$IREN Sunday Night Space - Data Center Scrutiny In Texas - Anthropic Export Ban - What's Happening At IREN?

Hosted by @Frans Bakker · 2026-06-14 · Tags: IREN

TLDR

The speakers argued that AI export controls, Texas data-center regulation, and growing power constraints are unlikely to undermine IREN because of its geographic diversification, secured grid capacity, and community-focused development model. They also examined IREN’s construction progress, possible Microsoft expansion, RAISE Summit presence, Mirantis integration, financing strategy, and long-term opportunity to become a globally significant AI infrastructure platform.

Speakers

Notable quotes

Transcript

Frans Bakker: Hello, everyone. Welcome to our space this Sunday evening. we have a lot to go over. I don't really know where to start. I had a couple of headlines and maybe it's good to 1st welcome Bitcoin Butcher. Welcome Butcher, how are you doing today?

₿itcoin ₿utcher: Hey Prans, thanks for co-hosting or coming back, allowing me to co-host this week with you. And yeah, the first thought, We had a lot to talk about before this new peace deal, or it's really an extension, 60 days, but that kind of throws another wrench into what we're doing tonight. So plenty to discuss and let's get it going.

Frans Bakker: Yeah, so I haven't seen much of the last couple of... hours in terms of news, but I did see iron ripping overnight, or while ripping like $62. That's a welcome change from the typical overnight redness we've seen in the last weeks. It does seem to me that something good has happened, and if I hear you talk about a 60-day extension, I guess that's a small win for markets. I see Bitcoin is also up. So let's hope we get a good counter to the more seemingly bearish news around the export control directive where Anthropic was basically forced to limit the use of Fable 5. And yeah, well, it was a directive targeted at foreign users and Anthropic decided to just pull the whole model. So yeah, I was expecting the markets would be maybe, volatile because of this, because it was specifically targeting large language models globally. And of course, there's a counter to say for that if you have data centers globally, for example, Iran has or is planning to have, that could also be perceived as a good sovereign counter to that. So, well, here we are. It's apparent that markets care more about the price of oil and stability in the region than an export control by the United States towards the entropic model. So yeah. I'm curious, Butcher. I saw you had some thoughts on this as well. I saw you consider this very early on to be a China targeted directive. What is your current stance on this? I don't know if I'm the only one, but I can't hear you. Maybe some people can give a thumbs down if they... Okay, I guess we lost Butcher. Well, then I will just try to give my version of it. I've started off with... Okay, maybe he's back. It's always glitchy with these spaces. but it seems to be particularly glitchy when I'm the one who's hosting it. I'm not sure why that is. And of course, okay, so I watched the All In podcast and it seems to me that there is two ways to look at this. There is the version where the United States wants to prevent the use of powerful models by foreign foreign citizens by foreigners and particularly China. That was basically Bitcoin butcher's initial thought. And then there is the version where Anthropic basically wants to wants governments to control the content of LLMs and they made on purpose a model that would be viable to be conceived as a weapon and therefore they are now taking a hit just for morale, for pushing their agenda. So welcome back, Butcher. I'm not sure what happened, but maybe you can give us your thoughts on this. given that you were very early on mentioning this was probably aimed at China.

