1
00:00:02,160 --> 00:00:07,280
Speaker 1: Zone Media burd Up. I'm at Zetron and this is

2
00:00:07,320 --> 00:00:18,360
Speaker 1: your weekly better offline monologue. So last week I put

3
00:00:18,360 --> 00:00:21,479
Speaker 1: out a newsletter about core Weave, an a cloud provider

4
00:00:21,520 --> 00:00:24,840
Speaker 1: that sells GPU compute to AI companies looking to run

5
00:00:24,960 --> 00:00:27,680
Speaker 1: or train their models, and they recently filed the paperwork

6
00:00:27,720 --> 00:00:31,680
Speaker 1: to go public. Now, the newsletter explained the shaky fundamentals

7
00:00:31,680 --> 00:00:34,519
Speaker 1: of the business, which is putting it mildly and ask

8
00:00:34,600 --> 00:00:37,960
Speaker 1: the question how the fuck does core Weave survive. This

9
00:00:38,080 --> 00:00:41,040
Speaker 1: is a company that's drowning in debt and relying on one

10
00:00:40,960 --> 00:00:43,599
Speaker 1: company for more than half of its revenue. And said company,

11
00:00:43,640 --> 00:00:46,600
Speaker 1: Microsoft is actively pulling away from building new data center

12
00:00:46,640 --> 00:00:49,680
Speaker 1: capacity and has reportedly pulled back from some contracts with

13
00:00:49,800 --> 00:00:53,440
Speaker 1: core Weave, though it's not clear what those contracts are. Now,

14
00:00:53,479 --> 00:00:56,080
Speaker 1: this is an important story and one that raises questions

15
00:00:56,120 --> 00:00:58,560
Speaker 1: not just about the viability of core Weave, but the

16
00:00:58,600 --> 00:01:02,360
Speaker 1: generative AI industry. At that said, the subject matter is

17
00:01:02,400 --> 00:01:04,880
Speaker 1: a little dense to turn into a hulking two part episode,

18
00:01:04,920 --> 00:01:07,480
Speaker 1: so I decided to condense it into a shorter monologue

19
00:01:07,520 --> 00:01:10,720
Speaker 1: for you. Now, Core We've had intended to go public

20
00:01:10,800 --> 00:01:12,920
Speaker 1: a week ago. They might go public and a week

21
00:01:12,959 --> 00:01:15,960
Speaker 1: from now, and they're apparently going with like a twenty

22
00:01:16,319 --> 00:01:19,840
Speaker 1: five to thirty billion dollar valuation. While it's hardly a

23
00:01:19,880 --> 00:01:23,800
Speaker 1: recognizable name like say open AI or Microsoft or in Vidia,

24
00:01:23,840 --> 00:01:26,479
Speaker 1: this company is definitely worth taking a look at, if

25
00:01:26,520 --> 00:01:28,680
Speaker 1: not for the fact that it's arguably the first major

26
00:01:28,760 --> 00:01:31,720
Speaker 1: IPO that we've seen from the current generative AI hYP bubble,

27
00:01:32,040 --> 00:01:36,440
Speaker 1: and it's undoubtedly the biggest. Moreover, corwave is a company

28
00:01:36,440 --> 00:01:39,360
Speaker 1: that deals in the infrastructure aspect of AI, where one

29
00:01:39,400 --> 00:01:41,960
Speaker 1: would naturally assume is where all the money really is

30
00:01:42,240 --> 00:01:45,200
Speaker 1: putting up service for hyperscals to run their hallucination pro

31
00:01:45,280 --> 00:01:48,800
Speaker 1: and nonprofitable models. If in Vidia, the company that makes

32
00:01:48,800 --> 00:01:51,360
Speaker 1: the GPUs for generative AI, is selling the pick axes

33
00:01:51,360 --> 00:01:53,920
Speaker 1: for the current gold rush, cor we've well, they own

34
00:01:53,960 --> 00:01:56,560
Speaker 1: the land, or maybe they're the shovels. I'm not going

35
00:01:56,600 --> 00:02:01,080
Speaker 1: to get too far into that metaphors side up. Until

