnect.

215
00:22:27,318 --> 00:22:31,938
It's all electrically connected. The entire Western part of the United States, you know,

216
00:22:31,938 --> 00:22:40,898
everywhere from Washington state down to like New Mexico over to Colorado and maybe like the

217
00:22:40,898 --> 00:22:46,477
bordering Nebraska or whatever, is electrically connected. So they have all that generation,

218
00:22:46,638 --> 00:22:52,457
all that spinning mass. They've got nuclear here, you've got solar there. There's a lot of support

219
00:22:52,457 --> 00:22:58,857
from the entire grid. Like, okay, how much could it actually support? It's so far away. Well,

220
00:22:58,857 --> 00:23:05,057
electrons travel at the speed of light and not to get too technical, they don't actually travel,

221
00:23:05,257 --> 00:23:09,998
blah, blah, blah, but it's very quick. And so with all of that system-wide generation,

222
00:23:09,998 --> 00:23:16,938
all that system-wide load, the interconnect can support fluctuations, bolted fluctuations,

223
00:23:17,217 --> 00:23:23,578
generator trips, load trips, just because of its inherent size. ERCOT is completely electrically

224
00:23:23,578 --> 00:23:30,957
isolated. It is essentially, you could imagine in an island, it has like small DC ties, like direct

225
00:23:30,957 --> 00:23:34,757
current. That's what I, my understanding is that there were some ties, but I don't know enough to

226
00:23:34,757 --> 00:23:41,318
know how or to what extent yeah so a couple gigawatts of dc but the thing is our grid doesn't

227
00:23:41,318 --> 00:23:47,857
run on dc it runs on alternating current you know it's three magnets spinning around generating

228
00:23:47,857 --> 00:23:54,457
current it's like that spinning mass is what generated our electricity pre-wind and solar

229
00:23:54,457 --> 00:24:03,217
um and so that creates the alternating current that is is like the heartbeat the 60 hertz of our

230
00:24:03,217 --> 00:24:08,357
grid. Uh, so the DC doesn't help support that. So it gives us a little bit, but it doesn't give us

231
00:24:08,357 --> 00:24:16,617
the, uh, the support. Um, and so why, why ERCOT was so unique is because of that islanded situation.

232
00:24:16,617 --> 00:24:21,918
Like it has to rely on itself. And so because it has to rely on itself, it needs to prioritize

233
00:24:21,918 --> 00:24:29,337
reliability, right? You said it's, it's, it's in the name. Um, I got in the weeds and I forgot

234
00:24:29,337 --> 00:24:40,737
a little bit of the question but so like talking about the function within ERCOT of why or how

235
00:24:40,737 --> 00:24:49,857
loads how they ramp up or down from an actual grid reliability standpoint is a concern beyond just

236
00:24:49,857 --> 00:24:58,297
price volatility and bitcoin miners responding to price to come off that um how they do that

237
00:24:58,297 --> 00:25:04,318
right is consequential but just talk about that concept and i and i think something to highlight

238
00:25:04,318 --> 00:25:12,737
too is that you know bitcoin mining as a load might get a lot of criticism or like oh it goes

239
00:25:12,737 --> 00:25:19,097
off too quick like it's not following a ramp right or it's um you know previously the conversation

240
00:25:19,097 --> 00:25:25,037
was this is even real like is this just going to go away um i don't i don't think we get enough

241
00:25:25,037 --> 00:25:31,998
credit because we are able to like very, very fine tune our load. Most load that is on the grid is

242
00:25:31,998 --> 00:25:39,018
like heating load. It's totally blind to ERCOT. They have no visibility into this. Bitcoin miners,

243
00:25:39,018 --> 00:25:47,697
if they qualify as a controllable load resource, they're able to submit their load curve into

244
00:25:47,697 --> 00:25:55,258
into ERCOT so they have complete visibility all generation submits their their generation stack

245
00:25:55,258 --> 00:26:02,837
into the curve on the flip side submitting the load side into into the curve allows ERCOT to match

246
00:26:02,837 --> 00:26:12,318
help a system reliability and ensuring that the frequency is in that and it has the range of 60

247
00:26:12,318 --> 00:26:17,638
hertz it has to be within a very tight band at all times correct yeah and like that would be

248
00:26:17,638 --> 00:26:26,457
that would be supported via um ancillary services like fast frequency response regulation up

249
00:26:26,457 --> 00:26:32,537
regulation down it basically will request say frequency goes too high meaning the generators

250
00:26:32,537 --> 00:26:39,277
are spinning too much you need to bring load up bring the frequency down say frequency goes too low

251
00:26:39,277 --> 00:26:43,578
means not enough generation on the system.

252
00:26:43,717 --> 00:26:45,258
You bring load to match.

253
00:26:45,578 --> 00:26:47,898
And so that's something that Bitcoin miners

254
00:26:47,898 --> 00:26:50,617
are participating in.

255
00:26:51,418 --> 00:26:53,938
However, batteries are able to do it pretty well.

256
00:26:54,318 --> 00:26:56,438
So they have kind of eaten up most of that market.

257
00:26:57,037 --> 00:27:00,758
But Bitcoin miners are in other services

258
00:27:00,758 --> 00:27:04,957
like non-spin or ERS.

259
00:27:05,857 --> 00:27:06,638
And these are like all,

260
00:27:06,918 --> 00:27:08,418
basically ERCOT has,

261
00:27:08,418 --> 00:27:12,857
because there's no capacity market, it's just energy only. They have some different

262
00:27:12,857 --> 00:27:19,318
ancillary demand response is basically what this gets lumped under. But the technical term is like

263
00:27:19,318 --> 00:27:26,597
an ancillary market. The ancillary market has different products that can be bid into as a load

264
00:27:26,597 --> 00:27:33,537
or as a generator. And so Bitcoin miners can provide these ancillary services to ERCOT.

265
00:27:33,537 --> 00:27:52,377
And this is, you know, like, for example, the controversial riot filing in like September during an EEA event, they committed their load to, they committed their load to be an ancillary service.

266
00:27:52,597 --> 00:27:57,258
They said, we will stay on and we will only be curtailed when you tell us.

267
00:27:57,258 --> 00:27:58,998
So they ignored the price.

268
00:27:59,117 --> 00:28:00,537
You have to have a hedge to do this.

269
00:28:00,597 --> 00:28:02,237
Otherwise, it's not economically rational.

270
00:28:02,237 --> 00:28:09,658
but they use their bitcoin mining as a way to support grid frequency or voltage or whatever it

271
00:28:09,658 --> 00:28:16,638
was and by doing so they are paid the thing is like the the bucket of ancillary services

272
00:28:16,638 --> 00:28:23,398
is it's like it's it's on a bid basis they they clear as the lowest bid and so if riot was

273
00:28:23,398 --> 00:28:31,297
called as that ancillary service that means that they were one of the lower bids so as batteries

274
00:28:31,297 --> 00:28:39,438
and Bitcoin mining reduce the cost to bid into the ancillary services, all ERCOP stakeholders

275
00:28:39,438 --> 00:28:45,197
benefit from it. Right. Because if I'm interpreting that correctly, because if they weren't there,

276
00:28:45,377 --> 00:28:51,837
somebody with a higher bid. Correct. It would have been a steel mill or a heavy industrial that

277
00:28:51,837 --> 00:28:59,557
wouldn't have that fine nuance control of being able to like step down, you know, 20% of your load

278
00:28:59,557 --> 00:29:04,638
every minute over the five minutes to get down to certain targets, right? Like you can,

279
00:29:04,638 --> 00:29:08,398
you can fine tune Bitcoin mining, like each miner's three kilowatts, right? You can,

280
00:29:08,638 --> 00:29:17,297
you can selectively curtail to bring, you know, if they tell you to go, I want 1000 megawatts to go

281
00:29:17,297 --> 00:29:25,377
to 654, like Bitcoin mining can do that. Could a steel mill do this? No. Could a refinery do this?

282
00:29:25,377 --> 00:29:30,377
no batteries yes they could also do something like this and then generation on the other side

283
00:29:30,377 --> 00:29:37,758
can also ramp up to a specific capacity as long as they have that but um that's like something very

284
00:29:37,758 --> 00:29:45,837
very unique with bitcoin and maybe the highest level importance of this is that if those the

285
00:29:45,837 --> 00:29:54,037
demand for power and the generation of power is not lined up at all times and if that frequency

286
00:29:54,037 --> 00:30:02,957
is an imbalance then that's where you get blackouts or brownouts and if there was a crazy

287
00:30:02,957 --> 00:30:09,138
event where the potential potentially the whole grid could go down is that fair yes like voltage

288
00:30:09,138 --> 00:30:34,355
out of sync frequency out of sync yes it can all that that the the main concern And historically before Bitcoin mining the primary way that load and generation demand supply were balanced was by supply changing

289
00:30:35,255 --> 00:30:35,535
Correct?

290
00:30:35,735 --> 00:30:41,155
Like generators coming up, down, is that fair?

291
00:30:41,335 --> 00:30:41,555
Yes.

292
00:30:41,555 --> 00:30:52,275
it was by generation and then um bringing that to the way that that bitcoin miners come up and down

293
00:30:52,275 --> 00:31:01,015
is that because each rig could be controlled on an automated basis it can be very precise

294
00:31:01,015 --> 00:31:05,695
but even though it might be good from a market standpoint if

295
00:31:05,695 --> 00:31:13,775
right it's 500 megawatts came offline if it shut off unpredictably the ercot not knowing it or

296
00:31:13,775 --> 00:31:20,315
based on some ramp down schedule it could actually have a negative effect yes unreliability but

297
00:31:20,315 --> 00:31:24,835
because it is so flexible and because it can be controlled as long as they are

298
00:31:24,835 --> 00:31:33,415
as long as they are in sync with the grid operator then it can be a real asset it requires a

299
00:31:33,415 --> 00:31:41,895
partnership you can't just be flipping the breaker on a thousand megawatts on a 85 mega 85 000 megawatt

300
00:31:41,895 --> 00:31:50,855
system like it requires you to to be a good steward um it's it's it's every large industrial load

301
00:31:50,855 --> 00:31:55,255
has to do something like this like there's procedures to to be there like we're playing

302
00:31:55,255 --> 00:32:01,675
big boy games like we are four gigawatts in an 84 gigawatt system we are making a material impact

303
00:32:01,675 --> 00:32:08,255
we are a large large consumer of electricity in texas and like like it's time to put on your big

304
00:32:08,255 --> 00:32:14,835
boy pants and uh and act like it from your perspective and clearly urcott's paying attention

305
00:32:14,835 --> 00:32:21,995
to how large flexible loads aka bitcoin miners are performing in practice

306
00:32:21,995 --> 00:32:32,575
from your seat how how do you think that ercott views bitcoin mining and what what in your mind

307
00:32:32,575 --> 00:32:37,555
is their level of understanding not to say any express knowledge but just your general perception

308
00:32:37,555 --> 00:32:46,055
yeah i think overall the perception is continuing to trend more positively like it's it's on us as

309
00:32:46,055 --> 00:32:51,375
market participants and stakeholders to to be involved and be engaged and and be good stewards

310
00:32:51,375 --> 00:32:56,655
We need to be showing up to these large load working group meetings up at ERCOT.

