$iren Earnings Preview, Panel Discussion and open mic 🎤
Hosted by @₿itcoin ₿utcher 🥩 🐑 🐷 · 2025-08-24 · Tags: IREN
TLDR
The panel was strongly bullish on IREN ahead of earnings, expecting a meaningful AI-cloud, GPU-financing, or colocation announcement while debating whether management should close a deal immediately or preserve optionality for better terms. Participants highlighted IREN's power portfolio, Canadian air-cooled facilities, Horizon development, mining cash flow, management talent, and access to scarce AI infrastructure, while acknowledging execution, financing, litigation, location-specific demand, and dilution risks.
- Several speakers characterized the upcoming earnings call as a watershed moment and expected activity beyond routine guidance.
- The leading prediction was an expanded cloud strategy involving a large Blackwell GPU order, customer commitments, and new financing.
- Others expected a 50-megawatt Horizon colocation agreement, potentially including options for at least another 100 megawatts.
- The panel argued IREN should avoid leasing capacity to a competing neocloud when it can capture more upside through its own cloud platform.
- Participants disagreed over timing: some believed Horizon must already have a customer, while others favored waiting for construction risk to decline and megawatt values to rise.
- IREN's Bitcoin mining cash flow, growing cloud revenue, higher share price, and three business verticals were viewed as reducing near-term dilution pressure.
- Canadian sites were praised for their existing air-cooled infrastructure, low retrofit costs, and ability to support dense Blackwell deployments.
- Management's Macquarie and Goldman Sachs experience was viewed positively, although investor relations, the CFO function, and past convertible terms drew criticism.
- Settlement of the NYDIG dispute was framed primarily as a way to avoid management distraction, testimony, and discovery during a pivotal period.
- AI-driven electricity demand and tighter renewable development were seen as increasing the strategic value of IREN's existing power assets.
Speakers
- ₿itcoin ₿utcher — Hosted the discussion, compared IREN with peers, explored cloud and colocation scenarios, and emphasized financing access, optionality, renewable assets, and reduced dilution pressure.
- Speaker 2 — Argued that IREN is a premium operator with valuable power and data-center assets, expressed confidence in a major cloud or hyperscaler deal, and disclosed a long-term bullish position.
- Dolce — Contended that a 50-megawatt liquid-cooled facility coming online in the fourth quarter likely already has substantial customer coordination and predicted a near-term colocation announcement.
- Speaker 4 — Presented detailed predictions about cloud expansion, GPU financing, hyperscaler structures, Canadian air-cooled deployments, management talent, litigation, and renewable-power economics.
- John Austin — Discussed contract conditions, construction execution, capital-light colocation, management capabilities, CFO responsibilities, and the strategic rationale for resolving the NYDIG case.
- Speaker 5 — Declined to make specific predictions because of board-related constraints but praised the investor community and framed IREN as positioned at the convergence of major technological megatrends.
Notable quotes
- “I would even call it a watershed moment when it comes to their earnings call this Thursday.” — Speaker 4
- “But in terms of this announce-- I don't know if there's an announce, but if there's a deal to be announced or some kind of announcement made for this earnings call, I would suspect something related to the cloud side.” — Speaker 4
- “But after all this research, oh boy, I mean, I feel way more confident than ever.” — Speaker 2
- “So I would just have a real hard time believing that they don't have that in the bag right now.” — Dolce
- “I'm not, like I'm almost 100% sure that Iron has many, many offers, but for whatever reason, Iron believes that those offers are not commensurate with the strategic value of the site, right?” — ₿itcoin ₿utcher
- “Like there's, there's absolutely no point in doing a deal now.” — John Austin
- “That isn't particularly troubling because if the AI demand sustains, as we all expect, then power prices should be high enough for most of these products to come in.” — Speaker 4
- “I think we're at one of the most pivotal moments in the history of humanity and in the history of all technological development, where you have multiple major megatrends converging.” — Speaker 5
Transcript
₿itcoin ₿utcher: Sometimes you just throw the time.
Speaker 2: Don't spilling these rats as long as you're feeling to the day that I drop.
Dolce: I'll never say that I'm not killing them, 'cause why I'm not.
Speaker 2: Then I'm gonna stop killing them, man. I am not a whop, and I'm just not Eminem.
Dolce: Sublimin's thoughts when I'm gonna stop spinning them.
Speaker 4: Women are calling red cinnamon. Adrenaline shots.
Speaker 2: The life is still is just not killing up. The criminal's not killing hip-hop, selling the minimal swap to drop millions of black listeners.
Dolce: You come in with me, feel it or not, you're gonna fear it like I, stand in the spirit of God, that's enough.
Speaker 4: You hear it or not, living the shock, give it a miracle on. There's product of heart, that's enough. But this is the plot, that's enough.
Speaker 2: I'm done, feel the light, go out.
₿itcoin ₿utcher: All right, everyone. For those of you who don't know me, I'm Bitcoin Butcher, and I am from Detroit, Michigan, hence the Eminem background music. We're going to wait another minute or two. I have Kama as my co-host. If you guys could give me a thumbs up if you could hear me well, I'd appreciate it. shows that we're recording, and I'm excited to talk about Iron.
Speaker 4: Yeah, good evening, everybody. We're waiting for maybe Daltrey to join. And he has, it looks like, seen some technical issues. Franz might join later as well, but he's been away for the better part of the week. So I think he might be very tired. I don't know if he's going to be able to make it. Tonight, we're going to make it a very light evening, right? We're coming into one of the most important weeks for Iron. I would even call it a watershed moment when it comes to their earnings call this Thursday. There are a lot of signs that point to Iron doing something more than the usual earnings call of, you know, just giving guidance. And, you know, I think, you know, in previous, in some previous earning calls, Some of the expectations might have been muted. But given what the industry has seen in the past couple of weeks and the type of deals that we've seen, it leads me to believe there might be some momentum in this industry. So I would expect some sort of activity, some significant activity from Iron in this earnings call. So tonight, we're going to open it up, since we didn't really have any other guests on our program or space and all that. So we're going to open it up. We're going to let more people come onto the stage, and we're going to give predictions of what Iron is going to do for this earnings, and what kind of announcements that they can make, or what do you think Iron will make.
₿itcoin ₿utcher: All right, it looks like Dolce is almost ready. And yeah, anyone who wants to come up, send a request, Comma and I can approve it. I couldn't agree more with Comma. Super excited for this week. I thought it was interesting that earnings got pushed back a day to Thursday versus a Wednesday with Nvidia leading off on Wednesday the 27th. And I think they will kind of set the tone for GPU demand and kind of set the narrative straight as we head into Thursday earnings. And as Kamo was saying, I tend to agree that it's most likely going to be a cloud announcement, given the just moving parts with a co-location deal with a counterparty and agreeing on the sites takes a little longer where the cloud piece of it looks like they're a little closer on it from what we see from our bird's eye view. And looks like Dolce is on now. Welcome, Dolce.
Speaker 4: I think he's still trying to get on. Does anybody else want to come up? Just send the request and we'll approve you.
₿itcoin ₿utcher: Yep, sorry about that. Just to approve, Dolce, you should have access now.
Speaker 4: Okay, so while we're waiting for Dolce, I'm going to give my prediction first. Is that okay?
₿itcoin ₿utcher: Yeah, go for it.
Speaker 4: Okay, so we had TerraWolf sign what I called a hyperscaler proxy deal. So it's not direct with the hyperscaler. So the first deal that was signed was for the Bitcoin miners was Cores and Coreweave so that was not a hyperscaler deal that was a neocloud right a wannabe hyperscaler but a pretty big neocloud signing with a Bitcoin miner for the megawatts so now we have the next iteration and this is the hyperscaler proxy so it's not direct with the hyperscaler but the hyperscaler is being the backstop so Google is backstopping fluid stack in the deal with TerraWolf. So my question is, do we see, if Iron is going to sign a deal, do we see more of the same type of iteration that we're at now? I don't think Iron is going back towards the NeoCloud side, because I've written before on reasons why it doesn't make sense for Iron to sign with a NeoCloud because they are a NeoCloud themselves. So if they're taking that risk with a NeoCloud, so why take it with a NeoCloud when you know your own NeoCloud business the best. So that's one of the reasons why I never thought that it made sense for Ayren to sign with a NeoCloud. And that's probably you've never seen, that's probably why we've never seen a deal with Ayren involving Coreweave. Coreweave went to Applied Digital, Coreweave went to galaxy after cores, and they've never had anything downviron. So I think strategically, that was never an alignment there, right? Because in the future, you're going to be competitors, and you're weakening yourself strategically by signing your megawatts to a competitor. And you're assuming the same risk, but you don't get the same upside as the NeoCloud would. So if you sign with the NeoCloud, think about it this way, right? If you sign with Coreweave, they have they they can pay you they are able to pay you the the lease if they're successful they're able to pay you the you know the terms of the lease and they're still making hefty margins so if you view that this way an IronCloud is a NeoCloud why don't you sign IronCloud right if they're successful they can pay you the lease right and if IronCloud successful they will also get all the the profits of above the colocation lease. So that's why it never made sense for me to see Iron do something with Coreweave. So now the question is, we've gone away from NeoClouds and now we're at hyperscaler proxy deals. Does that continue? Right? The reason you have hyperscaler proxy deals is because the The hyperscalers are still not comfortable enough. They need an installation, a layer of installation before the risk gets to them. The execution risk right now would be squarely on Terawolf and FluidStack. So it doesn't come to Google yet. So Google is going to let them play out. If there's any-- if it comes to it and FluidStack defaults, then Google will be on the hook. But they have a layer of installation right now. So the question is, do we see those kind of deals moving forward, or do we go straight to hyperscaler-type deals now, right? For me, I looked at the different hyperscalers and other potential cloud partners. Right now, I don't see a path directly. I don't see it yet to a direct hyperscaler deal. And also, if it's a hyperscaler proxy deal, I can see one out there because I looked at the NVIDIA DGX Cloud left in partners and there is a candidate that Iron already has a good relationship with right that's the partner they're selling the white label GPU hours to but if it's a hyperscaler proxy then you also need the hyperscaler or a MAG 7 to backstop it so I don't know the relationships between Iron the you know their DGX left in partner and the the hyperscaler, if there is a hyperscaler relationship between those two companies to be able to pull that kind of deal off. But, you know, I also look at, you know, even though I say the hyperscaler deals are not here yet, I can see a path of iron signing with somebody like Meta or Oracle, right? That makes a lot of sense to me as well. So it's just a matter of what deals make sense the most. But in terms of this announce-- I don't know if there's an announce, but if there's a deal to be announced or some kind of announcement made for this earnings call, I would suspect something related to the cloud side. It would probably be a big order for more GPUs, just because we've highlighted so many times that Iron has such a big CapEx advantage of installing and deploying Blackwells, specifically the B200 series in Canada, because they can run it in the exact same configuration now in those data centers without any extra CapEx. The only thing they really have to do is rejig the way the racks are located and probably widen the racks slightly. And then you have essentially free infrastructure to plug in all these B200s into. So that's probably the way to go. And if you do that, your path to monetization is much quicker than a colocation deal or any other type of deal because Right now, when Iron is plugging in the Blackwells, those are accounted for right away, right? If you do a co-location deal, you're gonna have to wait until the data center is built and your tenant is in and operational before you see any revenue. But if you go with the CSP route, it's as fast as you can plug it in. And right now, I think Iron said they can transition, you know, these kind of, from A6 to GPUs, I think is, within a couple of months. So that's quite quick, right? That transition, that turnaround to revenue. So I would, if there's an announcement, I would be expecting it to be on that side. It would be a big, you know, a big order for GPUs where they have secured the financing. So on the financing side, it would be either something of the financial engineering route, whether it's debt financing, convertibles, or big line of credit to draw on. And there might be, if they do the financial engineering, then I suspect that much of the GPU hours or the GPUs they're buying is accounted for with a deal with their current DGX Lepton partner that they're selling the white label compute on. So that deal will be expanded. So they have probably two to three-year deals on a large portion of these GPUs, and they can use that to bring to the debt equity market to get the financing. So that's why I suspect if there's an announcement, that's what I suspect it will be.
