WEBVTT

1
00:00:00.080 --> 00:00:08.570
The shift that we're seeing from Bitcoin mining to AI reminds you a lot of the Chinese hash rate ban and the great hash rate migration.

2
00:00:08.600 --> 00:00:23.180
The China exodus was essentially the government saying, "Hey, we're, we're done with this," e-essentially excommunicating eighty percent of the global hash rate during twenty twenty, and a, a good percentage of it ended up migrating to the United States and really put US mining on the map.

3
00:00:23.330 --> 00:00:31.280
And so like, what I'm seeing now is like the shift of all the mega miners getting out and, and going into HPC AI, it's gonna have downstream effects.

4
00:00:31.340 --> 00:00:38.440
There's like, right now you would think ASIC brokers are boring, they don't have anything going on, and like, they're the busiest they've ever been. And it's like, wait a minute, how is that the case?

5
00:00:38.600 --> 00:00:48.040
Everyone thought when this whole emergence came that the traditional data center companies would just chew up all that demand, and that's not happening. The Bitcoin guys kinda understood the power game a lot better.

6
00:00:48.080 --> 00:00:52.440
In five or ten years when this market really matures, they're going to be the industry leaders.

7
00:00:52.760 --> 00:01:04.120
Right now, it kinda, you know, it's like, oh, they've abandoned Bitcoin, but in reality, all of these companies all have Bitcoin in their DNA, and now are gonna be part of this crazy emergence of where things are going with, um, compute.

8
00:01:06.700 --> 00:01:11.740
Welcome back to the Blockspace Pod, brought to you by CleanSpark. Is Bitcoin mining dead or dying?

9
00:01:12.140 --> 00:01:30.240
Well, as I record this, Bitcoin has plummeted to sixty-five thousand dollars, a level we haven't seen since twenty twenty-four, and as a result, hash price has also fallen to an all-time low of twenty-nine dollars per peta hash per day, and hash rate is cratering below a zeta hash as a result of the margin squeeze on miners.

10
00:01:30.650 --> 00:01:32.060
So where do we go from here?

11
00:01:32.100 --> 00:01:43.860
Well, for today's show, which I recorded two weeks ago live at the National Energy and Mining Summit, we have Tom Masero, the head of strategy at Cathedra, to discuss the structural shifts we're seeing in Bitcoin mining today.

12
00:01:44.220 --> 00:01:56.770
Tom shares his thesis about why what we're seeing with the AI pivots really mirrors the China mining ban of twenty twenty-one, and not just how it will change the industry, but how it will structurally change where Bitcoin's hash rate is flowing.

13
00:01:56.780 --> 00:02:07.160
And we also talk about dynamics in the ASIC market and where the opportunities still lie for Bitcoin miners and for the little guys who might be looking into getting into other forms of compute.

14
00:02:07.200 --> 00:02:09.840
You're listening to the Blockspace Pod. We'll be right back.

15
00:02:09.960 --> 00:02:20.280
Hey, this is Charlie and Colin from Blockspace Media, and you're listening to the Blockspace Podcast, a show about emerging tech in Bitcoin, AI, energy, and markets.

16
00:02:20.500 --> 00:02:33.230
We publish two interviews weekly with CEOs, investors, analysts, and anyone else of consequence within these spaces. Plus, we have a weekly news roundup for all the important stories you might have missed from that week.

17
00:02:33.560 --> 00:02:44.140
The show is perfect for retail and institutional investors, analysts, and really anyone who wants to keep their finger on the pulse of the stories that are moving Bitcoin, energy, and data markets.

18
00:02:44.480 --> 00:02:52.720
So if you've stumbled across us, make sure to search Blockspace wherever you get your podcasts, and on YouTube, hit the subscribe button, give us a rating.

19
00:02:52.960 --> 00:02:57.760
We produce bonus podcasts and other content on our main feed, so you don't wanna miss that.

20
00:02:57.800 --> 00:03:10.109
And if you have any feedback or comments to give us or shows that you would like to see and topics you would like us to cover, hit us up at hello@blockspace.media. Tom, welcome back to the Mining Pod, man.

21
00:03:10.140 --> 00:03:16.840
What's up, Colin? Man, it's been a long time. Yeah. It's, uh, I mean, it's been, uh, about a year, I think, which is like ten years in this industry. Yeah, absolutely.

22
00:03:17.560 --> 00:03:21.760
Um, so I think a good place to start off with this conversation is

23
00:03:22.880 --> 00:03:44.540
you just gave the closing talk on one of the clo-- You, you were a panelist on one of the closing panels for today at Bitcoin Park at the NIM Summit, and you mentioned this idea that the shift that we're seeing from Bitcoin mining to AI reminds you a lot of the Chinese hash rate ban and the great hash rate migration that resulted from that.

24
00:03:44.980 --> 00:03:57.300
Can you unpack that to give our listeners an idea of what you think is going on currently with this shift? Yeah. I mean, obviously, like, the, the reason or the rationale for this shift is obviously much different.

25
00:03:57.380 --> 00:04:15.980
You know, in, in the-- the China exodus was essentially the government saying, "Hey, we're, we're done with this," and e-essentially excommunicating eighty percent of the global hash rate during twenty twenty, and a, a good percentage of it ended up migrating to the United States and really put US mining on the map in twenty twenty-one.

26
00:04:16.399 --> 00:04:25.360
All of us here wouldn't even be doing what we're doing right now if that, if that didn't happen. We were essentially two big players in the market, uh, in the global mo-mining market then.

27
00:04:25.900 --> 00:04:36.780
And so like, what I'm seeing now is like the shift of all the mega miners getting out and, and going into HPC AI, and then now you're-- it's gonna have downstream effects.

28
00:04:36.840 --> 00:04:47.840
And like, you know, I didn't really start putting the pieces together till I got here yesterday and started talking to a lot of folks and hearing about like, you know, uh, talking to brokers, ASIC brokers.

29
00:04:47.900 --> 00:04:53.440
Like, right now you would think ASIC brokers are boring, they don't have anything going on, and like, they're the busiest they've ever been.

30
00:04:53.480 --> 00:04:58.650
And it's like, wait a minute, how is that the case of like mining, you know, difficulties, you know, like...

31
00:04:58.680 --> 00:05:07.280
Well, we've had some negative difficulty adjustments, but overall market is not great for mining, yet the brokers are as busy as they've ever been.

32
00:05:07.730 --> 00:05:19.760
And that's because you have a lot of these companies who have been sitting on millions of ASICs who are now dumping them to make room for their new clients and their new build-outs, and where are those machines going?

33
00:05:19.830 --> 00:05:24.800
And who are they going to? What use cases they're going to? They're getting spread out through the whole world.

34
00:05:24.860 --> 00:05:36.940
So there was this whole talk after twenty-one into like, you know, let's say from twenty-one to like twenty-three, twenty-four, we went from, "Oh, there's too much centralization in China" to "There's too much centralization in the United States or North America."

35
00:05:37.400 --> 00:05:47.620
Now, you're gonna see like almost a lot of that hash rate evaporate out of the United States, and where is it gonna go? So I, I don't think we actually know the ramifications of it.

36
00:05:48.320 --> 00:05:59.800
The one thing I think of on the top of the funnel is if you think about-- I was just-- That was what I kinda closed out with was, you know, it was like, w-w-what, like what is like an underrated thing that has happened this past year?

37
00:06:00.320 --> 00:06:10.060
And a lot of Bitcoiners say, "Oh, it's not g-" The traditional conventional thought is it's not good that all these b- traditional Bitcoin mining companies have now just pivoted into HPC AI.

38
00:06:10.720 --> 00:06:24.066
And on the surface, that makes sense. But-There's something that happened with HPC AI that's, I think, kind of revolutionary because it's an entirely new segment of an industry. It's not traditional data centers. Mm.

39
00:06:24.256 --> 00:06:31.336
Everyone thought when this whole emergence came that the traditional data center companies would just chew up all that demand, and that's not happened.

40
00:06:31.776 --> 00:06:42.946
The, the, the, the Bitcoin guys kinda understood the power game a lot better, especially, like, the density side of things on a computing side and have really jumped in and have, like...

41
00:06:42.976 --> 00:06:48.576
They're, like, in five or 10 years when this market really matures, they're going to be the industry leaders.

42
00:06:48.976 --> 00:07:01.136
So right now it kinda, you know, it's like, oh, they've abandoned Bitcoin, but in reality, all of these companies all have Bitcoin in their DNA, and now are gonna be part of this crazy emergence of where things are going with, um, compute.