311
00:32:57,315 --> 00:33:09,055
We need to be having the conversations with the system operators that are in the room operating the system and talking about what Bitcoin mining can actually do.

312
00:33:09,355 --> 00:33:10,555
We're in the weeds, right?

313
00:33:10,955 --> 00:33:11,375
We're Bitcoiners.

314
00:33:12,075 --> 00:33:13,615
We've gotten into the Bitcoin mining industry.

315
00:33:14,255 --> 00:33:15,435
We know what we can do.

316
00:33:16,775 --> 00:33:19,755
But it's still a matter of education.

317
00:33:19,755 --> 00:33:27,375
You still say mining Bitcoin and you'll get mixed responses.

318
00:33:28,215 --> 00:33:34,235
But I do think really it's an ongoing conversation.

319
00:33:35,515 --> 00:33:37,355
We can't rest on our laurels.

320
00:33:38,355 --> 00:33:48,495
Bitcoin mining in Texas has succeeded because the Texas market was the most primed to receive

321
00:33:48,495 --> 00:33:54,615
the Bitcoin mining shift post-China ban, but it's not guaranteed.

322
00:33:54,975 --> 00:34:01,155
Like the rules can change in ERCOT, new markets can, like a new rule in the market could be

323
00:34:01,155 --> 00:34:01,915
implemented.

324
00:34:01,915 --> 00:34:11,095
It is on us to make sure that we are still successfully selling ourselves and performing

325
00:34:11,095 --> 00:34:13,895
to the standard that we are talking about.

326
00:34:13,895 --> 00:34:22,815
um i do think that you know overall the value is seen like they there's like you you can see right

327
00:34:22,815 --> 00:34:30,155
there on the chart that you know bitcoin miners large flexible loads are responding to price like

328
00:34:30,155 --> 00:34:35,275
we are doing what we said we were doing is there room for growth and is there room for improvement

329
00:34:35,275 --> 00:34:41,075
and can we get those like last 100 or 200 megawatts down like yes like that should be the

330
00:34:41,075 --> 00:34:48,235
the goal of the industry to um to optimize as much as we can but like time and time again

331
00:34:48,235 --> 00:34:55,495
bitcoin mining has proven to be a mutually beneficial addition to the texas grid because

332
00:34:55,495 --> 00:35:01,855
we're consuming the hours that are cheap they're consuming the hours where like we are incentivizing

333
00:35:01,855 --> 00:35:08,835
more generation to come online you know we're on depending on your break even efficiency 75 to 95

334
00:35:08,835 --> 00:35:15,315
5% of the time, maybe even 99, depending, but you're off on those stress events.

335
00:35:15,315 --> 00:35:20,455
If you're off on those stress events, you are net positive through and through to the

336
00:35:20,455 --> 00:35:20,655
grid.

337
00:35:21,115 --> 00:35:25,475
And like, it's, it's starting, it's starting to feel like we're, we are getting some respect

338
00:35:25,475 --> 00:35:28,395
and we are, they do understand like what's going on.

339
00:35:28,975 --> 00:35:30,035
And yeah, that's okay.

340
00:35:30,055 --> 00:35:32,495
That's a good place to move to next.

341
00:35:33,375 --> 00:35:36,475
Before I do, I just want to reinforce something that you said.

342
00:35:36,475 --> 00:35:40,855
I think it was, I don't know what year it was, maybe it was 2023.

343
00:35:42,115 --> 00:35:49,175
There was a, because ERCOT, I don't know what the right way to say this is.

344
00:35:50,275 --> 00:35:57,375
ERCOT is regulated by the PUC, Public Utility Commission, and then the PUC answers to the

345
00:35:57,375 --> 00:35:58,035
state legislature.

346
00:35:58,035 --> 00:36:05,015
So ultimately, state legislature regulates both the PUC and ERCOT.

347
00:36:05,015 --> 00:36:23,215
And there was a bill that Senator Lois Colchorz put out that would have capped Bitcoin mining participating to a certain percentage of one of the ancillary service programs.

348
00:36:23,215 --> 00:36:35,835
And again, the summary of ancillary services is a tool in the toolbox for ERCOT to help balance supply and demand to keep the grid reliable.

349
00:36:36,295 --> 00:36:41,895
And that's a perfect example of why education is important.

350
00:36:41,895 --> 00:36:54,435
Because even if ERCOT is understanding it and being able to see in the market and being able to interact with the actual stakeholders, people at the statehouse need to understand as well.

351
00:36:55,175 --> 00:37:02,155
Because in the example you gave, if it wasn't for Riot being there, then the cost of the system would have been greater.

352
00:37:02,155 --> 00:37:07,775
And there might be nuance in terms of having a diverse market to ensure that the market is competitive.

353
00:37:07,775 --> 00:37:13,055
but having that education informs policy.

354
00:37:14,415 --> 00:37:17,955
But for somebody, because I think this happened,

355
00:37:18,095 --> 00:37:19,435
or I know this happened,

356
00:37:19,815 --> 00:37:23,635
when Riot went to have their second site in,

357
00:37:24,335 --> 00:37:25,475
I don't know if it's their second site,

358
00:37:25,535 --> 00:37:27,095
but their second large site in Corsicana,

359
00:37:27,715 --> 00:37:30,195
there were a group of concerned citizens who,

360
00:37:31,355 --> 00:37:33,475
if they do not understand Bitcoin,

361
00:37:33,475 --> 00:37:37,395
look at four gigawatts of power,

362
00:37:37,775 --> 00:37:48,175
Coming online, consuming power for a reason that they think is waste or they think is waste, they do not understand.

363
00:37:50,055 --> 00:37:58,195
Go into greater detail about how four gigawatts of power could come online.

364
00:37:59,415 --> 00:38:07,475
And somebody might be able to conceptualize why coming offline at those peaks might save at the peak.

365
00:38:07,775 --> 00:38:15,715
but the general dynamic of why or why not without new generation coming online,

366
00:38:15,935 --> 00:38:20,755
how it doesn't increase the power, the cost of power overall, or maybe it does.

367
00:38:21,315 --> 00:38:28,715
You just talk about that dynamic of, of maybe how power is priced, how it's absorbing power

368
00:38:28,715 --> 00:38:34,775
that might not be being utilized in off peaks and then coming offline, just that general dynamic to

369
00:38:34,775 --> 00:38:44,515
help articulate for somebody who wouldn't otherwise reconcile a lot of demand 24-7 or

370
00:38:44,515 --> 00:38:56,095
mostly 24-7 coming on with a fixed capacity of generation yeah it's and you touched on it and

371
00:38:56,095 --> 00:39:01,555
you mentioned it it's just a better utilization like these nat gas turbines these solar farms

372
00:39:01,555 --> 00:39:10,275
these wind farms they're already in the ground they already exist but i the 70 or something like

373
00:39:10,275 --> 00:39:18,315
that of all energy is is wasted um is that 70 energy but it's not electricity but yeah and what

374
00:39:18,315 --> 00:39:25,335
i was going to ask is that 70 of electricity capacity yeah like ercott's capacity is much

375
00:39:25,335 --> 00:39:30,115
much hot like and forgive me for not knowing the number but it's over 100 gigawatts worth of

376
00:39:30,115 --> 00:39:36,755
capacity. We're only hitting 85 gigawatts of system wide peak, right? And so-

377
00:39:36,755 --> 00:39:38,735
For short periods of time for certain days of the year.

378
00:39:38,735 --> 00:39:45,795
You know, for 0.5% of the time is when that's actually happening. And, you know, overnight,

379
00:39:46,015 --> 00:39:53,895
maybe your load is only down, you're down to 50 or 60 gigawatts or even lower. And so basically

380
00:39:53,895 --> 00:40:02,635
what Bitcoin does is it increases the floor. It monetizes those hours that you wouldn't be

381
00:40:02,635 --> 00:40:08,195
running that generation otherwise, making the generation more economic. But the price that

382
00:40:08,195 --> 00:40:14,315
the generator bids in on the curve, it stays the same. And so as long as coming back to that price

383
00:40:14,315 --> 00:40:23,635
responsiveness, if price is going too high, Bitcoin goes off and the low disappears. And

384
00:40:23,635 --> 00:40:29,255
it's not increasing anybody's prices because actually like, like the studies that have been

385
00:40:29,255 --> 00:40:38,075
that have been done around, uh, cost of increase of power, uh, for residential customers has been,

386
00:40:38,075 --> 00:40:43,815
it's largely around the distribution cost, transmission, distribution cost, um, transmission,

387
00:40:43,935 --> 00:40:49,175
a smaller portion of the distribution and distribution is getting, you know, power to

388
00:40:49,175 --> 00:40:55,615
residential buildings, apartment complexes. It's the wires that you see,

389
00:40:55,995 --> 00:41:01,835
not the massive overhead wires. It's the smaller stuff. That's the lion's share of

390
00:41:01,835 --> 00:41:08,515
the... That's the lion's share of why our electricity prices are going up in Texas,

391
00:41:08,515 --> 00:41:16,175
specifically. Doug Lewin had a podcast talking about this. It's not the actual input of electricity

392
00:41:16,175 --> 00:41:23,755
costs. There's times where there's pricing on the grid in West Texas where it's negative $30

393
00:41:23,755 --> 00:41:30,835
a megawatt hour, but the residential consumer doesn't see that benefit because they're still

394
00:41:30,835 --> 00:41:35,795
getting hit with distribution charges and transmission charges and a fixed price.

395
00:41:35,995 --> 00:41:39,235
The incentives need to be aligned and they're working on that for retailers.

