1
00:00:10,640 --> 00:00:15,000
Speaker 1: Hello, and welcome to another episode of the Odd Lots podcast.

2
00:00:15,160 --> 00:00:20,720
Speaker 1: I'm Joe Wisenthal. Unfortunately, my colleague Tracy Alloway, she is

3
00:00:20,800 --> 00:00:24,560
Speaker 1: off today and is particularly unfortunate because this is gonna

4
00:00:24,600 --> 00:00:27,720
Speaker 1: be a big one. So, you know, like on this podcast,

5
00:00:27,880 --> 00:00:31,360
Speaker 1: obviously we talk about chips a lot, and I think

6
00:00:31,400 --> 00:00:35,279
Speaker 1: now we've done like eight different episodes on semiconductors and

7
00:00:35,640 --> 00:00:38,400
Speaker 1: who makes them and the shortages and the challenges of

8
00:00:38,400 --> 00:00:41,279
Speaker 1: building them and so forth. So Odd Lots listeners will

9
00:00:41,320 --> 00:00:44,320
Speaker 1: know that we also do a fair number of episodes

10
00:00:44,320 --> 00:00:48,040
Speaker 1: about cryptocurrencies, and of course that's for obvious reasons why

11
00:00:48,280 --> 00:00:50,880
Speaker 1: cryptocracy has been going up a lot and it's a

12
00:00:50,920 --> 00:00:55,240
Speaker 1: fascinating area. Well, this episode, I actually I'm gonna say

13
00:00:55,240 --> 00:01:00,240
Speaker 1: it actually combines the two because it's slowly becoming or

14
00:01:00,800 --> 00:01:05,319
Speaker 1: that you can't really talk about the semiconductor market right

15
00:01:05,360 --> 00:01:08,760
Speaker 1: now without talking about cryptocurrencies because you need chips to

16
00:01:08,800 --> 00:01:11,919
Speaker 1: mind them. And when they said, when the cryptocurrency market

17
00:01:12,040 --> 00:01:15,880
Speaker 1: is booming, people want to acquire more chips to mind them,

18
00:01:15,880 --> 00:01:19,440
Speaker 1: and it's becoming a sort of more significant source of demand.

19
00:01:19,840 --> 00:01:22,840
Speaker 1: If you look in the video gamer community, there's a

20
00:01:22,920 --> 00:01:27,200
Speaker 1: huge frustration about their inability to buy various graphics cards

21
00:01:27,240 --> 00:01:30,800
Speaker 1: from the likes of Nvidia because etherory of miners snap

22
00:01:30,880 --> 00:01:35,000
Speaker 1: them up. Recently, there have been stories out of Asia,

23
00:01:35,400 --> 00:01:39,640
Speaker 1: China in particular, where people are are buying hard drives

24
00:01:39,680 --> 00:01:43,959
Speaker 1: because they want to mind the new cryptocurrency Chia Bitcoin

25
00:01:44,160 --> 00:01:48,120
Speaker 1: is mind using specialized chips, but ultimately that takes a

26
00:01:48,320 --> 00:01:52,840
Speaker 1: chip production capacity to so really the stories are interlocked,

27
00:01:53,120 --> 00:01:55,600
Speaker 1: I think in a way that I hadn't up until

28
00:01:55,640 --> 00:01:59,000
Speaker 1: recently fully appreciated, and I not I would not say

29
00:01:59,040 --> 00:02:02,880
Speaker 1: that cryptos main source of reason why there is a

30
00:02:02,960 --> 00:02:06,640
Speaker 1: shortage or difficulty to get a semiconductors, but everything matters

31
00:02:06,760 --> 00:02:09,840
Speaker 1: at the margin, and so we're going to explore this

32
00:02:10,040 --> 00:02:13,160
Speaker 1: nexus today between what is going on in the world

33
00:02:13,160 --> 00:02:17,600
Speaker 1: of semiconductors, the fight to acquire semiconductors between different players.

34
00:02:17,639 --> 00:02:20,880
Speaker 1: Because you have the minors, you have the gamers, people

35
00:02:20,960 --> 00:02:24,120
Speaker 1: use computers, you have data centers, which of course have

36
00:02:24,280 --> 00:02:28,000
Speaker 1: tremendous chip needs, and they're all fighting it out for

37
00:02:28,080 --> 00:02:32,600
Speaker 1: this semiconductor capacity and it all comes together and it's

38
00:02:32,600 --> 00:02:36,000
Speaker 1: a really fascinating facet of all of these different stories

39
00:02:36,160 --> 00:02:40,359
Speaker 1: brought together. I'm very excited about this episode. I am

40
00:02:40,440 --> 00:02:42,920
Speaker 1: going to be speaking with Brian Venturo. He is the

41
00:02:43,000 --> 00:02:47,320
Speaker 1: CTO and co founder of core Weave, which is uh

42
00:02:47,320 --> 00:02:50,000
Speaker 1: it's a cloud computing company, and we'll get into the

43
00:02:50,040 --> 00:02:52,680
Speaker 1: details of specifically what they do, but as a cloud

44
00:02:52,680 --> 00:02:56,120
Speaker 1: computing company, of course, it's specifically has a lot of

45
00:02:56,200 --> 00:02:58,959
Speaker 1: chip needs itself, and so he's gonna walk us through

46
00:02:59,000 --> 00:03:01,400
Speaker 1: what the world looks like right now. He's also in

47
00:03:01,480 --> 00:03:03,560
Speaker 1: the ethery of mining space, sees it from all the

48
00:03:03,800 --> 00:03:07,400
Speaker 1: all the different angles. We're going to talk about that. So, Brian,

49
00:03:07,680 --> 00:03:10,600
Speaker 1: thank you so much for coming on odd lots. Thank

50
00:03:10,600 --> 00:03:13,200
Speaker 1: you so much for having me. This is a topic

51
00:03:13,280 --> 00:03:16,519
Speaker 1: that I am near and unfortunately dear too, so I

52
00:03:16,680 --> 00:03:19,200
Speaker 1: really excited to dig into it. So why don't you

53
00:03:19,240 --> 00:03:22,160
Speaker 1: just get us started? I mean cloud computing. You know,

54
00:03:22,200 --> 00:03:26,280
Speaker 1: people think of Amazon Aws, people think of Microsoft perhaps,

55
00:03:26,360 --> 00:03:29,320
Speaker 1: but obviously they're not the only players. What is core

56
00:03:29,400 --> 00:03:35,760
Speaker 1: Weave and how is it situated within the cloud computing ecosystem?

57
00:03:35,840 --> 00:03:40,280
Speaker 1: We typically talk about ourselves as a specialized cloud right,

58
00:03:40,360 --> 00:03:45,800
Speaker 1: so we're primarily providing highly scalable, highly paralyzable burst compute too.

59
00:03:46,480 --> 00:03:50,200
Speaker 1: Companies that are in the VFX production and rendering space.

60
00:03:50,440 --> 00:03:54,680
Speaker 1: That's VFX. That's video videographics correct, Like, so pretty much

61
00:03:54,720 --> 00:03:57,480
Speaker 1: anything that you see that's episodic on a streaming channel,

62
00:03:57,920 --> 00:04:01,160
Speaker 1: for that's being produced in long form content like that's

63
00:04:01,200 --> 00:04:03,600
Speaker 1: being touched by a VFX artist somewhere, whether it's on

64
00:04:03,600 --> 00:04:06,280
Speaker 1: the compositing layer or it is actually on the computer

65
00:04:06,320 --> 00:04:10,240
Speaker 1: generated graphics layer. That stuff needs a significant amount of compute,

66
00:04:10,480 --> 00:04:12,480
Speaker 1: and that's one of our focus areas as a business,

67
00:04:12,480 --> 00:04:15,600
Speaker 1: so we're there for them. And then on in addition

68
00:04:15,600 --> 00:04:18,120
Speaker 1: to that, it's AI machine learning, so it's model training

69
00:04:18,160 --> 00:04:21,359
Speaker 1: and serving. There's one additional one that kind of sits

70
00:04:21,640 --> 00:04:25,960
Speaker 1: between the media and entertainment space and the kind of

71
00:04:26,000 --> 00:04:28,400
Speaker 1: batch processing space, which is called pixel streaming, which is

72
00:04:28,920 --> 00:04:32,479
Speaker 1: serving real time experiences in your browser. Right, So if

73
00:04:32,520 --> 00:04:35,000
Speaker 1: you ever go to configure a car online, they don't

74
00:04:35,040 --> 00:04:38,080
Speaker 1: have seven million versions of that car render. They actually

75
00:04:38,080 --> 00:04:40,120
Speaker 1: do it in real time on a GPU someone in

76
00:04:40,200 --> 00:04:42,520
Speaker 1: the data. And then the last piece is kind of

77
00:04:42,520 --> 00:04:46,680
Speaker 1: the legacy HPC, right, So it's highly parallelizable batch processing.

78
00:04:47,080 --> 00:04:50,240
Speaker 1: It's people who are doing drug discovery, if they're doing

79
00:04:50,360 --> 00:04:53,680
Speaker 1: cancer research, kind of everything under the sun. What makes

80
00:04:53,760 --> 00:04:56,440
Speaker 1: us a little bit different is that we're not really

81
00:04:56,440 --> 00:04:59,840
Speaker 1: competing for people to host their word press blugs. Right,

82
00:05:00,000 --> 00:05:03,200
Speaker 1: So we're building for we're building scale compute for large

83
00:05:03,200 --> 00:05:06,000
Speaker 1: customers of compute Um, and then we're kind of working

84
00:05:06,000 --> 00:05:07,480
Speaker 1: with them all along the way to kind of meet

85
00:05:07,480 --> 00:05:09,359
Speaker 1: their needs. And we do a lot of the spoke

86
00:05:09,400 --> 00:05:12,320
Speaker 1: builds for people. So we're constantly in the in the

87
00:05:12,400 --> 00:05:15,640
Speaker 1: chip market and in the component market. And just because

88
00:05:15,640 --> 00:05:18,159
Speaker 1: every industry kind of has their own unique needs. So

89
00:05:18,720 --> 00:05:21,880
Speaker 1: is there such a thing as a sort of a

90
00:05:21,920 --> 00:05:24,880
Speaker 1: normal chip market? I mean right now, and we'll get

91
00:05:24,880 --> 00:05:27,120
Speaker 1: we're going to get very deep into this, and right now,

92
00:05:27,160 --> 00:05:30,679
Speaker 1: we know that there's incredible amount of demand for chips

93
00:05:30,760 --> 00:05:35,040
Speaker 1: for cryptocurrency mining. But you know, go back, I don't know,

94
00:05:36,520 --> 00:05:39,080
Speaker 1: it was a bear market for crypto There wasn't a

95
00:05:39,080 --> 00:05:42,360
Speaker 1: lot of enthusiasm at the time. Talk to us about

96
00:05:42,400 --> 00:05:46,840
Speaker 1: what a sort of normal chip market looks like and

97
00:05:47,040 --> 00:05:50,320
Speaker 1: what your needs are as a company on a stay

98
00:05:50,360 --> 00:05:53,600
Speaker 1: annual basis to acquire chips such that you can provide

99
00:05:53,600 --> 00:05:56,279
Speaker 1: cloud services for your customers. Let's start with the normal

100
00:05:56,279 --> 00:05:59,200
Speaker 1: world first. Yeah, So let's let's start off by framing

101
00:05:59,240 --> 00:06:02,799
Speaker 1: the kind of ore and all needs perfect. We're probably

102
00:06:02,839 --> 00:06:05,320
Speaker 1: buying between like let's call it seven and ten thousand

103
00:06:05,360 --> 00:06:08,800
Speaker 1: servers a year. That's seven ten thousand servers is going

104
00:06:08,839 --> 00:06:11,039
Speaker 1: to be built two different ways, right, One portion of that,

105
00:06:11,080 --> 00:06:14,120
Speaker 1: we're going to build a house, which is typically like

106
00:06:14,160 --> 00:06:17,000
Speaker 1: a really dense GPU compute, and then the second piece

107
00:06:17,080 --> 00:06:19,480
Speaker 1: is going to be kind of more generic CPU builds

108
00:06:19,520 --> 00:06:22,520
Speaker 1: that we buy for ancillary services and other things that

109
00:06:22,560 --> 00:06:25,320
Speaker 1: may not be GPU focused, and that's typically from regular

110
00:06:25,400 --> 00:06:29,320
Speaker 1: like O E m's like super Micro, HP, Dell, whoever

111
00:06:29,400 --> 00:06:32,279
Speaker 1: that may be. What does the typical market look like? Well,

112
00:06:32,480 --> 00:06:34,880
Speaker 1: you know, like sitting here today, I would I would

113
00:06:34,880 --> 00:06:37,680
Speaker 1: die to go back to like six months ago. The

114
00:06:37,680 --> 00:06:40,039
Speaker 1: typical market is a place where you can where things

115
00:06:40,040 --> 00:06:43,240
Speaker 1: are in stock, and you may have trouble finding them,

116
00:06:43,279 --> 00:06:45,599
Speaker 1: but it may take you a couple of hours and

117
00:06:45,680 --> 00:06:47,520
Speaker 1: you can find the stock that you need and haven't

118
00:06:47,520 --> 00:06:50,520
Speaker 1: delivered the next day. It's not really like that right now.

119
00:06:50,880 --> 00:06:54,080
Speaker 1: I spend most of it's a little while that half

120
00:06:54,160 --> 00:06:56,520
Speaker 1: my job here is really scouring the internet to try

121
00:06:56,520 --> 00:06:58,640
Speaker 1: to find people that may have stocked. Like if I'm

122
00:06:58,680 --> 00:07:00,680
Speaker 1: looking for a certain item, whether it's a CPU, whether

123
00:07:00,720 --> 00:07:04,280
Speaker 1: it's RAM, whether it's a special motherboard that we're looking

124
00:07:04,320 --> 00:07:06,240
Speaker 1: for you know, I go through a kind of our

125
00:07:06,279 --> 00:07:09,240
Speaker 1: regular partner channels first, which is some of the large distributors,

126
00:07:09,880 --> 00:07:12,040
Speaker 1: and then typically right now, I get told there's no

127
00:07:12,120 --> 00:07:14,200
Speaker 1: chance to get that within sixteen weeks, and then I

128
00:07:14,240 --> 00:07:16,760
Speaker 1: go to eBay, right and I don't go to eBay

129
00:07:16,760 --> 00:07:19,120
Speaker 1: to buy the stuff. I go to eBay to just say, Okay,

130
00:07:19,200 --> 00:07:21,520
Speaker 1: is there a supply out there? How do I source it?