₿itcoin ₿utcher: Yeah, we're still... There's going to be an ongoing battle back and forth with China for the foreseeable future and as a global superpower. And if the US wants to remain top dog, then Their job is to fend off the challengers, and the next foreseeable one would be China. So I don't, my initial reaction was we have export controls on NVIDIA chips for the same reason. And to the extent that you have probably our most valuable company, you could argue, besides Anthropic and OpenAI right now, if we've learned anything from this conflict overseas was the influence of Palantir and the U.S.'s strategic, or I should say military execution. So if I take a step back and I think about when China and in a parallel situation, when China and Russia tried to ban Bitcoin, how at first it freaked people out. when really it was a tell by those governments how powerful Bitcoin would eventually become, even if it's not fashionable right now, it's still a trillion dollar asset, where at the time it was a multi-billion dollar asset. So to me, I've kind of seen this story before and it's just repeating where, and then you pair that with the export controls on NVIDIA chips. And then you consider that what the Chinese have an edge on the US right now is on rare earth minerals. So it's this constant back and forth where one party has a strategic advantage over the other. And in this struggle, I don't see us ever fully getting along with them, but it's just this back and forth of leverage and kind of trying to assert power one party versus the other. And I think it's a combination of not allowing China to have access to it, to use against us, but also trying to get Anthropic to embrace its role as a US company and kind of play ball with the US government. There, I mean, there's already been for a few months, whether it was Secretary Hegseth kicking them out of the DOD or how, I mean, they can frame it that way, but whatever it was, I think at a high level, it just comes down to military superpowers usually have a technology edge on everyone else. And if the US is trying to preserve their space at the top, then it, is rational, whether it's right or wrong is a different conversation. That's not the point of this. But if the US is trying to stay in control, then it makes sense to me that whether it's Nvidia chips or frontier models created by OpenAI or Anthropic, that they would want to limit use to people we're friendly with.

Frans Bakker: Yeah, well, it does seem to me that this directive is pretty much reaching everyone globally. That is, of course, for now, the whole model is not available to anyone. But if there will be a version, if Anthropic will come with a new version that has more guardrails, that they will then roll out under the directive. That means that they will still try to make it very difficult for foreigners, for non-US nationals to use. Of course, that will probably be somehow able to get circumvented. So I guess what I'm saying, there will also be a way, possibly in the future, where Anthropic releases a new version of of Fable 5 with more guardrails that could potentially leak to foreigners, but under the directive would comply. But then you get that, for example, 99% of the non-US users could not access it, and then it would still be something that would be, sort of proprietary to Americans. And that is the question here as well, right? Is that something that we really want? Will this be a sovereign proprietary language model where governments decide who can use it and who can't? And companies together with governments in backrooms who will decide the content of these models, right? I mean, if markets work as they usually do, or as they should, then the winners will be the models that find the most users. That's typically how the market behaves. I read a lot of people are very dissatisfied with Anthropic now. They see, first of all, they saw a lot of token burn on the Mythos models. And now they're being confronted with being downgraded or even having no access to these models. So they're looking elsewhere. And competition is also not sitting still. At some point, you're going to ask yourself, will this be a first cannery in a coal mine? And will all the big large language models providers like OpenAI and others, will they all play the same game? And is it just the user who has no real way out anymore and will be caught in this system of overregulated government? And will it be like that? Because that's the whole thing I got from the All In podcast with Chamath, who was saying we need to push for open source. But at the current state that open source is in, it's just not a real viable alternative, especially if you see the infrastructure that's needed behind all of this. So yeah, I'm still trying to make up my mind of what I really think about this. But it seems to me that, at least for Iron in particular, it doesn't really do any harm for now. I mean, Iron works with Microsoft in the US. They have signed some kind of partnership that goes beyond just delivering compute for Horizon. But how will that work if Microsoft as a US company will be told that they cannot work with foreign controlled companies any further? You know, it just, there is of course a massive bear case that you could make of this if they are going to restrict infrastructure like this being used for language models training, right? So yeah, it's kind of like a, I'm kind of making no sense maybe, but that's because it just, it's all over the place. But I guess the point I'm trying to make here, the good, the good thing about Iron is they have data centers in the Canada, in Canada and in the US, and they have sites that they aim to develop in Europe and in Australia. So from a sovereign point of view, at least they have the optionality to build into this future where sovereign AI and sovereign language models become a real thing. And in particular, for example, Europe doesn't really have their own frontier model. And maybe this whole move that's been initiated by the US could lead to, a big increase in development of these models in geographic regions like Europe and Australia. I mean, it's too early to say what this is really about, but personally, I don't really like overreach from governments. And I certainly don't like these socialist statements from Dario with regards to the world is going to end unless the government steps in and regulates everything. But I guess for now, we will have to find out what this really means. But as a company that owns the power and the land, I guess you're still in somewhat of a safe spot because you're always able to go with the flow of the next directives that are coming out there. And maybe to that regard, it's good that Iron only has at least a couple of percentages of their full power portfolio. So I guess I'll just pause there. Otherwise, it's going to be a we lost Butcher again. So it was a one person rant because I was the only speaker. So let's see if we can get Butcher back on. I'm I was actually planning to talk more about the taxes, the new proposal by the governor, Abbott. But I'll let Bitcoin butcher maybe respond to what I've said before if you heard everything.