396
00:41:39,235 --> 00:41:49,155
um but having these large loads like the one gigawatt and corsicana it's only increasing

397
00:41:49,155 --> 00:41:54,755
utilization and the economic output of a generator and if that generator is

398
00:41:54,755 --> 00:42:02,415
economically like on like it's producing electrons it's paying its tax revenue somebody's buying like

399
00:42:02,415 --> 00:42:08,975
it's creating jobs it's creating better utilization of the assets we already have on the grid

400
00:42:08,975 --> 00:42:15,315
and then also incentivizing new generation because like ERCOT's not contracting any generators

401
00:42:15,315 --> 00:42:23,435
generators are going at risk and responding to price signals on the grid in order to cite their

402
00:42:23,435 --> 00:42:29,035
generation and in order to in order to build it like they are they're not guaranteed anything

403
00:42:29,035 --> 00:42:36,335
there's no capacity market high prices are the signal to generators to go out and build that load

404
00:42:36,335 --> 00:42:44,055
and so as system-wide load and you can see this massive interconnect queue in ERCOT that they're

405
00:42:44,055 --> 00:42:49,095
talking about 150 gigawatts or something like that generators are looking at this like okay

406
00:42:49,095 --> 00:42:56,395
this is AI this is Bitcoin largely AI but some Bitcoin too we're going to go build generation

407
00:42:56,395 --> 00:43:03,415
in anticipation of this and then the generation bids into the market and that's how you get your

408
00:43:03,415 --> 00:43:08,615
energy price but the beautiful thing you know whether you like solar and wind or not is that

409
00:43:08,615 --> 00:43:15,535
the the input cost to wind generation and solar generation is zero dollars a megawatt hour it

410
00:43:15,535 --> 00:43:20,635
doesn't cost anything for the wind to blow or the sun to shine um the marginal cost the marginal cost

411
00:43:20,635 --> 00:43:29,335
correct it it ends up having during the solar hours and during windy hours very low pricing

412
00:43:29,335 --> 00:43:37,115
um so what does that do it shifts the peaks to when the sun's going down early in the morning

413
00:43:37,115 --> 00:43:43,235
during the winter uh and that's providing incentive maybe to batteries if the peaks

414
00:43:43,235 --> 00:43:48,595
are only two two hours three hours or if they're long sustained and we're starting to see high

415
00:43:48,595 --> 00:43:54,615
prices in the evening time in loads on west maybe a nat gas generation wants to go out to

416
00:43:54,615 --> 00:44:03,395
west texas take some very cheap uh waha gas put out peaker plant in there and they build out some

417
00:44:03,395 --> 00:44:09,495
more generation but the consumers are not paying for more generation to get built generations are

418
00:44:09,495 --> 00:44:16,335
going at risk to build this generation based off the price and so a way to think about it

419
00:44:16,335 --> 00:44:24,455
was part of what you mentioned is if the total capacity of generation is 100 gigawatts

420
00:44:24,615 --> 00:44:36,595
in ERCOT at any point in time. Peak is generally 85 gigawatts. That on a megawatt hour basis,

421
00:44:37,255 --> 00:44:45,655
it's 100 gigawatts times 365 days a year. And that if you, I don't know what the precise number

422
00:44:45,655 --> 00:44:51,995
is, but if you looked at every hour versus 365 days a year, 100 gigawatts capacity versus what

423
00:44:51,995 --> 00:44:58,895
is actually used it's probably something like 30 30 or 40 somewhere in there that if you can

424
00:44:58,895 --> 00:45:07,735
absorb more of that asset capacity and have greater utilization the transmission lines are there

425
00:45:07,735 --> 00:45:15,595
that it increases the profitability which allows the economics to improve and also flattens

426
00:45:15,595 --> 00:45:38,512
the curve and that a net benefit maybe shift to talk about because you brought up two things you brought up ai and batteries let Let talk about both of those But first Bitcoin demand for power and its characteristics relative to something like an AI the source of demand for the power

427
00:45:38,892 --> 00:45:40,773
How are those similar?

428
00:45:41,192 --> 00:45:42,112
How do those differ?

429
00:45:42,412 --> 00:45:50,872
And what would be the considerations for a grid system or grid operator like ERCOT looking at both of those?

430
00:45:51,212 --> 00:45:55,372
And they're definitely different loads, even though they may seem similar.

431
00:45:55,372 --> 00:46:02,392
But what I will say is that Bitcoin mining paved the way for these AI data centers to come to Texas.

432
00:46:03,192 --> 00:46:11,672
700 megawatts at Rockdale and one gigawatt at Corsicana, that is an unheard of level of interconnect.

433
00:46:11,672 --> 00:46:23,793
Data centers prior to this AI revolution were maybe high double digits, low 100s of megawatts.

434
00:46:24,412 --> 00:46:38,112
So I'm going to pause that thought because there was something else that I wanted to talk about that feeds into this, which is we talked about ERCOT being a deregulated energy market.

435
00:46:38,112 --> 00:46:52,313
And what that means is that generation is deregulated and then the retail market, people actually selling power to commercial loads, homes, that piece is deregulated.

436
00:46:52,852 --> 00:46:55,852
The transmission and distribution is regulated.

437
00:46:56,952 --> 00:47:03,672
And there's a process to being able to interconnect to the grid and that that's a long process.

438
00:47:03,672 --> 00:47:24,512
So talk, maybe distinguish between the regulated pieces and the deregulated pieces, and then how that factors into this discussion of how long of a lead time it might take to get a 500 megawatt site online and connected to the ERCOT grid or a gigawatt site.

439
00:47:25,852 --> 00:47:29,752
I mean, we'll tie it back to the reliability piece, right?

440
00:47:29,813 --> 00:47:31,672
Everything is under the guise of reliability.

441
00:47:31,672 --> 00:47:37,073
You see this massive load interconnection queue, hundreds plus gigawatts.

442
00:47:38,452 --> 00:47:49,052
Encore, AEP, the TDSPs, the transmission distribution service providers, the wires companies, they have to take these interconnect requests and be like, are they real?

443
00:47:49,712 --> 00:47:51,432
We've never seen something this big.

444
00:47:51,552 --> 00:47:53,012
We've seen a couple of Bitcoin mines.

445
00:47:53,672 --> 00:47:55,452
This is unheard of demand.

446
00:47:55,452 --> 00:47:58,273
and then they have to work with ERCOT

447
00:47:58,273 --> 00:48:00,492
to make sure that the system operator,

448
00:48:01,273 --> 00:48:04,032
like nothing breaks in their model.

449
00:48:04,273 --> 00:48:06,252
And so it's an aggregation of you.

450
00:48:06,372 --> 00:48:10,192
You've got your load, you've got your TDSP,

451
00:48:10,492 --> 00:48:12,172
and you have ERCOT all working together

452
00:48:12,172 --> 00:48:17,273
to try to fit 150 gigawatts plus

453
00:48:17,273 --> 00:48:20,252
into an 85 gigawatt peaking system.

454
00:48:20,252 --> 00:48:23,252
And so there's a lot of...

455
00:48:25,452 --> 00:48:26,313
Basically, it's unprecedented.

456
00:48:26,732 --> 00:48:29,752
They haven't seen something like this before.

457
00:48:30,372 --> 00:48:31,592
And this amount of growth is...

458
00:48:33,692 --> 00:48:36,652
Utilities grew at maybe single-digit percentages.

459
00:48:37,293 --> 00:48:41,832
Post-COVID, ERCOT was growing maybe 6%, 7%, 8%,

460
00:48:41,832 --> 00:48:45,532
which is a very, very, very high rate compared to everywhere else.

461
00:48:46,232 --> 00:48:52,712
We're talking like estimations of 15% of able to actually get done.

462
00:48:52,712 --> 00:49:03,892
but uh what they're requesting is is 150 of what the ERCOT total system-wide peak is right now

463
00:49:03,892 --> 00:49:11,112
uh and so that's wild it's it's absolutely insane and you know in their right to take

464
00:49:11,112 --> 00:49:15,972
take a step back and be like is this real like do we actually have to serve this serve this load

465
00:49:15,972 --> 00:49:24,073
um and so these ai data centers there's different types of data centers like you've got the

466
00:49:24,073 --> 00:49:31,372
you know the models the training model uh training the models that can be um you don't need as good

467
00:49:31,372 --> 00:49:38,532
of latency you don't need to be in the the city centers those might be the ones competing for

468
00:49:38,532 --> 00:49:43,912
competing with bitcoin miners like going out to deep west texas going to surrounded assets

469
00:49:43,912 --> 00:49:51,132
then you have inference inference compute what's actually running your chat gbt answer and query

470
00:49:51,132 --> 00:49:56,912
humans are okay they're all right with waiting a little bit we understand that like okay it's

471
00:49:56,912 --> 00:50:02,813
got to think for a second um that inference compute that's closer to the city centers

472
00:50:02,813 --> 00:50:12,852
uh and then you've got the the edge whatever but right now what we're seeing in texas is

473
00:50:12,852 --> 00:50:18,432
Dallas is a hotspot. South Dallas is a hotspot for AI compute. Then you've got north of Austin,

474
00:50:18,712 --> 00:50:25,793
you've got Samsung, you've got the Gigafactory. Data centers are coming there, but you do have

475
00:50:25,793 --> 00:50:32,132
them coming all the way out to West Texas. And that maybe is where the intersection of Bitcoin

476
00:50:32,132 --> 00:50:40,732
mining and AI compute will be. Galaxy just contracted with CoreWeave for their entire

477
00:50:40,732 --> 00:50:46,872
capacity and so what we're seeing now is galaxy digital yeah okay they're mining

478
00:50:46,872 --> 00:50:56,192
their mining business is um their helios their flagship site uh has completely decommissioned

479
00:50:56,192 --> 00:51:04,332
the the mining infrastructure and is working towards converting to strictly ai hpc or core weave

480
00:51:04,332 --> 00:51:10,952
and but say just for a single site say if a

481
00:51:10,952 --> 00:51:18,212
setting aside this crazy amount of forecasted growth based on all of the

482
00:51:18,212 --> 00:51:26,492
requests that have been put in you want to get a 500 megawatt site online in ERCOT

483
00:51:26,492 --> 00:51:32,592
what does that process look like and how long does it take good luck is uh is what i would say

484
00:51:32,592 --> 00:51:43,972
right now everything has more or less uh ground to a halt uh we came in at the right time and we

485
00:51:43,972 --> 00:51:49,032
were early and were able to secure these interconnect requests they will they still

486
00:51:49,032 --> 00:51:54,432
have like an obligation to serve and the tdsps have the obligation to serve what is tdsp

487
00:51:54,432 --> 00:52:02,172
transmission and distribution service provider however the time to interconnect now for a

488
00:52:02,172 --> 00:52:08,652
a large load we're talking years and years and years down the road like there is no easy access

489
00:52:08,652 --> 00:52:17,252
to power anymore that has been more or less taken up within the last year or two with with all this

490
00:52:17,252 --> 00:52:27,552
increased load interconnections and so the reason why the ai data a bitcoin mining data center

491
00:52:27,552 --> 00:52:34,152
might be so valuable to an AI operator

492
00:52:34,152 --> 00:52:40,012
is the ability to access large amounts of power

493
00:52:40,012 --> 00:52:46,152
without having to get a new interconnect approved through ERCOT.

494
00:52:46,612 --> 00:52:48,092
Yeah, I think we're going to see this.

495
00:52:48,352 --> 00:52:50,012
I mean, we're already seeing this.

496
00:52:52,692 --> 00:52:55,032
Building out generation and power infrastructure

497
00:52:55,032 --> 00:52:57,872
is a matter of national security at this point, right?

498
00:52:57,952 --> 00:52:59,572
Like we're betting on AI.

499
00:53:01,092 --> 00:53:02,172
It's an arms race.

500
00:53:02,552 --> 00:53:05,632
And it comes down to who has the most energy

501
00:53:05,632 --> 00:53:06,512
and power and electrons.

502
00:53:07,532 --> 00:53:12,632
And so it's imperative that we do so.

503
00:53:15,192 --> 00:53:18,612
Right now, they're trying to get all capacity

504
00:53:18,612 --> 00:53:20,032
that's left on the market.