₿itcoin ₿utcher: Thanks for that, Kama. I think what's interesting to add to what you had to say was while I tend to agree that it's most likely a cloud announcement given the purchase of GPUs announced in July and what we speculate as to ATM use most recently on days when it appears we're running away and then all of a sudden price action drops off and it's most likely to fund additional purchases. I would say, though-- and I saw a Pakistani hodl in here, and he's a big TerraWolf supporter. I think it's interesting, running in parallel to what Wolf did, that you saw that quasi-relationship with a hyperscaler, even though it's FluidStack. The power of that in the capital markets is immense, and they issued a convertible bond days later for, I believe, $400 million that was oversubscribed to either $650 or $850 million. But point being, even though a convertible bond is in some ways similar to an ATM in that there's eventual delayed dilution, albeit at a higher share price, the demand for that convertible debt was so much bigger once they had that brand name. And I do think it's important to, at a minimum, secure that relationship at Horizon with a brand name to enable better financing in the future. Because if we're going to build out two gigawatts of power at Sweetwater, you can bet that they can do it faster with the help of those kind of relationships, as opposed to self-funding via an ATM or waiting for the GPUs that they're installing to pay themselves off. But with that, Dalce, I don't know if you had anything you wanted to add, but we invited Super Saiyan up, I believe a fellow Canadian to Kama, who I've seen on Spaces more recently. So Dalce, unless you want to cut in, we're going to have Super Saiyan. I don't know if you had a question or comment, but the floor is yours. I have nothing to add, so you can go to Super Saiyan.
Speaker 2: Hey everybody, Butcher, thanks for hosting this again. Appreciate everybody coming in and listening in and great research by Comma again. He has been on a row. I mean, if you've been following him for the past few months, his deep dive, like, I mean, I remember since the earnings call, a lot of people clowned him for trying to... compare Core Reef to Iron, but now we see that Core Reef going down massively. It was at $148 now. I mean, the market cap, it's literally slashed, right? So I mean, they were clowning all of us, but then again, we stayed true to the data and the facts and everybody listening in also agreed and Mike especially holding the flag for us all. So it's nice to see. And also about Iron with the earnings coming out. Again, I do agree with Kamil that I think they might have something to announce with their cloud. It just makes sense. I don't think Iron is at a position to be rushing to get a data center deal. They just know they have the cards and it just doesn't make sense for them to make an announcement quick, get people rallied up. It's just they don't want to go the cheap way. I'm just going to put it that way. So Iron is a premium company right now, but the asset, the connection, the management that it has, it has a long way to go. And when I mean long way to go, it's just going to go higher, in my opinion, because Like if you just look at the demand, I would say I would put Iron first. I would put MBIS out there as well. I think they have good assets as well based on the demand that we're seeing. And I'm going to put Cypher up there as well. But I do think Cypher and Iron are very different in the category that Cypher could sign a deal faster and it would be good for them if they do. And Tyler was saying that they have tier one data center. And a lot of this demand are that whoever leases, they would want to build their own from tier one and make it customizable to however they want to run it. So that being said, I think iron is definitely my top one. I've been on iron and I love to say it all the time. I've been on iron since $5. I bought crazy call options at $9 and I'm still sitting in like my thing expires January 2027. So I've been my target's been around $50. But after all this research, oh boy, I mean, I feel way more confident than ever. And I'm not selling any of that. I still have those contracts. I might actually exercise some just to hold for the long run. But yeah, that's just my two cents. And if anybody wants to add in, please feel free.
₿itcoin ₿utcher: Yeah, we appreciate that. We all want unique perspectives. I think what I found interesting about what you were saying, Super Saiyan, was the contrast with Nevius. I believe they only have access to up to a gigawatt from what I've seen, and Iron has three times that. And then the cipher while I have a heavy call position for year end, hoping that Tyler Page closes the deal. I think what's interesting about Iron is Dolce, you know, with his legal background, we'll call it BATNA, which is best alternative to no agreement, meaning, as you're saying, super saying, like that you don't have to necessarily sign a deal out of desperation, whereas when you talked listened to Tyler on the cipher calls, there's more of an urgency to sign because they have joint ventures with Fortress and part of that joint venture with Fortress is in anticipation of closing a deal and then there's shared deal economics. And then Tyler Page has spoken to the franchise value of getting a hyperscaler and that given that Cipher won't operate as a NeoCloud and right now only Strictly wants to operate in a co-location facility as well as their Bitcoin mining. There's more of a significance to them signing someone, whereas Iron, in theory, could just buy more GPUs and plug them in. So I appreciate your take on that. I see we have Ed Vestor requested to speak. Ed Vestor, you're up.
Dolce: Thank you. How's everyone doing? Yeah, good deal. So, I don't know, I kind of go back to like the, 50-50 megawatt data center. I mean, that's a lot of that's a lot of load, right? I mean, I don't sometimes people don't realize, like, you know, how big 50 megawatts is, right? It kind of sounds small when you compare it to the overall pipeline and two big, but 50-50 megawatts is a substantial amount of load. And when you're building those out, right, you're not, you typically aren't just going to build that out without, without the, without some customer feedback and coordination along the way, right? And there's specifications that they built to. If you kind of look at the Terra Wolf deal when they signed with Core 42, there was a lot of design work, right, that went into preparing for that site and that build out. So, You could say on one hand that, Irons, they're building 50 megawatts and they're, they future-proofed it so that any customer could come in and do whatever they want. But I just don't see that likely, right? I kind of feel like they've been kind of working hand in hand with a customer to build out that data center. I think they're kind of to the point now where in their build phase where they're going to need that input. And so when you kind of look at the comments that Dan has made, right, he's spoke quite a bit about the demand for that site, right? And not just 50 megawatts, but beyond that. So I would just have a real hard time believing that they don't have that in the bag right now. And then if you kind of look at, you know, take Applied, for example, right? They built, they had 100 megawatt data center, that was coming online in October, right? And, I got the timing of that right, because when you look at the preparation needed to turn over a data center and go live, and, the CEO gave us a little bit better color in terms of, when it was going to be operational, right? So I think, So far, we've we've heard from Dan just to say, all right, Q4 Q4 delivery, but for applied, it was operational in October, so there's a little bit of difference in language there, but right, like that was about four months. They announced their core wave deal about four months ahead of October, right? So, you know, you got to think like there's you know, equipment ordering, you know, resource preparation, delivery schedule, like there's just so much to do to prepare for 50 megawatts that you don't really have time to wait to the last second to sign a deal. You just, you just don't, you don't do it right. And really, if, you know, I think we, I think we should probably also assume that, the data center isn't going to sit empty for a month, right? Well, while they, while they think it over, right? I think when it's ready, it's going live. The demand is just too great. So I don't think there's going to, when we say there's no urgency, yeah, there is, right? Like this is the only 50 megawatt liquid cooled capacity that's coming online probably for the rest of the year. I mean, Terra Wolf, I think they'll have theirs built, but like that's it. You know, I mean, it's very limited. So Somebody somebody's gonna take up that space, and I think they got to take it now, or at least, and I would say late by late December, or I mean, I'm sorry, late late September would be the absolute drop dead. Like, you'd have to have this thing signed and ready. So, that's kind of how I see that piece. You know, could they go the CSP route and, you know, with Horizon, maybe, but they've talked so much about the demand there. I just don't I just don't see that being likely either. Kind of going, look at my notes here. So kind of going back to the Google, and LiquidStack deal with TerraWolf. I mean, that was really ingenious, right? Because if you look at what Google was able to do is they were actually, you know, and I'm quite sure LiquidStack has promised them that capacity, but I mean, they were essentially able to get, what, almost 400 megawatts without committing any capital, right? Like they just backstopped it. They didn't have to put out any cash and still got a ton of capacity. So it was actually a brilliant, brilliant play because, you know, they, like one thing that's true about, you know, all these companies right now is they are under pressure with their CapEx spend, right? And the reason they lease these data centers out and the reason they, do deals like this is because they have to manage that budget. They have to manage that CapEx spend, right? I think if you see Microsoft, they were getting beat up for a long time because, they were spending so much. But once they showed they could monetize it, boom, all right, now all is well. But not all the hyperscalers have shown that they can do it as well. Right, so, there is a lot of pressure and there, I hear a lot of people say, Oh, why don't why don't they just build their own? Of course, they're all building their own, but the reason they the reason they lease buildings is because they can spread that cap out cost over time, right? That CapEx cost over time, right? So, and it gives you that scale, right? Like, you can scale much quicker if you if you're leveraging, you know, lease sites. So, yeah, I mean, I think, We got, I think, some, I think it's got, I think something's going to happen here, this earnings call with the deal. I think there's a deal. How that deal looks, I don't know. We could, I could, I think we could really see another iteration because I don't, you know, I don't think they were working, you know, I don't know how long it's been in the works, but I think Terra Wolf deals set the precedent on that front. And so I don't think they've been working on it maybe that long to, or I don't know, they could have been creative just the same and came up with the same thing, but to be seen.
₿itcoin ₿utcher: Great color. Yeah, Super Saiyan, I saw had a question or comment, sorry, what were you going to ask?