36
00:02:01,120 --> 00:02:03,720
Speaker 1: a few years ago, Corweave was a different company, a

37
00:02:03,720 --> 00:02:08,080
Speaker 1: cryptocurrency mining data center company to be specific. Now twenty

38
00:02:08,120 --> 00:02:11,240
Speaker 1: tens were turbulent for crypto and the company then pivoted

39
00:02:11,280 --> 00:02:14,320
Speaker 1: to providing high performance computing for third parties that didn't

40
00:02:14,320 --> 00:02:16,639
Speaker 1: want to run their own infrastructure for things like three

41
00:02:16,720 --> 00:02:19,760
Speaker 1: D modeling. Around the time of the launch of chair GPT,

42
00:02:20,240 --> 00:02:22,920
Speaker 1: cor We've pivoted again, this time towards providing compute to

43
00:02:22,960 --> 00:02:26,800
Speaker 1: hyperscalers and then generative AI interest. It gambled, and it

44
00:02:27,000 --> 00:02:30,680
Speaker 1: gambled successfully. You'd imagine that such a company would be

45
00:02:30,720 --> 00:02:34,639
Speaker 1: a thriving, healthy business though, right they I mean they

46
00:02:34,760 --> 00:02:37,680
Speaker 1: sell the surfer architecture. The demand is there, the growth

47
00:02:37,760 --> 00:02:41,680
Speaker 1: is there. I mean everyone wants generative AI. Right, it

48
00:02:41,680 --> 00:02:46,840
Speaker 1: wouldn't be another thing where the reality Jesus Christ. Even

49
00:02:46,880 --> 00:02:51,040
Speaker 1: a cursory glance at Corwey's financial disclosure documents reveals a

50
00:02:51,080 --> 00:02:55,680
Speaker 1: business that's precarious at best, and in my most uncharitable opinion,

51
00:02:55,840 --> 00:02:59,880
Speaker 1: it's utterly rancid. If this company was in any other industry,

52
00:03:00,240 --> 00:03:03,480
Speaker 1: be seen as a big pile of shit, except it's

53
00:03:03,560 --> 00:03:05,919
Speaker 1: one of the standard bearers of the generative AI boom

54
00:03:05,960 --> 00:03:09,959
Speaker 1: and so exists within its own reality distortion field. Now

55
00:03:10,000 --> 00:03:12,880
Speaker 1: Callw've makes its money by renting out GPUs to companies

56
00:03:12,880 --> 00:03:15,200
Speaker 1: looking to run or train their models, which means that

57
00:03:15,240 --> 00:03:18,120
Speaker 1: it's incredibly indicative of whether there's a real business in

58
00:03:18,200 --> 00:03:20,840
Speaker 1: generative AI and you'll be shocked to hear that there

59
00:03:20,919 --> 00:03:23,919
Speaker 1: isn't call. We've made just under two billion dollars in

60
00:03:23,960 --> 00:03:26,560
Speaker 1: twenty twenty four, yet somehow managed to lose eight hundred

61
00:03:26,560 --> 00:03:30,519
Speaker 1: and sixty three million dollars. Worse still, seventy seven percent

62
00:03:30,520 --> 00:03:34,359
Speaker 1: of its revenue comes from two companies, Microsoft sixty two percent, and,

63
00:03:34,480 --> 00:03:38,240
Speaker 1: although not said explicitly, most likely in Video fifteen percent.

64
00:03:38,280 --> 00:03:41,520
Speaker 1: Now the information is reported Project Gosprey, which is basically

65
00:03:41,640 --> 00:03:44,920
Speaker 1: Nvidia's stake in the fact that they're finaling money and

66
00:03:44,960 --> 00:03:47,840
Speaker 1: it's actually not immediately obvious what it is that in

67
00:03:47,920 --> 00:03:51,040
Speaker 1: Video is paying Core week four, though Invidio is both

68
00:03:51,080 --> 00:03:53,520
Speaker 1: invested in the company and provides it with priority access

69
00:03:53,520 --> 00:03:56,200
Speaker 1: to its chips. With Core, we've getting some products before