505
00:53:20,392 --> 00:53:22,112
When that capacity dries up

506
00:53:22,112 --> 00:53:25,392
and we're looking at three, four year interconnect times

507
00:53:25,392 --> 00:53:27,672
and they have nothing in their pipeline,

508
00:53:28,092 --> 00:53:30,212
this is where I would see like the Mag7

509
00:53:30,212 --> 00:53:33,012
picking off the large interconnects.

510
00:53:33,152 --> 00:53:35,172
So, you know, like Iron's doing this right now.

511
00:53:35,252 --> 00:53:38,892
Like they've become a preferred partner of NVIDIA.

512
00:53:39,492 --> 00:53:42,492
They started Bitcoin mining in Childress,

513
00:53:43,172 --> 00:53:45,172
West Texas, close to the panhandle,

514
00:53:45,512 --> 00:53:47,313
lots of wind, decent amount of solar.

515
00:53:48,432 --> 00:53:49,412
Is it near Lubbock?

516
00:53:49,412 --> 00:53:59,552
uh yeah it's on the way okay it's on the way or on the way to amarillo okay uh 287 or 187

517
00:53:59,552 --> 00:54:04,972
anyways yeah it's it's on the way up to like denver so it's on that fiber corridor actually

518
00:54:04,972 --> 00:54:12,632
of dallas to denver and so they're like a great prime example they're monetizing the electrons

519
00:54:12,632 --> 00:54:19,592
with Bitcoin mining to start and then marketing this to the AI and HPC compute companies like,

520
00:54:19,592 --> 00:54:24,432
hey, we have this power. If you want it, come and get it. We'll sell you the rack space.

521
00:54:27,572 --> 00:54:33,232
Maybe there's going to be total acquisitions. I know Riot's right now marketing the rest of

522
00:54:33,232 --> 00:54:38,592
their core Sakana site. They got 400 megawatts worth of Bitcoin mining currently energized,

523
00:54:38,592 --> 00:54:45,892
600 megawatts of latent capacity south dallas is completely locked up of course the canna is not

524
00:54:45,892 --> 00:54:52,972
that much farther south than dallas it's only an hour drive away very easily can see the sale

525
00:54:52,972 --> 00:55:00,632
of that of 600 megawatts whether it's a total acquisition of riot or it's um you know them

526
00:55:00,632 --> 00:55:11,012
selling that 600 megawatts or them hosting with that rack space so maybe talk about what is the

527
00:55:11,012 --> 00:55:18,313
difference in profile in terms of what an ai data center is willing to pay for power versus what a

528
00:55:18,313 --> 00:55:25,672
bitcoin miner might be willing to pay yeah we're talking orders of magnitude different uh capex

529
00:55:25,672 --> 00:55:32,732
spend one gpu something like eighty thousand dollars it's it's producing flops or something

530
00:55:32,732 --> 00:55:39,552
like that and they're well as one gpu is how much i've heard numbers of eighty thousand dollars i'm

531
00:55:39,552 --> 00:55:44,472
not in the weeds of this thing but like they're they're very expensive so they are incentivized

532
00:55:44,472 --> 00:55:52,012
to be running their compute as much as possible um i think their generalized break-even number

533
00:55:52,012 --> 00:55:58,512
is a thousand plus a megawatt hour go ahead you were gonna say something yeah so a bitcoin miner

534
00:55:58,512 --> 00:56:06,092
in asic is between four and five thousand pricing wise yeah yeah you can even get some cheap used

535
00:56:06,092 --> 00:56:16,192
ones but your price for the asic probably 500 bucks to 5 000 and then your break is 70

536
00:56:16,192 --> 00:56:21,432
per megawatt hour to $150 a megawatt hour currently.

537
00:56:21,872 --> 00:56:22,152
Okay.

538
00:56:22,813 --> 00:56:25,112
I'm trying to just think about some of the unit economics.

539
00:56:25,732 --> 00:56:26,752
So that's...

540
00:56:26,752 --> 00:56:27,552
Don't use my ADK.

541
00:56:27,572 --> 00:56:28,592
No, I'm not going to use your ADK,

542
00:56:28,692 --> 00:56:29,852
but like, let's say 50K.

543
00:56:29,972 --> 00:56:30,192
Yeah.

544
00:56:30,852 --> 00:56:32,332
To maybe be conservative.

545
00:56:32,332 --> 00:56:36,472
How much power does a,

546
00:56:37,232 --> 00:56:38,452
on a relative basis,

547
00:56:38,592 --> 00:56:44,512
does a Bitcoin ASIC draw versus a AI GPU?

548
00:56:44,512 --> 00:56:53,273
asics you know air-cooled asics we're looking at low 3ks hydro it's all up to 10k

549
00:56:53,273 --> 00:56:58,372
aw okay and then immersion somewhere in between like five or six thousand

550
00:56:58,372 --> 00:57:07,232
i know that right now the gpus so they're making it up as they go by the way like what bitcoin

551
00:57:07,232 --> 00:57:16,552
mining went through and like s9s to s17s to s19s like this really quick iteration and burn and

552
00:57:16,552 --> 00:57:23,852
churn is happening right now with nvidia like the racks for cloud compute like your your compute that

553
00:57:23,852 --> 00:57:30,852
ran your netflix that runs your google query like your traditional intranet infrastructure

554
00:57:30,852 --> 00:57:40,332
was like 10 or 20 kilowatts per rack and a rack is like a 48 unit um it's a it's a little taller

555
00:57:40,332 --> 00:57:50,112
than a person it's like seven feet tall maybe uh 10 to 20 kw they are now talking about trying to

556
00:57:50,112 --> 00:57:56,813
work towards one megawatt a rack like the density is getting absolutely insane they used to use like

557
00:57:56,813 --> 00:58:05,412
rear door heat exchangers now they're going strictly direct chip hydro and but contextualize

558
00:58:05,412 --> 00:58:11,293
that relative to bitcoin yeah to bitcoin the same the same rack and we were we were breaking barriers

559
00:58:11,293 --> 00:58:18,293
like 200 kw or rack is what you can put in um with bitcoin mining asics with like the what's

560
00:58:18,293 --> 00:58:23,712
minor hydro or the aridine hydro or the canaan or bit deer those two you hydro units that you see

561
00:58:23,712 --> 00:58:35,032
um they consume 10 kw uh and you can put about 200 kw inside of a rack and so i don't actually

562
00:58:35,032 --> 00:58:42,232
i don't think h100s are and that's like a common gpu that nvidia makes i don't think h100s are to

563
00:58:42,232 --> 00:58:47,152
that 200 kw yet i don't know if they've commercialized 200 kw but they've they're

564
00:58:47,152 --> 00:58:55,112
talking about in the future of having 400 600 kw racks and like they're going for as dense as

565
00:58:55,112 --> 00:59:03,052
possible as my understanding maybe this is a shift that's happening because of ai was that the shift

566
00:59:03,052 --> 00:59:10,832
from gpus to bitcoin a6 the power density of bitcoin a6 was far greater than the power density

567
00:59:10,832 --> 00:59:17,652
of a legacy gpu but maybe the gap is shrinking because of the nvidia chips and ai is that

568
00:59:17,652 --> 00:59:25,352
we were 100 more dense uh before like it like i was like i was saying like 20 kw is what your

569
00:59:25,352 --> 00:59:35,092
your normal racks uh heat capacity was or power capacity 200 kw is what bitcoin was this was you

570
00:59:35,092 --> 00:59:43,252
know like rewind three years 20 kw data center 200 kw bitcoin mining they're catching up i don't

571
00:59:43,252 --> 00:59:49,972
think they're at 200 kw a rack yet but they do plan on passing us in like rack density and

572
00:59:49,972 --> 00:59:57,952
my understanding is that the build out of an ai data center is far more expensive something like

573
00:59:57,952 --> 01:00:05,032
on the order of magnitude of 10x the cost that's less consequential than

574
01:00:05,032 --> 01:00:10,212
they're willing to pay way more for power

575
01:00:10,212 --> 01:00:19,732
and that is because whatever their end market is their their revenue source

576
01:00:19,732 --> 01:00:22,032
is greater than

577
01:00:22,032 --> 01:00:23,273
bitcoins on a

578
01:00:23,273 --> 01:00:24,132
power

579
01:00:24,132 --> 01:00:26,572
per unit

580
01:00:26,572 --> 01:00:27,293
economics

581
01:00:27,293 --> 01:00:28,152
is that fair

582
01:00:28,152 --> 01:00:30,172
yes

583
01:00:30,172 --> 01:00:30,652
I mean their

584
01:00:30,652 --> 01:00:31,412
capex is

585
01:00:31,412 --> 01:00:32,293
higher

586
01:00:32,293 --> 01:00:32,832
but that's their

587
01:00:32,832 --> 01:00:33,172
cost

588
01:00:33,172 --> 01:00:33,492
yeah

589
01:00:33,492 --> 01:00:34,212
but their

590
01:00:34,212 --> 01:00:34,932
but their

591
01:00:34,968 --> 01:00:40,388
Their willingness to pay, and you said potentially up to what per megawatt hour?

592
01:00:41,468 --> 01:00:46,288
Figures I'm hearing somewhere in like the mid $1,000 a megawatt hour.

593
01:00:46,588 --> 01:00:53,088
Versus Bitcoin mining effectively being breakeven somewhere between $70 a megawatt hour and $150.

594
01:00:53,388 --> 01:00:53,568
Yeah.

595
01:00:56,568 --> 01:00:57,248
What?

596
01:00:57,248 --> 01:01:04,488
we were talking at the beginning of this episode about real world examples where

597
01:01:04,488 --> 01:01:11,328
in total in terms of all the bitcoin mining that is in urkat going from 1.5

598
01:01:11,328 --> 01:01:20,248
gigawatts to 1.9 to 2.2 to somewhere around 4 gigawatts today and that flexibility of the load

599
01:01:20,248 --> 01:01:23,388
being a value to the system,

600
01:01:24,248 --> 01:01:28,748
is AI flexible?

601
01:01:29,308 --> 01:01:30,048
Why or why not?

602
01:01:30,868 --> 01:01:32,588
Can you talk about that in relation to the nature

603
01:01:32,588 --> 01:01:34,888
of the differences in the load?

604
01:01:35,848 --> 01:01:36,608
It's a good question.

605
01:01:37,248 --> 01:01:39,828
And I think it's still yet to be determined.

606
01:01:40,468 --> 01:01:42,248
Google has put out some white papers

607
01:01:42,248 --> 01:01:43,728
talking about load flexibility.

608
01:01:44,248 --> 01:01:46,708
They all know that flexibility is important.