43
00:07:01.336 --> 00:07:11.096
And I think long-term it's gonna be very good for Bitcoin, you know, for, for that from, like, a macro perspective. It just doesn't seem like it right now because everything's sort of- Right... unfolding.

44
00:07:11.196 --> 00:07:19.256
So yeah, that was, like, sort of the rift I had, and I hadn't really, like, put that together till I got here. Um, a f- there are a lot of things to latch onto there.

45
00:07:19.856 --> 00:07:26.176
And, and going back to that last point, it may not seem good right now, but if you're a miner plugged in right now, you're probably...

46
00:07:26.196 --> 00:07:34.656
I mean, you're not doing great 'cause hash price is still pretty compressed, but you're feeling a lot better about seeing, what was it, five out of the last six difficulty adjustments- Negative... being negative. Yep.

47
00:07:34.886 --> 00:07:43.326
I... That, that is... The last time we had a stretch like that was the China mining ban. Mm-hmm, mm-hmm. Just going back to what you were saying about the, uh, verisimilitude of these two things, right?

48
00:07:44.086 --> 00:07:53.826
The other thing about the brokerage landscape, I had the same thought. I talked to a broker here, and I asked them, "You must be kinda chilling right now, right?" Mm-hmm.

49
00:07:53.836 --> 00:07:58.716
And they said, "No, January has been one of the busiest months that I've ever had." Crazy.

50
00:07:58.856 --> 00:08:07.276
And what's incredible is that they are basically processing more used volume right now because all the public miners are liquidating their machines. Yep.

51
00:08:07.646 --> 00:08:17.436
And they told me, you know, some of those are going to places you might expect, like South America. Some of them are going to Northern Europe, which was something that I wasn't really anticipating, right?

52
00:08:17.516 --> 00:08:27.745
And I kinda nudged them and said, "Are those intermediaries for Russia?" Yeah. But I think it actually is legitimately going to just the Iceland... or going to Iceland, going to Sweden, going to Norway- Sure...

53
00:08:27.756 --> 00:08:29.486
going to those Scandinavian countries.

54
00:08:30.536 --> 00:08:42.776
Going back to what you were saying about Bitcoin miners being more primed to tackle this new computing industry than maybe the traditional data centers, you mentioned they understand power density better.

55
00:08:43.176 --> 00:08:51.036
Could you unpack why you think they're positioned to actually dominate this compared to the hyperscalers and the Mag 7 companies of the world?

56
00:08:51.456 --> 00:09:03.516
Is that a complacency with the en- entrenched tech companies to where they haven't really had to struggle the way Bitcoin miners had to? Mm. Or what do you think is the reasoning behind that thesis?

57
00:09:04.696 --> 00:09:10.316
I, I, I think there is, like, a complacency aspect to it. I think the biggest thing is just the market dynamics are...

58
00:09:10.336 --> 00:09:16.856
Like, when, when this whole kind of, like, push started, you know, we started getting RFPs, you know, to, "Hey, what would a... what...

59
00:09:16.896 --> 00:09:25.296
you know, what could you do to get, you know, a 50 megawatt setup or something like that?" And we're like... They're like, "Oh, we, we need this w- you know, tier three, all of these criterias done in a year."

60
00:09:25.356 --> 00:09:31.236
It's just like that's not, not gonna happen. You could do X, Y, Z in a year. And they're like, "No, no, no, we couldn't do that."

61
00:09:31.296 --> 00:09:42.006
And, and now 18 months later, they're all doing what I think Bitcoin miners originally thought, which was this, like, very light version of, like, a HPC, um, AI compute.

62
00:09:42.676 --> 00:09:51.676
And, you know, really I think the, the biggest thing is, like, they have chips that are essentially sitting in warehouses that they have to...

63
00:09:51.776 --> 00:10:00.416
So it's very similar to, like, when the machines came over here from China, they just started putting up sites like crazy because they had to get them plugged in. Mm.

64
00:10:00.426 --> 00:10:11.336
And you've got so many orders of these GPUs that are out, and they don't have, like, the option or the luxury of saying, "Oh, we'll just wait for our tier three data center to be ready and all these other things," 'cause it's just not gonna happen.

65
00:10:11.556 --> 00:10:22.245
So now they're like, "Okay, well, we gotta figure it out, and maybe we don't get the uptime that we have. Maybe we don't get as much redundancy. Maybe we'll do 20% backup battery to kinda keep the core systems going.

66
00:10:22.276 --> 00:10:30.856
Maybe we have to adapt and do some type of curtailment, and we're okay with it." Whereas, like, a year ago or even longer, that was n- no, no, no, non-negotiable.

67
00:10:31.136 --> 00:10:42.296
Has to be 100% redundancy, this, this, this and that, and now it's all coming. And then it... Same thing on the size. Size was massive mega scale, gigawatts. Half a gigawatt was, like, considered small.

68
00:10:42.316 --> 00:10:48.936
And now we're seeing things in the market where they're gonna come down to, like, the 10 and 20 megawatt sites because the chips are getting more efficient.

69
00:10:48.996 --> 00:10:58.416
So the same cy- kind of cycle that happened with ASICs from, like, 2016, 2017 to, like, you know, kinda where we are now, that's happening on the GPU side. The one...

70
00:10:58.476 --> 00:11:11.136
I think the one thing that there's inherent risk for for any traditional miner kinda getting into this is that the infrastructure side is still not fully developed. So there's, like, these two kind of divergent plays.

71
00:11:11.156 --> 00:11:18.516
You got, like, your shell provider, soft shell provider versus, like, someone who's fully building stuff out, which you and I talked about with, like, Iron and stuff. Right.

72
00:11:18.556 --> 00:11:28.526
You know, like, the fact that there's not even racks avail- commercially available for some of these GPUs that are y- gonna be coming out, like, next year, like, what do you do with that? Mm.

73
00:11:28.696 --> 00:11:31.916
I wouldn't wanna be exposed to those things, but I think that there are...

74
00:11:31.976 --> 00:11:46.135
That's where, like, the collaboration's kind of coming in with some of the traditional data center hardware partners who are finding good partners with Bitcoiners because, like, we understand the power game, and then we partner with them on what they have their expertise in.

75
00:11:46.156 --> 00:11:50.156
And so I think you'll see more of a convergence, um, you know, with, with that side of things.

76
00:11:50.176 --> 00:12:02.636
When you're talking about the racks, basically you're saying that miners and some of these companies are gonna have to build, like, bespoke, b- bespoke infrastructure because we actually haven't standardized anything for the bones of these operations.

77
00:12:02.736 --> 00:12:10.066
Yeah. Is that the idea? In some cases, yes. And then in, in some cases it's like, okay, well, what's the lifespan on this? Because if you have these new generations that are just...

78
00:12:10.076 --> 00:12:19.644
I mean, you see Nvidia's, like, launches, they're like... I mean, every single launch every, whatever, eight or 12 months-These huge jumps. Mm-hmm. And now the specs don't fit.

79
00:12:19.964 --> 00:12:30.004
So like you've got a company who's two years out with chips that are a couple generations old. They build out all of these infrastructures at like ten and fifteen million dollars a megawatt, and then, then what?

80
00:12:30.684 --> 00:12:41.094
What do you do when those are outdated? You can't-- like now the, the bones of it can't even handle the newer generation stuff. Mm-hmm. So as a miner, I think you have to be very careful with what's going on.

81
00:12:41.124 --> 00:12:52.984
And that's why if you look at like the public segment, the focus is on who's the end customer, who's the counterparty, because if it is like a, a Google, a Microsoft, like they can handle that. Right.

82
00:12:53.024 --> 00:13:09.944
Like, and so I think like as the market matures and there's some other kind of like opportunities there, and maybe there's more standardization around specific infrastructure or maybe just being solely inference, then you'll see, I think more build outs in like that sub fifty megawatt range.

83
00:13:10.604 --> 00:13:21.194
In going back to your point about the sub fifty megawatt range and some of the concessions that these companies are making, and then I want to get into some of the details of the opportunities that are available for the smaller guys- Mm-hmm.

84
00:13:21.194 --> 00:13:33.644
'cause I think that's an important point to double tap on that we see all of the dazzling and sparkling headlines of the like five hundred megawatt build that is being contracted out for, you know, a billion dollars a year over ten years, right?

85
00:13:33.814 --> 00:13:33.814
Mm-hmm.