₿itcoin ₿utcher: Yes, I'm bullish on iron, as I always am, but especially after considering the you already alluded to it, but the geographical diversification now that there's an Australian presence officially as well as Spain. So even if The US remains the primary market. No one will dispute that. And Canada might be less promising than we thought it once was. But if we're in a world that I think with President Trump's presidency and influence, I think the era of globalization, the past, you know, five to 10 years with his initial presidency, you kind of see the world beginning to silo off a little bit, and everyone seems to kind of be keeping their cards closer to their chests and home and embracing globalization left. So if that's the case with AI, Iron still stands to benefit by positioning itself in those unique markets. One question I would have for you, Franz, before we move on to Texas, I think another bear case, because we speak about Anthropic so much as if Iron was to partner with Anthropic, does that, now that there's this going on, does that influence anything? And I think there's credence to that, but there's also the Chinese open source models are being dumped into the American marketplace. I forgot who wrote it or what I read. I think it was an analyst report. I forgot which institution wrote it. But if these Chinese open source models are reducing token costs and the subsequent use of open AI and anthropic, would that hurt iron? And my initial reaction to that was that's like the whole point of What we've been trying to do the whole time when Dan speaks to Javan's paradox is if the token cost gets reduced, then these models, even if they're open source, if they're 80% as good as the private close-ended models, ultimately we're looking for adoption. And if there is adoption and increased token consumption, that Iron still has a place who whatever they're running, whether it's anthropic open AI, the point is you're still going to need the physical infrastructure and power. And so I was I wanted to hear your thoughts on that or if there was anyone else who wanted to come up and speak to that, because I don't pretend to be a frontier model expert, but this still is just about physical infrastructure for what I believe to be an exponential growth trend that isn't going anywhere. And every time I think the US in some ways, just like I referenced the Chinese and the Russian governments earlier, they can try and hold this thing down, but it's going to take on a life of its own. So I think it's a very telling that the US sees that much potential in anthropics models to show you where we could go from here.

Frans Bakker: I guess I'll just give my overview on 2 versions of two worldviews, basically. Then I just added the Microsoft's CEO article in the Nest. And I think I am happy that Iron has partnered with Satya and not yet. with Dario. First of all, personally, I'm not aligned with Dario's worldview. But more importantly, from an Iron's perspective, Iron has stated multiple times that they want to increase adoption. And I think a business model that Iron has with a, you know, full vertical integration from, you know, the power to the ultimately to the model. Right now, of course, Iron is a GPU as a service provider. I don't want to get into the whole thing about is Iron a NeoCloud or not, or are they just a infrastructure as a service provider? That doesn't really matter for this point. The point I'm trying to make is if you want to help with adoption, accelerate adoption, then spreading the compute globally while you're vertically integrated gives you an advantage on cost. And for adoption, typically to increase market share, you should not focus on, you know, margins initially. So Iron is potentially able to undercut other providers to spread their data center capacity that they are able to stand up globally. So I think a partner like Satya, who is arguing basically the opposite of Dario, where Dario says too many, you know, too many people have access to two powerful models, or such as not enough people have access to these AI models. So you could basically argue that they are aiming to achieve the opposite in terms of, the focus for Anthropic seems to be that the information within the LLM models and the usage of it is too powerful to give to the market. Satya is arguing that it's more important to spread adoption, to give as many people as possible access to AI. And I guess if you see tokens as revenue, then of course, the diversion that doesn't limit and doesn't, you know, over regulate is better for monetizing your portfolio. So, it's overly simplified, but I am more aligned with the world view that Microsoft and Iron share. And that's more a capitalist world view, ultimately. So I guess I'll just leave it at that for now, because we don't really know how this is going to play out. I would recommend everyone here to do some homework, read these posts by the people from the all-in pod, and read this article that I shared. It's good to, I mean, we are all here to make money. Investing doesn't have to be a political thing or an ethical theme all the time. But I guess it is somehow also relevant to be aligned to a certain degree with the investment you have. You know, if you don't have to agree with everything, then and we'll do, there's obviously a lot of people that are invested in iron that couldn't care about the ESG label, you know, the green electricity. But at the end of the day, you don't have to support it for, you know, for an ideological reason. But it's ultimately, it gives, you know, it gives less of a hurdle in some, you know, jurisdictions. if you can just say we've always been 100% renewable. And that's just, I guess we've seen this in the last five years as well, where all the Max 7 became more focused on their water consumption and on their, green energy with these targets for 2030 to be net zero with emissions. And, you can always, wonder if it's really ideological or if it's just a political agenda to comply with in order to advance your plans. So I guess this falls in the same category. And yeah, I cannot give you a conclusive answer as to how this will play out for Iron, but I can say that for now I am very satisfied that we have not signed anything with Entropic as of yet. Before this thing settles, I am more glad that we have a partnership with Microsoft, where Microsoft is signing MOUs with the Australian government. They're investing in Canada. Ultimately, Anthropic is sort of trying to do the same thing, but with a different thing in mind. And I really don't know if Iron ultimately will be able to sign both these companies to a meaningful anchor tenant or partnership, because honestly, it doesn't seem like these two companies, Microsoft and Anthropic, really carry the same values on a C-suite level. So yeah, that's just some thought I have. I mean, I'm still trying to process this, but I guess we can pause this topic here unless, Butcher, you have some final thoughts on this. Otherwise, I guess we can move to the Texas topic.