609
01:01:47,108 --> 01:01:48,608
They know you can't just be running,

610
01:01:48,608 --> 01:01:59,308
99.999, they call it the 5.9, so like 99.59% uptime is what the industry standard for a data

611
01:01:59,308 --> 01:02:06,888
center uptime was. And if they want to maintain that, they know that they can't rely on the grid

612
01:02:06,888 --> 01:02:14,908
to do so. Where we're at, because there's an insatiable demand for power, there's the opportunity

613
01:02:14,908 --> 01:02:23,088
you know for urcot to engage with these conversation engage these hyperscalers in

614
01:02:23,088 --> 01:02:29,128
this conversation like we need some flexibility like how can you build systems to do so and because

615
01:02:29,128 --> 01:02:34,848
the gpus are so expensive and because the capex spend of these data centers is so expensive

616
01:02:34,848 --> 01:02:43,308
they're able to add in ups systems like what's a ups system um uninterruptible power supply

617
01:02:43,308 --> 01:02:50,148
they can even put in grid scale batteries you can put in e-stat comms which helps with like

618
01:02:50,148 --> 01:02:58,368
voltage issues uh because because their capex spend is so high they can add reliability uh

619
01:02:58,368 --> 01:03:04,268
you're effectively saying backup generation well and then i was scared there and then they also have

620
01:03:04,268 --> 01:03:12,608
diesel backup generation too so like they have all of these different tools um that

621
01:03:12,608 --> 01:03:21,548
could be used to help reliability. You could switch over onto your battery in a grid scarcity

622
01:03:21,548 --> 01:03:29,268
event. They're consuming just like us for most of the time. Price goes high, EA event, something

623
01:03:29,268 --> 01:03:33,968
like that. Bitcoin miners are just turning off. Fine, whatever. We just don't need to sell our

624
01:03:33,968 --> 01:03:40,348
compute into the Bitcoin market. Data centers could switch over to their UPS,

625
01:03:40,348 --> 01:03:50,268
slowly ramp down what the grid sees they could then kick on their backup generation

626
01:03:50,268 --> 01:03:56,848
and then get them to like a checkpoint or to where they need to be on say their training model

627
01:03:56,848 --> 01:04:07,628
or you know reduce load to a certain extent like there are tools and ways to make all of this ai

628
01:04:07,628 --> 01:04:14,728
growth in texas not an issue on a reliability standard and that's what like we're trying to do

629
01:04:14,728 --> 01:04:18,328
what i'm because yeah where i'm getting to is

630
01:04:18,328 --> 01:04:21,648
is that

631
01:04:21,648 --> 01:04:25,308
a

632
01:04:25,308 --> 01:04:34,608
ai data center is right and everyone throws around the term ai and you know could be chat gbt someone

633
01:04:34,608 --> 01:04:41,268
Mid-journey, the end market is people typing in requests to GROC,

634
01:04:41,908 --> 01:04:48,468
to LGBT, mid-journey, and operations are happening at a data center

635
01:04:48,468 --> 01:04:49,768
and then push back out.

636
01:04:50,448 --> 01:04:57,708
That might not be as complex or sensitive as an operation as a foundry,

637
01:04:58,708 --> 01:05:04,008
working on ASICs, or a smelter, or a hospital.

638
01:05:04,608 --> 01:05:11,708
But it is still a source of demand that is more complex.

639
01:05:12,948 --> 01:05:18,948
And if they were just to go down, then the thing that's actually paying the thousand gigawatt,

640
01:05:19,128 --> 01:05:24,748
or sorry, thousand dollars a megawatt hour for power, that it would be interrupting.

641
01:05:24,748 --> 01:05:32,748
And so it's not so much a complex energy problem as it is a service delivery problem.

642
01:05:32,748 --> 01:05:39,008
problem whereas in bitcoin miners they're when when they shut down

643
01:05:39,008 --> 01:05:46,988
the bitcoin network works perfectly well right because everyone is providing power to one

644
01:05:46,988 --> 01:05:53,488
aggregated source of demand versus on the ai side it is customers paying for a service and

645
01:05:53,488 --> 01:05:58,128
if you typed into the grok and you were paying for the service and you didn't get

646
01:05:58,128 --> 01:06:03,308
response back because the service was down because texas was having a scarcity event

647
01:06:03,308 --> 01:06:10,448
then you're not going to be willing to pay thousand dollars a megawatt hour plus as the end customer

648
01:06:10,448 --> 01:06:18,088
so it might be possible in the future but it doesn't seem like

649
01:06:18,088 --> 01:06:27,048
the ai data centers are getting near the the concern trolling that bitcoin miners

650
01:06:28,128 --> 01:06:35,908
do or were and it might actually cause a problem if all of this because i can understand economically

651
01:06:35,908 --> 01:06:43,088
why if there's four gigawatts power in texas that's being consumed by large flexible loads

652
01:06:43,088 --> 01:06:49,348
and maybe not all four gigawatts large flexible loads that are above 75 megawatts let's say 3.5

653
01:06:49,348 --> 01:06:58,328
gigawatts are if bitcoin mine or sorry if ai data centers are willing to pay 10x for the power

654
01:06:58,328 --> 01:07:07,448
then the economically rational response to that is to sell your site or sell lease your site

655
01:07:07,448 --> 01:07:13,708
however you do it transfer it to an ai data center but then the benefit of having 3.5

656
01:07:13,708 --> 01:07:19,748
gigawatts to four gigawatts of large flexible load that can be hyper responsive to scarcity

657
01:07:19,748 --> 01:07:27,428
events in a very direct way goes away so is that part of the discussion that people are having

658
01:07:27,428 --> 01:07:35,728
out there in the market or is it largely not because the the power brokers that be are

659
01:07:35,728 --> 01:07:46,328
amazon microsoft google twitter grok what's the perception of that you know i think we don't know

660
01:07:46,328 --> 01:07:52,208
we don't know because abilene like the stargate campus the 1.2 gigawatts like it's it's still

661
01:07:52,208 --> 01:07:58,528
being built like there's there is data center load in texas but that's traditional data center load

662
01:07:58,528 --> 01:08:07,028
this new um ai compute it's it's it's not here in any sizable way uh but you're you're completely

663
01:08:07,028 --> 01:08:13,748
right like bitcoin mining completely pure economic signal like we will respond to the price signal

664
01:08:13,748 --> 01:08:21,708
ai's price signal is a lot higher right and they have a different more convoluted equation to

665
01:08:21,708 --> 01:08:27,008
like their their payback and like service to their customers and like downtime and all this type of

666
01:08:27,008 --> 01:08:34,828
stuff i think the i think the key is is that load growth is good for texas because load growth means

667
01:08:34,828 --> 01:08:41,228
more investment dollars like it's trillions and trillions of dollars are going to flow into texas

668
01:08:41,228 --> 01:08:51,368
if texas gets this right the key is flexibility like urkot has a position of power at this point

669
01:08:51,368 --> 01:08:58,628
Like they can, they have the best, most competitive market and grid system.

670
01:08:59,508 --> 01:09:01,968
Permissionless gen can just connect and then manage.

671
01:09:03,268 --> 01:09:04,668
We want the load to be here.

672
01:09:04,808 --> 01:09:08,748
We want Bitcoin mining to be here because it's a different load profile and it's going to be a different place.

673
01:09:08,868 --> 01:09:11,728
Bitcoin mining, it can be two megawatts.

674
01:09:11,828 --> 01:09:12,728
It can be three megawatts.

675
01:09:12,788 --> 01:09:15,208
Like you don't need scale to be a profitable Bitcoin mining.

676
01:09:15,408 --> 01:09:21,048
These AI data centers, they're looking for 500, 600, 700 megawatts plus.

677
01:09:21,368 --> 01:09:25,288
They have become a whole other beast in themselves.

678
01:09:25,928 --> 01:09:28,888
Bitcoin mining actually doesn't need to be that big.

679
01:09:29,588 --> 01:09:30,728
And honestly...

680
01:09:30,728 --> 01:09:32,048
Well, maybe explain that.

681
01:09:32,168 --> 01:09:34,488
Why does an AI data center need to be that big?

682
01:09:35,088 --> 01:09:39,728
It has something to do with just having shorter networking runs

683
01:09:39,728 --> 01:09:41,928
between the whole entire cluster of GPUs.

684
01:09:42,008 --> 01:09:42,848
I'm not an expert here.

685
01:09:43,288 --> 01:09:44,628
Yeah, I'm not looking for...

686
01:09:44,628 --> 01:09:51,728
The close proximity of all the GPUs in one location helps them on their training.

687
01:09:52,028 --> 01:09:56,248
And so trainings, what you were talking about are like querying Grok.

688
01:09:56,488 --> 01:09:57,448
That's inference.

689
01:09:58,788 --> 01:10:01,628
So that's one type of compute GPUs can give.

690
01:10:02,068 --> 01:10:06,768
The other compute is like training the model that Grok is querying.

691
01:10:07,488 --> 01:10:11,088
And so when you're doing these training runs, you want a lot of GPU.

692
01:10:11,088 --> 01:10:18,848
and um a lot of gpu compute and you and you really want it in one location and that's why people are

693
01:10:18,848 --> 01:10:25,608
looking at the you know the irons the galaxies the riot because there's a lot of light in megawatts

694
01:10:25,608 --> 01:10:33,968
in one location there was no other load on the grid that ever requested 600 800 megawatts a

695
01:10:33,968 --> 01:10:39,988
thousand megawatts like big coin mining paved the way yeah how large was a traditional data set like

696
01:10:39,988 --> 01:10:50,228
a large traditional data center 20 megawatts right so there are at least 10 that i know of 300

697
01:10:50,228 --> 01:10:57,268
megawatt sites in texas correct and so the bitcoin miners effectively

698
01:10:57,268 --> 01:11:11,623
built sites that could offload from the grid large amounts of power in very dense ways

699
01:11:11,803 --> 01:11:12,703
We paved the way.

700
01:11:13,463 --> 01:11:21,303
We gave ERCOT a stress test and going through the procedure of interconnecting one gigawatt

701
01:11:21,303 --> 01:11:22,523
worth of electrical load.

702
01:11:22,523 --> 01:11:23,143
It's compute.

703
01:11:23,283 --> 01:11:26,863
We're doing a different type of compute.