Speaker 4: Yeah, just before you go to Super Saiyan, I just want to jump in with a few points, right? So, you know, I really think Horizon is not going to be for CSP, right? It makes sense for them to deploy their CSP in Canada because there's just so much more a CapEx advantage because they can do the air cooled, right? The customers might not-- see, right now, there's a lot of demand for air cooling as well because a lot of people are surprised. But there is actually a lot of demand for air cooling because some people are not quite comfortable going with the liquid cooling, you know, quite yet, right? You know, you might hear a lot of people talk about liquid cooling, liquid cooling, but if you listen to other sources like Data Center Hawk, they talk a lot about flexibility, right? So having said that, I think it really makes sense for Horizon One to be a co-location site because that's going to showcase the rest of of Childress because then you have horizon two, horizon three, horizon four. So that's what I think is going to happen to Childress eventually. It's going to be all co-location unless somebody wants to step up and want to take all of it and do built to suit, right? But if you do all built to suit, it would probably make more sense to do it at a Greenfield site like Sweetwater. So Childress, it makes the most sense to be a co-location that is converted all to horizons. And then the CSP part will be done in Canada because it just has for the next three years or the life of the B300s that they can just throw into Canada. They just have such a big CapEx advantage of deploying black holes in Canada. If you're deploying Blackwells, if liquid cool at 80 kilowatt rack density, you're looking at $7 million per megawatt in CapEx for these data centers and up. But Iron can do these with air cooling. And these data centers are all built. You just have to convert the hot aisles and the racks the racks where they'll be, you know, mounted and that's about it, right? There isn't that much. And there's another point I wanted to touch upon, you know, is Iron and, you know, investor said, you know, they might have had something going on, you know, behind the scenes and he doesn't know. And the point is, I think Iron is very stealthy. So a lot of people just don't. see a lot of the things that Iron might be doing and they just um think oh these guys are behind you know they're taking the no they're taking the um the cue from some some other people but you know just an example of the stealth um that Iron has right if you know the I I wrote um something on um their new VP and it's not a new new person right their VP of commercial right his name is uh Trebini Lod, or I don't know how to pronounce his name, but it's L-O-D-H, Load or Lod, right? But the point is, he's been hired since October of 2024. So that is almost a year, right? Almost a year. And nobody's heard of him, right? And he's not a small fry because when I Google him, something very odd came up, right? No, the Google results keep on coming up with a article to the air trunk deal. And the air trunk deal is the biggest data center deal in Australian history, where air trunks got sold to Blackstone for, I think, 2.4 or $24 billion. I got to pull that article back up. But the point is, they have a lot of personnel and talent on the team and they don't really broadcast it, right? So a lot of people think, oh, these iron guys, they don't have a deal. They don't have the personnel. They don't know what you're doing. They're behind. I don't think that's the case at all. Because if you look at how much Dan Roberts, how many deals he's done in terms of monetary value, you add that to Kent Draper, who's done a lot of deals himself, right? And you add in, you know, now Trabini, I'll just call him Trabini, the VP of commercial. I'm pretty sure he was part of that deal when he was with Goldman Sachs. So he came from Goldman Sachs, right? And coming from Goldman Sachs and Macquarie, these guys have a lot of pedigree, right? If you look across some of the other Bitcoin miners, you don't see that kind of quality. Iron has a lot of quality in their personnel. But of course, I acknowledge they have some areas of weakness, right? Like we've talked about a certain position that they're weak and which is, it's not a secret, it's the CFO, right? We can openly say it. We are not thrilled at the CFO position, right? But they've addressed that issue. They've brought in somebody named Anthony Lewis, who came from Macquarie, and he was a co-treasurer of one of the divisions in Macquarie. So Iron's loaded with talent. They're able to make these deals, right? And going back to how far they've been doing, they had Morgan Stanley, right? I think they've been in a lot of talks for a lot of folks, right? It's just getting the right deal. So maybe the structure isn't correct or they can't get the type of financing that they want. But I think, you know, they're not starting from zero at this point, or I think they're very far ahead. It's just that we just don't know at what point they are in and what are the things that are causing the delays. But given how far back that Iron was looking to make these kind of deals, I think they're much further ahead or more in advanced stages than people realize. OK, then I'll hand the mic back to you guys.
₿itcoin ₿utcher: Yeah, so I had a I had a few thoughts on that and a follow up question for you, Kama, but let's get to Super Sane for a second because he's had his hand up.
Speaker 2: Yeah, like listening to you guys, I have a few points as well. Firstly, the 50 MW data center in Prince George for Iron. You guys also have to look at it in the way. that British Columbia, they're very, like, I know you guys are pretty well aware that some of the data centers are being rejected 'cause of community, political, environment, regulation, stuff like that, right? And I just feel like if you take in, Prince George is very close to Seattle and the Bay Area, right? And also the cost of electricity there is, Way cheaper. So now all these hyperscalers that are that and and if you listen to Coreweaves earnings. I'm just gonna cut it short. All all of it was OK. We have huge clients coming to us. They need this, this, this. And towards the end they kind of said it indirectly that how the cores deal will kind of fulfill that that demand that they're seeing, but now. We are at a point where cores might not even go through with that deal. And I feel like this chaos just creates more demand for companies like Cypher Iron. They're just sitting on gold mine. I mean, think about it, right? If you're a hyperscaler, everything, you want fast, cheap electricity to make sure your demand is met, right? And the demand that they're seeing is, it's crazy demand. I mean, So I genuinely think Iron's 50 megawatt and Prince George, and again, you also have to look at the political standpoint here. Canada and US has been on a back and forth with tariffs. And I can say living in Canada right now, our government put the tariffs down on the US side. So that gives a clue that things are getting better, right? And of course, also when companies, hyperscalers, and in US, they want to make a deal, they want to see the location, political side also matters. And when it calms down, you're just more reluctant, you're just, you just favor, it just feels like a more of a safer deal, right? And looking at the future. And also another thing, quarter four is coming alongside with earnings. And if you look at, again, core weaves earnings, They have to, I genuinely felt that they were heavily relying on the fact that the course deal will go and the next earnings, they're going to say, oh, the deal is going to go through. And again, when you are listening to the earnings, it's all about growth. Market doesn't care much about the revenue at the moment. It's all about what the guidance is going to be and the growth is going to be. That being said, all these hyperscalers right now in meta Google, Microsoft, Oracle, all of them know that, right? They know that there is a demand, but they somehow has to say they're working towards it that they can fulfill that demand. And that being said, this year end, I genuinely think you're going to see a bunch of hyperscaler deals. Wolf was one of the first one, but then again, right, like it was a good deal. Again, don't get me wrong, but The deal that Arin and Cypher are going to make, I truly believe it's going to be way bigger than the Wolf deal. Nothing to say that the Wolf deal was bad. I'm just saying based on the demand right now, it's just going to be higher. And also the rate cut situation is going down. Overall, I feel like it's just all leading up to a good position for Iron and Cipher to get a deal. And again, all of us here have been talking about it, but I just took it in with the political situation calming down, the Fed rate, and there's a high chance that it's going to get cut. That being said, and quarter four earnings are coming. for all these hyperscaler companies, and they want to make sure their guidance is great, not good, and going into the new year. So that being said, I'm very strong on a deal coming through for both Iron and Cypher.
₿itcoin ₿utcher: Totally agree. We're going to-- John Austin, we're going to get you next with Dolce had his hand up, probably related to SuperSaint's comment. So Dolce, why don't you take the spot here, and then, John, you're up next. Sure. Can you hear me? Just need to check. Yep, your connection is coming through. Yeah, it's pretty cool. All right, then. So if we're being honest, I think the main reason why we wanted a co-location deal earlier is because we were very afraid of dilution at a low share price. Probably just a couple months ago, we were $5, $6, and Iron actually diluted at $6 on average. So this was a legitimate concern. And we saw that in the short term, as in, you know, early 2026, Ironwood need several hundred million more dollars. And so if the market failed to acknowledge its AI HPC potential and give it a commensurate rerating, we would have all been diluted to health, right? So there was a significant amount of self-interest there. in wanting an early co-location deal, because if Bitcoin price did not rebound and if CSP did not amount to anything, which kind of seemed like it around March, then co-location was our only hope. Now the context and circumstances are very different, and that means that we are afforded the opportunity to think strategically about co-location timing. Do we want it in September? Do we want it in October? Do we want it in November? How does the supply-demand dynamic change over time? It's not so straightforward. And we're no longer so pressured because the stock price is 21 right now instead of six. I think it's actually really meaningful or valuable that you know, Iron has three different verticals, each of which is paying off to some extent. You have the Bitcoin mining, and Bitcoin is 115K, then I guess 112K. It's been around that price for, you know, quite a while now. And everyone seems to agree at this point that Iron's mining economics are really solid. And furthermore, Iron's CSP is growing. They've recently expanded their GPU fleet. And our ARR, once all the acquired black wells or purchased black wells are contracted, should amount to almost $100 million, if not more. Nine figures is a pretty important threshold. We also have, you know, peers like Coral Weave and Nevius who are receiving that rich valuation. And so when the entire market believes we belong in that peer set, we will start to pump. We don't need co-ocation in order to reach a fair valuation anymore. Now, I still strongly want it, but like we're no longer so pressed for it like we were in March. And so then if you have this luxury and you can wait, then is it better to have the co-ocation deal earlier or later? And my belief is that it should be later. It should be later because the, you know, availability of megawatts will only decrease over time. So in general, the value of the megawatt will grow and execution risk does matter. So with Wolf, they already were building, they already had the Wolf Den, which I think was a two megawatt pilot site. That was made probably a year ago, something like that. And then they've already been working on their Core 42 site, which is like, I don't know, 50 megawatts. I'm sure FluidStack had the opportunity to tour the Wolf Den and that 50 megawatt site, and that expedited the negotiation process. What about Horizon? Iron has been working on Horizon for several months, but I don't think it's you know, fully built. I don't think the electricals are finished. I don't think the liquid cooling is installed. So, there is still pretty significant execution risk there that can affect pricing and pricing really matters. If we're like, in order to have a closed negotiation or like a closed deal, you need the two parties to agree, right? There's a bid and there's an ask and they have to meet in the middle. And so, I'm not, like I'm almost 100% sure that Iron has many, many offers, but for whatever reason, Iron believes that those offers are not commensurate with the strategic value of the site, right? So if someone offered $1.4 million per megawatt or $1.2 million per megawatt, like that is still an offer. It's just not one that Iron would consider to be reasonable. So what about like 2 million? What about 2.5 million? Maybe Iron makes that counteroffer and then that's when the counterparty starts to bulk, right? And as long as execution risk is an overhang, then, you know, it's harder to reach that agreement. But over the next few months, you know, as people become more You know, as as it becomes more urgent to secure capacity and that execution risk evaporates because Verizon nears completion and then everyone can actually tour it and do their inspection to make sure it conforms to specifications. People are no longer as concerned about paying a fair value like you can. It's it's analogous to CSP where. before they have the GPUs commissioned, Irene can't find a customer. There's just not a single one. Then as soon as it's plugged in, they all pop out of the woodwork and say, I want a GPU, right? Because they don't want to deal with this like back and forth of. pre-leasing a GPU just to find out a month later that, you know, it never was installed because of delays, because of other issues, right? Like, no one doubts that Iron can build Horizon, but they might doubt Iron's ability to secure long lead items and install it on time. There are a number of things that can go wrong that go beyond whether Iron is technically capable of building a data center. And all of those factors, all of those risk factors disappear when the building is.