86
00:13:33.864 --> 00:14:06.824
But to your point, there are real resource constraints here, and there are real energy constraints here, and this is part of the reason I would imagine why we're seeing, you know, I think a lot of, I think a lot of folks maybe in-- I, I shouldn't sp- say a lot of folks, but some commentators will look at a company like Microsoft partnering with Iren, or a company like Amazon partnering with Cipher, or a company like AMD partnering with Riot, and they might think to themselves, "Why are they partnering with these Bitcoin miners who some investors consider, you know, like pink sheet companies?"

87
00:14:06.844 --> 00:14:14.444
Mm-hmm. Like, I'm not saying they are. Obviously, I respect- Right. a lot of like friends with a lot of management teams. I know they're really high quality people working at these companies.

88
00:14:14.484 --> 00:14:25.693
But there is-- there's this perception that you're playing with the JV squad, right? And that the varsity is having to go to JV for, for not just practice, but for, for playing the game, right? Yeah.

89
00:14:25.704 --> 00:14:39.274
How much of that is just in, in to your knowledge, like how much are all of those concessions just the fact that kind of the capital got way ahead of its skis here, and then everyone had to take a step back and say, "Hang on, guys,

90
00:14:40.504 --> 00:14:48.464
we don't actually have enough power. We don't have enough raw materials to go around." Yeah. I mean, that's definitely, that's definitely the formula that happened.

91
00:14:48.504 --> 00:14:56.814
Like when these guys all started putting these RFPs out, they went to their traditional data center partners, and then they realized- When you say, sorry, RFP, can you just unpack that for me?

92
00:14:56.824 --> 00:15:05.904
Like, um, like, so like Nvidia and some of their cloud partners would come out and say they'd, they'd go to power developers and say, "Hey, look, here's what we're looking for."

93
00:15:06.264 --> 00:15:14.264
And they would initially tap the, you know, their traditional data center builders or co-location companies to say, "Hey, here's what we're tasked with. Go build it."

94
00:15:14.924 --> 00:15:21.604
And then they realized the timelines didn't match up, and they knew the Bitcoiners were there. We had been raising our hands saying like, "We'll do this."

95
00:15:22.024 --> 00:15:28.184
And then it took like, you know, CoreSci to kind of break the, the mold- Mm-hmm. with the CoreWeave deal, and then it slowly started to happen.

96
00:15:28.204 --> 00:15:35.824
But like there was like a six or s- probably longer than that, almost a year span of companies you knew were sort of getting into it but didn't have a deal put in place.

97
00:15:35.853 --> 00:15:45.144
And there was a lot of-- those deals are very complex, first off, so there's a lot of due diligence, a lot of making sure like on risk analysis to make sure both sides are covered.

98
00:15:45.184 --> 00:15:57.104
And it was sort of like them vetting us out and then them realizing like, wait a minute, these guys have been doing this for a while at very large scale. And it's not like we just are plugging like popcorn machines in.

99
00:15:57.144 --> 00:16:06.884
[chuckles] Like there's a lot of capital at play and like, you know, we're, we're capable of doing things very quickly, and especially if the incentives are tied, y- you know, together that they can accomplish that.

100
00:16:06.913 --> 00:16:08.404
And that's, that's why they're doing it.

101
00:16:08.544 --> 00:16:19.104
It's just, it's just like a new market, and there's a completely new segment, and, you know, they don't-- I don't think they see the risk in it, especially like with the, a lot of the collaboration that's, uh, that's taking place.

102
00:16:19.164 --> 00:16:30.404
Now, there is going-- there is like a hype cycle obviously that is like part of this whole thing, but it's pretty clear like, you know, we have, you know, we got all our chat VMs and stuff like that.

103
00:16:30.444 --> 00:16:40.064
Like how many people in your chat or in your sphere are using like Claude Code or any of these tools? I have like five pro accounts with all of them. Right.

104
00:16:40.084 --> 00:16:48.304
And I'll use them interchangeably for different jobs all the time. And I know I'm like kind of like an early adopter. Right. But like I don't actually even have like a necessary need for a lot of these things.

105
00:16:48.344 --> 00:16:57.124
They just make my life more efficient in some ways, and we're like at the very nascent edge of it and like what's go- what's that gonna be in like five years? The GPUs are still gonna be needed.

106
00:16:57.504 --> 00:17:02.584
Now, like you could say some of these projects are like, are gonna get blown up. A lot of them are gonna get blown up.

107
00:17:02.664 --> 00:17:09.644
But like the actual reality of like this new paradigm that we're entering is very real, and it's not going away.

108
00:17:09.784 --> 00:17:20.984
I think that's an important point to zoom in on because I think one of the strongest arguments against comparing this to the dot-com boom is that the tools are already here.

109
00:17:21.504 --> 00:17:31.444
Now, the market saturation for those tools and potentially obviously the upper bound for the potential for those tools, we're not anywhere near that. Mm-hmm. The revenue models are still getting fleshed out. Yeah.

110
00:17:31.524 --> 00:17:40.774
Um, the revenue to actually cover all of the CapEx for these things and all of the spend for, uh, renting out this compute, you know, there might be a shortfall there.

111
00:17:40.784 --> 00:17:49.834
But the fact of the matter is, at the height of the dot-com boom, I could go to, you know, the fa- famous example. I could go to pets.com, and it was just HTML text on a page. Mm-hmm.

112
00:17:49.904 --> 00:17:59.964
But I can go to ChatGPT, I can go to Claude, I can Vibe code something. I can go to ChatGPT and have it do research that would've taken me an hour combing through Google.

113
00:18:00.684 --> 00:18:04.204
And the, the efficiencies and the use case are already there.

114
00:18:04.284 --> 00:18:14.334
So I, I, I think that that's what I-- whenever I talk to anyone in my, uh, personal life who is real skeptic about this stuff, I, I just say like, "Have you actually used these things?" Mm-hmm. Mm-hmm.

115
00:18:14.484 --> 00:18:20.624
So many are just like they're, they're not-- they don't want to, or, you know, like they'll go to the fact that like, look, a lot of this stuff isn't accurate.

116
00:18:20.664 --> 00:18:25.724
If I put something in, I've, you know, it like hallucinates or whatever, and it's like-Well, of course. Sure. Yeah, exactly.

117
00:18:25.824 --> 00:18:31.404
But, but then when you've actually, like, experienced some of the magic around it, you're like, oh man- Yeah... this is gonna be. I mean, exactly.

118
00:18:31.424 --> 00:18:39.894
And I mean, I- that's j- that's a good point not to say that it's perfect. I mean, I've had plenty of times where I was like, "This is wrong," and been frustrated because I'm just not getting the prompt right. Right.

119
00:18:39.924 --> 00:18:46.244
But then there are other times where I will have it spit something out, and I look at the... It'll give me the citations and everything, and I'm just like...

120
00:18:46.314 --> 00:18:57.094
'Cause that's what as a, as a journalist and a researcher, that's the thing that I find it most useful for is curating information. I don't need it to think for me. Right. I need it to find things for me. Yeah. And it is

121
00:18:58.184 --> 00:19:07.333
s- it is, it is like a freaking search engine on, on steroids. Yeah. It's incredible. Yep, and it hones you to... It, it can do two things.

122
00:19:07.384 --> 00:19:16.544
Like, I've seen this, like, thread of, like, oh, it's, it's, it, it basically gives you psychosis. Or it, like, gives you superpower 'cause, like, it helps you critical- critically think a little bit more too.

123
00:19:16.584 --> 00:19:22.324
It's like, like you said, how am I prompting? What's a better way of saying what I want to say? And so then you go back to, like, the basics.

124
00:19:22.364 --> 00:19:37.024
Like, the people who are the best prompters are people who have really good grasp of, like, English and literature, especially with Cod- Claude Code, you're, you're, you know, they've had the CEO of Anthropic on, you know, whatever, in Davos, and basically he was like, "Software engineers are gonna be completely extinct within 12 months."

125
00:19:37.624 --> 00:19:47.324
Like, what? [laughs] And I know, you know, like we're, we're, you know, this is probably to the, to the bane of, uh, um, some of our, our shareholders at Cathedra, but, like, I, I vibe coded a website.

126
00:19:47.364 --> 00:20:02.544
Now, it's not live yet because I've been, we've been dealing with a lot of, like, stuff here with some of our sites and trying to get them up and running and ready for these storms, but I vibe coded, like, a, a brand-new website and branding for Cathedra that we would've paid $50,000 to $100,000 for- Yeah...

127
00:20:02.564 --> 00:20:11.244
through a traditional agency, and we were actually, like, engaged with an agency, and I sent them this 'cause I was like, "Hey, I kinda did this. Like, what, like, what's wrong with it?"