704
01:11:26,863 --> 01:11:33,703
it's a sha 256 hash instead of a flop or whatever ai compute is but it's it's an electrical load

705
01:11:33,703 --> 01:11:42,043
on the system and what we did was pivotal to the future success of our cop with ai compute yeah i

706
01:11:42,043 --> 01:11:51,883
mean that that makes sense to me in terms of figuring out how large loads of that size and

707
01:11:51,883 --> 01:11:58,083
scale which hadn't existed before would interact with the grid under a certain set of conditions

708
01:11:58,083 --> 01:12:08,103
with a lot of flexibility and flexibility that was seen in practice now re-entering a world where

709
01:12:08,103 --> 01:12:17,103
that it might be in theory but not proven out yet and that the concerns over the nature of the large

710
01:12:17,103 --> 01:12:23,683
loads that bitcoin received were largely misplaced because of their flexibility but it might be the

711
01:12:23,683 --> 01:12:31,923
case that with the lack of flexibility or at least not having a a model proven out that allows for

712
01:12:31,923 --> 01:12:37,523
that flexibility so i couldn't imagine a diesel you know if there's a 500 megawatt site having a

713
01:12:37,523 --> 01:12:47,143
diesel backup generator that can materially replace a large amount of what might be going on

714
01:12:47,143 --> 01:12:54,823
at an ai data center site at any point in time um and it's not to say that hey there should be an

715
01:12:54,823 --> 01:13:04,823
incentive to service all large loads and to incentivize more generation to come online but

716
01:13:04,823 --> 01:13:11,943
you might be looking at a very different load profile because of the economic incentive model

717
01:13:11,943 --> 01:13:16,583
for ai companies that differ significantly from bitcoin companies

718
01:13:16,583 --> 01:13:26,383
how now pivoting now to so we talked about ai you mentioned also batteries if

719
01:13:26,383 --> 01:13:36,743
bitcoin mining is a form of demand of power load ai form of load batteries can be both

720
01:13:36,743 --> 01:13:41,683
demand and it's not generation it's not actually generating power but it can

721
01:13:41,683 --> 01:13:48,663
stores power and can supply power back to the grid talk about bitcoin mining

722
01:13:48,663 --> 01:13:57,623
and how it can be a service to the grid relative to how batteries have a place where they complement

723
01:13:57,623 --> 01:14:04,503
each other what one can do that the other cannot vice versa and where they potentially compete

724
01:14:04,503 --> 01:14:11,863
yeah so you see now that pretty much any solar farm that is being built is being

725
01:14:11,863 --> 01:14:17,283
co-located with batteries batteries help improve the economics of solar farms

726
01:14:17,283 --> 01:14:24,263
uh bitcoin mining would do so too where the where i see the main difference of bitcoin mining

727
01:14:24,263 --> 01:14:32,283
before you go into that yeah we had one of your choya interns yes in here a month or two ago i

728
01:14:32,283 --> 01:14:36,763
believe a texas a&m student yeah robert yeah he was explaining something about

729
01:14:36,763 --> 01:14:48,423
why there's a benefit to having a battery co-located with solar not just to have over

730
01:14:48,423 --> 01:14:52,863
capacity to be able to serve a consistent amount but like actually from a technical perspective

731
01:14:52,863 --> 01:15:01,443
that there's something about frequency that was a little bit over my head that allows for

732
01:15:01,443 --> 01:15:08,543
frequency management by first putting solar to a battery and then connecting the battery to a grid

733
01:15:08,543 --> 01:15:15,523
is that my wrong there i could have totally misinterpreted you are 100 right like the

734
01:15:15,523 --> 01:15:22,223
batteries have such fine nuance to the amount of power that they they can inject that's why they're

735
01:15:22,223 --> 01:15:27,043
taking the lion's share of the ancillary market because they respond so well so yes it is true like

736
01:15:27,043 --> 01:15:36,643
like it will smooth out any fluctuations and that's why fluctuations in frequency like output

737
01:15:36,643 --> 01:15:42,683
generation output okay and so that that's why data centers are also implementing innovative battery

738
01:15:42,683 --> 01:15:48,103
solutions because sorry to go back to ai real quick but like the inference and the compute it

739
01:15:48,103 --> 01:15:53,703
can go up and down it's bitcoin mining when we're on we're kind of just flattened going for it uh

740
01:15:53,703 --> 01:15:58,283
these training models they can go up they can go down like it could be a mess like it would

741
01:15:58,283 --> 01:16:04,563
absolutely not be a good thing to to sub subject the grid to like instant on instant off at large

742
01:16:04,563 --> 01:16:12,243
loads of 1.2 gigawatts plus plus plus um and so they're using batteries to do that like basically

743
01:16:12,243 --> 01:16:17,343
spoof the grid and say like hey we're actually just flat and so i'm assuming the same thing can

744
01:16:17,343 --> 01:16:21,883
be done for solar i don't know the technicalities behind it but yes i believe that's what it was

745
01:16:21,883 --> 01:16:27,343
yeah yeah so like the grid wants you know nice nice and stable bitcoin mining nice and stable

746
01:16:27,343 --> 01:16:34,323
battery output nice and stable uh ai inference and compute not so stable but back to back to

747
01:16:34,323 --> 01:16:40,963
batteries uh the thing about batteries is that they only have a they're constrained by like how

748
01:16:40,963 --> 01:16:46,923
many hours of charge that they can actually charge and so a battery might only be able to consume

749
01:16:46,923 --> 01:16:52,543
power for two hours worth of the day, or I think they might be getting to four, you can only

750
01:16:52,543 --> 01:16:58,923
discharge for four. So it's got eight hours of total in and out capacity. Bitcoin mining is 24

751
01:16:58,923 --> 01:17:05,743
seven, right? We can consume the entire day. If prices, say it's on a shoulder month and it's

752
01:17:05,743 --> 01:17:12,943
super windy, super sunny, and prices are $0 the entire day, the battery will have been charged

753
01:17:12,943 --> 01:17:16,883
and stay at the same state of charge the entire time.

754
01:17:17,323 --> 01:17:20,123
They'll probably, in this case, bid into the ancillary market

755
01:17:20,123 --> 01:17:25,723
and provide those frequency support services.

756
01:17:25,723 --> 01:17:29,003
Whereas Bitcoin, if it's $0 the entire day,

757
01:17:29,083 --> 01:17:33,463
we're going to be blasting full-on monetizing those electrons.

758
01:17:34,243 --> 01:17:37,103
And so I think they're both flexible.

759
01:17:37,103 --> 01:17:44,823
I would say batteries are better at like very, very nuanced.

760
01:17:45,003 --> 01:17:46,283
It's electronics, right?

761
01:17:46,343 --> 01:17:51,903
You can set the voltage or set the output to like a very nuanced degree.

762
01:17:52,983 --> 01:17:57,743
Bitcoin mining, also flexible, maybe not as nuanced, but the benefit that Bitcoin brings

763
01:17:57,743 --> 01:18:05,123
is that you can oversize it considerably and you can run it for a larger amount of time.

764
01:18:05,123 --> 01:18:14,663
and explain i didn't fully grok what you were saying about the um the charging the amount of

765
01:18:14,663 --> 01:18:18,843
time like what what is a large battery how large is a large battery first off

766
01:18:18,843 --> 01:18:30,083
ooh um battery installations nowadays are i think a couple hundred megawatts okay there's 15

767
01:18:30,083 --> 01:18:37,263
gigawatts so 15 000 megawatts worth of batteries in ercott right now wait repeat that amount 15

768
01:18:37,263 --> 01:18:44,143
000 megawatts or 15 gigawatts of batteries in ercott would that how much would that have been

769
01:18:44,143 --> 01:18:50,903
five years ago order of magnitude zero okay pretty much effectively zero uh they've they've come on

770
01:18:50,903 --> 01:18:58,183
super strong in the last couple years and so take a hundred megawatt battery yeah

771
01:19:00,083 --> 01:19:06,663
is that actually like one just massive battery or is it a site that has it's it's a series of

772
01:19:06,663 --> 01:19:14,163
batteries and so like 100 megawatt battery has 100 megawatt hours so it's megawatts like being

773
01:19:14,163 --> 01:19:19,223
able to deliver to the grid and then there's like total capacity so like maybe it can only

774
01:19:19,223 --> 01:19:27,043
inject 100 megawatts at a time but it can hold 200 megawatts so it's 100 megawatt two hour battery

775
01:19:27,043 --> 01:19:38,663
okay i think i follow that so if it's got 200 megawatts of capacity it can take 100 megawatt

776
01:19:38,663 --> 01:19:47,063
charge at any point in time but then there's a certain number of hours that a 200 megawatt

777
01:19:47,063 --> 01:19:56,263
battery could supply say 20 megawatts of power yeah i think it's based off of and forgive me

778
01:19:56,263 --> 01:20:00,043
for not knowing this, but I think it's based off of the inverter size. I don't expect you to be an

779
01:20:00,043 --> 01:20:06,623
expert on everything. I'm looking for general knowledge here. Yeah. I believe the way it works

780
01:20:06,623 --> 01:20:13,783
is that it's based off of the inverter, how much you can convert from DC to AC and send through it.

781
01:20:13,783 --> 01:20:22,083
So say you have a 400 megawatt pack and you can send 200 megawatt hours through it at a time.

782
01:20:22,083 --> 01:20:29,023
that's a 200 megawatt two-hour battery got it and then once it was depleted it need to be

783
01:20:29,023 --> 01:20:34,903
recharged correct and it can only maximally charge 200 megawatts at a time they could

784
01:20:34,903 --> 01:20:40,403
probably go less than that if they wanted to spread it out and provide more even uh charge

785
01:20:40,403 --> 01:20:47,903
but it can only take 400 megawatts total 200 megawatts an hour whereas if you have a you know

786
01:20:47,903 --> 01:20:50,983
a 200 megawatt Bitcoin mine,

787
01:20:51,403 --> 01:20:53,423
you're consuming 200 megawatts

788
01:20:53,423 --> 01:20:55,183
for 24 hours,

789
01:20:55,763 --> 01:20:56,963
significantly more.

790
01:20:57,143 --> 01:21:01,443
The utilization of the entire grid

791
01:21:01,443 --> 01:21:04,883
is much better on a Bitcoin mine

792
01:21:04,883 --> 01:21:07,223
than just exclusively a battery.

793
01:21:07,963 --> 01:21:08,403
And then,

794
01:21:08,943 --> 01:21:12,923
if you have batteries distributed,

795
01:21:13,683 --> 01:21:15,523
it's basically got stored power

796
01:21:15,523 --> 01:21:22,943
that can that can provide incremental supply to the grid at a point in time

797
01:21:22,943 --> 01:21:30,063
so talk about that so you talked about one circumstance where battery could

798
01:21:30,063 --> 01:21:49,638
in a much more fine balance not necessarily the grid but some other demand for power talk about the scenarios where bitcoin mining could help balance the market for power

799
01:21:49,638 --> 01:21:59,378
more effectively in a broad like at scale versus batteries like what i think like conceptual

800
01:21:59,378 --> 01:22:10,278
scenarios you know i think the unique thing about both of them is that they're location agnostic

801
01:22:10,278 --> 01:22:25,458
like you you have well as a as a battery you are injecting as a generator and like injecting to a

802
01:22:25,458 --> 01:22:32,258
they consider batteries generators yes i think they're considered both they get the treatment of

803
01:22:32,258 --> 01:22:37,658
both and so and like we're getting back into the weeds of our cut here but like as a battery you're

804
01:22:37,658 --> 01:22:48,118
injecting at the nodal price as a load bitcoin miner you're in you're drawing based off of a

805
01:22:48,118 --> 01:22:55,238
zonal price and so actually the thing that um works to the battery's favor here is that

806
01:22:55,238 --> 01:23:00,898
the nodes a better signal right if you if you aggregate all the load the nodes together you

807
01:23:00,898 --> 01:23:10,218
get the zone the zone could have how many zones are there in our cut there's west north south

808
01:23:10,218 --> 01:23:16,878
and houston okay four and those are the deregulated zones like within those zones like you have the

809
01:23:16,878 --> 01:23:23,378
co-ops and the municipalities that aren't deregulated and i said co-ops well anyways you