70
00:03:56,240 --> 00:04:00,040
Speaker 1: even Amazon or Microsoft. It's bonkers. And again, yeah, I

71
00:04:00,120 --> 00:04:04,400
Speaker 1: think with all this preferential accents Core, we've just printing

72
00:04:04,440 --> 00:04:07,760
Speaker 1: money right versus burning it, burning it by the hundreds

73
00:04:07,800 --> 00:04:10,839
Speaker 1: of millions full. I wish I could get this much

74
00:04:10,920 --> 00:04:13,600
Speaker 1: money and just I don't know, I would be doing

75
00:04:13,600 --> 00:04:16,320
Speaker 1: a far weirder podcast. It would be exactly the same thing,

76
00:04:16,320 --> 00:04:18,880
Speaker 1: better offline would exist. I would just have hundreds of

77
00:04:18,880 --> 00:04:21,480
Speaker 1: millions of dollars we'd have just like Samuel L. Jackson

78
00:04:21,839 --> 00:04:25,480
Speaker 1: coming in for some episodes. Anyway, I don't have that money.

79
00:04:25,520 --> 00:04:27,279
Speaker 1: If you'd like to give it to me, please please

80
00:04:27,320 --> 00:04:30,200
Speaker 1: do now. A few weeks ago, The Financial Times reported

81
00:04:30,240 --> 00:04:32,799
Speaker 1: that Microsoft had pulled out some of their core Weave contracts,

82
00:04:32,839 --> 00:04:36,479
Speaker 1: though Core We've denied the allegations, and mysteriously, a week later,

83
00:04:36,600 --> 00:04:39,360
Speaker 1: open Ai announced that they had an eleven point nine

84
00:04:39,440 --> 00:04:44,160
Speaker 1: billion dollar contract with Corewave to provide compute services. Now,

85
00:04:44,200 --> 00:04:47,320
Speaker 1: I can't say for certain. Can you think of another

86
00:04:47,360 --> 00:04:50,320
Speaker 1: reason Microsoft needed a whole bunch of GPU compute other

87
00:04:50,360 --> 00:04:55,039
Speaker 1: than to host open AI's models. Hmm, Now, I'm just

88
00:04:55,760 --> 00:04:58,360
Speaker 1: guessing here, but one has to wonder if it wasn't

89
00:04:58,400 --> 00:05:01,560
Speaker 1: so much open Ai se a new deal and more

90
00:05:01,600 --> 00:05:04,920
Speaker 1: at taking over the future compute than Microsoft was formally handling.

91
00:05:05,640 --> 00:05:09,080
Speaker 1: And yes, you're remembering correctly. Microsoft has recently canceled more

92
00:05:09,080 --> 00:05:12,400
Speaker 1: than the gigawater future compute capacity, though from what I've heard,

93
00:05:12,440 --> 00:05:15,640
Speaker 1: its core Weave contracts future or otherwise weren't included in

94
00:05:15,680 --> 00:05:18,800
Speaker 1: that number. Anyway, For core Weave to make that eleven

95
00:05:18,839 --> 00:05:21,440
Speaker 1: point nine billion dollars in revenue from open Ai, they're

96
00:05:21,440 --> 00:05:23,760
Speaker 1: going to need compute, which will cost them tens of

97
00:05:23,760 --> 00:05:26,680
Speaker 1: billions of dollars to build, which will be difficult as

98
00:05:26,720 --> 00:05:28,800
Speaker 1: the company has burdened with about eight billion dollars of

99
00:05:28,839 --> 00:05:31,800
Speaker 1: debt with horrifying interest rates, which made lead it to

100
00:05:31,839 --> 00:05:34,360
Speaker 1: pay upwards of one point five billion dollars a year

101
00:05:34,360 --> 00:05:36,880
Speaker 1: in loan payments, with one of them requiring Core Weave

102
00:05:36,920 --> 00:05:39,560
Speaker 1: to repay the loan with any other debt they raise

103
00:05:39,640 --> 00:05:42,599
Speaker 1: in the future. Core Weave, by my estimates, needs at

104
00:05:42,680 --> 00:05:45,440
Speaker 1: least thirty billion dollars to expand. It has about one

105
00:05:45,480 --> 00:05:47,400
Speaker 1: point three billion dollars in the bank and just under

106
00:05:47,440 --> 00:05:49,520
Speaker 1: four billion dollars left on its loans that they can draw.