Speaker 5: Basically finished.
₿itcoin ₿utcher: So I think Dan is considerate of all of this. Like there's no longer this huge necessity to worry about dilution. And like he understands that the value of the megawatts and the value of Horizon's megawatts specifically are just going up. Why rush, right? Like it's not going to, they're not gonna realize any revenue from Horizon any faster from closing a deal in September. It's not going to protect them in any meaningful way in the event of like a natural disaster or an economic catastrophe, right? Like these contracts don't protect against those contingencies. So it really does very little. And then, and so what Dan is really thinking about, in my opinion, is how does this set Iron up to leapfrog Nabius and Coreleaf? How does, how can you maximally use Childress and Sweetwater, so that they become, you know, the biggest possible version of themselves that they can be, because they only get one shot, right? Like there's only one child, just one Sweetwater. If they **** this up, you know, that's it. All right, we're going to go to Ed Vestor who had a question, and John Austin, you're up next. Ed Vestor, what do you got?
Dolce: Yeah, no, just to kind of follow up with some of the points, right? I mean, The bottom line in the way I look at it is they got a 50 MW data center coming online in Q4. That's it, right? Like, and so you're not going to wait and to do negotiations until it's fully built and all that, right? Like, there has to be things going on well before that. when you look at the specifications of the data center, you're going to have to, you're going to have to build it to the customer to some extent, right? I mean, some things will be standard, but not all of them. And so, yeah, I don't see it. There's no way that they finish up a data center in Q4 and be like, okay, you know, let's talk, right? I mean, you kind of go back to Wolf Den and, you know, with Wolf, I mean, that was a 2 MW. I mean, you don't get much of A feel from 2 megawatts, right? They, and Dan even said, right, like he doesn't, none of his customers believe he can't execute, right? And so you look at kind of those comments and I don't know, to me, you got to believe that, they're confident that they can build a data center with the customer and deliver it in Q4 to somebody. So, but anyway, I hear you, I hear you though.
₿itcoin ₿utcher: Thanks for that. John Austin, it's your time.
John Austin: Yeah, great. Couple of, I guess, comments. Firstly, I mean, I agree with everything Dolce said there. I think optionalities clearly favours Iron with this. The other thing I'd say is there probably there would be some conditions to one of these contracts about execution. So, I mean, whilst there could be something well progressed, even if they were to sign something in advance of the the actual build and commissioning of the site, there'd be conditions favoring whoever the tenant is around making sure that liquid cooling is satisfied. They would get construction certificates, all that sort of stuff. So there's probably not much benefit in them actually going early on a deal anyways, because it'd still be get outs for whoever was involved prior. And I think that's probably the case with all these deals that have been announced in the sector, to be fair. No one's going to sign up, whether it's FluidStack or otherwise, no one's going to sign up to an unconditional deal that doesn't have get-outs for them not executing on the construction of these things. So whilst they can all be announced, whether they're actually going to proceed will depend on each of the parties' ability to execute. I think one thing Karma raised as well earlier that I was keen to talk about was the CFO's seat, which is a contentious one. And look, I agree with you, Karma, for what it's worth. I think the challenge for anyone taking on a CFO role in the organisation is there's a lot of CFO types in the business. So what is the role of a CFO? It's, you know, it's not potentially an attractive CFO role for a lot of strategic CFOs because you got a lot of that capability held elsewhere in the business. So really what you need is a financial controller and bean counter. which is effectively what they've hired. So they've got fit for purpose and someone who's pretty good on the statutory reporting. I think what everyone's realised is maybe investor relations isn't in a strong suit. That's fine. All you need is someone that's going to help them do the stat accounts and manage the financial control in the business, which is fine. The rest of them are effectively, everything's strategic from a capital standpoint. I mean, there's probably 10 people in there. that could do the capital side of the business. I think knowing these guys as well and, you know, a variety of the team in there, there's clearly a focus on preserving optionality on these deals. And there's no need to rush. And I think that's been, obviously there's a lot of people saying, You know, when's the deal coming, et cetera. Like there's, there's absolutely no point in doing a deal now. And you know, I think that the question is what, what is the mix in their portfolio going to look like 6, 12 months out, you know, with a colo deal at one, clearly CSP. I mean, I'm probably a little bit different to Dolce. I'd probably much prefer them to go deeper into colo just because. it's more more capital light. But you know, I think you look at the skills that are around the table now, obviously Anthony Lewis, you know, group treasurer at Macquarie, there's, you know, what's the balance sheet in Macquarie, 100 and something billion in kind of proprietary stuff and almost a trillion in third party funds if you add everything together. So there's clearly a capital bit of capital being managed there. And then I think about what else is sitting around the table. I mean, obviously, Triveni did data centers in Australia and Goldman Sachs obviously not only had a advisory role on that deal, but they also held some of the equity in that deal. So clearly he knew a fair bit about Air Trunk before it sold. So there's good experience there. I guess the question is, you know, how involved or otherwise was he? Clearly Macquarie was very involved because they held a majority stake, but I look at it all of the guys. I look at it with all the guys that are in there and it's kind of getting the band back together, which is no doubt means it's probably got a pretty good culture inside the business at the moment because getting the band back together and working with your mates is ultimately what everyone wants to do, right? So by getting all the mates back together, you know, I'd say it's good culturally and it means they'll all run hard. So it's good for shareholders.
Speaker 4: Yeah, I just love it.
₿itcoin ₿utcher: Tama, would you? Sorry, Tama.
Speaker 4: Yeah, I was just going to jump in there. Yeah. For the people who are not aware of who John Austin is, he's one of the seed investors in Ireland. So he's been in Ireland before it went public. So he has, you know, quite of, you know, good knowledge of Ireland from the early days and their history. And yeah, I agree with you. The way that the company came about and the personnel with, you know, it's very stacked heavy with people from Macquarie. And now we see with Goldman Sachs, you know, there's a lot of people who are very adept, you know, doing the CFO kind of work, like you said, right? But I guess the thing I had to say on that was Look, some people may have been disappointed on how, the terms, it seems, like the converts, the rates weren't quite as good. So I don't know if that's has more to do with, the assets that they hold or who they negotiate with. But with the addition of Anthony Lewis, I think it really changes some of that. Because I think the reason he was brought in was more specifically to look at the funding and the financial engineering side, that if there was any weakness, they're shoring that up. With that many people with finance backgrounds, I don't think Iron was really weak at that, but it might have been a perception. And if they were weak, they certainly have addressed that now with some of the personnel they have. on the team. It's just such a stacked team when you look at all the personnel on the team.
John Austin: Agreed. Karma, one thing I was keen to get your thoughts on while I'm here is, what do you, I mean, I see on Twitter X or whatever it is, all of this kind of talk about settling a, settling the Nidig case in advance of a deal. I mean, I've got a different thought to probably what's flying around X there. I think it's probably more selfishly that they don't want to put the team on the stand, as opposed to them actually thinking that it would preclude a hyperscaler from doing a deal. I mean, it's not massive contingent liability risk for them. We're not talking about a settlement that could derail the company. We're talking about a relatively small settlement in the context of the cash flow and revenue profile of the business. So I was just keen to get your thoughts, Karma, on what you think the reason behind a settlement would be other than typically no board wants their CEO and some of the other senior members of the management team standing up in court.
Speaker 4: You raised a very good point because I looked at the press release and having some background knowledge of the timeline of the case, the on the Canadian side, right, you know, that was already blocked. So I think they couldn't they couldn't question the the CEO or the board. Right. That was blocked in Canada. So that's why PwC went over to Australia. And I at that time, you know, we were speaking with the the Iron Bowes and we were really surprised that PwC was able to to get that in Australia to get the the testimony from Iron. Iron appealed that and they were denied. So it was coming very close to having Dan and the rest of the C-suite and possibly the board members testify. Like you said, I don't think any company would want to go through that. But you know, having said that, if you know, I think like Dolce and some of the other guys brought up first, right? They might think it's more of just getting this closure and moving on and reducing that counterparty risk to a future customer now, just get it over with. It could be because, okay, we're having this deal come up now and the battle is over 100 something million. And now we're expected to get more funds from a deal, right? So why don't we just settle out now and not having to go through that process and drag it on? No, that's what I think, you know, it's probably likely, but probably don't you have more? Yeah.
John Austin: It raises an interesting point though. I mean, the settlement, even if they lost in its entirety, probably still, I mean, it would be material, but it wouldn't be. you know it wouldn't destroy them um but I think the the points the points right I think it's more from a you know no public board wants their management team in the court um having to and I guess then what follows on from that is discovery uh I mean from what I understand is it's a pretty standard non-recourse asset backed facility that went bust and ultimately served the purpose of why you put it in there in the first place, which is non-recourse. So it's pretty standard infrastructure type deal that happened here that obviously, you know, they brought to that sort of structuring to Bitcoin mining when it wasn't in the market at that point. And they were the one of the only ones that really got away with the entire non-recourse. and then obviously it went belly up, NIDIG didn't like it, and has tried to pursue them in every court, possible pretty much globally, that it could have dealt with. So I think about it as more... Look, there's probably not a lot of risk. I think there's there's two things. If I was sitting on the board of this thing that I'd be focused on, it's how much distraction am I taking away from management at a key time? You know, the next six months, probably the most pivotal, pivotal, pivotal six months for the business. And then you've got, you know, Dan, Will, others potentially having to. sit in court for however long it is and give testimonies, et cetera. But then what flows on from that as well is discovery. You know, they when you go through these litigation cases, you sit, you sit there and you then get discovery. So NIDIG can then look at WhatsApp chats, you know, emails, anything that they really want. And who knows what they find then, which might have been pointless discussion, but it just opens up to NIDIG potentially finding more angles to pursue a case and then repursue it again and clearly just waste time. You know, if you can settle and get management free of those distractions, obviously that's a good thing, good thing for everyone, particularly if the settlement was relatively immaterial. So I think it was, in my view, it feels. more likely that it's that. But, you know, I think the point's still right, that clearly they're going to become more creditworthy. But at the same time, I mean, if anything, it goes to the structuring was spot on here because when it goes bust, you want it to be non-recourse and therefore you can kind of give back the keys and say, okay, it's your problem now, which is exactly what they did, and then 90 got the ****.