810
01:23:23,378 --> 01:23:27,098
have those you have those four zones and so i'll talk about load zone west is where we're located

811
01:23:27,098 --> 01:23:35,658
and i know i know best and and maybe there's there's many nodes in a zone yes just for some

812
01:23:35,658 --> 01:23:41,178
context even though different zones probably a different number of nodes how many nodes per zone

813
01:23:41,178 --> 01:23:52,918
oh um i believe each generator will create a node uh so load zone west for example i'm pretty sure

814
01:23:52,918 --> 01:24:01,678
there's got to be 50 to 100 nodes and the the nodes themselves will have a will have a price

815
01:24:01,678 --> 01:24:10,258
point and then the aggregation of those nodes will create your zonal price that your load responds to

816
01:24:10,258 --> 01:24:17,198
okay because yeah before you had made some comparison between nodal pricing and zone pricing

817
01:24:17,198 --> 01:24:24,478
now reconnect that to the idea of batteries how they might benefit from one or the other

818
01:24:24,478 --> 01:24:33,678
sure like the for example deep in deep deep west texas talking like pecos for stockton area um

819
01:24:33,678 --> 01:24:40,178
a generator could trip or it could it just couldn't it might not be windy and so that deep west

820
01:24:40,178 --> 01:24:46,278
node will get blown out it'll go to five thousand five hundred bucks a megawatt hour

821
01:24:46,278 --> 01:24:55,798
and then based off of that high price the zonal price for the load increases but it might not

822
01:24:55,798 --> 01:25:02,318
actually reflect the electric like the electric grid conditions on the eastern part of the zone

823
01:25:02,318 --> 01:25:09,938
in Abilene, right? You're 400 miles away, multiple, multiple buses away. They're not

824
01:25:09,938 --> 01:25:16,018
very electrically connected. And so where batteries are nice in this situation is that they

825
01:25:16,018 --> 01:25:21,418
operate on the node. And so by operating on the node, you get a clearer price signal.

826
01:25:22,578 --> 01:25:31,178
But to be fair, or to circle around, Bitcoin mining, loads can operate on nodes to get better

827
01:25:31,178 --> 01:25:39,218
price signals it just requires uh qualification to become a controllable load resource this

828
01:25:39,218 --> 01:25:45,118
controllable load resource allows you to operate based off of nodal conditions once you qualify

829
01:25:45,118 --> 01:25:54,018
with ERCOT's conditions um and then at that point like bitcoin mining truly is like an apex predator

830
01:25:54,018 --> 01:25:57,018
for responding to grid conditions

831
01:25:57,018 --> 01:25:58,458
because you have your node

832
01:25:58,458 --> 01:26:00,358
that you're on,

833
01:26:00,478 --> 01:26:02,098
your most electrically close node.

834
01:26:02,438 --> 01:26:04,798
If you're responding to the prices

835
01:26:04,798 --> 01:26:06,918
and conditions at that node,

836
01:26:07,058 --> 01:26:09,298
that is the clearest signal

837
01:26:09,298 --> 01:26:11,898
for scarcity or abundance that there is.

838
01:26:12,138 --> 01:26:13,458
And so if prices are cheap,

839
01:26:14,058 --> 01:26:15,238
there's abundance of electricity

840
01:26:15,238 --> 01:26:16,138
and you're consuming.

841
01:26:17,898 --> 01:26:20,098
Price gets blown out just at your node

842
01:26:20,098 --> 01:26:21,538
and you can turn off.

843
01:26:22,058 --> 01:26:23,298
That helps like

844
01:26:23,298 --> 01:26:31,358
spread spread it back out and like reduce that nodes um blow out and then that improves the

845
01:26:31,358 --> 01:26:38,678
prices for every single person in that zone because not everybody can be a clr and so

846
01:26:38,678 --> 01:26:40,858
if i'm thinking about this correctly

847
01:26:40,858 --> 01:26:46,538
if a generator trips

848
01:26:46,538 --> 01:26:57,298
then either a bitcoin mine could come down if it caused prices to spike or a battery could

849
01:26:57,298 --> 01:27:04,178
replace that power or both or both or combination it's not mutually exclusive like they can both

850
01:27:04,178 --> 01:27:12,078
respond to the the proper price signals or as demand is increasing a battery could provide

851
01:27:12,078 --> 01:27:17,218
more power where a bit you know often not oftentimes there are people that will say bitcoin

852
01:27:17,218 --> 01:27:26,738
is a battery my view it's a bad analogy because it can't provide incremental power to the market

853
01:27:26,738 --> 01:27:35,558
correct it can turn down and allow a generator that was providing it it power to have that power

854
01:27:35,558 --> 01:27:43,118
go somewhere else, but can't replace the power. So it's that if you need more power or need more

855
01:27:43,118 --> 01:27:51,018
generation, a battery could provide that more efficiently where a Bitcoin miner could not.

856
01:27:51,018 --> 01:27:56,538
Yeah. Just for two hours. But the Bitcoin miner could stay off as long as the economic signal

857
01:27:56,538 --> 01:28:02,398
signals that the generation asset has state tripped, some type of issue.

858
01:28:02,398 --> 01:28:08,938
and so it might be the case where and i'm not saying in a specific circumstance but

859
01:28:08,938 --> 01:28:16,658
a a battery might provide the first relief to price signals but then if it's needed for longer

860
01:28:16,658 --> 01:28:24,358
the bitcoin miner would be more responsive i think it would be you know depending on the price of

861
01:28:24,358 --> 01:28:29,918
of power right i think they would both respond to the price um but

862
01:28:29,918 --> 01:28:36,978
Bitcoin, like the battery is capped to its level of response, although it is, you know,

863
01:28:36,998 --> 01:28:41,958
it is injecting, right? It is providing that power support. Bitcoin mining can stay off for longer.

864
01:28:41,958 --> 01:28:50,658
Like another, another really good example, um, of this was, this was this summer. I can't remember

865
01:28:50,658 --> 01:28:57,618
the month, but prices, the batteries discharged for the solar ramp in the evening and the afternoon

866
01:28:57,618 --> 01:29:05,758
as the sun is discharged injected into the grid like they sent power into the grid as the sun was

867
01:29:05,758 --> 01:29:11,758
setting and solar was coming off of its its ramp and so for those two hours three hours batteries

868
01:29:11,758 --> 01:29:19,898
were injecting prices were maybe 120 150 bucks a megawatt hour 300 whatever it was um overnight

869
01:29:19,898 --> 01:29:29,818
it wasn't windy and so prices stayed in that maybe 80 to 100 a megawatt hour range which is

870
01:29:29,818 --> 01:29:35,078
significantly elevated versus where which is elevated comparatively yeah and then in the

871
01:29:35,078 --> 01:29:42,178
morning there was another peak like there was a there was a morning peak of demand the batteries

872
01:29:42,178 --> 01:29:46,978
had not charged because of that $80 to $100 price range.

873
01:29:47,638 --> 01:29:51,358
And as it peaked to maybe $200 a megawatt hour,

874
01:29:51,558 --> 01:29:54,198
the batteries were sitting there twiddling the thumbs like,

875
01:29:54,258 --> 01:29:55,178
oh, we're out of capacity.

876
01:29:55,278 --> 01:29:56,018
There's no discharge.

877
01:29:56,458 --> 01:30:00,398
Whereas Bitcoin mining, you're off at that solar shoulder.

878
01:30:01,118 --> 01:30:02,838
And then as long as your break even is like,

879
01:30:03,018 --> 01:30:05,478
you've got some newer efficiency machines at 120,

880
01:30:05,738 --> 01:30:08,238
you're soaking up all that electricity at night.

881
01:30:08,718 --> 01:30:11,838
And then when that peak goes to 200 bucks,

882
01:30:11,838 --> 01:30:21,538
Bitcoin responded and so we have a we have that ability to always respond batteries have to worry

883
01:30:21,538 --> 01:30:27,538
about their state of charge and it's it's more of a guessing game it's less guaranteed so they

884
01:30:27,538 --> 01:30:34,698
they can both be compliments and they can both be responding at the same time and be

885
01:30:34,698 --> 01:30:42,258
or you could think about that as being competitive to how they're providing or how they're responding

886
01:30:42,258 --> 01:30:47,218
to price signals you know urkot needs everything we need gas we need solar we need batteries we

887
01:30:47,218 --> 01:30:53,658
need wind like the more it grows the more load grows the better texas does um you know batteries

888
01:30:53,658 --> 01:31:01,758
provide things like like voltage uh ride through support that electrical loads can't do like it's

889
01:31:01,758 --> 01:31:08,218
all going to work together and because we're in an island system a variety is key like it's and

890
01:31:08,218 --> 01:31:16,358
you mentioned batteries increasingly being co-located with solar are are most batteries

891
01:31:16,358 --> 01:31:23,418
co-located with some form of generation or do they do they typically i mean you want to put it next to

892
01:31:23,418 --> 01:31:30,298
intermittent generation right because if your gen stack is built for if you have a one gigawatt

893
01:31:30,298 --> 01:31:37,038
interconnect and you size your nat gas generation to one gigawatt like you couldn't inject any more

894
01:31:37,038 --> 01:31:41,238
than that one gigawatt so it wouldn't really make sense to charge your battery if you're just using

895
01:31:41,238 --> 01:31:47,438
a gas peaker plant where it does make sense is that you're generating your one gigawatt solar

896
01:31:47,438 --> 01:31:56,578
plant you put one gigawatts of batteries you charge during your um you know your did i say

897
01:31:56,578 --> 01:31:57,118
Solar or wind?

898
01:31:58,318 --> 01:31:58,798
Solar.

899
01:31:59,138 --> 01:31:59,438
Solar.