128
00:20:11.344 --> 00:20:17.844
And they were like, "This is, like, incredible. You did this with?" Like, I, I think I used, like, Gemini, and it was unbelievable.

129
00:20:17.884 --> 00:20:29.304
There's so much, like, tax for businesses that if you get smart people and making things more efficient and better, like, now it's like it has my touch, like, the things I would want are there. Whereas, like,

130
00:20:30.224 --> 00:20:38.804
not to say that it completely replaces a person, but there's a lot of people who are very creative but don't have the ability to actually, like, write out or whatever. Right. Now you actually...

131
00:20:38.844 --> 00:20:46.584
It's al- almost as good as saying, like, you know, I would go and task, like, a freelancer website before and say, "This is what I'm thinking. This is what I want."

132
00:20:46.604 --> 00:20:56.084
And the turnaround, like you said, would be, like, a week or two, and now it's like, "Okay, look at this." I did this in a weekend. Yeah. It's crazy. [gentle music] We are CleanSpark,

133
00:20:57.004 --> 00:21:09.484
America's Bitcoin miner, a publicly traded company with the largest operating hash rate, powered entirely by self-operated infrastructure across four states.

134
00:21:11.304 --> 00:21:22.724
This is our proof of work, and we are setting the standard for what's next. Learn more about the intersection of energy and Bitcoin at cleanspark.com.

135
00:21:23.604 --> 00:21:36.944
Going back to the opportunities for smaller AI and HPC developments, I've had conversations with some miners here that, uh, you know, guys that I've never met before who are like, "Oh, I have, like, this site in this part of the US.

136
00:21:37.524 --> 00:21:48.853
I am, you know, outfitting it and trying to get fiber optic put in," or, "I'm trying to get generators put in so that I can market it to a, a data center company- Mm... to buy for AI and HPC." Mm.

137
00:21:48.884 --> 00:21:59.484
So and y- there, there's a part of me that wants to be a little bit critical and think, like, can this really work? And so I'll just kick that to you. I mean, is it...

138
00:21:59.544 --> 00:22:05.404
Obviously, there are a number of factors you need to make sure that you have, uh, a number of variables checked off the box.

139
00:22:05.464 --> 00:22:19.144
But how feasible is it for smaller competitors, for smaller miners to retrofit or expand into AI, and in what ways would their business maybe be marketing that compute a little bit differently?

140
00:22:19.204 --> 00:22:26.274
Because I feel like we're marketing what it can deliver differently because we've kind of, we lump AI and HPC. It's like blockchain. Mm-hmm.

141
00:22:26.304 --> 00:22:33.474
You kind of create this huge umbrella, but it's like you could be do- y- you know, you could be an HPC company that's basically just doing graphics rendering- Right...

142
00:22:33.484 --> 00:22:44.174
which is something that we've had for decade, you know, decades, and, like... Or you could be actually doing inference or, or training models for an OpenAI, for an xAI, things like that. Yep.

143
00:22:44.224 --> 00:22:53.924
So, like, what opportunities do exist for some of the smaller miners or smaller operators in the space for this, for this boom? I think it's emerging. I think the mo- like if...

144
00:22:54.004 --> 00:23:04.064
I always try to take, like, the caveman approach because, like, everything's happening so fast that there is really, no one knows. Uh, it is the Wild West, so it's the power. At the end of the day, it's the power.

145
00:23:04.364 --> 00:23:14.524
If you have power, and you have it reserved, that market is gonna come to you. It might not be at one megawatt. It's not at 10 megawatts right now, but it's gonna come. It's already coming down market.

146
00:23:14.874 --> 00:23:24.504
So if you can, you know, if you're somewhat, like, initially it was like you've gotta be within, you know, a 20-mile radius of a major metropolitan hub, then that's gone a little wider.

147
00:23:24.564 --> 00:23:33.534
Now, it's like, okay, are you along an interstate? Um, you know, are you in between, like, some big cities? And then that, that, that little egg is or that onion's growing out further from there.

148
00:23:33.564 --> 00:23:49.604
So I just think it's, like, having power, and then I think we're Bitcoin miners, and this is sort of like, like, my thesis on it, maybe not necessarily like the companies or whatever, but I, I think, like, if, if we think, like, what our advantage is, is it's go and get that power reserved.

149
00:23:49.804 --> 00:23:57.064
Mm-hmm. Now, what the utilities don't want is somebody just sitting, like, basically an, a domain squatter for power. Right. They don't want that. It doesn't help.

150
00:23:57.344 --> 00:24:03.804
So they actually would rather do business with somebody who's already doing the mining because they know they're a good customer. Interesting.

151
00:24:03.864 --> 00:24:11.424
And you go there, you set your stuff up, and then you just play the waiting game of, hey, that real estate is gonna become extremely valuable in the future.

152
00:24:11.844 --> 00:24:20.424
And you don't have to worry about, oh, well, do I have all the prerequisites that, that I need? Because eventually, like, if they want it bad enough, they'll make sure the fiber gets there.

153
00:24:20.704 --> 00:24:29.624
They'll make sure gensets get there or battery packs or... The battery packs thing is sort of like a cheat code that I think is gonna become like a, like a really big piece moving in the future.

154
00:24:29.664 --> 00:24:30.414
Are we, are we starting to...

155
00:24:30.504 --> 00:24:46.044
I, I've heard a few operators here talk about-I was talking to a guy who works, uh, for a supply chain firm here in Nashville, actually, and he was saying he was doing a deal where this company was asking for generators, and they were asking for battery packs, and he didn't-- he couldn't really piece it together.

156
00:24:46.084 --> 00:24:53.774
It's like, what are you, what are you doing? Mm. Like, wait, so wait, you're getting the battery pack, but you're also getting the thing to produce more power. Like, what's going on here? Mm.

157
00:24:53.824 --> 00:25:04.004
So could you unpack the battery pack play? I mean, 'cause I mean it seems, like, very intuitive, but does that reach in the early a- stages of that meadow where people are actually deploying batteries on site?

158
00:25:04.304 --> 00:25:14.584
I think, I think you're starting to see that, where the battery packs essentially act as not an entire redundant backup, but a medic- a critical mission backup.

159
00:25:14.864 --> 00:25:22.444
So, like, let's say they've got a fifty-megawatt load, and then maybe they do have to... Part of their deal with the utility is they do have to curtail a certain percentage of their load.

160
00:25:22.724 --> 00:25:35.964
Instead of getting gens- gensets, and maybe they don't have access to natural gas or diesel readily available, they're getting these battery packs, and they're, let's say, ten megawatts of battery power, where they know they have to commit to X amount of curtailment.

161
00:25:35.974 --> 00:25:42.964
They know that that ten megawatts can be mission-critical stuff that needs to stay on in order for, like, everything to function, so when the power comes back on,

162
00:25:43.904 --> 00:25:49.704
you know, maybe they have a cluster a hundred miles away or fifty miles away or ten miles away that can still, like, do its thing.

163
00:25:49.744 --> 00:26:01.324
And that's the type of, like, uh-- You're starting to see deployments that are taking more of that approach. Honestly, I think Grok or xAI is a company to watch when it comes to innovating.

164
00:26:01.484 --> 00:26:07.904
Uh, they, they're blowing the doors off of everybody when it comes to building right now. They have the, the biggest cluster right now there is.

165
00:26:08.284 --> 00:26:16.564
They were extremely fast in putting it together, way faster than everybody else. They put that site in Memphis together in four months. They were extremely-- They thought like a Bitcoin miner.

166
00:26:16.574 --> 00:26:20.154
More greenfield in six month-- In four months. They basically took over an abandoned warehouse.

167
00:26:20.434 --> 00:26:27.024
They have, like, twelve hundred megawatts live, and only a hundred and fifty to two hundred of those out of the twelve hundred are fed by the grid.

168
00:26:27.244 --> 00:26:36.304
The rest is all n-natural gas and gensets, and they're just-- they basically retrofitted a warehouse, and they got it working. That's wild. 'Cause it's very, like, first principles.

169
00:26:36.364 --> 00:26:44.564
They were like, "Oh, you know, you need this, X, Y, and this," and they were like, "No, we're, we're gonna do it this way." And they have Bitcoiners. I think Asher's been pretty vocal about it.

170
00:26:44.604 --> 00:26:54.564
He lost two of his engineers to them. Mm. And, like, these guys are running fast with that stuff, a-and they're not playing by the traditional data center rules when it comes to this.

171
00:26:54.724 --> 00:27:02.493
It, it, it reminds me of the, the fact that they stood up, or I shouldn't say they, but Tesla, so, like, you know, it's all in the Musk umbrella. Mm.