₿itcoin ₿utcher: I like that view, John. I was one of those people saying online that it might have been prohibitive to a hyperscaler signing a deal. But the more when you say that out loud and you consider, say it was $100 million and they're producing 50 to 60 million a month in cash flow just from mining right now, it would have been like a bad weekend out for them, whereas I think, to your point, the opportunity cost of time, given you're trying to sign new CSC clients and/or close the co-location deal, that it makes more sense for management to manage and not be witnesses. While we're on that thought, unless someone else had a question, just some other topics that I found interesting this week. We touched on the NYDIG settlement. Previously, just more background for people that are unaware, more recently, the head of sales for Iron Cloud had a tweet about moving to San Francisco, which is closer to Nvidia's headquarters. I thought that was interesting. That was a few weeks back. But what caught my attention this week and what I thought was interesting, given the company's emphasis on renewables, was the Trump administration, as of right now, whether it's a negotiating tactic or if it's an actual position stance, given their special interest donations from the oil and gas industry, recently came out with the intent to restrict permits for new solar and wind farm projects. and their explanation being the energy-dense nature of fossil fuels. But why I find this interesting, especially for a company like Iron, is they do have renewables under their belt. And given that there are a lot of corporate boards in their mandate that want carbon-neutral footprints for their businesses, And then I had sent a question to Brian Fry about this, and he confirmed what I thought to be a potential opportunity while there is this uncertainty if Burgum's going to approve, Burgum being the energy secretary, who was a former governor of North Dakota, if he toes the party line and denies solar and wind for political purposes, or if they actually believe that's best for the country, that's a whole other conversation. the remaining renewables that are available on the market become that much exponentially more valuable unless all these corporate boards want to change their mandate. And then there's the optics from the outside that they reverse and that they're dirtier companies, and then they'll get political pressure on the outside from environmental groups. So just thought that was interesting. And Kama, I wanted to circle back. You were You were referencing Canada earlier. If everyone's unfamiliar, the air-cooled facilities being able to run 70-kilowatt machinery as that for 80-kilowatt capacity is an attractive opportunity for iron, given the lower CapEx spend and the ability to retrofit. Can you detail, Kama? I caught the tail end of it, but you were chatting with someone about what Meta released today. And does that lead you to believe Meta might have interest in the Canadian sites?
Speaker 4: Well, I mean, if Meta is still interested in air cooling, then I definitely think there might be a match there, right? The air cooling in Texas would be a lot harder to get to-- 80 kilowatts rack density without more additional hardware such as no rear door he's exchangers or of that nature. Right. For air cooling anyways. So if Meta does want to still do it, you know, you know, with the air cooling, the way that I saw the picture was only a 20 kilowatt rack density. Right. That's that's not very good. Right. If they want all of Canada, I think I would be open to it. There's 160 megawatts total. Right now, the deployments for Iron's own CSP is less than 10 megawatts. So there is 150 megawatts that's available for co-location if Meta wanted those megawatts. And it would be very easy for Meta, right? It's very low risk. And that's what I brought. I wrote in one of my pieces. With Iron, they talked about this, the air cooling interest when Deepsea came about. So when Deepsea came out, a lot of the companies they were worried about the CapEx risk. I'm spending this much in deployment, and what if I don't need it? And especially when it's liquid cooled, you're spending upwards, minimum of $7 million per megawatts to build it out. And that's before GPUs. So that's a big CapEx commitment to do. If you're able to, if you're meta and you see what you're doing now, right, they are looking at different leases. They want to mitigate risk. And if you can co-locate at a location in Canada where you're essentially, you know, not having to worry about the capex of the build and you can get 80 kilowatt rack density into those facilities. I think it's a pretty, that's a no-brainer for me. If I'm interested in air cooling, that's a no-brainer for me to co-locate because those facilities have been proven. They've, no, the uptime's excellent. They had a customer in poolside, which left not because of the performance, right? They were raving about the performance of the clusters. They left because they needed bigger clusters and Iron at that time would be, buy more hoppers, and it would be suicide to buy hoppers at that time, right? They did not buy any more hoppers, and they started buying Blackwells now because they made sense. So if anybody's looking for air cooling for 80 kilowatt rack density, it's a no-brainer. But on the flip side, if you look at those facilities, those are so attractive for iron to keep, right, as the own CSP. that I don't know if they're willing to give that up. People used to laugh at the air cooling, but that's, for me, looking at the economic value of that now, those are very valuable infrastructures pieces that they have now because it's already built out. You don't have to wait. They're ready and they're proven commodities, right? That's the way I look at it.
₿itcoin ₿utcher: Thanks, Kama. I saw that, CryptoNic, I want to actually go to Kosh. Kosh, if you don't mind, I had a question I tweeted earlier today regarding the U.S. government position on renewables currently with respect to additional permitting by Secretary Burgum on wind and solar. And I just wanted to get your thoughts on those developments. If you're willing to come up or if you're unable to, we'll go to Crypto Nick. I'm just sending an invite right now to see if he's available.
Speaker 4: Yeah, I just approved them. Okay, cool. He accepted and I just approved them. We'll wait for him to get on.
₿itcoin ₿utcher: Perfect. Kash, thanks for coming back on.
Speaker 4: Yeah, thanks for having me, guys. Butcher.
Speaker 5: Pleasure to be on.
Speaker 4: I saw your
Speaker 5: question earlier today, and I've been busy touching grass for the better part of the weekend.
Speaker 4: So finally getting around to it. Yeah, look, I think federal, so offshore wind falls under federal jurisdiction, and that is why Trump will be, you know, a force to reckon with. on the wind generation side, offshore wind generation side. I think in many ways he is right or the narrative that renewables has been detrimental to the grid is true to some degree. In fact, we've seen that play out in Germany and the rest of Europe where power prices have actually gone negative because wind is generating Or and and yeah, wind primarily, it was offshore wind that drove it.
₿itcoin ₿utcher: But wind power prices
Speaker 4: went negative because they were generating so much with during periods where there was really no demand. And so power prices go negative. And what happens is that's, you know, they're priced to the marginal cost of production. And when you take that and you apply it to natural gas fired power generation or nuclear or or historically coal. Those are all really, really high CapEx projects that can't really sustain negative power prices because they're not subsidized in the way that the government subsidizes subsidizes in in North America. For example, your PTC production production tax.
₿itcoin ₿utcher: Not sure if we lost Kash. Kama, do you hear him?
Speaker 4: No, I do not hear him. Why don't we go over to the...
₿itcoin ₿utcher: There we go. He's back.
Speaker 4: Sorry, guys, headset issues. So the production tax credits effectively allow generators to produce at negative power prices and effectively get a full offset from the government for that. So on a grossed up PTC basis, that wind farm could technically bid at minus $50 these days and still be, quote unquote, profitable, if you can call it that,
₿itcoin ₿utcher: as a function of tax
Speaker 4: credits, which is completely insane. For an industry that is as mature as it is today, technically, wind and solar doesn't need those subsidies. Thankfully, The war on renewables, if you will, is coming at a time exactly when AI is putting upward pressure on power prices to a degree that hasn't been seen for several decades. That's all part of the iron thesis as well. But that is now going to be the primary economic driver for these projects, not the PTCs. So the fact that the recent OBB, the big, beautiful bill, The fact that it eliminates PTCs and ITCs beyond a certain period, it phases out in sort of 2027 to 2030. That isn't particularly troubling because if the AI demand sustains, as we all expect, then power prices should be high enough for most of these products to come in. So to answer your question succinctly, Butcher, offshore wind Federal jurisdiction sump has been anti- offshore wind for a long time because it ruins the view from his mansions. But look, it is what it is, and I think he's maintained his stance for over 20 years. That's going to probably stay. At the state level, the impacts aren't as bad. I mean, the PTCs are being phased out under the current administration. the chances of that actually coming back in the next administration, we'll have to see. If the renewables industry needs that support, there is probably a good chance it comes back. But again, that's a bit of a guessing game. It's been so for administration, administration for the better part of 10 years now.
₿itcoin ₿utcher: Thanks for that, Kash. I think as a follow-up question, it's just strange to me. It feels political in nature. I mean, everyone's heard the sound clips about him complaining about the birds flying into the wind farms. But when you have people like Elon Musk, and to be fair, Elon Musk has self-interest with his battery storage systems. I think what I'm most surprised by all this, and maybe You can, I'd like to hear your thoughts on this, but given the success of Texas ERCOT and companies like Cipher and Iron tapping into this negative price energy, like there's already a proven use case to monetize it and grow the grid and take advantage of this. And if power is the bottleneck of all this AI productivity, the idea that we're going to abandon wind permitting and solar permitting, like subsidies are a whole other conversation, but to just flat out say that we're not going to use either of those seems short-sighted in my humble opinion. So I guess that's not a question, but if you had any thoughts, Kash, otherwise, we had Crypto, Nick, who was going to come up, and then I saw Mike Alfred join the room. If he wanted to speak, I can invite him up as a speaker.
Speaker 4: So just on that point, offshore wind is a mega infrastructure project. These are long gestation cycle projects, and I don't know if offshore wind will be a panacea to our problems. But to your point, really, we need every bit of power we can get to solve our issues. And so thankfully, on the battery front, the incentives haven't fully fallen away. They fall under a separate class. It's wind and solar that have been the ire of the Trump administration and totally makes sense. Or rather, it's happenstance, but it would make sense to do away with the production tax credits and ITC's investment tax credits at the current stage. I do think power prices, you know, The best solution for low prices is low prices. And the best solution for high prices is high prices. So inevitably, as power prices tick up, it'll solve a lot of problems. And then once they get to an unsustainable point, they will yield another set of.
₿itcoin ₿utcher: Solutions to bring the price back down.
Speaker 4: So it's going to be dynamic.