131
00:07:22,080 --> 00:07:24,240
Speaker 1: Because it's really hard to kind of connect buyers and

132
00:07:24,280 --> 00:07:26,200
Speaker 1: sellers in this market because there's so many people that

133
00:07:26,280 --> 00:07:30,160
Speaker 1: provide services and our distributors and resellers. So it's really

134
00:07:30,200 --> 00:07:32,480
Speaker 1: like okay, like get a feel for what's out there,

135
00:07:32,520 --> 00:07:35,640
Speaker 1: and then go hunt it down. So it's it's pretty wild.

136
00:07:36,320 --> 00:07:38,320
Speaker 1: Like last year, it would be no problem for me

137
00:07:38,360 --> 00:07:41,000
Speaker 1: to go out and get ten terabytes of RAM right

138
00:07:41,040 --> 00:07:43,400
Speaker 1: and like sixty four gig dems of ten of RAM

139
00:07:43,440 --> 00:07:45,560
Speaker 1: and build up ten terabytes and get it overnight it

140
00:07:45,600 --> 00:07:48,440
Speaker 1: to me if we needed it for a build. And like,

141
00:07:48,480 --> 00:07:51,080
Speaker 1: I mean, there was a time late last year where

142
00:07:51,080 --> 00:07:53,240
Speaker 1: we actually had somebody get on an airplane from California

143
00:07:53,240 --> 00:07:55,400
Speaker 1: to Chicago because we needed something done in like three hours.

144
00:07:55,920 --> 00:07:58,600
Speaker 1: That's just not possible today. Wait, did that person have

145
00:07:59,760 --> 00:08:02,000
Speaker 1: a chip in or some sort of chip set in

146
00:08:02,040 --> 00:08:04,560
Speaker 1: there that they just traveled. They just got on the

147
00:08:04,560 --> 00:08:07,200
Speaker 1: plane and brought it to Yeah, we needed that. We

148
00:08:07,240 --> 00:08:09,160
Speaker 1: had to deliver something for a client. It was a

149
00:08:09,240 --> 00:08:12,400
Speaker 1: late late on a Friday night and we needed I

150
00:08:12,400 --> 00:08:14,440
Speaker 1: think it may have been like twenty terabytes of ramp okay,

151
00:08:14,520 --> 00:08:17,120
Speaker 1: and we found a supplier in California that was still open,

152
00:08:17,200 --> 00:08:19,800
Speaker 1: that had it in stock, and we we paid them

153
00:08:19,800 --> 00:08:22,720
Speaker 1: extra to put them on a plane. Wow. Right, So

154
00:08:23,280 --> 00:08:24,960
Speaker 1: a lot of what we do is very much just

155
00:08:25,040 --> 00:08:28,800
Speaker 1: in time. Yeah. Right, So we we've done dumb stuff

156
00:08:28,840 --> 00:08:31,080
Speaker 1: like that before. But if I needed to do that today,

157
00:08:31,120 --> 00:08:32,800
Speaker 1: like I don't even think I could source the ramp.

158
00:08:33,200 --> 00:08:35,959
Speaker 1: So again, we'll get to the today. But I want

159
00:08:35,960 --> 00:08:39,000
Speaker 1: to keep diving further into the normal market. When you

160
00:08:39,040 --> 00:08:45,160
Speaker 1: talk about GPUs, CPUs, RAM, talk to us a little

161
00:08:45,200 --> 00:08:48,600
Speaker 1: bit about the difference between the two GPU and CPU

162
00:08:48,800 --> 00:08:52,080
Speaker 1: and how they serve different types of applications for your customers.

163
00:08:52,400 --> 00:08:54,800
Speaker 1: Let's start with the VFX industry. Okay, right there, it's

164
00:08:54,880 --> 00:08:57,800
Speaker 1: really usable there for two different things. The first is

165
00:08:58,000 --> 00:09:01,800
Speaker 1: for virtual workstations where artists are actually now fully migrated

166
00:09:01,840 --> 00:09:03,800
Speaker 1: to the cloud, and working in the cloud on a

167
00:09:04,480 --> 00:09:07,000
Speaker 1: on a desktop that that we've helped them set up.

168
00:09:07,559 --> 00:09:09,720
Speaker 1: So you know, we have hundreds and hundreds of artists

169
00:09:09,760 --> 00:09:13,440
Speaker 1: doing that today. They connect to us via a low

170
00:09:13,520 --> 00:09:16,600
Speaker 1: latency application from their laptop or from their desktop at

171
00:09:16,640 --> 00:09:18,720
Speaker 1: home or whatever it may be, and then they're connected

172
00:09:18,760 --> 00:09:21,800
Speaker 1: to a hardware accelerated desktop in our cloud. That's one

173
00:09:21,880 --> 00:09:24,160
Speaker 1: use for the GPU, right, The GPU provides the hardware

174
00:09:24,200 --> 00:09:26,400
Speaker 1: acceleration of the desktop as well as the graphics for

175
00:09:26,440 --> 00:09:28,880
Speaker 1: what they're doing. And then the second piece for for

176
00:09:28,920 --> 00:09:31,480
Speaker 1: GPU compute on the VFX side is for the actual

177
00:09:31,520 --> 00:09:34,640
Speaker 1: rendering itself. For us, we actually spend kind of two

178
00:09:34,679 --> 00:09:36,800
Speaker 1: different markets in the rendering side, Like one is for

179
00:09:36,800 --> 00:09:39,840
Speaker 1: for large studios where they have people rendering every day,

180
00:09:40,240 --> 00:09:42,920
Speaker 1: and then we also operate a director consumer render farm

181
00:09:43,160 --> 00:09:47,240
Speaker 1: for small freelancers and artists that has like just over

182
00:09:47,280 --> 00:09:50,480
Speaker 1: twenty users at this point. On both sides, though, the

183
00:09:50,480 --> 00:09:54,400
Speaker 1: demand there is is is bursty and crazy, where hey,

184
00:09:54,440 --> 00:09:56,520
Speaker 1: they need to get a shot done, um, they need

185
00:09:56,520 --> 00:09:58,360
Speaker 1: it done in the next two hours to meet a deadline,

186
00:09:58,480 --> 00:10:00,280
Speaker 1: and if they were to do it on our own

187
00:10:00,280 --> 00:10:02,720
Speaker 1: computer in their own farm, and may take them forty

188
00:10:02,800 --> 00:10:05,520
Speaker 1: days we really become the only option at that point

189
00:10:05,559 --> 00:10:08,160
Speaker 1: for some folks. UM, and you know, it's my job

190
00:10:08,200 --> 00:10:10,440
Speaker 1: here to make sure that I have the GPS, the

191
00:10:10,480 --> 00:10:13,800
Speaker 1: scale of GPS available to be able to handle a

192
00:10:13,880 --> 00:10:16,760
Speaker 1: large number of concurrent users like that. I'm looking at

193
00:10:16,880 --> 00:10:20,199
Speaker 1: your website and you know, scrolling down, one of the

194
00:10:20,200 --> 00:10:22,800
Speaker 1: thing is is the GPUs you need when you need them,

195
00:10:22,840 --> 00:10:26,960
Speaker 1: And you say, we currently have over in video GPUs.

196
00:10:27,120 --> 00:10:30,000
Speaker 1: And so when we talk about this battle for chips

197
00:10:30,080 --> 00:10:34,240
Speaker 1: between gamers and miners, these and I don't know if

198
00:10:34,240 --> 00:10:38,680
Speaker 1: they're precisely the same, but generally speaking, these chips that

199
00:10:38,800 --> 00:10:42,320
Speaker 1: you have are like the chips that everyone covets. No,

200
00:10:43,400 --> 00:10:46,280
Speaker 1: they're not, right, So they're very similar in nature, but

201
00:10:46,400 --> 00:10:49,920
Speaker 1: they're similar. Right. So we're in video cloud services provider partner. Okay,

202
00:10:50,000 --> 00:10:53,280
Speaker 1: so we're restricted from offering consumer based GPUs in our cloud,

203
00:10:53,480 --> 00:10:56,000
Speaker 1: got it, got it right. So we're actually we're required

204
00:10:56,040 --> 00:10:59,839
Speaker 1: to buy datas energy pus. So, but from a technological standpoint,

205
00:11:00,600 --> 00:11:03,160
Speaker 1: we're they're talking, we're talking about roughly the same family

206
00:11:03,200 --> 00:11:06,320
Speaker 1: type of technology. Yeah, you know, yes, it's okay. Like

207
00:11:06,360 --> 00:11:09,640
Speaker 1: the data center GPUs typically have a little bit higher quality,

208
00:11:09,880 --> 00:11:11,840
Speaker 1: like they may have error correcting RAM instead of just

209
00:11:11,920 --> 00:11:15,600
Speaker 1: regular video RAM on them, and then typically the frame buffers,

210
00:11:15,600 --> 00:11:17,720
Speaker 1: so the actual amount of video memory on the GPU

211
00:11:17,720 --> 00:11:19,240
Speaker 1: is going to be higher for the data center cards.

212
00:11:19,800 --> 00:11:23,440
Speaker 1: Got it. But technologically what people are lining up for

213
00:11:23,559 --> 00:11:26,640
Speaker 1: in stores outside of video games supply stores and what

214
00:11:26,720 --> 00:11:28,560
Speaker 1: you have a deck of for do you I have thousand?

215
00:11:28,640 --> 00:11:32,480
Speaker 1: Probably more? By now? It's a it's the same kind

216
00:11:32,520 --> 00:11:37,080
Speaker 1: of technology. And it's the artists that would the animation

217
00:11:37,200 --> 00:11:40,000
Speaker 1: artists who probably worked for big studios or television shows

218
00:11:40,040 --> 00:11:42,680
Speaker 1: on Netflix and so forth. They all want access to

219
00:11:42,920 --> 00:11:46,760
Speaker 1: roughly that same technological capability. Yeah, it went from having

220
00:11:46,840 --> 00:11:50,280
Speaker 1: the access to that in their office now that they

221
00:11:50,320 --> 00:11:52,360
Speaker 1: may not have i T. Staff that's sophisticated enough to

222
00:11:52,400 --> 00:11:54,360
Speaker 1: set it up for them, or their office may no

223
00:11:54,440 --> 00:11:57,560
Speaker 1: longer exist. Right, And we've seen this huge rush to

224
00:11:57,600 --> 00:12:00,040
Speaker 1: the cloud, the wild things we get on the on

225
00:12:00,320 --> 00:12:02,719
Speaker 1: with studios and they're like by the end of the call,

226
00:12:02,760 --> 00:12:05,559
Speaker 1: they're like, Okay, when can we start to be honest

227
00:12:05,600 --> 00:12:07,400
Speaker 1: with you. The problem that we have as a company

228
00:12:07,440 --> 00:12:10,319
Speaker 1: is we're actually limited in our ability to onboard, right,

229
00:12:10,400 --> 00:12:13,000
Speaker 1: is there's just so much demand for this cloud migration. Yeah,

230
00:12:13,280 --> 00:12:15,240
Speaker 1: and then when I look at that from a capacity

231
00:12:15,240 --> 00:12:18,280
Speaker 1: planning perspective, it's like you're looking at this environment that

232
00:12:18,320 --> 00:12:20,960
Speaker 1: we have today where the world is in this one

233
00:12:21,000 --> 00:12:24,600
Speaker 1: big shortage of everything possible. Doing the capacity planning and

234
00:12:24,600 --> 00:12:41,920
Speaker 1: scaling for that is this is kind of hard. You

235
00:12:41,960 --> 00:12:44,560
Speaker 1: said something, you said birth capacity, and again I'm looking

236
00:12:44,559 --> 00:12:46,960
Speaker 1: at your website. It says the GPS you need when

237
00:12:47,000 --> 00:12:49,760
Speaker 1: you need them. And so what I take from this

238
00:12:50,640 --> 00:12:53,960
Speaker 1: is that perhaps there are some you know, consumer startups

239
00:12:54,000 --> 00:12:57,160
Speaker 1: or whatever, and they might use AWS cloud and they

240
00:12:57,280 --> 00:13:01,080
Speaker 1: use a sort of predictable on a day to day

241
00:13:01,120 --> 00:13:05,960
Speaker 1: basis roughly share of cloud capacity. Well you're offering, though,

242
00:13:06,440 --> 00:13:08,880
Speaker 1: I guess part of your niche. It sounds like, is

243
00:13:09,360 --> 00:13:13,640
Speaker 1: here's someone who doesn't normally need computational power, but when

244
00:13:13,640 --> 00:13:17,440
Speaker 1: they need crazy computational power, a burst of it. It's

245
00:13:17,440 --> 00:13:21,480
Speaker 1: not regular, you have huge, you have huge scaling ability

246
00:13:21,520 --> 00:13:23,640
Speaker 1: for them. Yeah, you know that. That's the crazy thing.