172
00:27:02.524 --> 00:27:11.564
Tesla stood up that lithium refinery w- and, and, and, and breakneck time. I mean, people were saying that his timeline was impossible. Yep.

173
00:27:11.664 --> 00:27:21.104
I don't know how long it took them, but I remember seeing some marketing materials about it, and you look at the facility, and you're just like, "Wait, hang on. This took, like, a year or two." Mm.

174
00:27:21.114 --> 00:27:29.584
And it, and it just looks absurdly complex. For, for xAI specifically, do you think that they're... Like, what do, what do you attribute that to?

175
00:27:30.064 --> 00:27:43.993
Th-there's kind of a- Their, like, their business, their business model is they already have a built-in business that will benefit from that compute being on, whether it's Tesla, whether it's Starlink, whether it's obviously their consumer products with xAI, I mean, with, with Grok and with Twitter.

176
00:27:44.814 --> 00:27:55.284
They already are-- like, they're able to get something out of it, where, like, a lot of these other companies that are doing it are sort of like, you know, it's like they're, they're acting like, you know, the, that the market's gonna come to them.

177
00:27:55.384 --> 00:27:57.993
They already have a market that's already paying for these products. Mm.

178
00:27:58.004 --> 00:28:04.664
It's already being subsidized by multiple aspects of their business, so, like, they can just continue to burn through it, where everyone else is like, you know...

179
00:28:05.204 --> 00:28:12.364
You see the stories about A- OpenAI and where they're just burning billions and billions of dollars because they're spending way more than what's actually coming in the door. Mm.

180
00:28:12.433 --> 00:28:18.784
And their moat is not that big considering, considering... I mean, they've grown. I mean, the fastest growing company in, in history.

181
00:28:18.864 --> 00:28:29.944
So I'm not, like, crapping on them, but their moat is a lot smaller because they're competing against Google and Microsoft and some of these other guys who have similar competing products, whereas, like, it's sort of, like,

182
00:28:30.944 --> 00:28:39.844
em-em-embedded into all of these aspects of, like, the X kind of, like, architecture. Yeah, and it's, these are synergies maybe between all of these different- Mm...

183
00:28:39.854 --> 00:28:47.504
companies in Elon's orbit, you know- Yeah, like that compute can get- Particularly for this Tesla... that compute can get used for Teslas. It can get used for the robots. Yeah. It can g-get used for Twitter.

184
00:28:47.644 --> 00:29:03.164
I mean, there's probably four or five different use cases for it and then, you know, at the same time, they have the ability to maybe put things in sites or put sites in areas that would not traditionally be viable for a lot of other companies because of their Starlink stuff.

185
00:29:03.564 --> 00:29:10.824
Ah, yeah. It's their Starlink bases you could probab-- You know. So, like, they're just, they're thinking about things. They're playing chess while everyone else is playing checkers.

186
00:29:10.864 --> 00:29:30.314
Yeah, that's interesting, and it kind of proves out what you were saying about companies are going to have to start writing new rule books and making concessions where it-it-- you're kind of in this weird space where it's like the traditional data center world is saying, "Well, this is the way we usually do it," and then all of these companies breaking into this new form of computer saying, "Hold my beer.

187
00:29:30.324 --> 00:29:31.754
We don't have time for that." You know? Mm-hmm.

188
00:29:31.784 --> 00:29:44.704
It's like, it's almost like, uh, the rush to, to find El Dorado or something, and, like, everyone knows where it is, and so as long as you get there first before everyone else, you're gonna have a, a, a larger share of the hoard of gold there, right?

189
00:29:45.434 --> 00:29:47.664
And it's really just access to power. Like- Yeah...

190
00:29:47.684 --> 00:29:59.204
there's a lot of push for these off-grid, or I call them off-grid, more behind-the-meter projects that are taking place, so you're seeing a lot of natural gas-type projects where, you know, they're targeting areas that meet their criteria.

191
00:29:59.244 --> 00:30:07.883
There's, like, fiber, there's a big natural gas pipeline, and then they're ordering, you know, generators, and they'll just make their own power and do their own thing.

192
00:30:08.584 --> 00:30:12.124
Now, with the EPA ruling that came out, like, a couple weeks ago, it was kind of...

193
00:30:12.164 --> 00:30:26.904
It basically ruled against what xAI is doing as far as, like, there are certain standards that you have to meet for some of these generators for on the EPA side that are, I think are gonna slow that stuff down, and that's where all these big guys were running 'cause they're like, "Oh, there isn't five hundred megawatts available on grid."

194
00:30:27.724 --> 00:30:31.284
So they're like, "Well, we're just gonna do these behind-the-meter things, and we're just gonna make it happen."

195
00:30:31.564 --> 00:30:39.504
But now, because that's gonna slow down, I think, like, the real on-grid opportunities are gonna be clusters. They're not gonna be these mega sites.

196
00:30:39.544 --> 00:30:46.908
Now-That removes Texas from this conversation because Texas can have probably dozens and dozens of these mega sites because of their power profile.

197
00:30:46.968 --> 00:31:00.788
But throughout the rest of the country, there's gonna be-- I think what's gonna happen is there's just gonna be clusters of, you know, maybe they'll still have five hundred megawatts worth of stuff, but there's gonna be fifty here, fifty here, fifty there, and they're all gonna be kind of, you know, within the same area.

198
00:31:00.828 --> 00:31:08.528
And that's just y- and w- you-- Why the clusters specifically? Because if like, uh, 'cause when I- Because A, they can get them up and running very quickly.

199
00:31:08.788 --> 00:31:17.318
More, more quickly than the current scenario, the current options. Just because it's easier to get the infrastructure in place for like five twenty-five megawatt sites versus one one hundred megawatt site.

200
00:31:17.348 --> 00:31:25.938
Or just the power availability. You can't go anywhere and get power. Like, if you went to-- There's just no way you're gonna be able to go in and plug in five hundred megawatts anywhere on grid right now.

201
00:31:25.968 --> 00:31:41.158
And, and, and I guess what my question for that is, is 'cause like just thinking about this like power for dummies, if we look at power as a reservoir, if it's five hundred megawatts split between, you know, twenty sites versus one site with five hundred megawatts, is that because of the transmission lines?

202
00:31:41.188 --> 00:31:49.538
Is it because of the infrastructure to actually deliver that power to the end, to, to the end user? It's not necessarily constraints for the production of that power, it's everything else that goes into it.

203
00:31:49.588 --> 00:31:59.428
Yeah, I mean, it, it, it's a network at the end of the day, so like you've got areas that have pockets of available power. It's just not as much as the five hundred.

204
00:31:59.568 --> 00:32:06.448
So it's sort of like the Bitcoin min-- that's where Bitcoin mining flourished, 'cause like we went to where- Right... no one wanted to go for that power.

205
00:32:06.498 --> 00:32:16.208
'Cause there's power available in those areas, um, that would be not available in one centrally located place. I, I was talking to, uh, uh, Braden Stover.

206
00:32:16.328 --> 00:32:22.808
He works at the Tennessee Department of, of Economic and Community Development. Mm-hmm. And he was talking about this specifically.

207
00:32:22.848 --> 00:32:38.008
It was like when they work with Bitcoin miners, they would tell them, you know, "Well, we have this rural area in Eastern Tennessee that is really sparsely populated, but maybe there used to be a factory there or a mill there, and it's no longer there."

208
00:32:38.048 --> 00:32:48.347
So there's kind of this excess block of potential power for that area that no one's using, and they would try to like, you know, nudge people in that direction and say, "You know, if you wanna set up, this is the place to go."

209
00:32:48.487 --> 00:32:57.288
Yep. Our profile mix is, are those sites. And, uh, those are the sites that nobody wanted in like maybe the last epoch because they were considered too small.

210
00:32:57.388 --> 00:33:03.088
But as machines became more efficient, you know, a ten or a twenty megawatt site now can produce a large amount of hash.

211
00:33:03.488 --> 00:33:14.508
And then same thing on the AI side, like eventually that would be desirable for someone who's building an AI cloud- Mm.... um, because the machines are getting more powerful and becoming more efficient.

212
00:33:14.548 --> 00:33:22.468
So it's just, it's just really it's a power availability play and a relationship game with where those, where that power is.

213
00:33:22.548 --> 00:33:32.698
Because, you know, it's different than just going directly to, you know, one, you know, going to TVA's boss and saying, "Oh, I need five hundred megawatts." That's not how the power gets distributed. Right.