₿itcoin ₿utcher: On the Texas.

Frans Bakker: All right, so let me just share this post. in the nest as well. It is a letter from Governor Abbott where he is saying a bunch of things that are supposed to be, I mean, it's supposed from Shanu Matthew and I had some interactions with him recently when it was he was talking about this project from Crusoe. You may have seen this or not, but I'm not really Bullish Crusoe. But this topic that he shared is the letter where he's proposing or signing the intent to relieve ratepayers. And basically, I would argue that this is a similar version as the House Bill, I guess it was, House Bill 2992 in Oklahoma. It is, I had a note about this, just, yeah, so House Bill 2992 is currently signed into law in Oklahoma, and it basically says, large load must bear more grid cost. It protects residential ratepayers. It voices concerns about grid reliability. It says there needs to be more oversight on large loads and future projects and a potential requirement for dedicated generation. And the ABOS directive, if I can call it like this, is a little worse to some degree, but it's not actual law right now. So there is a distinction to make between these two, which says that data centers should pay all infrastructure costs. It should reduce residential electrical bills. It also asks for more oversight of data centers. And the one sentence that's sort of a red flag or sort of worrying is ensure data centers add to Texas electricity capacity, not just demand. So of course, this seems not very pro-business on a first glance, but I would just say that this is not a big deal. for Iron. Obviously, Iron is already compliant with the House bill in Oklahoma. They have paid for all the cost towards PSO, including a massive down payment for, I believe it was 20 years of firm power. It is, you know, they're had a bunch of meetings in the county in Macalester with locals where they explained that they are paying for all the costs to connect to the grid, for all the infrastructure upgrades, substations, et cetera, et cetera, as they basically always done. That's to say I don't think that they literally paid for the switchyard in Childers, but That's a couple of years back, so it's not very relevant now. But for Oklahoma, they have stated that they will pay for all the infrastructure costs associated with them connecting, including transmission. And so when Abbott says something similar, do you know what does it really mean for future projects of iron? It may mean that Iron needs to pay a little bit more CapEx for a utility-owned switchyard, maybe some transmission cost, some, you know, some lines and poles. I don't think that this is very significant on a project level for these sites that Iron has, which, you know, their future sites. For example, Sweetwater 2 is currently seeing a a utility switchyard being built, it's currently in progress to interconnect them next year to the grid. So I guess this is the kind of thing that Iron should be paying going forward if this thing becomes law. So, you know, Iron has paid for their own bulk substation in in Sweetwater One, they paid for all the primary substation, all the electricity lines and electrical infrastructure there. I don't think that a directive like this is going to move the needle on other sites going forward. I guess you could say like their high voltage infrastructure CapEx is going to increase, you know, but is it going to make is it going to move the needle on a one GW site that they may have somewhere in Texas in the future? I don't think that is that this is the problem. Reduce residential electrical bills. Well, I don't think that they will be able to really get this through, you know, with the public utility commission, because ultimately The rates for residents are based on many factors that cannot be controlled by data centers connecting to the grid. I think this is sort of making a political statement more than actually trying to get anywhere because There's just too many factors in the pricing of electricity rates. I mean, it's easy, it's one way to say that the, a data center cannot increase the rates, but it's another thing to say that they should reduce residential electric, electric, electrical bills. That's just, you know, I think that's just grasping. I think he's just basically making a statement to get in favor with voters, I guess. The other thing, most of the things he's talked about are already within Iron's, community package. Don't build near residential areas, low noise, setback zones. This is what they do. Communities typically like Iron. That's also something I can Yeah, I can explain you a little bit about that. Iron had a public hearing in McAllister in their Oklahoma, future Oklahoma site last week, I believe it was. And they basically talked about all the topics that are part of the House Bill 2992. And here they also said that, they want to be there for the community. They don't want to be a strain on local water sources. They don't want to create data centers that are too close to people's homes. Of course, the people are still not entirely convinced. So there will be another public hearing in the coming weeks. I don't know exactly when, but I guess I could talk about this for hours, but the point I'm trying to make here, this directive by Abbott or this proposal, it's not something that will deter Iron in particular. This is just, is there bread and butter to be good to communities, to care about locals, to there's Sweetwater One site is in the middle of nowhere. There are literally no houses there. There's maybe one small lost farm or something, but obviously those people are well informed of the data center coming. Typically what Iron does is already in line with this, and that's mostly because they, you know, the philosophy of the company has always been, we want to be green and we want to be supportive of local communities. They had the community grants program for years now where they're supporting businesses with $100,000 every year spread between volunteer projects and non-profit companies. This thing that Abbott is trying to create, I think it is not going to be it's not going to be difficult for Iron to comply with if it becomes law. And I think part of it is never going to, hold up. He said something else about reducing the tax, sales tax exemptions. Let me see if I can find the exact line. But One second. Yeah. Repeal sales tax exemptions and other outdated or necessary incentives for data centers. Well, that's a proposal that he will work on. So that's even further away than this initial directive. So I think this is not very relevant for Iron. First of all, Iron is mostly doing tax abatements with with counties, and that is not with regards to sales tax. Those are property improvement taxes. So I don't think he's going to tarnish with the counties making deals with these data centers to increase, you know, the local, you know, jobs that are being created and prosperity that's coming from the data center. So That's obviously not going to happen, I think. So yeah, I guess to summarize it, Iran is probably not going to have a very difficult task with this. It's just a bit of a political statement that Abbott's trying to make towards the communities in Texas where he just wants to say, okay, the governor is here for you to protect you. your electricity bills are not going to go up. I think at the end of the day, this is very helpful for Iron because now they can just point at the bill if it will be a bill in the future and just say, look, this is the directive from your own governor and we are compliant with this. So, you know, we've done this in Childers before and yada, yada, yada. So I think that this is just going to be something, you know. It's not going to hurt them. It's going to help them. It's not going to cost them much more. So I guess concluding, I would deem this for now a pretty much a nothing burger. It's not really pushing back on data centers or saying the business is not welcome here. So I guess I'll pause there.