247
00:13:24,440 --> 00:13:26,720
Speaker 1: When we hired a director of sales last year, I

248
00:13:26,760 --> 00:13:28,760
Speaker 1: had known him for like ten years, and I was like, dude,

249
00:13:28,800 --> 00:13:31,880
Speaker 1: like like every customer we talked to tells us they

250
00:13:31,920 --> 00:13:33,679
Speaker 1: just cannot get the capacity they need at the big

251
00:13:33,679 --> 00:13:36,120
Speaker 1: clouds and he's like, yeah, okay, like that's a bunch

252
00:13:36,120 --> 00:13:39,040
Speaker 1: of b s whatever. Okay, you're being like hyperbolic. And

253
00:13:39,080 --> 00:13:41,319
Speaker 1: then he joined us. In the first like fifty meetings,

254
00:13:41,320 --> 00:13:43,520
Speaker 1: he did every one in the exact same thing. So

255
00:13:43,840 --> 00:13:47,079
Speaker 1: I think that the industry really struggles with that access

256
00:13:47,120 --> 00:13:49,719
Speaker 1: to scale. Yeah, right, And the reason for that is

257
00:13:49,760 --> 00:13:51,839
Speaker 1: just I don't like a w S g C P

258
00:13:51,920 --> 00:13:55,600
Speaker 1: ANDAGE or are building these. They're kind of hyper scale

259
00:13:55,640 --> 00:13:59,360
Speaker 1: data center regions two handle every single use case under

260
00:13:59,400 --> 00:14:02,440
Speaker 1: the sun, right, right, So they've got X number of

261
00:14:02,440 --> 00:14:05,599
Speaker 1: types of CPU compute, they've got all different types of

262
00:14:05,640 --> 00:14:08,000
Speaker 1: optimized compute, and you know, at the end of the day,

263
00:14:08,000 --> 00:14:11,160
Speaker 1: there's only a certain portion that's available for the GPU stuff, right,

264
00:14:11,200 --> 00:14:13,640
Speaker 1: And when that's our entire business, like, we can really

265
00:14:13,640 --> 00:14:16,400
Speaker 1: approach it very differently. You know. What we want people

266
00:14:16,440 --> 00:14:17,839
Speaker 1: to do is to come in and say, hey, I

267
00:14:17,880 --> 00:14:20,160
Speaker 1: need to get two thousand GPUs. I need them for

268
00:14:20,200 --> 00:14:22,880
Speaker 1: three hours. If they were to do that at AWUS

269
00:14:22,960 --> 00:14:24,960
Speaker 1: like one, it may not even be possible for them

270
00:14:24,960 --> 00:14:27,080
Speaker 1: if they don't have a good relationship. And then too,

271
00:14:27,200 --> 00:14:30,000
Speaker 1: they probably have to do a lot of like manual

272
00:14:30,120 --> 00:14:34,280
Speaker 1: engineering and reservation planning to get there, we have people

273
00:14:34,320 --> 00:14:36,880
Speaker 1: come in all the time that they they'll burst the

274
00:14:37,080 --> 00:14:39,960
Speaker 1: like to three four or five tho GPUs for like

275
00:14:40,000 --> 00:14:41,760
Speaker 1: six hours and then we won't hear from again for

276
00:14:41,800 --> 00:14:45,760
Speaker 1: three weeks. My commitment to them is that I want

277
00:14:45,800 --> 00:14:47,800
Speaker 1: them to have that capacity without having to talk to me.

278
00:14:48,280 --> 00:14:50,440
Speaker 1: So one thing that and I guess I kind of

279
00:14:50,480 --> 00:14:53,760
Speaker 1: intuitively knew this before, but one thing that's making this

280
00:14:53,840 --> 00:14:57,400
Speaker 1: clear is that cloud computing. You know, you think of

281
00:14:57,400 --> 00:14:59,560
Speaker 1: a cloud like a literal clouds at the cloud up there,

282
00:14:59,560 --> 00:15:03,120
Speaker 1: but cloud computing it's just not a commodity, both in

283
00:15:03,240 --> 00:15:08,160
Speaker 1: terms of the specific types of applications, but also in

284
00:15:08,680 --> 00:15:12,400
Speaker 1: UH scaling and the difference between sort of like standard

285
00:15:12,480 --> 00:15:17,920
Speaker 1: run time consistent demand versus cloud that can be turned

286
00:15:17,920 --> 00:15:21,600
Speaker 1: on at high scale for a client at a moment's notice.

287
00:15:22,320 --> 00:15:25,120
Speaker 1: That's a really good point. I like to think that

288
00:15:25,640 --> 00:15:28,480
Speaker 1: our scale helps us break the commodity label. Yeah, but

289
00:15:28,640 --> 00:15:30,720
Speaker 1: for anybody who goes in then needs like one virtual

290
00:15:30,760 --> 00:15:33,560
Speaker 1: machine with like four CPU corps and sixteen RAM like,

291
00:15:33,840 --> 00:15:36,560
Speaker 1: all right, go get in the cheapest place, right, same same,

292
00:15:37,080 --> 00:15:38,760
Speaker 1: But that's not even the people that we go after

293
00:15:38,840 --> 00:15:40,920
Speaker 1: or talk to, right, I can't help them. I can't

294
00:15:40,920 --> 00:15:45,080
Speaker 1: help them there, right, Why is RAM hard to come by? Okay,

295
00:15:45,160 --> 00:15:47,520
Speaker 1: so let's go, let's go to crypto money, all right,

296
00:15:47,680 --> 00:15:49,360
Speaker 1: I guess where I was. I was trying to like

297
00:15:49,440 --> 00:15:52,320
Speaker 1: hold off the crypto part for as long as we could,

298
00:15:52,960 --> 00:15:54,680
Speaker 1: but I knew eventually I would ask you a question

299
00:15:54,720 --> 00:15:56,800
Speaker 1: and we'd stumble into the crypto parts. So let's I

300
00:15:56,840 --> 00:15:59,760
Speaker 1: guess we're there now. Yeah, so that there's a lot

301
00:15:59,800 --> 00:16:02,520
Speaker 1: of supply chain stuff here is going to be driven back,

302
00:16:02,520 --> 00:16:04,600
Speaker 1: it's going to be kind of linked back to crypto. Well, actually,

303
00:16:04,640 --> 00:16:07,800
Speaker 1: then in that case, let me so before then, you

304
00:16:07,840 --> 00:16:09,960
Speaker 1: mentioned how you would love to go back like six

305
00:16:10,040 --> 00:16:12,920
Speaker 1: months ago because your job was really easy. So when

306
00:16:12,960 --> 00:16:15,760
Speaker 1: did you notice that things were starting to change and

307
00:16:15,800 --> 00:16:20,960
Speaker 1: that sourcing compute, whether it's GPUs, CPUs or RAM, that

308
00:16:21,120 --> 00:16:23,880
Speaker 1: was suddenly starting to get tighter. You know, every February

309
00:16:23,920 --> 00:16:27,360
Speaker 1: it gets pretty rough because of Chinese New Year. Okay, Right,

310
00:16:27,560 --> 00:16:29,320
Speaker 1: so you kind of rushed to get all your orders

311
00:16:29,400 --> 00:16:31,640
Speaker 1: in before the end of because this is where it's

312
00:16:31,640 --> 00:16:34,800
Speaker 1: all being manufactured. Yeah, before before the middle of January, right,

313
00:16:34,880 --> 00:16:36,080
Speaker 1: and then if you get it in before the middle

314
00:16:36,080 --> 00:16:37,800
Speaker 1: of January, they're gonna do everything they can to shift

315
00:16:37,800 --> 00:16:40,240
Speaker 1: it out for the start of their holiday. If you

316
00:16:40,240 --> 00:16:42,440
Speaker 1: sent it by air, you'll have it early February, sent

317
00:16:42,480 --> 00:16:45,320
Speaker 1: it by boat end of February, but like if you wait,

318
00:16:45,800 --> 00:16:49,040
Speaker 1: there's there's nothing that gets processed in February in China. Okay,

319
00:16:49,320 --> 00:16:52,480
Speaker 1: So that was kind of like normal operating conditions in

320
00:16:52,600 --> 00:16:54,840
Speaker 1: middle of January, and you know, Crypto was kind of

321
00:16:54,840 --> 00:16:57,080
Speaker 1: doing his thing and it was kind of ripping the

322
00:16:57,160 --> 00:17:01,160
Speaker 1: roof off and you saw it coming right, But a

323
00:17:01,200 --> 00:17:03,520
Speaker 1: lot of that was, you know, that was really in

324
00:17:03,560 --> 00:17:06,520
Speaker 1: the consumer side, and so consumer GPUs are really tight,

325
00:17:06,840 --> 00:17:12,200
Speaker 1: didn't really impact us. And then starting in March, orders

326
00:17:12,240 --> 00:17:14,240
Speaker 1: that we have placed in January that we're supposed to

327
00:17:14,240 --> 00:17:16,960
Speaker 1: be in stock in the us UM and we're supposed

328
00:17:16,960 --> 00:17:19,680
Speaker 1: to be delivered the next day, they didn't show up right.

329
00:17:19,800 --> 00:17:22,560
Speaker 1: And this is like, let's call it medium quality server

330
00:17:22,600 --> 00:17:25,959
Speaker 1: grade CPUs, so like a md epic rome platform, and

331
00:17:26,080 --> 00:17:28,320
Speaker 1: we had maybe five hundred or seven fifty at least

332
00:17:28,320 --> 00:17:31,399
Speaker 1: things on order, and it was like, okay, fine, like

333
00:17:31,440 --> 00:17:33,040
Speaker 1: the other components that we need to do our bills

334
00:17:33,040 --> 00:17:35,119
Speaker 1: aren't here yet, Like they arrived in early March, and

335
00:17:35,160 --> 00:17:37,640
Speaker 1: we were actually waiting a GPUs at the time, and

336
00:17:37,800 --> 00:17:39,520
Speaker 1: I kind of wasn't really pressing about it. And then

337
00:17:39,560 --> 00:17:41,959
Speaker 1: the middle of March came and I'm like, where's my stuff?

338
00:17:42,600 --> 00:17:45,040
Speaker 1: And you reach out and they go, oh, we didn't

339
00:17:45,040 --> 00:17:46,639
Speaker 1: first see the semi conductor short. I'm like, what are

340
00:17:46,640 --> 00:17:48,200
Speaker 1: you talking about? You told me these were in stock

341
00:17:48,240 --> 00:17:52,680
Speaker 1: inventor warehouse yesterday like that. It was fine. And this

342
00:17:52,720 --> 00:17:54,679
Speaker 1: is kind of what you start to see when the

343
00:17:54,760 --> 00:17:57,280
Speaker 1: market gets stretched like this, is you have conversations like

344
00:17:57,320 --> 00:17:59,119
Speaker 1: that and people just flat out lie to you in

345
00:17:59,160 --> 00:18:02,399
Speaker 1: the supply chain. That was the first real telltale sign.

346
00:18:02,920 --> 00:18:05,359
Speaker 1: The second one was when I went to buy I

347
00:18:05,400 --> 00:18:07,800
Speaker 1: think it was like eight d sticks of sixty d

348
00:18:07,880 --> 00:18:10,720
Speaker 1: d R four you know for that, like for for

349
00:18:10,800 --> 00:18:13,480
Speaker 1: d d R four what d d R four it's

350
00:18:13,760 --> 00:18:17,600
Speaker 1: System Ram okay, right, so it's the it's the most

351
00:18:17,600 --> 00:18:20,640
Speaker 1: recent commodity system rahm um. There's new versions coming out,

352
00:18:20,680 --> 00:18:23,080
Speaker 1: but this is really what's used today. You reached around

353
00:18:23,119 --> 00:18:24,719
Speaker 1: for System Ram because a lot of that stuff comes

354
00:18:24,760 --> 00:18:27,159
Speaker 1: out of liquidation, right, so you can get it for

355
00:18:27,200 --> 00:18:29,800
Speaker 1: this particular component for Ram, you typically buy it used.

356
00:18:30,480 --> 00:18:33,880
Speaker 1: You're not going to go out and spend sixty premium

357
00:18:33,880 --> 00:18:36,400
Speaker 1: on this stuff. Just because you know, that doesn't really

358
00:18:36,400 --> 00:18:39,639
Speaker 1: wear down. So I reached out to to one of

359
00:18:39,640 --> 00:18:41,639
Speaker 1: our main suppliers and I said, hey, man, like I

360
00:18:41,680 --> 00:18:43,560
Speaker 1: need eight hundred sticks and he's like, dude, I got nothing,

361
00:18:43,960 --> 00:18:45,240
Speaker 1: Like what are you talking about? Like what do you

362
00:18:45,280 --> 00:18:48,440
Speaker 1: mean you give nothing? He's like, China? But when was that? What?

363
00:18:48,440 --> 00:18:50,920
Speaker 1: What month? This was in March? Middle of March? Yeah,

364
00:18:51,560 --> 00:18:53,800
Speaker 1: and he goes he goes, dude, like I had some

365
00:18:53,840 --> 00:18:56,040
Speaker 1: guy in China buy all of it last night, Like

366
00:18:56,080 --> 00:18:57,360
Speaker 1: what do you mean all of it? And he's like, yeah,

367
00:18:57,440 --> 00:19:01,200
Speaker 1: he bought like ten thousand sticks. Stupid question and stupid question.

368
00:19:01,240 --> 00:19:04,880
Speaker 1: I don't know if this is you know, proprietary sticks,

369
00:19:04,920 --> 00:19:06,800
Speaker 1: Like what are we talking about from a dollar amount

370
00:19:08,160 --> 00:19:10,240
Speaker 1: two and a half maybe I think it's maybe two

371
00:19:10,280 --> 00:19:13,640
Speaker 1: and a half million bucks? Okay, okay, right, but ten

372
00:19:13,640 --> 00:19:16,920
Speaker 1: thousand six a ram is not really a small order,

373
00:19:17,080 --> 00:19:18,399
Speaker 1: right you don't just like buy them and throw it

374
00:19:18,440 --> 00:19:20,720
Speaker 1: in the closet for rainy day. Well I guess maybe

375
00:19:20,720 --> 00:19:22,880
Speaker 1: this year you do. And I said, okay, like what's

376
00:19:22,920 --> 00:19:24,960
Speaker 1: going on and then said that's he was like, you know,

377
00:19:25,040 --> 00:19:26,720
Speaker 1: some one buyer came in from China and bought it

378
00:19:26,800 --> 00:19:29,680
Speaker 1: all and and you know, I kind of give him

379
00:19:29,680 --> 00:19:32,040
Speaker 1: the the expletive late and you gotta be kidding me.