214
00:33:32.888 --> 00:33:45.668
So if you have a relationship in a rural area where, you know, you've got your-- That's where I think there's a lot of opportunities for Bitcoin miners in those, in that, those smaller segments, because incrementally there will be more load at those places.

215
00:33:45.708 --> 00:33:56.328
If you've got twenty megawatts from them, you'll get another five. Mm. Like that's actually how we've, we've grown our, our fleet down, not our fleet, but our capacity down here was we had a ten megawatt site.

216
00:33:56.368 --> 00:34:05.588
The utility said, "Hey, we had a factory pass, pass and went to another area. We've got an additional fifteen. Can you take it?" I'm like, "Yeah, I'll take it." [laughs] So we did it.

217
00:34:06.028 --> 00:34:16.018
Um, and then, you know, and then I get a call three, four weeks later, a site that's down the road that we operate, ten megawatts, and they say, "Hey, by the way, we finally got the approval from TVA to get an addition.

218
00:34:16.208 --> 00:34:25.488
Can you take an additional five?" Yeah. I'll take it. And then if you just take like the breadcrumb approach and just start scraping these things together, before you know it, you can have these larger clusters.

219
00:34:25.528 --> 00:34:34.888
And I think within like Middle America and like call it the Southern Belt, there's gonna be a pretty, you know, there's gonna be a lot of opportunities, and I would assume there'd be like that throughout the country.

220
00:34:35.728 --> 00:34:36.288
Yeah, interesting.

221
00:34:36.328 --> 00:34:44.548
This goes back to what you were saying about maybe the most prudent thing for the miners to do now is just get the power, get your foot in the door somewhere because you're going to keep that door open. Yeah.

222
00:34:44.608 --> 00:34:46.338
It might open more doors down the road.

223
00:34:46.368 --> 00:34:56.048
And, and that's what's gonna be unique with like the-- if you, if we rewind it back to the conversation around all these m-machines, because like you're just seeing like prices drop on ASICs.

224
00:34:56.648 --> 00:35:14.328
So you're like, okay, do you need to have like an insanely profitable Bitcoin mining facility, or do you just need-- Like now you have like advantage because you could buy insanely cheap ASICs and essentially get them in and subsidize what that power, uh, is gonna do for you over a period of time.

225
00:35:14.448 --> 00:35:28.628
Now, it's not extremely profitable, but at least you have the options to, to do that, because like you said, there's such an amount of liquidation that's taking place right now that, you know, you, you-- I mean, there's S21 selling for like five and six dollars a terahash.

226
00:35:28.788 --> 00:35:37.468
Yeah. It's crazy. On that note, I'm glad that you brought up the ASIC market because I want two more questions on the mining front and then a, a few closing questions- Mm-hmm... on AI.

227
00:35:38.768 --> 00:35:45.917
What are the secondary effects that you see outside of the ASIC market kind of contracting right now for this expansion?

228
00:35:45.968 --> 00:35:52.138
Like, what kind of secondary effects or knock-on effects do you see the Bitcoin mining network facing as a result of this?

229
00:35:52.178 --> 00:36:04.008
'Cause the ASICs are one, and there even is a question of, you know, I've heard rumors that Bitmain might be shelving the air-cooled S23. That's unsubstantiated. We have not had good reporting on that- Mm-hmm...

230
00:36:04.018 --> 00:36:11.908
or good sourcing on that. Mm-hmm. So I'm not saying that's definitely happening, but does raise this question of, will they even be making as many ASICs in the future? Mm.

231
00:36:11.978 --> 00:36:23.428
Uh, but open question, like what are some of the secondary and tertiary effects of this AI boom that you see for Bitcoin miners? Well, I would say one positive thing is that it's gonna decentralize the, the hash rate.

232
00:36:23.668 --> 00:36:26.208
It's already happening. So that, that's overall a good thing.

233
00:36:26.368 --> 00:36:37.748
I think there's this whole movement of, you know, there was this ton of, there was this-- I, I, I use this quote in the panel where like Jim Barksdale, who was like the original founder of Netscape, which was like the browser- Mm...

234
00:36:37.768 --> 00:36:47.404
that was competing with Microsoft back in the day, he said, "There's two ways to make money in business."Bundling and unbundling. And initially, you know, all the mega miners shoot up all the hash rate.

235
00:36:47.544 --> 00:36:51.784
Now there's a great unbundling that's happening, and I think overall that's really good.

236
00:36:51.824 --> 00:37:00.544
And then at the s- at the other micro end, you've seen this rise of the desktop miner, and I didn't actually like think that that was gonna become a thing, but it, it really is.

237
00:37:00.644 --> 00:37:02.734
You mean desktop miner, you mean like a Bitaxe? Yeah. Yeah.

238
00:37:02.904 --> 00:37:15.004
Like all those, like even like the Brains Unit, the Canaan, like there's, there's a bunch of them that I feel like those weren't like the chip efficiencies weren't good enough in, in like the past couple years to make those things work, but now, like they're viable, they're little...

239
00:37:15.044 --> 00:37:25.984
these awesome little lottery machines that you can plug in. They're quiet. Um, and so I think it will become a tool for people to-- that's how they're gonna buy their Bitcoin now.

240
00:37:26.044 --> 00:37:38.464
Like, that's actually a really good thing for no KYC Bitcoin. Uh, so that, that's a good thing. I think business model-wise, I don't know about the foundry side of things, and, you know, if that will happen. I think the

241
00:37:39.764 --> 00:37:45.664
manufacturers really, all of them, all made a concerted effort to push into hydro because I think there was this...

242
00:37:45.674 --> 00:37:51.424
It was sort of like they were playing against the ESG native narrative of like, you know, Bitcoin's w- w- boiling the ocean.

243
00:37:51.484 --> 00:37:58.564
So they're like, "Okay, we need to be better about this, and we're gonna make these more efficient machines and do water cooling, and then you can use the excess heat and all this other stuff."

244
00:37:58.584 --> 00:38:08.984
And they just got too cute with it, and there really isn't a market for it because at the end of the day, the machines are not much more efficient than an air-cooled machine, and then the infrastructure you need is much more expensive.

245
00:38:09.004 --> 00:38:13.704
You're like doubling your cap space [laughs]. Yeah, and, um, so there's just not demand for them.

246
00:38:13.844 --> 00:38:19.744
And so unless you're building new greenfield sites with this a- with that infrastructure, they're-- this is not gonna happen.

247
00:38:19.824 --> 00:38:35.264
So those-- I look at them as sort of strategic errors, um, o- on that front because I think they were banking on big mining becoming bigger and bigger, and because AI kind of stepped in-- 'Cause those were gonna be the ne- those were gonna be the natural customers of that next wave.

248
00:38:35.284 --> 00:38:45.344
Mm-hmm. All the big miners were gonna go to h- water cool or hydro, and now they're like, "No, we don't- we don't need any of that stuff." And the air-cooled stuff is still what sells. So I don't, I don't know.

249
00:38:45.404 --> 00:38:54.844
I mean, will they respond? Typically, they'll respond to, to the market with that stuff. Yeah, but I guess I'm just sitting here thinking about like how can they respond in a way that doesn't just like

250
00:38:56.064 --> 00:39:03.184
undermine their entire business in the sense that like their, their, their main buyer is gone now.

251
00:39:03.324 --> 00:39:12.574
It's the public companies in the US, like they are not buying, so much so that like the only volume that I've heard from brokers, I mean, obviously it's some new machines are being bought- Mm-hmm...

252
00:39:12.574 --> 00:39:24.064
but the majority volume is used. Mm-hmm. The coupon market is totally dead. Yep. No one... They're, they can't move their inventory. You know, they had a supply glut with the S21 that came probably at the worst time.

253
00:39:24.113 --> 00:39:35.854
Mm-hmm. They had to issue coupons and rebates on a lot of the, on the, on the, for the miners who bought the e- vanilla S21 because then they came out with the S21 XP and the S21 Pro. Mm-hmm.

254
00:39:35.854 --> 00:39:48.184
All these miners were mad, and they ended up, you know, getting these newer machines at a discount. It just seems to me like there is a huge oversupply of hash rate with no place to go.

255
00:39:48.393 --> 00:39:54.624
I-- and there are still going to be some places, obviously, and that kinda leads me to my last question about how this all impacts mining specifically.

256
00:39:55.684 --> 00:40:07.124
With a lot of miners, even the smaller ones in the US, trying to think of ways that they can cash in on the AI golden calf, what mining do you foresee being done in the US going forward?