₿itcoin ₿utcher: Thank you, Kash. Let's take it to CryptoNic. What do you got for us? Crypto Nick, you still there? Okay. Anyone else want to come up? So Mike was in the room. I saw Crypto Miami actually in the room and it's been nice to hear your views lately. Could you, if you don't mind Crypto Miami, if I invited you up, if you could detail kind of in parallel, add to some of our understanding of the Terra Wolf deal and how Iron might be able to utilize that going forward as they're trying to close their co-location deal. I'd be interested in hearing your take. Let's see if he's available to speak, just send an invite. Okay, it's not. Comma, what other topics do we want to touch on this evening going into earnings next week? I had said earlier, you know, I think it's important to see where GPU demand is from Navidia's standpoint on Wednesday night. And I think that will be a really big precursor into the excitement that heads into Thursday. And then we kind of detailed earlier our expectations of cloud announcements as opposed to co-location.
Speaker 4: I think we've covered a lot, quite a bit tonight already, right? I was hoping more people would come up here and give their predictions or their thoughts, but I think a lot of people are a little hesitant to come up. So why don't we ask, yeah, Investor or Cash, what kind of predictions do they have in terms of announcements if Iron was going to make an announcement? Investor.
Dolce: Yeah, no, I, you know, I think they announced a co-location deal. 50 megawatts with the option, to for another 100 at least. I think the timing of it is right and kind of as I pointed out before, they're bringing a they're bringing a 50 megawatt liquid cooled data center online in Q4. And really for me, that's all I need to know because it's not going to sit there empty. It's just not. You look at the schedule that they put forth, right, to get this done, right? They're trying to meet a deadline and they're trying to meet a deadline for a reason. So I'm expecting, you know, an announcement. If we don't get one now, I would say again at the latest by end of September. But I think the timing is prime to announce now.
Speaker 4: Well, regarding the timing, I think you do bring up a point. Because we have noticed, what we have noticed is Iran is hiring a lot of people for a night shift. And you don't hire a completely, it looks like from the personnel, a completely new night shift, right, with all those redundant personnel of the day shift. So they might be really stepping on the gas now to complete something. So there might be something there.
Dolce: Yeah, no, definitely. And just another comment, I've been hearing a lot of talk, right? And I just wanted to clear something up, right? Demand, just because there's demand doesn't mean there's demand everywhere for every. AI workload, right? So it is very specific to the company. It is very specific to what type of workload they're running, right? Whether it be training or inference or, you know, so just because there's 100 megawatts available, you know, in Canada doesn't mean. that, there's demand signals there for that, right? And so hyperscalers, they have to be very careful not to kind of over-provision, right? Because if, you over-provision, then the next thing you know, you're needing to retrofit those data centers, and that comes at quite a cost, right? So anyway, like there's a lot that goes into the timing and the location, right? Inference is very sensitive to latency. And so there's a lot to get right about that. So, it's, just, you might get one hyperscaler that has a demand signal in this region. there might be another one that doesn't, right? And so, it, again, it's highly variable when it comes to that. And that's it.
₿itcoin ₿utcher: All right, look, sorry, Ed. Guys, I want to go to Mike Alfred, given his position on the board, see what Mike had on his mind, and then we're going to do our best to get to everyone else. Mike, it's all yours.
Speaker 5: Hey guys, listen, as usual, I have a lot on my mind. I'm thinking about Napa cabs. I'm thinking about first growth Bordeaux. I'm shopping for white burgundies. I can't make any predictions for obvious reasons, so I won't be providing my predictions. I've greatly enjoyed reading everybody else's predictions that I'll just tell you. It's fascinating. You learn a lot about people. You start to categorize people in your mind in terms of the way they think about the world. And so I want to mostly come up and just say thank you to Butcher. Thank you, Kama Sutra. Thank you, Kash and all the others, Dulce and all of you guys. You guys are incredible people. incredible analysts, incredible investors. And look, I'm here for the same reason you guys are. And, you know, to the extent in which there's a difference, it's just because I signed up for a bit more, you know, a bit more than many of you probably would want to bargain for. And so like, I don't take that for granted. I don't think it's something that should be just pooh-poohed away. You know, it's a significant responsibility, but at the end of the day, I'm here for the same reason you are. I think we're at one of the most pivotal moments in the history of humanity and in the history of all technological development, where you have multiple major megatrends converging. And I think we all in this space are sitting right at the middle of it. And I think the market is largely starting to figure this out. I mean, I've been saying for two and a half years that Iron is operations oriented, operations first. I said repeatedly Bitfarms and others would, Riot, Marathon, et cetera, would all miss all of their forecasts. I got yelled at. I got told I didn't know what I was talking about. And I said, nope, I'll be vindicated. I will be validated. It may take a few years, but I have more endurance. more duration to the way that I work and think and run, et cetera, than almost anyone else I've ever worked with or been associated with. So I don't worry too much about people with really short attention spans. I don't worry too much about chart squigglers or kid analysts or people who draw lines on a chart and say, it violated the line on my chart, therefore I'm out. Like those people are out, right? How many of the people who primarily use charts are legitimately still in iron? right now after the last two and a half years. Like I would argue almost none of them, like there may be a few, right? But most people got blown out. And so we finally closed at a new 52 week high, perhaps the highest close in a full trading week since the IPO. If I remember correctly, we went public the week before Thanksgiving and then Thanksgiving was a partial week. And literally by the end of November of 2021, the stock was already like under the current closing price. So we're in kind of new territory. Like the stock didn't spend very much time above $20. Nobody knew what it was when it was above $20. And it was only there for a very short period of time during holiday shortened weeks, just before the market really started to crash. And so like, look, 2022 is a crap year. for mining. The cycle's been kind of a tough cycle for mining, but again, if you understood the value of the energy and the infrastructure, the megawatts and all these trends converging, there was never really any reason to be concerned, right? Like if you understood the balance sheet of these companies and you understood how many, you know, potential use cases there was for the power, there was never really any reason to be concerned. So, you know, here we are. We've got we've got earnings this week. As I said, I'm not going to make any predictions. Obviously, that would be super dumb to do that. But what I'll say is, you know, look, I'm very excited about the future. I'm very excited about what what's being built. You know, I think I've said this before, like I think this can be a generational style company. This is something I've said repeatedly for a long time, so it's nothing new here, but. You know, in spaces when we're at two, $3 and I said it could be a $200 stock. And it's it's pretty surreal. You know, when you get to 20 and all of a sudden you start to see more people who realize what I've realized for a long time, which is that there's a tremendous shortage in power and infrastructure and that there's no reason rooted in physics why a company that's built correctly in this space can't get to a trillion dollar market cap. you know, it doesn't need to get to a trillion dollar market cap. I mean, even a 50 or 100 billion dollar market cap would be quite surprising to people who were around in late December of 2022 when, people forget this now, the stock was trading between 50 and 100 million market cap. Now, there were only 54 or 55 million shares outstanding at that time, so it was a very different world. But like, honestly, not that much has changed. The strategy is still the same. The core team You know, the people that have made a lot of decisions to get the company to where it is today is still there. And the operating metrics and the operational capabilities haven't really changed that much. You know, the company's always delivered on time and largely within budget on almost everything it's done historically. And so the only thing that seems to have changed is that the market is catching up to where the company's mindset was two or three years ago. And some could argue five years ago, because yes, it takes a while to build physical infrastructure, but the concepts, the ideas, the strategies have been in place for four, five, six years now. And now you're just seeing the kind of manifestation, the materialization of those ideas, and the market is responding as such. And so, as I've urged for this entire cycle, patience is required. These businesses are harder to build, right, than like a frictionless, like launching a new token or, right, doing some sort of asset light business model where you just use the capital markets to accumulate an asset, try to monetize it directly. Like those business models look easier and they look sexier and they seemingly grow faster, but there's few barriers to entry. And what tends to happen is across the full cycle is the flaws in these sort of, asset light, no moat style business models get revealed over time. And so I'd much rather do the hard thing first, right? Build the hard thing first, take in the capital, deploy it into hard CapEx style stuff that can't be replicated easily, and then monetize it along a longer arc where it's really hard to dislodge you because it's almost impossible to do it in the physical world. So I like that strategy. It's slower. It seems inefficient initially, but then later it's not. And it mirrors a lot of other similar industries across the decades. Again, as I've said many times in these spaces, like most people just don't have the perspective. These haven't been around long enough. They haven't seen this style business model in another sector adjacent play out. So they don't realize that like what somebody says in any given quarter when they don't have any life experience and they're relying primarily on a chart, like it doesn't really matter. Like it doesn't have anything to do with like what's going to happen. So again, I'm quite bullish about the next three, four, five, seven, 10 years. I think AI and Bitcoin are both multi-decade trends. So I say seven or 10 years just because like it sounds crazy to say you're going to be here for 50 years, but there's no reason to believe that this isn't a 50-year story. right? Like show me one shred of evidence that says Bitcoin is going to be less important to society in 50 years from now. Show me one shred of evidence that says AI is getting less important and will be less important 10, 20, 30 years from now. I suspect society may change so dramatically that our mental models today largely won't apply at that point. But the one thing I think we can say with certainty is that power will be required to manifest any of these things across the full arc of time. And so again, I don't know how to pick lottery tickets, right? I don't know how to do scratch offs. I don't know whether OpenAI beats perplexity long term or Meta catches up to Google or Google catches up to Meta. I don't know how to pick horses in that race, but I am very confident that as long as anyone's participating in that race, then someone's going to need power. and someone's going to need infrastructure. And it appears that that shortage is becoming more clear to the rest of the market in real time. And, you know, I like what I'm seeing. So I wish everybody the best of luck along the journey. Again, I'm largely here just to say thank you to the group of you guys who run these spaces and uncover stuff. As I've said many times, guys like Franz and Kama in particular, you guys, are constantly posting tweets where I'm reading stuff and I'm literally telling you, like, I know pretty much everything you could know and you still educate me, which I love. And again, a lot of that's just like level of granularity, right? Level of fidelity. There's just like, there's just a lot to do, right? Like you see the end result, but there's a lot to do along the journey to get there. And everybody has their own role, right? I'm not an expert on satellite imagery. Right. I'm not an expert on like every piece of electrical infrastructure. Right. I'm I'm not an expert. I'm becoming an expert, but I'm not an expert on every model of ASICs. Right. And every model of GPUs. Like, I'm just not. And that's OK. Right. Like, I think I think it's important for people to admit what they are good and what they aren't good at. But what I am good at is is highly valuable. and has been an important part of getting us here and hopefully an important part of getting us to where we're going in the future. And we all play a part in that journey, right? Each and every one of us who's engaged. It may feel small at different points, right? Even when you're on the board, it feels small at certain points because the task in hand is so big that there's only so much any one person can actually add to the equation. It really does take a team, and that team is both inside and outside. And as I've said before, Like there's significant advantages to being outside, right? Especially when armed with the alpha that Dulce and Comma and Franz and Butcher and guys like that are dropping. I mean, I read their tweets, I'm telling you guys, these guys have significant alpha. But the nice thing is they're not on the inside. So there's no SEC regulations affecting what they do, right? There's no rules, there's no form fours, there's no compliance approval, right? There's no trading restrictions, there's no blackout windows. et cetera. And so count your blessings, right? You're at a generational moment, potentially generational company. You're getting generational alpha from these guys and you can do whatever you want with that information. Not everybody can. So anyway, I hope everybody has a great week. I'll stick around and listen and continue to support the space. But, you know, I wish everybody the best of luck on the journey.