380
00:19:32,320 --> 00:19:33,679
Speaker 1: A couple of days later he called me and he

381
00:19:33,720 --> 00:19:35,359
Speaker 1: was like, hey, I got like four hundred sticks, Like

382
00:19:35,359 --> 00:19:36,719
Speaker 1: do you want him? And I said, yeah, I'll take

383
00:19:36,880 --> 00:19:39,199
Speaker 1: everything you have. So that that was kind of my

384
00:19:39,280 --> 00:19:41,919
Speaker 1: reaction to that first one. Then a week later on

385
00:19:41,960 --> 00:19:43,280
Speaker 1: the fact of the same guy and said, hey, I've

386
00:19:43,280 --> 00:19:45,359
Speaker 1: got to I've got a need for like Intel's on

387
00:19:45,440 --> 00:19:49,240
Speaker 1: V four, which is chips that probably stopped production like ten,

388
00:19:50,000 --> 00:19:52,159
Speaker 1: Like what do you got? And he's like, dude, like,

389
00:19:52,200 --> 00:19:53,600
Speaker 1: are you gonna kill me if I tell you that

390
00:19:53,600 --> 00:19:57,240
Speaker 1: I sold them all to China last night? Like okay.

391
00:19:57,320 --> 00:19:59,560
Speaker 1: So this is at the point when when final coin

392
00:19:59,640 --> 00:20:04,040
Speaker 1: was going from like bucks. So just for so just

393
00:20:04,080 --> 00:20:06,399
Speaker 1: for listeners that don't know, and why do you explain

394
00:20:06,440 --> 00:20:09,040
Speaker 1: it further? But file coin is a new cryptocurrency is

395
00:20:09,080 --> 00:20:11,880
Speaker 1: actually I think first conceived of back in like seventeen.

396
00:20:11,920 --> 00:20:17,760
Speaker 1: It only recently launched, but the idea is essentially cloud files,

397
00:20:17,800 --> 00:20:22,000
Speaker 1: distributed file storage in the cloud. Rather than hosting your

398
00:20:22,200 --> 00:20:25,400
Speaker 1: documents with a company like dropbox or box, you hosted

399
00:20:25,400 --> 00:20:28,840
Speaker 1: on a network of a decentralized blockchain network of computers

400
00:20:28,880 --> 00:20:31,600
Speaker 1: all around the world. Yeah, and all around the world.

401
00:20:31,640 --> 00:20:34,480
Speaker 1: There is a little bit disingenuous at this point. I'd

402
00:20:34,480 --> 00:20:38,840
Speaker 1: say it's centralized in China, okay, but the the ecosystem

403
00:20:38,840 --> 00:20:42,840
Speaker 1: in China is they are insane for this project, okay. Interesting,

404
00:20:42,920 --> 00:20:46,280
Speaker 1: And the demand for this project spans it spans every

405
00:20:46,280 --> 00:20:49,200
Speaker 1: component in the supply chain, right, So you you need

406
00:20:49,240 --> 00:20:53,560
Speaker 1: GPUs for the actual computation of proofs, you need CPUs

407
00:20:53,600 --> 00:20:56,199
Speaker 1: because you need high high core clocks for some of

408
00:20:56,200 --> 00:20:59,960
Speaker 1: the processing. You need RAM because because of the storage operations,

409
00:21:00,000 --> 00:21:02,399
Speaker 1: and actually putting stuff in cash. So they're kind of

410
00:21:02,400 --> 00:21:06,000
Speaker 1: buying across the board. And like if you if you

411
00:21:06,040 --> 00:21:08,720
Speaker 1: search real quickly, you'll find mentions in the news of

412
00:21:08,760 --> 00:21:12,560
Speaker 1: people putting one point three billion dollars into file coin mining.

413
00:21:12,600 --> 00:21:15,399
Speaker 1: And I mean there's real moneyed over there. So that

414
00:21:16,000 --> 00:21:19,120
Speaker 1: this file coin driven demand has kind of sucked up

415
00:21:19,480 --> 00:21:23,119
Speaker 1: beyond what the typical component demand. And crypto bubble or

416
00:21:23,160 --> 00:21:27,200
Speaker 1: crypto bubble or crypto expansionary times, it's called that instead

417
00:21:27,200 --> 00:21:30,199
Speaker 1: of a bubble, it's going to drive, right. So it

418
00:21:30,240 --> 00:21:32,720
Speaker 1: went from you know, two thousand eighteen or two thousand

419
00:21:32,840 --> 00:21:35,280
Speaker 1: late two thousand seventeen early two eighteen, where the demand

420
00:21:35,280 --> 00:21:39,200
Speaker 1: was only for GPUs two. Now you have Ethereum driving

421
00:21:39,200 --> 00:21:41,639
Speaker 1: demand for GPUs, you have file Coin driving demand for

422
00:21:41,680 --> 00:21:44,520
Speaker 1: everything under the sun. And the amounts of capital that

423
00:21:44,560 --> 00:21:47,200
Speaker 1: are being thrown at this in China is just draining

424
00:21:47,200 --> 00:21:52,440
Speaker 1: the supply chain across the world. Wow. Right, So that

425
00:21:52,480 --> 00:21:54,600
Speaker 1: was really when I was like, okay, like my life

426
00:21:54,640 --> 00:21:57,720
Speaker 1: just changed, and we have a risk here where we

427
00:21:57,760 --> 00:22:00,520
Speaker 1: may actually be governed in our ability to grow by

428
00:22:00,560 --> 00:22:03,800
Speaker 1: access to the components. I just want to point out,

429
00:22:03,840 --> 00:22:06,439
Speaker 1: like I'm looking at a price chart right now of

430
00:22:06,840 --> 00:22:11,320
Speaker 1: file coin, the coin even as recently in January, like

431
00:22:11,400 --> 00:22:15,639
Speaker 1: say January one, one coin, one file coin was trading

432
00:22:15,640 --> 00:22:20,000
Speaker 1: at twenty two dollars. Fast forward to March thirty one,

433
00:22:20,240 --> 00:22:22,119
Speaker 1: a file coin coin to chart I'm looking at, was

434
00:22:22,160 --> 00:22:24,600
Speaker 1: at So that's almost like a ten bagger in the

435
00:22:24,600 --> 00:22:28,320
Speaker 1: price a file coin over the span of about three months,

436
00:22:28,680 --> 00:22:31,639
Speaker 1: which really, it seems to me then just makes the

437
00:22:31,720 --> 00:22:35,640
Speaker 1: demand to acquire all of these chips even that much more.

438
00:22:35,640 --> 00:22:38,120
Speaker 1: If if acquiring one coin is worth ten times more,

439
00:22:38,680 --> 00:22:41,119
Speaker 1: then what you're going to be willing to pay for

440
00:22:41,720 --> 00:22:45,200
Speaker 1: the compute to mind a file coin that just went

441
00:22:45,240 --> 00:22:47,159
Speaker 1: up a lot. Yeah, you know, And one of the

442
00:22:47,200 --> 00:22:50,159
Speaker 1: things that is unique or weird about file coin is

443
00:22:50,200 --> 00:22:53,919
Speaker 1: the staking requirements, right, so you know, a lot of

444
00:22:53,920 --> 00:22:57,359
Speaker 1: the cost of actually building the storage isn't in the

445
00:22:57,400 --> 00:23:00,960
Speaker 1: storage for file coin, h it's the amount of coin

446
00:23:01,000 --> 00:23:02,919
Speaker 1: you actually have the stake to be able to provide

447
00:23:02,920 --> 00:23:05,720
Speaker 1: your storage to the network. And at the time, like

448
00:23:05,720 --> 00:23:07,199
Speaker 1: when it was like a hundred nine bucks, like, we

449
00:23:07,240 --> 00:23:09,240
Speaker 1: had a lot of internal conversations about this, like why

450
00:23:09,240 --> 00:23:11,040
Speaker 1: aren't we going crazy on this, and why aren't we

451
00:23:11,080 --> 00:23:13,479
Speaker 1: doing this? For us to provide a one Paeder bid

452
00:23:13,520 --> 00:23:16,080
Speaker 1: of storage, which isn't a ton of storage to the network,

453
00:23:16,359 --> 00:23:17,879
Speaker 1: we would have had to stake a million dollars with

454
00:23:17,960 --> 00:23:21,520
Speaker 1: a file coin, okay, right, it costs me maybe like

455
00:23:21,560 --> 00:23:23,879
Speaker 1: a d k to build up one pad of storage,

456
00:23:24,520 --> 00:23:27,160
Speaker 1: so ten times the hardware value I have to put

457
00:23:27,240 --> 00:23:30,120
Speaker 1: up in staking risk and a coin that just went

458
00:23:30,200 --> 00:23:33,280
Speaker 1: up like ten x. I think at the time, like

459
00:23:33,359 --> 00:23:36,560
Speaker 1: it was this this positive feedback loop. People were going

460
00:23:36,600 --> 00:23:39,439
Speaker 1: into it because they want the prices going up, but

461
00:23:39,560 --> 00:23:41,119
Speaker 1: to get into it, they had to buy more. So

462
00:23:41,280 --> 00:23:45,680
Speaker 1: you buy wait, see, you buy hardware to mind file client.

463
00:23:46,240 --> 00:23:49,200
Speaker 1: But then to actually be able to plug your hardware

464
00:23:49,280 --> 00:23:52,679
Speaker 1: into the network and actually you know, use it, you

465
00:23:52,760 --> 00:23:56,040
Speaker 1: also have to buy the coin itself. So, as you say,

466
00:23:56,080 --> 00:23:59,800
Speaker 1: there's like this like weird like positive feedback loop that emerges.

467
00:24:00,000 --> 00:24:03,359
Speaker 1: Are you make this capital investment in semi in chips,

468
00:24:03,960 --> 00:24:06,800
Speaker 1: but then to use the chip to acquire coin, but

469
00:24:06,880 --> 00:24:09,879
Speaker 1: then to actually make the chips worthwhile you have to

470
00:24:09,920 --> 00:24:12,240
Speaker 1: buy the coin itself. And so do you get this

471
00:24:12,320 --> 00:24:15,480
Speaker 1: like crazy network effects? Yeah, you know, it's so like

472
00:24:16,160 --> 00:24:18,240
Speaker 1: I'm not gonna throw the P word out there, but okay,

473
00:24:18,359 --> 00:24:20,440
Speaker 1: in terms of schemes, it's a pretty good one. It's

474
00:24:20,480 --> 00:24:23,040
Speaker 1: a project that I think that internally like we're we

475
00:24:23,119 --> 00:24:25,399
Speaker 1: still look at as you know, it fits our infrastructure

476
00:24:25,400 --> 00:24:28,119
Speaker 1: really well. But right now it's just so hard to

477
00:24:28,119 --> 00:24:29,959
Speaker 1: get into it because you have to really have access

478
00:24:29,960 --> 00:24:33,560
Speaker 1: capacity for it, and getting access to that access capacity

479
00:24:33,640 --> 00:24:35,400
Speaker 1: is just so hard. Like kind of bringing us back

480
00:24:35,440 --> 00:24:38,760
Speaker 1: to the broader conversation here and now another now I

481
00:24:38,760 --> 00:24:40,800
Speaker 1: want to get into some of the other sort of

482
00:24:40,840 --> 00:24:43,320
Speaker 1: coins that are driving this dynamic. You and I had

483
00:24:43,320 --> 00:24:44,840
Speaker 1: a chat like on the phone like two or three

484
00:24:44,840 --> 00:24:47,600
Speaker 1: weeks ago before we did the podcast, and now and

485
00:24:47,640 --> 00:24:50,080
Speaker 1: I hadn't like sort of seen any articles about it.

486
00:24:50,160 --> 00:24:53,480
Speaker 1: But then there's this other coin, Chia that's in the mix,

487
00:24:53,680 --> 00:24:56,920
Speaker 1: So what's happening with that? Yeah, so that one is

488
00:24:57,520 --> 00:25:00,399
Speaker 1: is pretty crazy too. Okay, this is a I guess

489
00:25:00,280 --> 00:25:02,480
Speaker 1: a different take on proof of work, right, so I

490
00:25:02,480 --> 00:25:05,600
Speaker 1: guess they're calling it proof of space time. And effectively,

491
00:25:05,640 --> 00:25:08,080
Speaker 1: what you do is you use CPU compute to create

492
00:25:08,080 --> 00:25:11,359
Speaker 1: what they call plots, and then once a plot is created,

493
00:25:11,400 --> 00:25:14,359
Speaker 1: it's like a a couple hundred gigabyte space on a

494
00:25:14,359 --> 00:25:16,399
Speaker 1: hard drive. You can move that you can move at

495
00:25:16,440 --> 00:25:19,320
Speaker 1: the cold storage, and then for up to five years

496
00:25:19,400 --> 00:25:22,040
Speaker 1: you can use that as a as a mining mechanism

497
00:25:22,080 --> 00:25:25,119
Speaker 1: to gain to gain rewards on the network. That's another

498
00:25:25,119 --> 00:25:28,480
Speaker 1: one that I didn't really see coming. And like we

499
00:25:28,560 --> 00:25:31,639
Speaker 1: build large like multi petal bite level storage clusters in

500
00:25:31,680 --> 00:25:34,800
Speaker 1: each of our data centers and they're STEFF based, so

501
00:25:34,840 --> 00:25:37,919
Speaker 1: it's a software defined storage technology. And one of the

502
00:25:37,920 --> 00:25:39,440
Speaker 1: things that we do there is we we build them

503
00:25:39,480 --> 00:25:42,760
Speaker 1: to have triple replication. Right, so for every one pet

504
00:25:42,760 --> 00:25:45,359
Speaker 1: of it of storage usable for my clients, I have

505
00:25:45,440 --> 00:25:47,840
Speaker 1: to have three petabytes like on the ground that installed.

506
00:25:48,359 --> 00:25:50,960
Speaker 1: That makes me a pretty big buyer of hard drives.