₿itcoin ₿utcher: Nothing burger. I word. One way to look at it, there's two things. I put in the nest a post by Ryan, but it came from, it was either, I think it was a Wells Fargo Securities estimate of when Iron's going to build certain sites and it's on a cumulative level. But if you read that chart, Iron, for purposes of Texas, is going to be building Sweetwater One through 2029 and not start Sweetwater 2 until 2030. Now there might, maybe that's their base case and they could build faster. But my point in that is in itself, like if you have one gigawatt of critical IT in Sweetwater 1, With Vera Rubin, it's conservatively $15 billion of ARR, and it's probably closer to $20 billion of ARR. And I haven't even included Sweetwater too. And meanwhile, that's going to all be constructed in the next two or three years. So to weigh ourselves down on something that just like the other topic we discussed, probably will just play out over time and they already can put their heads down and just work on Sweetwater gives me comfort personally. What I think is more telling in the market is the fact that like I watched the HUD CEO with his interview on eight gigawatt pipeline or whatever the figure was. Well, I'm willing to bet Hutt has a lot of exposure in Texas. I own Cypher. Cypher has Texas exposure. And I don't think long-term, I think Texas remains pro-business. But what I would say is I think it's telling that someone like Tyler from Cypher is going behind the meter now, and I'm beginning to think it's less by choice and more out of necessity because this politicization of the grid going into an election in November, I think that's going to flow into ERCOT a little. And even if that backs things up a few months or six months, that creates more uncertainty that can weigh on your equity, even if it works out eventually. And I, like I said, I'd rather just be, it's in a good position as an iron shareholder to know that they can just focus on the largest site in the U.S. and arguably the world and take care of business and. where they're at in two or three years, then we can worry about that. But by the way, if if I'm correct and I have reason to believe that I am, that we're looking at $20 billion of ARR that's built out just from Sweetwater alone. And today we're currently projected at 4 billion. Like, are we really going to be upset about five X saying if if we couldn't have any other power in Texas in the future, would would you really be that upset? And I would say Probably not. And they're and they will always Dan and Will have always found the ability to find stranded power opportunities for iron. So I don't think anything's going to happen necessarily in Texas, but it's if you're a cipher or one of these other competitors and colocation that was relying on a specific development pipeline, I do think I personally want to hear a little more from Tyler as it relates to Cipher, because a lot of their pipeline is scheduled for 2028, 2029. And I just I'm not sure what assurances he's been given through the batch process, which I believe there's supposed to be a meeting this week. But the point in all that's not to fear monger as much as like get some clarification for people. But with iron, we just said like the power secured as of May 1st. So like who really cares? Like we're going to build and hopefully there is some more power from batch zero that gets approved. But even if it's not like it, I don't think it impacts what they're going to do the next two or three years in construction.