900
01:31:59,598 --> 01:32:13,973
You charge during your sunny time and then you have one gigawatt worth of capacity to sell on those times where you literally can be injecting into the grid when it in the evening time um

901
01:32:13,973 --> 01:32:23,673
and then when you if you talk about right here but just if we look at the landscape from

902
01:32:23,673 --> 01:32:32,533
yeah somewhere between 2017 and 2019 gideon and cholla went out and

903
01:32:32,533 --> 01:32:41,393
from my knowledge built the first greenfield substation that was 50 to 100 megawatts

904
01:32:41,393 --> 01:32:46,093
ultimately to be 100 megawatts in terms of the phase but that was the first greenfield

905
01:32:46,093 --> 01:32:53,533
substation for bitcoin mining virtually there was there was bitcoin mining on the ericot grid

906
01:32:53,533 --> 01:32:57,293
in 2020 but in terms of getting to four gigawatts today

907
01:32:57,293 --> 01:33:07,193
five six years ago it was very de minimis similar with batteries you said it was something like 15

908
01:33:07,193 --> 01:33:15,193
gigawatts now today of battery power so a lot has changed over the last five years

909
01:33:15,193 --> 01:33:19,553
if we look five years into the future

910
01:33:19,553 --> 01:33:24,413
specifically on the bitcoin mining side

911
01:33:24,413 --> 01:33:28,133
where do you see the scale going

912
01:33:28,133 --> 01:33:38,273
and where's choya most focused today in terms of participating in how you guys see the market

913
01:33:38,273 --> 01:33:48,573
changing i'll hit on bitcoin mining like i bitcoin mining is a unique load profile to

914
01:33:48,573 --> 01:33:53,793
AI compute that we are competing for power against. I do think that you're going to have

915
01:33:53,793 --> 01:33:59,713
less of these mega miners, hundreds of megawatts, and you're going to have more distributed 50,

916
01:34:00,273 --> 01:34:06,973
maybe even sub 75 to avoid the ERCA large load classification. And then you're just opportunistically

917
01:34:06,973 --> 01:34:13,273
citing these around areas that have an excess of power generation. You're going to have Bitcoin

918
01:34:13,273 --> 01:34:20,893
mining behind the meter at wind farms to that are just taking up putting like a floor ppa on that

919
01:34:20,893 --> 01:34:26,393
wind farm helping to modify monetize the excess energy there and then just dealing with like a

920
01:34:26,393 --> 01:34:37,293
60 or 70 percent uptime i think the days of the hunt you know one gigawatt bitcoin mining uh

921
01:34:37,293 --> 01:34:42,793
ventures or at least in texas are probably over we just can't command that premium like you could

922
01:34:42,793 --> 01:34:48,253
take of one gigawatt interconnect and sell to a hyperscaler of course this is all like given the

923
01:34:48,253 --> 01:34:54,573
data that i'm i'm seeing now with ai like is a bubble like we're not going to get into that like

924
01:34:54,573 --> 01:35:00,293
like will things change who knows but if it's on the current trajectory i think that bitcoin is

925
01:35:00,293 --> 01:35:04,773
going to just continue to be like the pioneering species like we're going to go collate locate

926
01:35:04,773 --> 01:35:12,733
next to the intermittent generation we're going to go co-locate um you know wherever these large

927
01:35:12,733 --> 01:35:19,953
data centers don't want to go um and then how cholla wants to position himself you know like

928
01:35:19,953 --> 01:35:26,013
like you were saying gideon gideon saw this flexible load like monetization back in 2017

929
01:35:26,013 --> 01:35:33,573
2018 and built that that spec built hoddle one substation in uh in pyot and then just

930
01:35:33,573 --> 01:35:38,773
everything that happened the last eight years was a confirmation of like what what he was talking

931
01:35:38,773 --> 01:35:45,573
about. You have the mecca of energy in West Texas. You've got oil and gas, the most prolific field in

932
01:35:45,573 --> 01:35:51,693
the world. You've got wind and you've got solar. You probably have geothermal too, but it hasn't

933
01:35:51,693 --> 01:35:57,593
really been commercialized, but you have so much energy right there. And if you can respond to the

934
01:35:57,593 --> 01:36:04,393
prices, which was the thesis with Bitcoin mining, you can capitalize on it. Load Zone West, where

935
01:36:04,393 --> 01:36:13,673
Gideon first built that substation and sold that project was the start of the West load zone

936
01:36:13,673 --> 01:36:19,293
tripling in size over the last eight years. I'll say it right here, he started that.

937
01:36:19,293 --> 01:36:23,773
Yeah. Of course, there's some oil and gas load in there too, but we're also an oil and gas

938
01:36:23,773 --> 01:36:31,013
company. So we'll put our hat into that too. Where Choya sees opportunity, what I do at Choya

939
01:36:31,013 --> 01:36:33,733
is run our innovation lab

940
01:36:33,733 --> 01:36:36,153
where we take low serial number units

941
01:36:36,153 --> 01:36:39,393
and test it out.

942
01:36:39,573 --> 01:36:41,673
We'll take on the technology risk.

943
01:36:42,373 --> 01:36:44,493
And so with all these AI data centers

944
01:36:44,493 --> 01:36:45,533
coming onto the grid,

945
01:36:45,853 --> 01:36:46,873
like, are we going to go,

946
01:36:47,033 --> 01:36:48,673
like, is off-grid the solution?

947
01:36:49,413 --> 01:36:49,853
Like, and, you know,

948
01:36:49,893 --> 01:36:51,173
is it the traditional off-grid?

949
01:36:51,393 --> 01:36:53,453
Is it, like, flare gas mining?

950
01:36:53,573 --> 01:36:54,833
Is it putting nat gas,

951
01:36:55,073 --> 01:36:56,373
you know, turbines on a pipeline

952
01:36:56,373 --> 01:36:58,433
and creating your own energy?

953
01:36:58,433 --> 01:37:09,013
or is it some combination of uh solar batteries and gas because solar and batteries the cost curves

954
01:37:09,013 --> 01:37:14,333
are just continuing to go down they're they're being built as consumer electronics like we all

955
01:37:14,333 --> 01:37:20,813
technology stacks and electronics price trends down over time uh and it's they're just going to

956
01:37:20,813 --> 01:37:28,133
become more and more competitive. So we can imagine a world where you can have massive

957
01:37:28,133 --> 01:37:37,253
off-grid solar battery farms powering Bitcoin mining. Will it be Bitcoin mining at first?

958
01:37:37,533 --> 01:37:43,293
No, of course not. It's going to be AI compute first because they're generating a higher dollar

959
01:37:43,293 --> 01:37:46,533
per megawatt hour basis. So the union economics work better for them.

960
01:37:46,533 --> 01:37:54,493
um where choy is participating right now is like we've we've got expertise in the oil and gas side

961
01:37:54,493 --> 01:38:04,573
the land side and the um electrical side and are trying to speed things up in our god like it's

962
01:38:04,573 --> 01:38:12,973
i talked about it earlier like we see it as a matter of national security in order to have as

963
01:38:12,973 --> 01:38:19,753
much AI compute in the United States as possible. And the more that comes to Texas, the more benefits

964
01:38:19,753 --> 01:38:27,833
Texas stakeholders. So our focus is accelerate that in any way possible. Gideon was super

965
01:38:27,833 --> 01:38:37,753
involved at the legislative last year, lobbying for flexible load and using Bitcoin as an example

966
01:38:37,753 --> 01:38:40,353
to basically troll these AI companies like,

967
01:38:40,513 --> 01:38:41,813
hey, we can do this.

968
01:38:42,693 --> 01:38:43,753
Y'all should figure it out too

969
01:38:43,753 --> 01:38:45,633
because we've got four gigawatts

970
01:38:45,633 --> 01:38:46,893
and we're turning off

971
01:38:46,893 --> 01:38:49,913
and like sending that power back to families and homes.

972
01:38:50,913 --> 01:38:52,753
Y'all can't tell us that you're too important

973
01:38:52,753 --> 01:38:53,673
to not do so.

974
01:38:54,513 --> 01:38:57,093
And so long way to say,

975
01:38:57,233 --> 01:38:59,453
like we're involved in all things,

976
01:38:59,453 --> 01:39:00,433
energy, power,

977
01:39:01,093 --> 01:39:04,473
and just excited for the future of Texas

978
01:39:04,473 --> 01:39:05,893
and how much it's going to grow

979
01:39:05,893 --> 01:39:07,473
and how much it's going to benefit all of us.

980
01:39:07,753 --> 01:39:22,653
When you mentioned that you guys view more AI demand as a matter of national security, I was going to say, are you talking about Texas or are you talking about the United States?

981
01:39:22,973 --> 01:39:28,773
And then is more Bitcoin mining on that same par?

982
01:39:29,773 --> 01:39:34,293
And this is the last question to wrap up and then also want you to share where people can find you.

983
01:39:34,293 --> 01:39:39,493
if they're interested in what you're doing in the innovation lab or outside that,

984
01:39:40,133 --> 01:39:42,293
where they can find Brad Cuddy.

985
01:39:43,433 --> 01:39:46,713
You know, of course, it's good to have AI compute in the Republic of Texas.

986
01:39:47,653 --> 01:39:53,673
But overarchingly, in the United States, having this compute and owning this compute is important.

987
01:39:53,673 --> 01:39:59,453
It's like we are obviously competing against China, state-owned utilities, state-owned generation.

988
01:39:59,873 --> 01:40:01,473
They can just build transmission.

989
01:40:01,653 --> 01:40:03,833
They can just build generation as much as they want.

990
01:40:03,833 --> 01:40:10,833
they can put old chips on and run these models. It is their model versus us. Will America's

991
01:40:10,833 --> 01:40:19,153
will to innovate the capitalism structure that incentivizes growth and innovation and

992
01:40:19,153 --> 01:40:25,453
thinking outside the box and allowing entrepreneurship, will it win? I don't know.

993
01:40:25,453 --> 01:40:32,773
I hope so. I believe so. But how do we ensure that we do what we can to make it win?

994
01:40:32,773 --> 01:40:36,733
access to electrons, which is the input to all this compute.

995
01:40:37,313 --> 01:40:40,393
That is like our North Star at Choya.

996
01:40:41,393 --> 01:40:45,133
And then for Bitcoin mining,

997
01:40:45,633 --> 01:40:49,173
of course I want as much Bitcoin produced in the United States as possible

998
01:40:49,173 --> 01:40:54,573
because I want Americans to prosper from energy-backed freedom money.

999
01:40:55,573 --> 01:41:00,573
I can see a world as AI continues to

1000
01:41:00,573 --> 01:41:10,633
gobble up jobs, maybe their AI agents start transacting with each other. Bitcoin will be

1001
01:41:10,633 --> 01:41:16,553
a part of this. And we all know it. How do we make sure it's a part of it? So one,

1002
01:41:16,553 --> 01:41:25,233
on the Bitcoin mining side is what I'm trying to do. We are a complement to the grid. We're

1003
01:41:25,233 --> 01:41:32,553
important for ERCOT. We're here to stay. We're not competing against AI. It's a different class

1004
01:41:32,553 --> 01:41:39,593
at this point. But of course, decentralization, not all Bitcoin mining needs to go to Texas.

1005
01:41:40,293 --> 01:41:45,393
It's a pioneering species. It should go where it's needed. Bhutan, great example, my favorite one.

1006
01:41:45,813 --> 01:41:50,713
The government of Bhutan having 12,000 Bitcoin on their balance sheet because they said yes to

1007
01:41:50,713 --> 01:41:57,973
bit deer mining on their stranded hydro is the coolest thing ever um but long live urcott long

1008
01:41:57,973 --> 01:42:04,173
live texas and then you can uh you can find me at i think it's underscore brad cuddy underscore

1009
01:42:04,173 --> 01:42:10,373
on twitter linkedin too i got a pirate flag and laser eyes so pretty easy to find uh but really

1010
01:42:10,373 --> 01:42:15,813
appreciate it parker yeah appreciate you coming on running the innovation lab at choy inc in many

1011
01:42:15,813 --> 01:42:24,833
ways ERCOT at a great scale is an innovation lab for the rest of the nation or the world to pay

1012
01:42:24,833 --> 01:42:30,393
attention to so appreciate all that you're doing individually that Troy is doing and thank you for

1013
01:42:30,393 --> 01:42:35,573
coming on the show glad to have you in Austin great to be here all right thanks Parker yep