507
00:25:51,000 --> 00:25:52,960
Speaker 1: This is another piece that kind of showed up in

508
00:25:52,960 --> 00:25:56,320
Speaker 1: in January that I didn't really understand yet, is we

509
00:25:56,040 --> 00:25:58,840
Speaker 1: we were trying to get access to I guess a

510
00:25:58,960 --> 00:26:02,000
Speaker 1: quasie non public EXCU and it's it's one that really

511
00:26:02,000 --> 00:26:04,520
Speaker 1: meant for data centers. It would reduce our latency in

512
00:26:04,560 --> 00:26:09,440
Speaker 1: our storage access pattern. You sk you just mean a product? Yeah, yeah, okay, no, no, no,

513
00:26:09,440 --> 00:26:12,320
Speaker 1: it's fine, just making sure so we we we went

514
00:26:12,320 --> 00:26:14,639
Speaker 1: through this process of kind of qualifying with the with

515
00:26:14,720 --> 00:26:16,919
Speaker 1: the manufacturer, and they were asking me all these questions

516
00:26:16,920 --> 00:26:18,600
Speaker 1: like what are they being used for? But I'm like, dude,

517
00:26:18,600 --> 00:26:20,959
Speaker 1: they're hard drives. What do you want be selling them

518
00:26:20,960 --> 00:26:22,359
Speaker 1: to me or not? Like if you want to increase

519
00:26:22,400 --> 00:26:23,959
Speaker 1: the price, increase the price, I'll blind, I don't care.

520
00:26:24,520 --> 00:26:27,480
Speaker 1: And it wasn't until March when like this, when she

521
00:26:27,680 --> 00:26:30,479
Speaker 1: launched that I was like, oh, like these guys had

522
00:26:30,520 --> 00:26:32,440
Speaker 1: a run on their stock and they weren't able to

523
00:26:32,440 --> 00:26:36,800
Speaker 1: provide like product to their real customers. So the demand

524
00:26:36,840 --> 00:26:40,120
Speaker 1: for Chia for hard drives with Chia, it's twofold, right.

525
00:26:40,160 --> 00:26:42,280
Speaker 1: So the first is you really need like high speed

526
00:26:42,359 --> 00:26:44,720
Speaker 1: NV and ME or SSD discs that are local to

527
00:26:44,760 --> 00:26:48,040
Speaker 1: your CPU compute to run the plotting, and then once

528
00:26:48,080 --> 00:26:49,560
Speaker 1: the plotting is done, you can push it off to

529
00:26:49,640 --> 00:26:52,600
Speaker 1: whatever cold storage device, whether it's like spinning rust like

530
00:26:52,680 --> 00:26:55,720
Speaker 1: hard drives or or whatever it may be. But then

531
00:26:55,720 --> 00:26:57,680
Speaker 1: it just sits there and the access pattern is pretty

532
00:26:57,680 --> 00:27:00,400
Speaker 1: low in that data. So this is another or thing

533
00:27:00,480 --> 00:27:03,040
Speaker 1: where I have conversations say hey, like what have you

534
00:27:03,080 --> 00:27:07,320
Speaker 1: gotten in like SASK and hard drives, and they're like,

535
00:27:07,520 --> 00:27:09,760
Speaker 1: would you believe that somebody bought all of these last night?

536
00:27:10,400 --> 00:27:12,880
Speaker 1: I didn't expect to be getting dragged into she in, Yeah,

537
00:27:12,920 --> 00:27:16,199
Speaker 1: And I was only really dragged into it by the

538
00:27:16,280 --> 00:27:18,880
Speaker 1: problems I was have acquiring I was having acquiring things,

539
00:27:18,880 --> 00:27:20,040
Speaker 1: and like you know, every once in a while, like

540
00:27:20,080 --> 00:27:21,800
Speaker 1: the bells go off, like you know, AD eight go

541
00:27:21,920 --> 00:27:24,240
Speaker 1: look at this, Like what is this thing? Yeah, so

542
00:27:24,280 --> 00:27:26,600
Speaker 1: now we've got a pretty large she of mining operation.

543
00:27:27,080 --> 00:27:29,200
Speaker 1: So this is something that I want to get into.

544
00:27:29,240 --> 00:27:32,640
Speaker 1: And it's actually pretty interesting in the way you're set

545
00:27:32,720 --> 00:27:35,680
Speaker 1: up Core. We've prior to like you know, all this

546
00:27:35,840 --> 00:27:37,920
Speaker 1: like crypto going nuts over the last over the last

547
00:27:37,920 --> 00:27:40,640
Speaker 1: several months, so particularly it seems like soon as January

548
00:27:40,640 --> 00:27:43,800
Speaker 1: in March, when all these all condit just started going bananas.

549
00:27:44,440 --> 00:27:48,719
Speaker 1: Your clients were people who actually had what most normal

550
00:27:48,800 --> 00:27:51,840
Speaker 1: humans considered to be real world in need. So like

551
00:27:52,240 --> 00:27:56,120
Speaker 1: animating a TV show or animating a movie. They're actually

552
00:27:56,119 --> 00:28:00,679
Speaker 1: making things, actually actually making things. And so you are

553
00:28:00,800 --> 00:28:04,400
Speaker 1: your go out into the market and buy chips and

554
00:28:04,560 --> 00:28:08,360
Speaker 1: it wasn't that hard. Now you're also involved in the mining,

555
00:28:08,480 --> 00:28:10,879
Speaker 1: but you do it in such a way that it

556
00:28:10,920 --> 00:28:14,639
Speaker 1: doesn't it compliments the service you provide to your clients

557
00:28:15,160 --> 00:28:18,159
Speaker 1: as opposed to compete with them. I guess, So, can

558
00:28:18,200 --> 00:28:22,200
Speaker 1: you explain a little bit about how you use cryptocurrency

559
00:28:22,200 --> 00:28:27,800
Speaker 1: mining yourself. At core, we've to compliment your business, compliment

560
00:28:27,840 --> 00:28:32,160
Speaker 1: your existing business model. The permissionless revenue from cryptocurrency mining

561
00:28:32,160 --> 00:28:34,199
Speaker 1: has really been what's allowed us to build to the

562
00:28:34,280 --> 00:28:38,640
Speaker 1: scale where we can provide these first compute services. Right

563
00:28:38,720 --> 00:28:42,000
Speaker 1: like with without crypto, especially over the last like six months,

564
00:28:42,440 --> 00:28:45,360
Speaker 1: we would have had to raise significant, significant amounts of

565
00:28:45,360 --> 00:28:47,560
Speaker 1: money to do it. To run this business. I think

566
00:28:47,560 --> 00:28:49,880
Speaker 1: that like life, to date, we've probably sold like we

567
00:28:49,960 --> 00:28:52,000
Speaker 1: may maybe raise like six or seven million bucks in

568
00:28:52,000 --> 00:28:54,840
Speaker 1: this whole thing. Yeah, every time that we're looking for

569
00:28:54,880 --> 00:28:57,560
Speaker 1: capital or capital, we have capital requirements like we're always

570
00:28:57,560 --> 00:29:00,400
Speaker 1: trying to protect ourselves. I guess founders like you don't

571
00:29:00,400 --> 00:29:02,360
Speaker 1: want to dilute it, especially don't want to dilute to

572
00:29:02,360 --> 00:29:05,040
Speaker 1: buy a depreciating asset. So we've been able to get

573
00:29:05,080 --> 00:29:08,520
Speaker 1: pretty creative, and some of that's been from the crypto side,

574
00:29:08,920 --> 00:29:11,600
Speaker 1: some of it's been kind of from from just founder

575
00:29:11,600 --> 00:29:14,760
Speaker 1: and partner commitment to fund things internally, what crypto does

576
00:29:14,880 --> 00:29:16,880
Speaker 1: for us, Like, say that our clusters at a steady

577
00:29:16,920 --> 00:29:21,120
Speaker 1: state or steady state and we've got let's call it GPS,

578
00:29:21,120 --> 00:29:24,480
Speaker 1: and that cluster are available for clients if we only

579
00:29:24,480 --> 00:29:28,520
Speaker 1: have like three thousand GPUs of baseload demand in that cluster, Yeah, right,

580
00:29:28,640 --> 00:29:32,200
Speaker 1: what are we doing with the other seventeen? And if

581
00:29:32,320 --> 00:29:34,280
Speaker 1: I don't have to talk to a client, and I

582
00:29:34,360 --> 00:29:36,520
Speaker 1: don't have to do anything besides start a minor on

583
00:29:36,560 --> 00:29:39,720
Speaker 1: those things, and I can make six of the revenue

584
00:29:39,840 --> 00:29:42,480
Speaker 1: to make from my cloud clients, like that seems pretty good.

585
00:29:43,280 --> 00:29:46,520
Speaker 1: So we we mind on everything everything possible when we

586
00:29:46,560 --> 00:29:48,760
Speaker 1: don't have higher end workload running. So this is this

587
00:29:48,840 --> 00:29:51,880
Speaker 1: is actually this part is super fascinating to me. So

588
00:29:52,600 --> 00:29:55,200
Speaker 1: if you want to have the ability to offer your

589
00:29:55,240 --> 00:30:00,680
Speaker 1: clients burst computers, first compute. Then by definition at any

590
00:30:00,720 --> 00:30:05,200
Speaker 1: given time you have to have significantly more capacity then

591
00:30:05,400 --> 00:30:09,200
Speaker 1: is you're sort of like normal baseload. And the way

592
00:30:09,280 --> 00:30:12,960
Speaker 1: then from your perspective, to make the investment in that

593
00:30:13,040 --> 00:30:16,560
Speaker 1: capacity worthwhile so that you can go out and buy

594
00:30:16,680 --> 00:30:19,600
Speaker 1: chips that you only expect to offer to your clients

595
00:30:19,640 --> 00:30:25,200
Speaker 1: that maybe you know normal state capacity to utilization is

596
00:30:25,280 --> 00:30:28,880
Speaker 1: if during those times the other seventy percent are earning

597
00:30:28,960 --> 00:30:32,479
Speaker 1: revenue for you from cryptocurrency mining. Yeah, you know, I

598
00:30:32,520 --> 00:30:35,480
Speaker 1: think that the payback calculations is a little bit different. Sure,

599
00:30:36,200 --> 00:30:38,760
Speaker 1: that's se with crypto mining is what allows me to

600
00:30:38,760 --> 00:30:44,320
Speaker 1: aggressively expand my business that pays the bills, m right.

601
00:30:44,360 --> 00:30:47,000
Speaker 1: It allows me to kind of operate. But the grand

602
00:30:47,000 --> 00:30:50,200
Speaker 1: plans over here is really to kind of just continuously

603
00:30:50,240 --> 00:30:53,840
Speaker 1: grow this this compute fleet. You know, we have customers

604
00:30:53,880 --> 00:30:55,480
Speaker 1: that come in all the time that are I can't

605
00:30:55,480 --> 00:30:57,080
Speaker 1: get what I need, I can't get this, I can't

606
00:30:57,080 --> 00:30:59,240
Speaker 1: get that, like a can you promise me that I

607
00:30:59,240 --> 00:31:00,920
Speaker 1: can get a hunter to the GPUs? And like dude,

608
00:31:00,920 --> 00:31:02,640
Speaker 1: it's fine, I don't like whatever. You don't have to

609
00:31:02,640 --> 00:31:05,120
Speaker 1: talk to me. So another way to think about it

610
00:31:05,160 --> 00:31:07,400
Speaker 1: is like if all these people are playing the game

611
00:31:08,120 --> 00:31:13,000
Speaker 1: of buying all different kinds of chips. To mind your

612
00:31:13,080 --> 00:31:17,520
Speaker 1: essentially forced to play the game too. Yeah, you know what,

613
00:31:17,640 --> 00:31:21,600
Speaker 1: we try to operate in different segments, right, So typically

614
00:31:21,640 --> 00:31:23,840
Speaker 1: when mining like etherory and mining rings are being built,

615
00:31:23,840 --> 00:31:26,960
Speaker 1: they're being built for the lowest cost possible. Right, So

616
00:31:27,320 --> 00:31:30,640
Speaker 1: the GPUs are really the only expensive thing in the setup,

617
00:31:31,200 --> 00:31:33,600
Speaker 1: like the host of ICE may have, whether it's somebody

618
00:31:33,640 --> 00:31:36,440
Speaker 1: building it out of like an egg carton or an

619
00:31:36,480 --> 00:31:39,960
Speaker 1: like an eggshell burton from in their garage, or putting

620
00:31:39,960 --> 00:31:41,960
Speaker 1: it on an open air rack, or putting it on

621
00:31:42,000 --> 00:31:45,280
Speaker 1: a wood frame, or actually putting in like in a

622
00:31:45,320 --> 00:31:48,880
Speaker 1: closed chassis with forced airflow. The motherboard there is gonna

623
00:31:48,880 --> 00:31:51,680
Speaker 1: be super cheap. It's gonna have a really like low

624
00:31:51,800 --> 00:31:55,480
Speaker 1: end CPU. Um, it's gonna have like four gig a

625
00:31:55,560 --> 00:31:58,040
Speaker 1: RAM that these aren't the types of systems that we're

626
00:31:58,040 --> 00:32:01,320
Speaker 1: I'm competing to build, right, which is why it's without

627
00:32:01,440 --> 00:32:03,600
Speaker 1: the last five months here, like it's always been kind

628
00:32:03,600 --> 00:32:06,080
Speaker 1: of reasonable, right. I never had to compete for high

629
00:32:06,120 --> 00:32:08,720
Speaker 1: end CPUs before. I never had to compete for server

630
00:32:08,800 --> 00:32:11,040
Speaker 1: chastis before, never had to compete for like network cards

631
00:32:11,080 --> 00:32:13,960
Speaker 1: and stuff. And this is where file coin really throws

632
00:32:14,000 --> 00:32:17,320
Speaker 1: things into the into a loop. But on the mining side,

633
00:32:17,320 --> 00:32:21,520
Speaker 1: you know, everybody talks about how miners and gamers are

634
00:32:21,560 --> 00:32:24,160
Speaker 1: opposed to each other, and you know, I think there's

635
00:32:24,200 --> 00:32:27,000
Speaker 1: been some vilification of some of the larger mining operations

636
00:32:27,480 --> 00:32:29,600
Speaker 1: around the world that they're taking all these GPUs, And

637
00:32:30,440 --> 00:32:32,960
Speaker 1: I'll tell you that even on our mining stuff, when

638
00:32:32,960 --> 00:32:35,040
Speaker 1: the nvideo thirty eight came out, we looked at it

639
00:32:35,080 --> 00:32:36,160
Speaker 1: and was like, oh my god, this thing is a

640
00:32:36,200 --> 00:32:38,680
Speaker 1: three slot, Like we can't like, we can't deploy that.