Frans Bakker: Yeah, I think it comes down to we we keep being proven right about things we have said one or two years ago about the way iron operates. And this goes much broader than what we've discussed so far. This includes, but it's not limited to, social license to operate large sites over small sites, you know, non-dense residential, you know, areas without people. I'm just going to say the very simple, simplified. So remote areas with large stranded excess electricity, it's just all coming together, grid connected and not behind the meter. You know, being very early with securing land and power, this is really the thesis of Iron. And If you remember, like a year or 18 months ago, we were discussing about the with peers about the remoteness of Iron's site. I don't know if you remember that Dan was writing the amount of milliseconds that it was from their Childress and Sweetwater site to Dallas. And now data centers are being prepared to go into space. I mean, I don't know how many miles those are. I don't know what the latency is with space data centers, but it just becomes obvious that we have, we went from, they need to be in residential areas to they should be in space. And that just happened in less than two years. I mean, Obviously, the distance is not so important anymore. The capacity is more important and the economics around that. So if it is about capacity, you have to be with iron. If you are looking for remote areas without people living nearby, you have to be with iron. If you want to look at grid-connected economical megawatts, you have to be with iron. I mean, This is just a confirmation of what we have seen early on, and it's being proven to be right today. So I think we are somewhere in a mass confusion of what is the best way to run a data center company. The market doesn't really differentiate much yet. they're giving SpaceX the benefit of the doubt. They're giving Nebius the benefit of the doubt. And now they're starting to give behind the meter projects the benefit of the doubt. But so far, all we've seen is behind the meter projects are delaying or they're failing. I mean, just it is a fact that the ship engines in Finland didn't get approved. Oracle's site in Aveline had issues with their gas turbines. These turbines are being, you know, the backlog on those is reaching into the next decade. I think they are economically much more expensive to run on a MW basis than grid-connected capacity is. I mean, we are we just get proven right all the time. And then comes the regulatory train, first with SB6, and then with batch 0, and now with Abbott. And everything is just, you know, Aaron is shrugging it off as, yeah, we were prepared for this. Oh, we already comply with this. I mean, we are going to see a culling of data center forecast for Texas after batch one is, sorry, batch 0 is launched. I don't think people are really anticipating this, but it's going to happen. If you see now in the interconnection queue, there is like almost 300 gigawatts of power until 2031 or 2032. I mean, it's complete ********. That's never going to happen. And I think we are going to conclude that people that are scrambling for gas turbines right now or saying that space DCs are the only viable one in two or three years from now. It's just, it's completely disconnected with reality. So yeah, if I see this table that you shared, I think it's, I don't think, you know, there was some discussion that IR has leaked this to Wells Fargo. I don't think so, because It's not accurate at all with regards to Bundy, for example, the Australian data center site. It's not going to be built out like in that cadence. I think there was a post that said Iron is going to build it out over 8 years with 100 megawatt a year. And given that they have only 500 workers projected to work at that site, I think it's not going to go much faster than that. because if you look at currently iron has almost 3000 people in Childress, they are constructing horizon one, give or take 12 months for for 200 megawatt. So let's say it's 300 megawatt of gross power for 3000 people. That means 500 people doing 100 megawatt a year in Australia is a pretty big improvement compared to children's right now. So I don't think this table is very accurate, but it's true that the build out is just beginning. So yeah, I'll pause there.