₿itcoin ₿utcher: Mike, we really appreciate you. I've learned a lot from you. know, I had a finance background, but I jokingly talk about my family business or butcher shop. And if a butcher can understand the need for power that Mike Alford brought to my attention, then the least I can do is... do my best to make more people aware of that, because we're at a crossroads right now where we have unsound money, and we have a productivity gap that needs to be filled by AI, and iron is the perfect solution to both of those problems. We got CryptoNick. Let's see if you can come back up. Otherwise, that might be the fitting way to close the space. CryptoNick, one last chance. Going once, twice. Okay, everyone, we appreciate you. Thank you so much for attending. We'll try and recirculate. It's going to be a great week. We're excited. Mike, I know you can't say anything given your position, but something tells me You're super bullish and you're trying to contain yourself right now, but I don't want to jeopardize your position.
Speaker 5: You guys cracked me up to no end. I said FAIR. I said FAIR as a response to a tweet and someone said, Oh, that's an acronym for Facebook. You guys are just like wild. You're just hilarious. You read into everything and I just love it. But I've been, to be 100% clear, I've been bullish. on iron since 2021. And technically, I've been wrong because I was bullish on iron at $28 a share, right? And I was bullish at iron at $1 a share. And the haters will remember that I was bullish at $28 a share. And the people who love me will say, Mike was super bullish at $1 a share. And they're both true. But again, along a long enough arc, it doesn't matter. So anyway, I'm just as bullish as I was last week and as bullish as I was three months ago and six months ago. I do prefer to buy stocks when they're really out of favor. I think everybody knows that. But I'm just not sure it's going to matter, right? Like if companies execute, their stocks can go up for a long time. And I think we've been conditioned as an industry to expect every rally to retrace because it's been such a brutal cycle like that. Every time there seems to be a breakaway rally, everything gets slapped back down again. But as I've said over the last six months, I think at some point that regime flips and then you do start to get the Nvidia or Palantir style run where like it just doesn't retrace. And I don't think the cycle can top this time around without something like that. So that's just like my view on the regime and the technicals. Like I just don't see how we top with like this sort of like heavy overhang negative sentiment like we've seen on this whole sector. and then I also don't see how that happens when like core weave is still valued at 50 billion and Companies that have a better line of sight to growth and sustainability in the future are training in a tenth of the valuation like I just don't I think a lot of that's hyper air or or You know everything is just undervalued. I don't know but we'll find out I mean I've been talking about this for a while and since since I really hammered on it, Corweave's gone from 80, 90 billion to under 50, and iron's gone from 1.5 to five. So it may be slow initially, but I think the market is catching on to that thematic.
Speaker 4: Mike, you're the Eric Jackson of Corweave, or the inverse Eric Jackson for Corweave, and how poorly that stock has or rather, well, it's retraced to, you know, in line with your expectation.
Speaker 5: So well done. I think that's a compliment, but it feels kind of backhanded.
Speaker 4: No, no, there is no backhanded there. Like your view on the Coreweave delta is very prescient. So no, well done on that one.
Speaker 5: Thanks. Well, I try not to be too negative, but sometimes I see stuff that's just excessive. Right? It's just very highly unusual. And I don't short or even try. I mean, you couldn't short very easily because there's no borrow, not with like most brokers right now. So that wasn't really something that I was going to trade around, but just intellectually, I found it very odd. And I think sometimes you're wrong about this thing. Think of all the people who thought they were smart shorting Tesla over the last 10 years. And they may not have been wrong from a classical fundamental perspective. There were things about Tesla's financials that didn't look quite right, but it hasn't mattered because the overall space is so big for them. And then they keep kind of morphing beyond people like they want to pigeonhole them as an auto company, but are they an auto company? You could probably make a legitimate argument that they're more of a. AI and robotics company now. And if that's true, then maybe there is justification for saying it's undervalued. And I think the same thing is possible here with even in a Coreweave where I'm not like necessarily bullish on the current iteration of the business model at 90 billion market cap. There's still a legitimate chance that Coreweave is a trillion dollar company. Like it's not impossible for that to happen. And the main driver of that will just be the size of the ultimate size of the AI compute industry. And actually, in a sense, even though I'm short-term bearish and I'm a short-term bearish at the extreme valuations, I'm not necessarily mid to long-term bearish because it's hard to be bearish on a sector that has the tailwinds that AI compute has, right? And in a sense, like as an iron investor, we should all be sort of rooting for Coreweave because the higher Coreweave goes And the better iron executes relative to Core Weave, the more certain future valuation increases for iron become, just statistically. And as a value investor who thinks very systematically like that about how these things tend to play out with a lot of patience required, again, I'm not necessarily rooting for Core Weave to go down a lot more. I just wouldn't be surprised if it happened. I could see a scenario where CoreWeave goes all the way down to $25 or $30 billion and then still ends up at a trillion. I could see a scenario where CoreWeave turns on a dime right here at $85 or $90 a share and goes straight to $500. I could see a scenario where it just bleeds out all the way through the rest of the cycle. And the AI market turns over in six or 12 months and CoreWeave never makes a new all-time high. There are many different paths that that could take. But again, as long as they don't blow themselves up with debt at the wrong time, which is, it's a real risk for them because they have borrowed a lot and they probably need to borrow more. As long as they don't blow themselves up with debt, I think long-term, anybody provisioning compute to large customers would be fine. And so that's just the risk. I think, you know, a good comparison might be Core Scientific last cycle. Core Scientific had some good assets. Clearly they had some good assets because Coreweave is probably going to end up owning them eventually. But Core Scientific managed to go bankrupt because of the timing of their borrowings relative to their cash flows and their forecast around the key dependency, which was the Bitcoin price, ended up being wrong. So they were overlevered at the wrong time. And so even though they were the largest producer with some of the best physical assets, they ultimately ended up having to effectively be put through bankruptcy, although the common shareholders survived somehow. It's a miracle, actually. But it happens, right, especially when you have founders or a lot of shareholders that have bankruptcy expertise. I mean, Barry Silver to DCG is another good example of that. He's a bankruptcy expert. And in spite of what happened with Genesis and all that, DCG is still standing. And I actually like Barry quite a bit. And I think it's DCG is a great organization in a lot of ways, but I think a lot of people are surprised by that. And they're surprised that, of course, scientific common shareholders still have stock. But that happens, right? But Coreweave is sort of, in my view, like that player, this cycle where like they've got to dodge whatever the next downturn is, they've got to dodge that. And if they do, then like they'll be fine because I think the space is big enough for a lot of winners.
John Austin: Mike, just one for me question. If you were Coreweave, why did they stumble so much on their first really a company making acquisition that they need to make in Core Scientific to give them the vertically integrated stack. I mean, obviously they're debt laden, but surely it's important enough to raise equity to become vertically integrated. And the deal that they put to Core Scientific was absolute rubbish and was never really going to get voted up in the context of a volatile. And as you mentioned, the probability is to the downside, not the upside on the core stock price at the moment.
Speaker 5: I mean, I think a lot of CEOs of companies are sort of optimistically delusional, like by design. And so I think the CEO of Coreweave legitimately thought he was offering a reasonable deal to the CEO of Core Scientific. And the CEO of Core Scientific is very young. It's his first time being the CEO. And he's probably not as rich as he wants to be. And he's got this guy who's going to be worth a lot more than them, probably wining and dining him and telling him how rich and successful he's going to be when they accelerate his stock. So look, I don't know if it's botched yet because we don't know what the final outcome is. What I'll tell you is that when I saw the structure right away, I said, holy crap, like Corewe screwed over the core scientific shareholders and actually the largest Shareholder of Core Scientific came to the same conclusion like weeks and weeks later and put out a letter basically saying, you know, look, we're shareholders of both, but we just feel wrong about this from the Core Scientific perspective. Like we know we're getting ripped off as Core Scientific shareholders. And while we applaud the deal Coreweave made, like good for you and I applauded them live, like the day it was announced, I said congrats to the Coreweave team. Like they basically like took, pulled a fast one over on. of course, scientific. I think they legitimately thought the deal was going to get done. And I think part of that is that the CEO legitimately thought their stock was worth $165 a share. But as a professional investor and as somebody who has to handicap these things more correctly, I knew at the time that that was completely ridiculous. Like the stock probably never should have gone above 75 or 90, right? And so we're at 165 and you're doing a deal with no collar. no downside protection, like really good job by Coreweave, really bad job by Core Scientific. I think the CEO of Coreweave didn't contemplate a scenario where their performance would be so bad from that point forward, such that it would probably require some sort of change to the deal structure. The reality is, though, if they get it at anywhere near the price that was proposed, it's still a good deal for Coreweave. Like, let's just say they adjust the terms now and it still closes, they sort of still win. But I think unfortunately for Coreweave, they revealed their vulnerability, which again, I've been calling out since before the IPO, which is that they're going to end up having to acquire this infrastructure directly because they will not. They absolutely will not be able to hit those revenue forecasts that Wall Street is putting out there, which justifies this inflated valuation unless they control the entire stack. It's just like it's a virtual inevitability. It happens in every one of these infrastructure spaces. It's the same thing that happened in the Bitcoin mining space last cycle, right? Like the companies who didn't own any of the infrastructure stack were the worst performers in terms of their operating outcomes. And you can't do that in AI because you need like near 100% uptime. You're serving large scale customers with large contracts with SLAs, it's not Bitcoin mining, it's something much more competitive. And it's something that requires much more operational excellence. And there's no way that CoreWeave will be able to maintain the uptime and the operational excellence required to service these kind of tier one customers with these massive contracts if they offload the sort of management of the physical infrastructure to Galaxy and Core Scientific and Applied Digital. Again, these are firms in a lot of cases that have never even plugged in a GPU before. At this moment, they don't own any GPUs. And so, sure, you could make an argument, well, we're going to control everything inside the data center. But I tend to think that if you don't control the data center itself and everything outside the data center, that you're going to have conflicts, especially as you... discover over time that you need to like remake the inside of the data center every three to five years because of the cadence of the release and the new chips and the rack densities changing. That's going to have a follow through impact on the power prices, the amount of power that you need to suck at different times of the day, different parts of the year. It's going to have some sort of reverberation into the physical infrastructure operators relationship with the utility. and the grid and some of those dynamics are changing. And so it's just unclear to me how you can go to a customer and give them a warranty that you're going to be able to provide this really high level white level white glove service consistently over a long period of time when you don't actually own anything and you've got a hodgepodge of operators and different geographies with different issues with different power providers, different power pricing. And magically, that's all just going to work out and justify like a massive multiple. Again, in this case, there's no earnings yet, but a massive multiple on what is perceived as future growth, which is, again, largely predicated on the delivery of that physical infrastructure. So it's a huge circle. And they're admitting that when they go out to buy Core Scientific, again, something that was highly predictable. But now they're in a tough pickle because now they've admitted the market they need it, and the market's waking up to the fact that, oh, wait, maybe this physical infrastructure is more valuable than we thought it is. And so who's going to be a seller now? And who's going to be willing to look dumb underselling the value of the infrastructure the way Core Scientific did? Like that's a huge, it's a huge liability. It's going to cause litigation, but more importantly, like you're reputationally as an executive, if you make the mistake of selling the assets for a third of their value or something, right? Like you'll never live that down. And I think that's what we have happening here, I did think there was more likelihood of like a Microsoft or Meta or something like that coming in with a cash offer. Because you imagine how easy it would be right now if Meta just said, you know what? Screw it. Like we'll buy you Core Scientific for just $10 billion cash, right? Like that deal would be done quickly in Core, we would lose it. And I think right now you could make the argument that that would be like a rounding error for Meta, but it would secure their short-term power needs, or at least partially secure some of their short-term power needs that otherwise they'd have to rely on somebody else for.