641
00:32:38,880 --> 00:32:41,640
Speaker 1: What does that mean? A three slot? So it takes

642
00:32:41,720 --> 00:32:43,760
Speaker 1: up three slots on the motherboard, right, So it's just

643
00:32:44,040 --> 00:32:47,280
Speaker 1: a different spacing than what they had done previously and

644
00:32:47,320 --> 00:32:50,400
Speaker 1: typically kind of the mining GPUs and mining variants they

645
00:32:50,400 --> 00:32:53,160
Speaker 1: had released in the past, where we're dual slots, so

646
00:32:53,240 --> 00:32:55,360
Speaker 1: we take up two slots on the motherboard, then what

647
00:32:55,560 --> 00:32:57,400
Speaker 1: these crazy things that were three and they're like, okay,

648
00:32:57,440 --> 00:33:00,520
Speaker 1: like we can't deploy that a scale and our focus

649
00:33:00,560 --> 00:33:02,120
Speaker 1: at the time was just so different that we didn't

650
00:33:02,120 --> 00:33:04,600
Speaker 1: even buy any There are some distributors in the US

651
00:33:04,640 --> 00:33:07,800
Speaker 1: that show like real time availability and real time and

652
00:33:07,960 --> 00:33:11,400
Speaker 1: how many back orders they have in their queue and

653
00:33:11,480 --> 00:33:13,960
Speaker 1: you see them and they have like we expect five

654
00:33:14,720 --> 00:33:18,520
Speaker 1: video RTX on September one, and like pre orders, like

655
00:33:20,640 --> 00:33:23,320
Speaker 1: I'll tell you, like the forty seven thousand pre orders.

656
00:33:23,360 --> 00:33:25,920
Speaker 1: I don't think that's sophisticated mining operations. I think that's

657
00:33:25,960 --> 00:33:28,360
Speaker 1: like like you and me personally going and doing and

658
00:33:28,360 --> 00:33:31,680
Speaker 1: putting in our garage. Where do gamers fit into the mix?

659
00:33:31,760 --> 00:33:33,760
Speaker 1: Because I mean like a discord group and I don't

660
00:33:33,760 --> 00:33:36,840
Speaker 1: really understand half the jargon in it, but it's basically

661
00:33:36,840 --> 00:33:41,560
Speaker 1: a group for gamers who are looking to acquire in

662
00:33:41,680 --> 00:33:44,120
Speaker 1: video GPUs because they just want state of the art

663
00:33:44,160 --> 00:33:47,920
Speaker 1: graphics for their systems. How do they even stand a

664
00:33:48,040 --> 00:33:51,040
Speaker 1: chance in this world? And how does how does the

665
00:33:51,080 --> 00:33:55,000
Speaker 1: industry think about sort of making sure that they can

666
00:33:55,000 --> 00:33:58,440
Speaker 1: get access to this technology? There's two steps here that

667
00:33:58,440 --> 00:34:00,800
Speaker 1: there's two pieces here that are really a problem. Okay, yes,

668
00:34:00,840 --> 00:34:03,320
Speaker 1: the mining demand is a problem, but the mining demand

669
00:34:03,400 --> 00:34:06,680
Speaker 1: is a problem because the economics warrant it, right. And

670
00:34:06,920 --> 00:34:08,560
Speaker 1: you know, one of the crazy things about proof of

671
00:34:08,600 --> 00:34:11,480
Speaker 1: war cryptocurrencies like Bitcoin and I guess ethereum and it's

672
00:34:11,480 --> 00:34:14,640
Speaker 1: in its current life, is that the early on mining

673
00:34:15,080 --> 00:34:17,960
Speaker 1: component of it is like the greatest user acquisition strategy

674
00:34:17,960 --> 00:34:20,560
Speaker 1: in the history of the world. Right, It's like, hey,

675
00:34:20,760 --> 00:34:23,200
Speaker 1: you have a computer. Your computer may already have this device.

676
00:34:23,400 --> 00:34:24,799
Speaker 1: If you turn it on, we're gonna pay you ten

677
00:34:24,840 --> 00:34:27,960
Speaker 1: dollars a day, right, And that ten dollars is now

678
00:34:28,040 --> 00:34:30,680
Speaker 1: denominated in ethereum. And now you say, Okay, I've got

679
00:34:30,680 --> 00:34:33,160
Speaker 1: all this ethereum, what do I do with it? And

680
00:34:33,239 --> 00:34:35,759
Speaker 1: now you as the user, are now researching everything in

681
00:34:35,800 --> 00:34:39,040
Speaker 1: ethereum because you now have this financial asset. The hundreds

682
00:34:39,040 --> 00:34:41,080
Speaker 1: and hundreds of thousands of people that have gotten into

683
00:34:41,120 --> 00:34:44,640
Speaker 1: cryptocurrency from mining, it's it's like, it's crazy, and I

684
00:34:44,640 --> 00:34:46,440
Speaker 1: think it's one of the things that's like super underlooked

685
00:34:46,680 --> 00:34:50,040
Speaker 1: or super overlooked. Excuse me. The mining economics definitely drive

686
00:34:50,080 --> 00:34:52,800
Speaker 1: a lot of this, right, and and videos taking steps

687
00:34:52,840 --> 00:34:56,000
Speaker 1: to put their hash rate limitters on and to to

688
00:34:56,120 --> 00:34:58,360
Speaker 1: kind of enforce that in a certain in such a

689
00:34:58,400 --> 00:35:01,799
Speaker 1: way for the ethereum mining to say, anybody who's going

690
00:35:01,840 --> 00:35:04,600
Speaker 1: to buy this stuff in real size, like by the

691
00:35:04,600 --> 00:35:07,880
Speaker 1: crypto mining focus g peeps. To me, the bigger problem

692
00:35:07,920 --> 00:35:10,280
Speaker 1: is really the scalpers. It's the guys who are running

693
00:35:10,280 --> 00:35:12,839
Speaker 1: bots on Amazon and they're buying the GPUs as soon

694
00:35:12,880 --> 00:35:15,680
Speaker 1: as they list, and they're selling them money bay, right,

695
00:35:15,800 --> 00:35:17,799
Speaker 1: is the supply isn't making it to the end user,

696
00:35:17,920 --> 00:35:20,640
Speaker 1: and it's I think that that supply is probably artificially

697
00:35:20,640 --> 00:35:23,520
Speaker 1: over priced right now. And you know, the crazy thing

698
00:35:23,560 --> 00:35:25,520
Speaker 1: is that the ethere of mining economics kind of warranted.

699
00:35:25,560 --> 00:35:28,520
Speaker 1: So maybe it's not artificial. But but for the gamers,

700
00:35:28,560 --> 00:35:30,400
Speaker 1: they're going, what the heck? Like this thing is supposed

701
00:35:30,400 --> 00:35:33,640
Speaker 1: to be six hundred bucks and it's twenty bucks between

702
00:35:34,040 --> 00:35:36,719
Speaker 1: like retail and that the gamer is actually getting it.

703
00:35:36,719 --> 00:35:38,960
Speaker 1: It's almost like there's there's two huge obstacles for them.

704
00:35:39,400 --> 00:35:41,719
Speaker 1: I don't see a way to solve that. Well. I

705
00:35:41,719 --> 00:35:45,799
Speaker 1: guess if the cryptocurrency, I mean the crypto market crashes

706
00:35:45,840 --> 00:35:49,239
Speaker 1: every few years or but booms every few years. Hey

707
00:35:49,239 --> 00:35:51,960
Speaker 1: it's different this time, man, Come on. But in theory

708
00:35:52,040 --> 00:35:56,319
Speaker 1: it could crash. From your view, would that ease up

709
00:35:56,400 --> 00:35:59,879
Speaker 1: capacity for all these things that we're talking about? Probably not?

710
00:36:00,480 --> 00:36:02,960
Speaker 1: How come right? I think I think what you would

711
00:36:03,000 --> 00:36:05,879
Speaker 1: see is you'd see a lot of GPUs that come

712
00:36:05,880 --> 00:36:10,359
Speaker 1: back to market. Is kind of like the stressed assets, okay, right,

713
00:36:10,440 --> 00:36:12,560
Speaker 1: But I don't think that those are going to go

714
00:36:12,560 --> 00:36:14,799
Speaker 1: in large blocks. They're not gonna want to be piecemealed

715
00:36:14,800 --> 00:36:16,759
Speaker 1: out to the end users. And you know there's some

716
00:36:16,840 --> 00:36:18,799
Speaker 1: services and some companies in the US that actually do

717
00:36:18,840 --> 00:36:22,000
Speaker 1: specifically that as they buy, like use crypto minds and

718
00:36:22,040 --> 00:36:25,520
Speaker 1: they piecemeal them out. I guess one gp lots on eBay. Yeah,

719
00:36:25,920 --> 00:36:29,040
Speaker 1: But what I meant aside the selling of views, I mean,

720
00:36:29,160 --> 00:36:32,720
Speaker 1: there are new new chips of all these different stripes

721
00:36:32,760 --> 00:36:37,000
Speaker 1: are always being manufactured. So theoretically, like in Vidia and

722
00:36:37,600 --> 00:36:40,640
Speaker 1: dozens of other companies that make hard drives and so forth,

723
00:36:43,600 --> 00:36:46,439
Speaker 1: if let's say we're in a crypto bear market, does

724
00:36:46,520 --> 00:36:50,560
Speaker 1: that free up capacity or does that free up don't?

725
00:36:50,760 --> 00:36:52,600
Speaker 1: I don't think so. I think this is this is

726
00:36:52,600 --> 00:36:54,320
Speaker 1: where you have to look at the structural change that

727
00:36:54,360 --> 00:36:57,480
Speaker 1: we're seeing in the compute market. Everything is more compute intensive,

728
00:36:57,920 --> 00:37:01,040
Speaker 1: everyone's going to the cloud data center. Demand for GPUs

729
00:37:01,120 --> 00:37:04,640
Speaker 1: just exploding, Okay, even if crypto were to disappear. Yeah,

730
00:37:04,719 --> 00:37:06,880
Speaker 1: Like the demand the growth pattern of this of this

731
00:37:06,960 --> 00:37:09,000
Speaker 1: market as well. It's kind of home gaming and everything.

732
00:37:09,000 --> 00:37:12,120
Speaker 1: Like the global foundry space for STEMI conductors is just

733
00:37:12,400 --> 00:37:16,200
Speaker 1: too small for all the industries that are relying are

734
00:37:16,239 --> 00:37:19,120
Speaker 1: reliant upon it. For it's going to have to cancel

735
00:37:19,200 --> 00:37:20,960
Speaker 1: how much like not deliver how many trucks this year

736
00:37:21,000 --> 00:37:25,000
Speaker 1: because the semi conductor shortages like that, like wild and

737
00:37:25,040 --> 00:37:27,400
Speaker 1: it's impacting everything. And when when you get kind of

738
00:37:27,400 --> 00:37:31,120
Speaker 1: in that structural dislocation like this, yeah, like the whole

739
00:37:31,200 --> 00:37:33,480
Speaker 1: like saying in commodity trading is like high high prices,

740
00:37:33,480 --> 00:37:36,040
Speaker 1: cure high prices. Yeah, the timeline to cure these high

741
00:37:36,040 --> 00:37:41,520
Speaker 1: prices maybe several liters. Interesting how many players? How many

742
00:37:41,600 --> 00:37:46,600
Speaker 1: people are essentially in this game of I mean it's interesting,

743
00:37:46,640 --> 00:37:49,160
Speaker 1: like I can go on stock x, like the sneaker

744
00:37:49,200 --> 00:37:51,800
Speaker 1: website and they actually have a few categories of chips

745
00:37:51,800 --> 00:37:54,359
Speaker 1: and you can buy buy in video chips on there

746
00:37:55,000 --> 00:37:58,799
Speaker 1: and they sell like above M S r P and

747
00:37:58,840 --> 00:38:01,520
Speaker 1: you can see the price chart. How many different players,

748
00:38:01,520 --> 00:38:05,239
Speaker 1: whether they're sort of large brokers or individual scalpers, have

749
00:38:05,480 --> 00:38:10,080
Speaker 1: essentially like gotten in this game of chip price arbitrage

750
00:38:10,800 --> 00:38:13,279
Speaker 1: and trying to get themselves in the middle between in

751
00:38:13,320 --> 00:38:16,640
Speaker 1: the middle between the chip manufacturers or the chip distributors

752
00:38:17,360 --> 00:38:21,120
Speaker 1: or somewhere between Amazon and eBay and gotten in this

753
00:38:21,200 --> 00:38:24,480
Speaker 1: game of sort of like chip reselling as if it's

754
00:38:24,520 --> 00:38:28,560
Speaker 1: a commodity or a stock. Yeah, you know, it's not

755
00:38:28,640 --> 00:38:31,080
Speaker 1: just the it's not just the scalpers soo either. What's

756
00:38:31,120 --> 00:38:33,920
Speaker 1: unique about the GP supply chain within video in particulars

757
00:38:33,960 --> 00:38:36,839
Speaker 1: and video that they make the GP chips and they

758
00:38:36,840 --> 00:38:38,799
Speaker 1: actually they sell them too. I guess they call AD

759
00:38:38,920 --> 00:38:41,719
Speaker 1: and board partners, right, And this is people like it's

760
00:38:41,760 --> 00:38:44,239
Speaker 1: kicka Bite, it's m s I, it's e v g A,