107
00:05:49,760 --> 00:05:52,000
Speaker 1: One of them is a term loan where you can

108
00:05:52,080 --> 00:05:55,680
Speaker 1: draw more up until about June twenty twenty five. Although

109
00:05:55,920 --> 00:05:58,880
Speaker 1: Core We've will raise capital in its IPO, the amount

110
00:05:58,880 --> 00:06:01,160
Speaker 1: will be anywhere near any to meet its needs. And

111
00:06:01,200 --> 00:06:03,719
Speaker 1: when I say its needs, I mean like the ability

112
00:06:03,760 --> 00:06:07,120
Speaker 1: to service the revenue that they need to spend more

113
00:06:07,120 --> 00:06:09,640
Speaker 1: money really does not make a lick of sense when

114
00:06:09,680 --> 00:06:12,120
Speaker 1: you think about it, and you may wonder if things

115
00:06:12,120 --> 00:06:14,559
Speaker 1: can get worse, and the answer is that they can.

116
00:06:14,920 --> 00:06:18,359
Speaker 1: Coreweave has planned one point three gigawats of expansion and

117
00:06:18,480 --> 00:06:21,320
Speaker 1: its partner to build it is a public company called

118
00:06:21,360 --> 00:06:26,840
Speaker 1: Core Scientific. Different company, similar name, genuinely different. Here are

119
00:06:26,839 --> 00:06:30,200
Speaker 1: some facts about Core Scientific. They went public in twenty

120
00:06:30,240 --> 00:06:33,119
Speaker 1: twenty two in a disastrous spack merger and then filed

121
00:06:33,120 --> 00:06:36,600
Speaker 1: for bankruptcy in the same year. Their entire business has

122
00:06:36,640 --> 00:06:40,239
Speaker 1: been focused on mining bitcoin, which requires specialized acic computer

123
00:06:40,360 --> 00:06:44,000
Speaker 1: chips that are entirely different both in their construction and

124
00:06:44,040 --> 00:06:47,400
Speaker 1: their maintenance to GPUs, and they can only really be

125
00:06:47,480 --> 00:06:51,039
Speaker 1: used for mining crypto. They're not easily repurposed for other tasks,

126
00:06:51,040 --> 00:06:53,239
Speaker 1: and indeed their data centers are not just a plug

127
00:06:53,240 --> 00:06:56,440
Speaker 1: and play thing. You basically have to demolish them. And

128
00:06:56,680 --> 00:06:59,360
Speaker 1: another thing about Core Scientific they only made twenty four

129
00:06:59,400 --> 00:07:02,320
Speaker 1: million dollars twenty twenty four from selling AI compute related

130
00:07:02,360 --> 00:07:05,200
Speaker 1: services to their customers. That's around five percent of its

131
00:07:05,200 --> 00:07:07,560
Speaker 1: total revenue, with the rest coming from mining and selling crypto.

132
00:07:08,160 --> 00:07:12,240
Speaker 1: Also really dumb ass mistaken. Just make that I said customers,

133
00:07:12,640 --> 00:07:17,040
Speaker 1: I meant customer. Course Scientific has one customer and that

134
00:07:17,080 --> 00:07:21,080
Speaker 1: customer is it's core Weave. As of right now, Corewave

135
00:07:21,080 --> 00:07:25,080
Speaker 1: has approximately three hundred and sixty megawatts of compute power. Somehow,

136
00:07:25,120 --> 00:07:28,119
Speaker 1: it intends to build another one point three gigawatts worth

137
00:07:28,120 --> 00:07:30,040
Speaker 1: of compute using a partner that has never built an

138
00:07:30,080 --> 00:07:32,120
Speaker 1: AI data center and does not be able to have

139
00:07:32,200 --> 00:07:35,200
Speaker 1: any meaningful compute right now, this is all so good.