257
00:40:07.164 --> 00:40:17.244
Is it just on the margins with stranded energy? Is that basically it? What, where do you think the opportunities will exist, uh- Yeah. You know, two, three, four, five years down the road? Oh, gosh.

258
00:40:17.404 --> 00:40:25.544
I don't know anything that far. [laughs] I, I, I actually think-- Okay, okay, okay. So let's just, just, you know, we'll play like, you know, we- we'll, we'll... I'll tap my inner Marty Bent, and we'll get cosmic.

259
00:40:26.224 --> 00:40:42.924
I think like on the-- if you look historically at what's happened prior at times with ASIC manufacturers, specifically when they were in China, when the market cooled down, they were still producing ASICs, but they had the ability to just go plug them in and self-mine till the market picked up, and then they'd clean them up and sell them.

260
00:40:43.244 --> 00:40:54.584
That's what Bitmain did, a lot of companies did. Then they get kicked out of China, and that kinda went away. But then the market slowed down in like whatever, '22, something like that, and what did they do?

261
00:40:54.914 --> 00:41:04.604
They were, Bitmain was sitting on hundreds and hundreds of thousands of XPs that they couldn't sell. So what'd they do? They partnered with people who had infrastructure sites like ours.

262
00:41:04.644 --> 00:41:14.624
We were among dozens and dozens of like independent just hosting providers that they said, "Here, host our machines. We'll bring our ops team. You just provide power, and you're good."

263
00:41:14.944 --> 00:41:25.464
So there's gonna be all the big-- I, I look at least short-term, a lot of the mining will be self-mining from a lot of these m- mining manufacturers. Mm. That will happen short-term.

264
00:41:25.524 --> 00:41:35.914
Then maybe five or six years down the road, I think a lot of these HPC AI companies who are gonna say that they're pivoting towards that, I think will eventually will come back around. Because- Mm-hmm...

265
00:41:35.994 --> 00:41:44.084
it's still gonna be the power and the access to that, and I think there's gonna be some probably interesting opportunities that they're gonna say, "You know what? Maybe we don't want battery backup.

266
00:41:44.244 --> 00:41:51.444
Maybe we wanna use, utilize Bitcoin mining as a sort of like way to get a energy mix that makes sense for us."

267
00:41:51.844 --> 00:42:01.354
And that's gonna be like natural because you're gonna have embedded inside all of these AI infrastructure companies, Bitcoiners. So they're eventually gonna figure it out.

268
00:42:01.504 --> 00:42:10.384
Like, 'cause right now they're competing with each other. At some point, it'll be sort of like oil and gas, where when we used to talk, when I was at Great American Mining, we used to talk to the oil and gas producer.

269
00:42:11.354 --> 00:42:17.624
They're so, like they had one pre... You know, they were like Spiegel. They had the one precious, which was- [laughs]... the oil in the ground, and that's what they cared about.

270
00:42:17.664 --> 00:42:29.204
They didn't care about the Bitcoin or anything, and typically, the relationships that worked or the scenarios that worked is when it was like a supplemental or complementary component to their business, meaning like the flare mitigation.

271
00:42:29.264 --> 00:42:39.504
That's a easy thing. There's other parts where it's just like, well, maybe I'll get fined if I flare at all, so I have to consume this gas in some way, and if I can mine some Bitcoin, great.

272
00:42:39.524 --> 00:42:42.904
Maybe I can make liquids and the gas that comes out of the liquids, and we can do that.

273
00:42:42.944 --> 00:42:53.048
So I think eventually for some of theseThese companies that are emerging, when they sort of get more, you know, like grown up, they're gonna be like, "Oh, wait a minute.

274
00:42:53.088 --> 00:43:03.188
We can use this to, like, increase, like, flip a lever on efficiency for how our whole companies operate and give us an advantage." And, you know, then we'll see kind of like a shift back towards those.

275
00:43:03.208 --> 00:43:11.548
But I also just think, like, there's gonna be a glut of cheap ASICs that are gonna be spread out all over the place.

276
00:43:11.628 --> 00:43:22.888
And, you know, you've got all these different firmwares that are out there, and people will always find a way to find some cheap or free, you know, electricity. And that's overall good for the network. Yeah.

277
00:43:23.088 --> 00:43:32.138
I keep coming back to this thought that the first time I stepped into my first, like, huge cathedral for one of the big pubcos, I won't name them.

278
00:43:32.568 --> 00:43:37.868
But I walked in, and I just kinda-- It felt a little icky, and I'm not a communist or anything.

279
00:43:37.928 --> 00:43:44.848
You know, I don't think that you shouldn't be able to deploy capital to do what you wanna do as long as you're not screwing anyone over. Mm-hmm.

280
00:43:45.068 --> 00:43:55.778
But I just thought, you know, this is probably not long-term healthy for the Bitcoin, right? Like I, I would rather see this site disintermediated into a hundred other sites where, you know, every- Can end up.

281
00:43:56.048 --> 00:44:00.288
Yeah, exactly. [chuckles] Right? That's it. It, it's interesting to see that, that pan out, you know.

282
00:44:00.368 --> 00:44:08.248
I, I really do think-- I, I really liked leading with your insight about this being kind of like shadows of the China mining ban.

283
00:44:08.288 --> 00:44:21.508
It's almost like not, not the US mining ban, but the great, uh, reshuffling of capital towards other forms of compute. Hey, Charlie here. Guess what? We just announced our next Bitcoin technical conference, OpNext.

284
00:44:21.768 --> 00:44:24.828
That's right, y'all. OpNext is back for twenty twenty-six.

285
00:44:24.888 --> 00:44:35.128
We're running it back after a successful event at Strategies HQ in Tysons, Virginia, last year, and this year we are bringing it to the Big Apple at the iconic Times Center in Midtown Manhattan.

286
00:44:35.468 --> 00:44:47.168
We're hosting the big names and projects that you recognize, like Robin Linus of BitVM, Nick Jonas of Blockstream, Antoine Ponce of Chaincode Labs, and Calay of BitChat will also be present.

287
00:44:47.228 --> 00:44:59.017
And this isn't just for the devs. We have institutions talking with the developers. That's what OpNext is all about. We have Robert Michnik, head of digital assets for BlackRock, in the building.

288
00:44:59.328 --> 00:45:08.948
We've got folks from mining pools, investor funds, Bitcoin startups, and other groups. With a ticket, of course, you'll get access to all the high signal programming and networking you could want.

289
00:45:09.028 --> 00:45:20.258
You'll also get coffee, catered lunch, and access to the after-party at Pub Key. If you wanna go VIP, you'll also get access to the speaker dinner following the event and an investor brunch on Friday.

290
00:45:20.548 --> 00:45:31.908
Tickets are capped at three hundred, and early bird tickets are already sold out. If you wanna save yourself a spot, go to opnext.dev. That is O-P-N-E-X-T dot D-E-V.

291
00:45:32.608 --> 00:45:40.948
Use code PODCAST to save twenty percent off a GA ticket to the event. Ticket prices go up every few weeks, so don't wait, y'all. Lock in that ticket today.

292
00:45:41.248 --> 00:45:44.628
We'll see you April sixteenth at the Time Center in New York City.

293
00:45:47.168 --> 00:45:56.888
Last question for you on this, and I just-- I think we touched on this when I had you on, either at the end of twenty twenty-four or early twenty twenty-five, so I want your take on it again.

294
00:45:57.548 --> 00:46:06.058
What are you more bullish on, the PowerShell model or the Neocloud model for, for Bitcoin miners specifically? Um- And, and for listeners who don't know what I mean by that- Yeah.

295
00:46:06.068 --> 00:46:27.128
-PowerShell is what you see Core Scientific pioneering and a bunch of the other, basically every other miner, TeraWulf, Cipher, Hut 8 going for, where you build the infrastructure and then you have a hyperscaler or a Neocloud come and host the GPUs, versus the Neocloud model where you're actually building the site or, and running the GPUs, or running the GPU specifically.

296
00:46:27.268 --> 00:46:39.288
Eiren is currently building a site and hosting the GPUs for Microsoft- Mm. -in their most recent deal. So out of those two models, which are you more bullish on? I think there's a hybrid model too where you're not...

297
00:46:39.928 --> 00:46:52.368
I think they might call it like a soft shell model where, like, there's us providing the power and maybe say like the baseline, you know, foundation, and then there's an additional partner that comes in and does the actual infrastructure.

298
00:46:52.688 --> 00:47:00.208
Mm. Then that gets bolted on top with the, the end user. I think that one is probably the safest. It might not have the biggest margins.