761
00:38:44,640 --> 00:38:47,799
Speaker 1: it's Zotec, which is like a PC partner's brand, And

762
00:38:47,840 --> 00:38:50,839
Speaker 1: they actually take those boards, they fabricate them, they put

763
00:38:50,840 --> 00:38:52,680
Speaker 1: the cooling devices on them, and they sell them to

764
00:38:52,760 --> 00:38:56,160
Speaker 1: end users. And you know, I think that the ultimate

765
00:38:56,200 --> 00:38:59,359
Speaker 1: pricing on that stuff is really controlled by the ai

766
00:38:59,440 --> 00:39:02,520
Speaker 1: B s what's a I B stand for? Again, the

767
00:39:02,560 --> 00:39:05,680
Speaker 1: added board partner, Right, So this is like the kind

768
00:39:05,680 --> 00:39:08,200
Speaker 1: of fabricator or distributor at the end of the line there,

769
00:39:08,760 --> 00:39:11,680
Speaker 1: and there're no dummies, right. They see crypto demand, they

770
00:39:11,719 --> 00:39:15,399
Speaker 1: see where it is, and price definitely changes with where

771
00:39:15,440 --> 00:39:18,279
Speaker 1: crypto is. Right, is that they know they have a

772
00:39:18,320 --> 00:39:22,879
Speaker 1: limited supply. It's virtually unlimited demand. So it's like this

773
00:39:22,920 --> 00:39:25,799
Speaker 1: is really just supply and demand, like in its most

774
00:39:25,880 --> 00:39:28,239
Speaker 1: less a fair state. Yeah, and you know, I think

775
00:39:28,239 --> 00:39:29,640
Speaker 1: that one of the things that the US had, Like

776
00:39:29,719 --> 00:39:31,120
Speaker 1: one of the problems in the West here is that

777
00:39:31,480 --> 00:39:33,080
Speaker 1: I don't think a lot of the stuff even gets

778
00:39:33,120 --> 00:39:36,880
Speaker 1: out of China. What do you mean by that? Oh,

779
00:39:37,280 --> 00:39:39,759
Speaker 1: if you know a guy at the AI B and

780
00:39:39,800 --> 00:39:43,960
Speaker 1: shen Jen and your local and you want to come

781
00:39:44,000 --> 00:39:46,640
Speaker 1: in and pick up a thousand GPUs um, that's a

782
00:39:46,719 --> 00:39:50,360
Speaker 1: much easier transaction. And like in terms of like political capital,

783
00:39:50,400 --> 00:39:52,080
Speaker 1: just on a personal basis, like I think it's a

784
00:39:52,080 --> 00:39:53,480
Speaker 1: lot easier for them to do than it is for

785
00:39:53,520 --> 00:39:57,880
Speaker 1: them to shift something to Brian Vinturo in the United States.

786
00:40:00,080 --> 00:40:16,879
Speaker 1: M ramping up here a little bit. You know, you've

787
00:40:16,920 --> 00:40:20,319
Speaker 1: explained how your life kind of has changed in the

788
00:40:20,360 --> 00:40:24,880
Speaker 1: last six months in terms of the difficulty of acquiring compute,

789
00:40:24,880 --> 00:40:28,799
Speaker 1: whether it's GPUs, CPUs, RAM and so forth. What is

790
00:40:28,840 --> 00:40:30,840
Speaker 1: it like now on a day to day basis, And

791
00:40:30,880 --> 00:40:34,719
Speaker 1: how have you had to change your business strategy so

792
00:40:34,760 --> 00:40:40,680
Speaker 1: that you can grow your cloud capacity alongside customer demand. Yeah,

793
00:40:40,760 --> 00:40:42,600
Speaker 1: you know that it used to be that GPUs were

794
00:40:42,640 --> 00:40:45,240
Speaker 1: the problem for us to get okay um and GPUs

795
00:40:45,360 --> 00:40:47,399
Speaker 1: right now, like the data center cards, Like I've got

796
00:40:47,800 --> 00:40:50,359
Speaker 1: shelves and shelves and shelves packed with data center cards

797
00:40:50,360 --> 00:40:53,279
Speaker 1: waiting for builds and they're all blocked by CPUs and

798
00:40:53,360 --> 00:40:57,160
Speaker 1: ram right CPUs in particular, I had a conversation with

799
00:40:57,200 --> 00:40:59,719
Speaker 1: my CEO I've been I've been with for seventeen years

800
00:40:59,719 --> 00:41:01,560
Speaker 1: now and in kind of different businesses, and I called

801
00:41:01,560 --> 00:41:04,080
Speaker 1: on I said, hey, I just found a bunch of

802
00:41:04,120 --> 00:41:07,840
Speaker 1: amd Epic Milan CPUs in stock ready for ship tomorrow.

803
00:41:08,360 --> 00:41:09,640
Speaker 1: I bought a hundred of him and he was like,

804
00:41:09,760 --> 00:41:10,959
Speaker 1: how many they have? And I was like, I don't

805
00:41:10,960 --> 00:41:13,319
Speaker 1: know two. He's like, why didn't you buy the rest?

806
00:41:13,719 --> 00:41:15,560
Speaker 1: And I'm like, well, I didn't really want to spend

807
00:41:15,600 --> 00:41:17,040
Speaker 1: the money, like I didn't have a place to put

808
00:41:17,080 --> 00:41:19,799
Speaker 1: him right now. He's like, dude, this is literally all

809
00:41:19,840 --> 00:41:22,320
Speaker 1: you complain about every day is not getting access to stuff.

810
00:41:22,560 --> 00:41:24,839
Speaker 1: Go back and buy the rest. So I went back

811
00:41:24,880 --> 00:41:27,200
Speaker 1: to buy the rest, and in the fifteen minute period,

812
00:41:27,239 --> 00:41:30,200
Speaker 1: the rest are going. So you know, this is a theme.

813
00:41:30,239 --> 00:41:32,319
Speaker 1: It's not just chips, obviously, and we talked about this

814
00:41:32,360 --> 00:41:36,439
Speaker 1: with lumber, we talked about this with industrial commodities. This

815
00:41:36,560 --> 00:41:40,120
Speaker 1: idea that you buy more when there's a sort of

816
00:41:40,160 --> 00:41:43,279
Speaker 1: sense of scarcity. And so I'm curious about, like how

817
00:41:43,360 --> 00:41:47,480
Speaker 1: much of the tightness of the chip market is essentially

818
00:41:47,880 --> 00:41:50,360
Speaker 1: it's like, well, I don't know where the next shipment

819
00:41:50,480 --> 00:41:52,040
Speaker 1: is going to be available. I don't know what I'm

820
00:41:52,040 --> 00:41:54,640
Speaker 1: gonna be able to buy these m D chips and

821
00:41:54,719 --> 00:41:56,799
Speaker 1: the size. I don't know when whether I'm next time

822
00:41:56,840 --> 00:41:59,399
Speaker 1: I'm going to buy RAM and size. So how much

823
00:41:59,440 --> 00:42:02,000
Speaker 1: does that have the effect of causing people to then

824
00:42:02,239 --> 00:42:05,080
Speaker 1: over order so that they can at least give get

825
00:42:05,120 --> 00:42:08,680
Speaker 1: some assurance of the chips that they're going to need

826
00:42:09,560 --> 00:42:12,680
Speaker 1: and then for which were then further exacerbates the shortage

827
00:42:12,719 --> 00:42:15,960
Speaker 1: causing more over ordering. I can answer for us, like

828
00:42:16,000 --> 00:42:19,000
Speaker 1: I mentioned earlier, that we're pretty just in time operation okay,

829
00:42:19,280 --> 00:42:21,480
Speaker 1: on the inventory side, So when I'm talking about if

830
00:42:21,480 --> 00:42:23,759
Speaker 1: I go out and by like two CPUs, like that

831
00:42:23,800 --> 00:42:27,200
Speaker 1: gets me like one month of of supply, okay, you

832
00:42:27,239 --> 00:42:29,960
Speaker 1: know the CPUs that I did buy that were delivered

833
00:42:30,000 --> 00:42:32,720
Speaker 1: the next day, like which was even more of a surprise,

834
00:42:33,320 --> 00:42:35,319
Speaker 1: Like they're already in systems, are already online that this

835
00:42:35,360 --> 00:42:37,400
Speaker 1: is a couple of weeks ago. Where when I'm talking

836
00:42:37,400 --> 00:42:39,920
Speaker 1: about buying things in stockpiling them, like my my usage

837
00:42:39,920 --> 00:42:42,759
Speaker 1: time frame is much that there's no like there's no

838
00:42:42,840 --> 00:42:45,960
Speaker 1: speculation on component values here, and it's like, hey, we're

839
00:42:45,960 --> 00:42:47,560
Speaker 1: gonna need him in the next thirty days. I typically

840
00:42:47,560 --> 00:42:49,320
Speaker 1: wouldn't buy it until I needed in the next seven

841
00:42:49,680 --> 00:42:52,439
Speaker 1: but here we are. The world was on fire six

842
00:42:52,520 --> 00:42:55,280
Speaker 1: months ago. You might not have made an order until

843
00:42:55,360 --> 00:42:57,960
Speaker 1: seven days before you needed to build them out, but

844
00:42:58,080 --> 00:43:01,160
Speaker 1: that is now lengthened out to more close to thirty days. Yeah.

845
00:43:01,239 --> 00:43:03,040
Speaker 1: And on the O E M side, so we've got

846
00:43:03,080 --> 00:43:05,400
Speaker 1: a bunch of stuff, that stuff that we're waiting on

847
00:43:05,480 --> 00:43:11,440
Speaker 1: deliveries for right now. We've probably like high tens, maybe

848
00:43:11,200 --> 00:43:14,839
Speaker 1: fifty million capex that that's out there kind of being fabricated. Yeah,

849
00:43:15,360 --> 00:43:17,759
Speaker 1: and this is what O E M s. And you know,

850
00:43:17,840 --> 00:43:19,719
Speaker 1: when we were doing the deals, it was, hey, we're

851
00:43:19,719 --> 00:43:22,359
Speaker 1: committing to do this within four weeks, and now it's

852
00:43:22,400 --> 00:43:24,839
Speaker 1: like out the twelve weeks. And and you go through

853
00:43:24,840 --> 00:43:26,759
Speaker 1: the conversation with them and you're like, all right, like

854
00:43:26,800 --> 00:43:28,640
Speaker 1: I get it, Like I know what's going on to

855
00:43:28,760 --> 00:43:32,240
Speaker 1: you taking me to like I get the understand problem.

856
00:43:32,360 --> 00:43:35,280
Speaker 1: All ability to plan in real time is like completely

857
00:43:35,280 --> 00:43:39,200
Speaker 1: out the window. Huh wow. Yeah, that's the thing. I

858
00:43:39,200 --> 00:43:42,080
Speaker 1: mean that seems to be the theme of so many

859
00:43:42,120 --> 00:43:46,320
Speaker 1: different industries, whether we're talking about shipping, logistics, trucking, lumber,

860
00:43:46,440 --> 00:43:52,520
Speaker 1: and so forth, everybody planning ability has kind of gotten obliterated. Yeah, yeah,

861
00:43:52,560 --> 00:43:54,480
Speaker 1: I mean the shipping thing is crazy too, because like

862
00:43:54,520 --> 00:43:56,719
Speaker 1: stuff that I would you like normally put on a

863
00:43:56,719 --> 00:43:58,840
Speaker 1: boat and got three weeks four, I'm like, put it

864
00:43:58,880 --> 00:44:00,640
Speaker 1: on an airplane. In the boat may ink, like may

865
00:44:00,719 --> 00:44:02,400
Speaker 1: get stuck. I have no idea what's gonna happen, Like,

866
00:44:02,440 --> 00:44:04,520
Speaker 1: get me my stuff. I just want to go back

867
00:44:04,520 --> 00:44:06,960
Speaker 1: to one question because I still feel felt a little

868
00:44:06,960 --> 00:44:11,960
Speaker 1: bit unsatisfied by the answer. Let's say cryptocurrency disappeared. I

869
00:44:11,960 --> 00:44:15,040
Speaker 1: don't know how that would happen. Maybe the prices crash,

870
00:44:15,680 --> 00:44:18,680
Speaker 1: Maybe somehow there's a coordinated national effort to ban it,

871
00:44:18,880 --> 00:44:21,400
Speaker 1: or maybe financial institutions are banned from dealing with it.

872
00:44:21,920 --> 00:44:25,040
Speaker 1: Somehow that industry, which is extremely big and growing and

873
00:44:25,040 --> 00:44:27,640
Speaker 1: worth like two trillion dollars in market cap, and just

874
00:44:27,680 --> 00:44:30,600
Speaker 1: say somehow it gets obliterated. Well, we can we can

875
00:44:30,640 --> 00:44:34,040
Speaker 1: put a real scenario in this, like actually ship proof

876
00:44:34,040 --> 00:44:37,279
Speaker 1: of steak, right right, Okay, right, So I mean that's

877
00:44:37,320 --> 00:44:39,759
Speaker 1: what like, that's the biggest question that I get from

878
00:44:39,760 --> 00:44:42,640
Speaker 1: investors is what about proof of steak. Yeah, it's a

879
00:44:42,640 --> 00:44:45,720
Speaker 1: proof of steak for people who don't know it. Basically, currently,

880
00:44:46,480 --> 00:44:50,800
Speaker 1: ethereum operates with the same security model as bitcoin mining,

881
00:44:50,840 --> 00:44:53,279
Speaker 1: more or less, with a sort of high chip and

882
00:44:53,680 --> 00:44:58,240
Speaker 1: electrical spend in order to secure the network and acquire coins.