Speaker 4: I think one of the things you raised there, Mike, I think is a very important thread for, certainly it reinforces it for me every time I talk about it, but also for those on the line here, which is that the rate of change of things, of our world as we know it, has never been so great. And so when we talk about, you know, core scientific and the Coreweave deal being struck at a certain level, in a vacuum, there was actually a world in which you could have justified it. And that world was pre sort of OpenAI, ChatGPT moment where enterprise data centers, the growth rates hadn't, they've been, better than a lot of industries, but not nearly at the pace at which AI is set to grow. And so to your point exactly around how that impacts operationally, if you're coming from a world where enterprise data centers, which is, I mean, is exactly how you described it, it's widely dispersed geographically, it's fragmented across operators.
₿itcoin ₿utcher: You know, and you can delegate.
Speaker 4: Majority of the operations, even the build elements of it, because there wasn't too much complexity to it. But now we're facing, we're in a world where the chips are changing, the power requirements are changing, the, the, the rack density, the, the design configurations of the data center themselves are changing so rapidly that it's impossible to actually peg value. And this goes back to the, the exponential themes that Dan always talks about. certainly around crypto, around AI, high performance compute. Those things, again, looked in a vacuum. Yeah, we've seen other growth sectors like that. But when you actually connect it to the rest of the world, it's affecting everything. So I think a lot of the deals that are being struck have a base rate problem where the base rate is based on a historical that is pretty irrelevant when you look out for the next 20, 30, 50 years. And so I think there's gonna be so much of dislocation and value to your point about Core Weave, you know, tanking literally from its highs and still being able to survive on the other side. The rate of change of that particular high, the rate at which it's growing.
₿itcoin ₿utcher: There's enough space for everybody to play, provided that the jockeys in the game.
Speaker 4: Are well seasoned and very capable and willing to navigate those complexities. Core Scientific, for example, I think the deal they struck to, it was a bad deal for them.
₿itcoin ₿utcher: It is a bad deal for them.
Speaker 4: And I sit there and put myself in the CEO's shoes and I go, man, that guy's got a lot of scars on his back from the bankruptcy.
₿itcoin ₿utcher: And so having to demonstrate and deliver value for the shareholders now, this might be an elegant solution for.
Speaker 4: The ego and the resume for whatever has to come on the other side of it. But ultimately, it's a base rate issue. And I think we're going to see so much of opportunity here, certainly from the infrastructure.
₿itcoin ₿utcher: Side, but also as the space evolves, there'll be multiple operators, multiple players.
Speaker 4: And I don't think it certainly started with iron for many of us here.
Speaker 5: But I don't think it ends there.
Speaker 4: I think there's so many more runs of this with other companies and other avenues. It's going to be pretty exciting. So I'm pretty excited for this group as a whole being able to ride that wave.
Speaker 5: Yeah, I think there's something there might be something else going on with Core Scientific. Like, you know, the argument was, well, look, we did this. Amazing deal with core we've 15 16 billion dollars we get to keep five billion dollars of capex and we have the optionality to develop additional sites and it's like well if that's true, then why would you just sell before you actually like pull that thread and figure out whether there's optionality there and so what I what I wonder is Did core we've have such like a bear hug? around core scientific where like there was some sort of like implied threat that like, look, if you don't pair up more closely with us, this is what we're going to do in turn, such that it like jeopardized the value of the existing deal. Because like you could imagine a scenario where they just say, screw it, like we're just going to default on this whole thing and like leave you holding the bag, like good luck unless you take our crappy, you know, acquisition offer. Again, I'm speculating here. I don't I don't know. It just it's so odd. The whole thing was so odd that it does beg these questions, right? It does raise these questions. And then from the CoreWeave side, like I sense a tremendous amount of hubris around the first mover advantage, which I've seen before in other spaces. I mean, look, they were in the right place, right time, and were able to offer Microsoft what Microsoft needed right at the moment they needed it when no one else could. because they were aggressively building out the kind of GPU fleet into that demand when no one else had positioned. And look, I'm not saying it was luck, like there's tremendous amount of skill to pivot out of Ethereum mining into the GPUs at that time, and then to partner up with Microsoft. But that's not the end of the story, right? Like Microsoft's not going to allow themselves to be dependent. on any one provider long term. And of course, there won't only be one provider. And if you got big with a specific strategy that worked at a specific timeline, you may delude yourself into believing that strategy, that exact strategy will work again and work in the long run. And that's kind of what I think the market sees in their earnings report. Like there's some growth, there's some backlog, but there's also just like really expensive financing for GPUs that are depreciating quite quickly and a customer base that has more and more options. And what's sustainable in my mind is the longer-term control of the physical infrastructure. And I think, again, if you were alone on the field and you could borrow money at any price if you could get Microsoft as a client a few years ago, I'm not sure that that strategy as a standalone works as well when you look out three years from now, and that's what I think is pushing them to to lock up more of this infrastructure. So I think that's the right thing. I'm just not sure that if your legacy kind of DNA and ethos is to BS at light, that that transition is going to be as easy for them as it is for somebody who's developed a mastery in physical infrastructure to just integrate off the shelf, increasingly commoditized off the shelf middleware. and software that everyone's going to have access to. So I'd much rather come at it from the perspective of the verticalized, the fully verticalized stack from the beginning, where maybe you're a little light on software, right? You're a little light on the middleware initially, and it's more bare metal, but you can add that relatively easily because you control the underlying power and infrastructure. It's all integrated. So the whole thing makes sense because it wasn't piece together hodgepodge from a bunch of different like totally different styles of infrastructure builds across multiple geographies with multiple utilities and power providers and construction techniques and right power contracts like there's so there's so many nuances and complexities the the broader you go and you see this with anyone who's largely assembled infrastructure portfolios through M&A versus people that have largely assembled them through a greenfield development. Again, back to my original point, some strategies appear to be more roundabout and convoluted and complex and expensive initially, but the real cost long-term is the complexity of managing a hodgepodge of unrelated assets that like nobody really knows where all the bodies are buried. Like there's no institutional knowledge. You take over an asset, you give the CEO a consulting contract And you try to retain as many people as you can, but those people aren't going to care as much about it as you do. And then they leave, and then they leave you holding the bag afterwards. And so I think like, look, the more M&A you try to do in the sector, the more problems the investors are going to discover down the road when they realize that the management team didn't actually understand all the complexity of the different problems that they acquired when they acquired all these different things. versus, again, somebody whose strategy from the go is to develop it yourself using a uniform process. So again, I'm not saying that it won't work long term. I just think, as I've been saying for months now, I think the market is way too bullish and not skeptical enough about that asset-light model. And they've been way too bearish and too skeptical of the asset-heavy infrastructure-first model. And I don't think it's even like, I don't even think it's in question, right? Like, I don't think it's in doubt. I actually think it's pretty much certain that that's the way it's going to play out. The only question is just how long does it take, right? And we're starting to see some movement in public markets in terms of the pricing of assets relative to these couple of different approaches to building out companies in this space. But I think over a long enough period of time, like it's almost certain that you have to have control of the infrastructure. You have to have developed some of the infrastructure. And that's why, look, I wouldn't rule out Coreweave and, you know, Nebius and companies like that who have done some development. I just don't think they have any long term structural advantage against somebody who's an expert at developing that. Yeah, for what's worth, I 100% agree with that.
Speaker 4: That perspective, I don't think I don't think there's there's going to be any players left that are trying to do one thing only. I think to your point, it's about vertical integration and it is those players and I think the mixes will be different, some a little bit more asset heavy than others. But yeah, if you don't control the power, if you don't control the substation, you don't control the land, you don't control the construction, and then you don't control how power prices are managed on a day-to-day basis while they operate. Having chips alone and a bit of a software stack on top of it doesn't actually mean that much.
₿itcoin ₿utcher: What do you control?
Speaker 4: You're basically a conduit for NVIDIA at the end of the day. So I fully agree with that perspective.
John Austin: The interesting thing from when I when I think about it is obviously their full start on buying Core Scientific in the first place meant that the price of the infrastructure has gone up, which is ultimately going to mean that they are less and less likely to be able to actually acquire that infrastructure as their stock deflates and the infrastructure providers rerate upwards. And so therefore it becomes harder and harder for them to actually get to the end goal, which is ultimately owning this and you know then. I think it gives ultimately a great boost to the infrastructure providers that will be able to get a leg up, as you said, Mike, on the software in the middleware stack. It's just interesting that that whole deal with they thought was absolutely a rockstar deal for Coreweave could ultimately be their undoing in the long run if they don't do something.
₿itcoin ₿utcher: Great stuff, everyone. We're pushing on a little over 2 hours here. Mike and Kosh, John, it's a pleasure. Really appreciate everyone. A lot to look forward to this week. It's 11 o'clock here on the East Coast, so I think it's a fitting point to close this out, and I look forward to interacting with everyone this week. Thank you so much. Have a great night. Thanks, guys.
Speaker 5: Bye.