299
00:47:00.517 --> 00:47:10.178
So you can see the risk tolerance that some of these groups have 'cause like, you know, maybe they have more experience in, in doing some of these other things. But I think, like, if you're focusing on,

300
00:47:11.208 --> 00:47:17.108
like, I don't know, may-maybe it's just 'cause, like, you know, if you build in Bitcoin, if you're not building, like, what are you doing? You're dying.

301
00:47:17.228 --> 00:47:30.428
So I think in order to, to continue to grow quick, you're gonna ha-- Like, doing those type of builds is really long. It's really sus-sustained, like, focus. You can't cut any corners.

302
00:47:30.928 --> 00:47:37.028
And so I think, like, we're really good. We're energy pirates. We're really good about going and getting power and, and, and getting it quickly.

303
00:47:37.128 --> 00:47:42.928
So I would fall for that third one, which is really like a, a subset of like the, you know, the shell version.

304
00:47:43.028 --> 00:47:52.668
I think that's probably the safest one, but also I feel like it's the one you can-- you're best positioned to... 'Cause, like, at the end of the day, like, the power is what they actually need.

305
00:47:53.088 --> 00:47:59.378
So if you can still own that power, it's like sort of like real estate of that- You have all the leverage basically. Exactly. Exactly. Yeah.

306
00:47:59.388 --> 00:48:05.128
And then that, you know-- But I-- Listen, everyone thought Eiren is the CleanSpark of this cycle.

307
00:48:05.568 --> 00:48:13.608
CleanSpark came out of nowhere, whatever, three, four, five years ago, and no, no one-- Everyone thought that their-- the way that they were doing things wasn't the right way. They weren't HODLing.

308
00:48:13.648 --> 00:48:19.958
They were doing everything kind of like counter, like, contrarian to what everyone else was doing. And I-Eiren did the same thing in this AI, said, "Oh, what are you doing?

309
00:48:20.008 --> 00:48:23.888
You can't do this, all of these, like, vertically integrated things." And then he's just like, "Yeah. Yeah, we can."

310
00:48:23.947 --> 00:48:28.688
[chuckles] And, and, like, you know, there's still things need to be fleshed out a little bit, but, like, good for...

311
00:48:28.728 --> 00:48:38.908
Like, I think that goes to your, your point earlier is, like, there's still this, like, innovation around business models that are completely untapped, and I think people are still trying to figure that out.

312
00:48:38.948 --> 00:48:51.068
And so that's what's really interesting to, like, go and analyze some of these AI deals because even if they have the same counterparty in some case, 'cause it's sort of like-Four or five groups that are essentially sharing, you know, the same counterparty.

313
00:48:51.748 --> 00:49:02.508
What do those deals look like? How are they different? How are they structured? Because, like, ultimately, you wanna be set up where you can continue to go build quickly more, and what allows you to grow faster.

314
00:49:02.948 --> 00:49:14.718
And I think, like, doing these mega sites, I mean, you know, if you remember back in the last cycle, Riot was dominating. They were bigger than everybody. You had Chad in his hard hat. He's doing his videos- [laughs]...

315
00:49:14.728 --> 00:49:23.248
and people are like, "This thing is incredible." And you can just see, like, when you do these mega si- it takes a long time- Mm-hmm... and you can't really focus on doing, you know, five at once.

316
00:49:23.708 --> 00:49:31.318
And I think, like, that momentum thing is there's something to it. And not to say that there's a right or a wrong way, there's just... you're gonna be different avenues for people to take.

317
00:49:31.868 --> 00:49:33.518
Yeah, and going back to the innovations part,

318
00:49:34.708 --> 00:49:46.868
it kind of speaks to what you were saying about maybe miners being mentally more capable for this work in the sense that they have had to figure out things as they go along for a new industry. Mm-hmm.

319
00:49:46.938 --> 00:49:52.568
And they've had to be swift on their feet, and they've had to be creative about how they are going to execute.

320
00:49:52.748 --> 00:50:02.908
And there might be an argument to be made that the traditional, you know, fat and happy, you know, Mag 7 companies have not had the... They haven't had their feet to the fire in a way. Mm-hmm.

321
00:50:02.958 --> 00:50:13.628
And so maybe they've gotten a little complacent about, "This is the way we do things," and a- as, as evidenced by XAI, we need to rewrite the script a little bit- Yeah, for sure... if we're going to shoot this movie.

322
00:50:13.768 --> 00:50:17.058
I even think about, you know, we made a decision. We got a pro...

323
00:50:17.068 --> 00:50:25.888
Like I was telling you, we got, we got offered an additional 15 megawatts at, in one of our areas, and one, o- one of the things that came up when that was offered was, "Okay, great.

324
00:50:25.908 --> 00:50:29.098
Can we get the same rate schedule as we have at the other ones?" "Oh, yeah, yeah, you're good."

325
00:50:29.108 --> 00:50:39.068
And then, oh, wait a minute, they're not allowing, you know, existing cu- customers to get grandfathered into some of the curtailment programs, which really allows it economically to work for- Mm... for what we do.

326
00:50:39.128 --> 00:50:46.388
So then we're like, "Well, what do we do?" And then there, then, you know, as we looked at the rate structure, off-peak, they had an off-peak version.

327
00:50:46.708 --> 00:50:56.617
So instead of us doing the curtailment and getting credits and participating, 'cause there wasn't an opportunity to do that for new load, we just said, "Hey, we'll take the load, but we'll only do off-peak."

328
00:50:57.088 --> 00:51:03.218
And because we have machines that allow us to do that, it gave us the opportunity to, to take that 15 megawatts.

329
00:51:03.228 --> 00:51:10.878
And at the end of the day, that power will be viable for somebody in the future at a rate that's doable for them. Mm.

330
00:51:10.908 --> 00:51:19.608
But bec- Bitcoin gives you, or the mining gives you the flexibility to take, you know, kind of turn, you know, chicken something into chicken salad. [laughs] You know what I mean?

331
00:51:19.728 --> 00:51:27.028
Uh, 'cause a lot of people would be like, "Oh man, I can't make that work." And it was like, "Well, could we make 83% uptime work?" Yeah, if the numbers are, are right.

332
00:51:27.388 --> 00:51:36.048
And I don't think a lot of people would be willing to, to do that. Um, and in fact, it's funny because, like, Grid initially, before they were bought by CleanSpark, that was the business model.

333
00:51:36.088 --> 00:51:42.768
They targeted these off-peak opportunities in TVA, and that's what they went after, and they built a really good portfolio.

334
00:51:42.778 --> 00:51:51.177
And when they sold their business, it was valued on the megawatts that they were able to, uh, provide, not just the uptime. Mm.

335
00:51:51.508 --> 00:51:57.628
Because eventually, you're, you do business with someone long enough, "Okay, okay, we'll get you in, into these programs now."

336
00:51:57.668 --> 00:52:03.528
But yeah, so I think it's just being about being nimble, and Bitcoiners are, are naturally gonna figure out a way to plug machines in.

337
00:52:04.028 --> 00:52:14.128
And then eventually, that power's gonna be viable for maybe a, a, a more, a competitive, uh, load in the future. The, the cockroach is becoming the butterfly perhaps. [laughs] Yeah. That's right.

338
00:52:14.808 --> 00:52:19.468
Tom, thank you so much, man. Oh, thanks. This was great. Likewise, man. Appreciate you guys coming on here for this event.

339
00:52:21.008 --> 00:52:35.368
[upbeat music] Hey, this is Charlie and Colin from Blockspace Media, and you're listening to the Blockspace Podcast, a show about emerging tech in Bitcoin, AI, energy, and markets.

340
00:52:35.628 --> 00:52:48.348
We publish two interviews weekly with CEOs, investors, analysts, and anyone else of consequence within these spaces. Plus, we have a weekly news roundup for all the important stories you might have missed from that week.

341
00:52:48.648 --> 00:52:59.528
The show is perfect for retail and institutional investors, analysts, and really anyone who wants to keep their finger on the pulse of the stories that are moving Bitcoin, energy, and data markets.

342
00:52:59.588 --> 00:53:08.028
So if you've stumbled across us, make sure to search Blockspace wherever you get your podcasts, and on YouTube, hit the subscribe button, give us a rating.

343
00:53:08.068 --> 00:53:12.658
We produce bonus podcasts and other content on our main feed, so you don't wanna miss that.

344
00:53:12.877 --> 00:53:22.828
And if you have any feedback or comments to give us or shows that you would like to see and topics you would like us to cover, hit us up at hello@blockspace.media. [upbeat music]