883
00:44:58,719 --> 00:45:01,000
Speaker 1: They want to change the mall to something where there's

884
00:45:01,040 --> 00:45:03,799
Speaker 1: less mining. Theoretically they'd be greener and so forth. But

885
00:45:04,120 --> 00:45:06,759
Speaker 1: dr tell me what how you see proof of steake

886
00:45:06,760 --> 00:45:09,239
Speaker 1: affecting things. I'm not going to talk about the probability

887
00:45:09,239 --> 00:45:12,279
Speaker 1: of that happening, right, I mean they've shown like, well,

888
00:45:12,280 --> 00:45:15,000
Speaker 1: I'm gonna give one comment on it, um, you know,

889
00:45:15,160 --> 00:45:17,960
Speaker 1: starting in two thousand sixteen, proof of steake with shipping

890
00:45:18,040 --> 00:45:20,400
Speaker 1: every three months. Yeah. Um. And here we are in

891
00:45:20,800 --> 00:45:23,400
Speaker 1: the middle, And like, do I think they're eventually going

892
00:45:23,440 --> 00:45:26,239
Speaker 1: to get there? Yes, I do, But like, look at

893
00:45:26,280 --> 00:45:28,399
Speaker 1: the market, look at how much money has been spent

894
00:45:28,480 --> 00:45:32,640
Speaker 1: on mining hardware and building mining operations. Like the miners

895
00:45:32,640 --> 00:45:35,239
Speaker 1: are not going down without a fight, and without a fight,

896
00:45:35,280 --> 00:45:38,560
Speaker 1: could be launching new coins. It could be contentious forks,

897
00:45:38,680 --> 00:45:41,319
Speaker 1: it could be anything. But you know that this idea

898
00:45:41,360 --> 00:45:43,319
Speaker 1: of GPU mining has been is going to die. It's

899
00:45:43,360 --> 00:45:46,160
Speaker 1: been going on since the i'd been to GAP mining, Right,

900
00:45:46,239 --> 00:45:48,800
Speaker 1: Like when Bitcoin went from GPUs to f p g

901
00:45:49,000 --> 00:45:51,520
Speaker 1: a S, the GPU mining was dead, and then like

902
00:45:51,560 --> 00:45:55,840
Speaker 1: six months later lightcoin showed up, right, So there's always

903
00:45:55,840 --> 00:45:59,560
Speaker 1: going to be this this demand for security driven by

904
00:45:59,600 --> 00:46:02,359
Speaker 1: GPU can pute. I think it goes back to that

905
00:46:02,840 --> 00:46:05,640
Speaker 1: kind of like fair launches and the idea of getting

906
00:46:05,640 --> 00:46:08,239
Speaker 1: coins out to the quote unquote to the people. Yeah, right,

907
00:46:08,320 --> 00:46:10,759
Speaker 1: is that as as soon as Etherorium goes prof ofstake, like,

908
00:46:11,160 --> 00:46:15,200
Speaker 1: that's gonna be rewarding the incumbents more so than new entrance. Right,

909
00:46:15,200 --> 00:46:17,040
Speaker 1: You're not gonna have a hundred thousand people gonn plug

910
00:46:17,040 --> 00:46:20,200
Speaker 1: in a GPU and learn what etherium is. I just

911
00:46:20,239 --> 00:46:22,600
Speaker 1: think that the idea that when Ethereum goes pro the stake,

912
00:46:22,600 --> 00:46:25,400
Speaker 1: the GPU mining dies is a little bit crazy. We

913
00:46:25,440 --> 00:46:27,120
Speaker 1: don't necessarily know what it's going to be yet, but

914
00:46:27,160 --> 00:46:30,520
Speaker 1: I'm pretty confident that there's going to be something. Now.

915
00:46:30,640 --> 00:46:33,480
Speaker 1: Let's like in a price crash environment. You know, this

916
00:46:33,520 --> 00:46:35,880
Speaker 1: is really where we built our whole business, is taking

917
00:46:35,880 --> 00:46:39,360
Speaker 1: the contrarium bet in the binary outcome environment, which was

918
00:46:39,760 --> 00:46:43,960
Speaker 1: really like January, like December of two thousand eighteen. January right,

919
00:46:44,080 --> 00:46:46,440
Speaker 1: is like, I don't know, but Crypto is either going

920
00:46:46,480 --> 00:46:49,319
Speaker 1: to zero or it's not. Like I think that in

921
00:46:49,440 --> 00:46:51,719
Speaker 1: the mining space, like where you make real money is

922
00:46:51,760 --> 00:46:53,880
Speaker 1: when you're in a position to take that contrary and

923
00:46:53,960 --> 00:46:57,400
Speaker 1: distressed asset. What we look at that as a company

924
00:46:57,520 --> 00:47:00,160
Speaker 1: is like, yeah, it would really suck if Crypto went

925
00:47:00,200 --> 00:47:02,879
Speaker 1: to zero again, but look how awesome it worked last time.

926
00:47:03,040 --> 00:47:04,799
Speaker 1: Right then. I don't think we're the only ones that

927
00:47:04,840 --> 00:47:07,480
Speaker 1: think that way. But bottom line still, if I'm a

928
00:47:07,520 --> 00:47:12,920
Speaker 1: gamer and Crypto goes to ten crypto crash, does it

929
00:47:12,960 --> 00:47:16,719
Speaker 1: get easier for me to acquire GPUs so that I

930
00:47:16,760 --> 00:47:19,760
Speaker 1: could play a game higher quality? Well, it'll get easier

931
00:47:19,800 --> 00:47:22,520
Speaker 1: for you to acquire used GPUs play games, but I

932
00:47:22,560 --> 00:47:25,520
Speaker 1: don't necessarily know if you're gonna want them. Why why

933
00:47:25,560 --> 00:47:27,319
Speaker 1: not new? Why not? Why why not the new ones?

934
00:47:27,320 --> 00:47:29,120
Speaker 1: Why wouldn't it be easier for me to then go

935
00:47:29,160 --> 00:47:32,920
Speaker 1: on Amazon? Because I mean, and Videos one of the

936
00:47:32,920 --> 00:47:35,120
Speaker 1: most valuable and forward thinking companies in the world for

937
00:47:35,120 --> 00:47:38,200
Speaker 1: a reason. Yeah, right, that they understand market demand. Um,

938
00:47:38,239 --> 00:47:40,719
Speaker 1: and if they see crypto crash, like they know that

939
00:47:40,920 --> 00:47:42,960
Speaker 1: a portion of their demand is from Crypto. Right there,

940
00:47:43,000 --> 00:47:46,200
Speaker 1: launching your crypto specific lines. They're going to book less

941
00:47:46,200 --> 00:47:49,680
Speaker 1: foundary space, why not to shift some of that foundry

942
00:47:49,680 --> 00:47:53,080
Speaker 1: space over to gamers. Now, well, I mean, but before

943
00:47:53,120 --> 00:47:56,120
Speaker 1: we already like concluded that it's the same it's the

944
00:47:56,160 --> 00:47:59,839
Speaker 1: same chip sets, right, right, They're just being packaged differently, right. So,

945
00:48:00,040 --> 00:48:02,160
Speaker 1: And one of the things that people don't necessarily understand

946
00:48:02,200 --> 00:48:04,000
Speaker 1: is that you know in a lot of the cards,

947
00:48:04,000 --> 00:48:06,040
Speaker 1: whether it's like a thirty sixty or thirty sixty t

948
00:48:06,200 --> 00:48:09,120
Speaker 1: I or thirty seventy, that that may be the same

949
00:48:09,160 --> 00:48:13,080
Speaker 1: chip set with just features disabled, right or maybe bempacked

950
00:48:13,080 --> 00:48:15,759
Speaker 1: differently because it has like the quality of the run

951
00:48:15,840 --> 00:48:18,319
Speaker 1: was was worse for some portion of it. And video

952
00:48:18,400 --> 00:48:20,520
Speaker 1: is not making all these decisions like hey, let's go

953
00:48:20,600 --> 00:48:24,839
Speaker 1: build like fifty million thirty sixties. Right, They've got their

954
00:48:24,920 --> 00:48:27,200
Speaker 1: their kind of foundry space and what that may be booked,

955
00:48:27,200 --> 00:48:29,360
Speaker 1: and they know what their overall yield is going to

956
00:48:29,440 --> 00:48:31,840
Speaker 1: be um and then it's they're just kind of optimizing

957
00:48:31,880 --> 00:48:34,160
Speaker 1: what they bring to market, right. It's just in a

958
00:48:34,239 --> 00:48:37,160
Speaker 1: down crypto year. I think that what they did with

959
00:48:37,200 --> 00:48:40,399
Speaker 1: the Turing line where they're the r t X two

960
00:48:40,480 --> 00:48:43,439
Speaker 1: thousand series, I think that the production of that line

961
00:48:43,520 --> 00:48:46,799
Speaker 1: was was way lower than the g t X one

962
00:48:46,840 --> 00:48:49,719
Speaker 1: thousand series, right, and that was kind of coming out

963
00:48:49,719 --> 00:48:52,640
Speaker 1: of crypto winter, during crypto winter, and I don't think

964
00:48:52,640 --> 00:48:56,040
Speaker 1: those gps are hard to find. Got it right? But

965
00:48:56,080 --> 00:49:00,399
Speaker 1: you couldn't, but you also couldn't buy got it well,

966
00:49:00,719 --> 00:49:03,319
Speaker 1: This is a fascinating discussion. Like I said, we've talked

967
00:49:03,320 --> 00:49:07,360
Speaker 1: a lot about chips in the past, but we really

968
00:49:07,400 --> 00:49:11,200
Speaker 1: haven't gotten into the actual fight to acquire them. Plenty

969
00:49:11,239 --> 00:49:14,080
Speaker 1: more to dive into on this topic alone, I feel like,

970
00:49:14,120 --> 00:49:17,200
Speaker 1: but Brian really appreciate you coming out on odd lots.

971
00:49:17,719 --> 00:49:19,680
Speaker 1: Thanks for having me. I'd love to do it against

972
00:49:19,680 --> 00:49:35,920
Speaker 1: some time. Yeah, for sure. Take care Bran Well. I

973
00:49:35,960 --> 00:49:38,120
Speaker 1: would love to be able to talk about that episode

974
00:49:38,160 --> 00:49:40,560
Speaker 1: with Tracy. Unfortunately should not hear. But I thought that

975
00:49:40,680 --> 00:49:45,719
Speaker 1: was completely fascinating. You know, this idea so many parallels

976
00:49:45,760 --> 00:49:48,840
Speaker 1: between everything else we're saying, like the inability to plan,

977
00:49:49,760 --> 00:49:53,719
Speaker 1: the requirement to buy further out into the future, This

978
00:49:53,840 --> 00:49:56,840
Speaker 1: idea that sort of just in time isn't working, And

979
00:49:56,880 --> 00:50:00,400
Speaker 1: it really does seem like the rise of these new coins,

980
00:50:01,040 --> 00:50:06,680
Speaker 1: file coin chio, like every every sort of like commoditized

981
00:50:07,000 --> 00:50:11,200
Speaker 1: technological capacity that you could think of somebody. It feels

982
00:50:11,239 --> 00:50:14,480
Speaker 1: like it's making a coin that can mine it. It's fascinating,

983
00:50:14,480 --> 00:50:16,520
Speaker 1: and I thought it was fascinating. The core we've owned

984
00:50:16,719 --> 00:50:20,000
Speaker 1: business model is kind of enabled, Like, on one hand

985
00:50:20,000 --> 00:50:22,160
Speaker 1: they're competing with all these miners for chips, but on

986
00:50:22,200 --> 00:50:26,120
Speaker 1: the other hand, their ability to use idle chips in

987
00:50:26,280 --> 00:50:28,799
Speaker 1: order to mind currency and get revenue when they're not

988
00:50:28,840 --> 00:50:32,160
Speaker 1: being used for the normal UH computing by their clients

989
00:50:32,239 --> 00:50:35,760
Speaker 1: is super interesting. So I loved I loved everything about

990
00:50:35,800 --> 00:50:37,239
Speaker 1: that chat. I loved the sort of I love the

991
00:50:37,280 --> 00:50:40,640
Speaker 1: merging of the crypto and chip story, and I increasingly

992
00:50:40,680 --> 00:50:43,480
Speaker 1: think that you can't talk about one without the other.

993
00:50:43,560 --> 00:50:47,840
Speaker 1: So really appreciated Brian coming on and just making a

994
00:50:47,880 --> 00:50:51,040
Speaker 1: plug here. By the way, if you haven't checked it

995
00:50:51,040 --> 00:50:53,680
Speaker 1: out yet, you should definitely check out the odd Lots

996
00:50:53,760 --> 00:50:57,480
Speaker 1: blog bloomberg dot com slash odd Lots. Tracy and I

997
00:50:57,560 --> 00:51:01,960
Speaker 1: have been writing there, We've been listing transcripts there. I

998
00:51:02,239 --> 00:51:04,719
Speaker 1: myself have been going back and reading transcripts because when

999
00:51:04,719 --> 00:51:06,839
Speaker 1: I do, I learn a lot. Each time. I found

1000
00:51:06,880 --> 00:51:09,480
Speaker 1: it to be incredibly useful. Check it out. Go to

1001
00:51:09,480 --> 00:51:13,480
Speaker 1: Bloombergy dot com slash odd Lots and read our blog.

1002
00:51:13,560 --> 00:51:17,440
Speaker 1: So this has been another episode of the Odd Lots podcast.

1003
00:51:17,719 --> 00:51:20,560
Speaker 1: I'm Joe Wisntal. You can follow me on Twitter at

1004
00:51:20,600 --> 00:51:24,040
Speaker 1: the Stalwart, following my normal co host Tracy Alloway on

1005
00:51:24,080 --> 00:51:28,759
Speaker 1: Twitter at Tracy Alloway. Follow our producer Laura Carlson. She's

1006
00:51:28,880 --> 00:51:32,360
Speaker 1: at Laura M. Carlson. Follow the Bloomberg head of podcast,

1007
00:51:32,440 --> 00:51:36,239
Speaker 1: Francesco Levie at Francesca Today, and check out all of

1008
00:51:36,280 --> 00:51:40,040
Speaker 1: the Bloomberg podcasts on Twitter under the handle at podcasts.

1009
00:51:40,200 --> 00:52:08,400
Speaker 1: Thanks for listening, e