140
00:07:35,280 --> 00:07:39,520
Speaker 1: I love this. Let's fucking go. This is the big IPO,

141
00:07:39,880 --> 00:07:43,400
Speaker 1: this is the big aiaper. I am going insane every

142
00:07:43,440 --> 00:07:48,000
Speaker 1: time I read about these goddamn companies. But let's summarize.

143
00:07:48,880 --> 00:07:51,400
Speaker 1: Core Weave is burdened by interest payments that may belloon

144
00:07:51,520 --> 00:07:53,600
Speaker 1: to as much as two billion dollars a year, and

145
00:07:53,640 --> 00:07:56,240
Speaker 1: lost eight hundred and sixty three million dollars on two

146
00:07:56,240 --> 00:07:59,400
Speaker 1: billion dollars of revenue in twenty twenty four. Sixty two

147
00:07:59,440 --> 00:08:02,000
Speaker 1: percent of that revenue is from Microsoft, which has materially

148
00:08:02,000 --> 00:08:04,280
Speaker 1: pulled back on data center buildouts and may have dropped

149
00:08:04,280 --> 00:08:06,560
Speaker 1: some core Weave contracts, though core We've denies this is

150
00:08:06,560 --> 00:08:09,680
Speaker 1: the case. Of course, Corby's expansion, which is critical to

151
00:08:09,720 --> 00:08:12,360
Speaker 1: servicing future revenue and growth, requires it to invest tens

152
00:08:12,360 --> 00:08:15,400
Speaker 1: of billion dollars the core Weave does not have. Corwy's

153
00:08:15,480 --> 00:08:18,080
Speaker 1: data center expansion is dependent on what is primarily a

154
00:08:18,080 --> 00:08:21,160
Speaker 1: bitcoin mining company called Core Scientific. It appears to have

155
00:08:21,280 --> 00:08:26,200
Speaker 1: no AI capacity building, and they're meant to build over

156
00:08:26,240 --> 00:08:28,679
Speaker 1: a gigawak capacity a time where they do not appear

157
00:08:28,720 --> 00:08:32,800
Speaker 1: to have done so. And as a reminder, converting cryptocurrency

158
00:08:32,840 --> 00:08:36,120
Speaker 1: mining data centers to HPC data centers is effectively starting

159
00:08:36,120 --> 00:08:40,480
Speaker 1: from scratch. It's not good. It's not good at all.

160
00:08:40,679 --> 00:08:43,000
Speaker 1: Corby should have been a positive signal for generative AI,

161
00:08:43,120 --> 00:08:46,200
Speaker 1: or at least the way for generative AI to kind

162
00:08:46,200 --> 00:08:49,160
Speaker 1: of go further than where it is today, or at

163
00:08:49,200 --> 00:08:51,319
Speaker 1: least a way for AI boosters to shut me up.

164
00:08:51,559 --> 00:08:55,040
Speaker 1: If GENERATIVEAI had this incredible demand, both from companies looking

165
00:08:55,040 --> 00:08:57,640
Speaker 1: to integrate it and us looking to use it, Corweve

166
00:08:57,679 --> 00:09:00,199
Speaker 1: would be making fat stacks of cash. I have a

167
00:09:00,240 --> 00:09:03,360
Speaker 1: far more diverse customer base and if I'm honest, not

168
00:09:03,480 --> 00:09:05,320
Speaker 1: have to take up more than five times its revenue

169
00:09:05,320 --> 00:09:07,920
Speaker 1: and burdens and loans with loan shark level interest rates

170
00:09:08,000 --> 00:09:12,079
Speaker 1: just to survive. But it is what it is, I guess.

171
00:09:12,880 --> 00:09:16,760
Speaker 1: And another note Jim Kramer of CNBC. He said that

172
00:09:16,840 --> 00:09:18,760
Speaker 1: Corweve was going to be one of the biggest IPOs

173
00:09:18,800 --> 00:09:20,880
Speaker 1: of the year. That should tell you about everything