₿itcoin ₿utcher: Sounds like we move on to the next topic. Do you want to talk about I see Lando's up and then we could go to Lando or you had the I saw the article I think you were referencing with Beth on that spoke to Corey even Nabius. So what do you want to do, Franz?

Frans Bakker: Yeah, I keep approving people to come up to speak, but they don't appear as speaker. I now we have Lando up, so maybe he can talk about the race summit.

Lando: Yeah, so. What's up everyone? I'm Lando. This is first space. I appreciate Butcher and Franz for all the work they do for this community and everyone else that's involved. But yeah, so yesterday I wrote an article or two days ago about the Race Summit that's coming in July. I initially thought when I saw it, I saw it on LinkedIn. about 15 hours, I think they posted it Friday, about 15 hours after they did it. And I came on X and was trying to find it and couldn't find it anywhere. So I took it upon myself to do a little bit of research on it. And it looks very promising, right? Iron's the headline sponsor, which in summary just means that they paid the most money to advertise and beyond the head of everything. But I don't believe that they would spend this much money or however, I assume it's in the seven to eight figures range for this event. But I don't believe that they would spend that much money, especially at an event where there's a lot of one, just to give the scale of who's all going to be attending this, pretty much every single hyperscaler will be attending this. Every single Frontier Lab like OpenAI, Anthropic, I'm trying to think of who else. Google DeepMind execs will be there. Eric Schmidt, former Google CEO and chair will. be speaking. It's one of the fastest growing names in or growing summits currently. But I'm trying to pull up my article.

₿itcoin ₿utcher: Well, so Lando, if you don't mind me cutting in for a second, I thought having read it earlier, what you highlighted that I thought made a lot of sense was previously the NVIDIA conference in Taipei, GCX was kind of more focused on DSX architecture, Vera Rubin, like hardware specific solutions and almost felt like similar to like a Steve Jobs pitch where it was talking about a specific product where I think you would agree and you can detail it more if you want. What's telling about this event in particular is the diversity among the attendees. And it's more of a B2B focus. So iron potentially, instead of talking about hardware capabilities, is more in a position to close business. And it's very rare to have such proximity to so many influential stakeholders and a variety of companies around the world. And I'd imagine Dan and his team will be in attendance for that reason to not only sponsor and give the presence of Iron and what they aspire to do with their cloud and infrastructure solutions, but also find some new customers that they can sign ultimately, which is what we've been trying to find. I mean, there's certainly no limit. I should say there's no trouble signing given their the demand and their relationship with Jensen now, but it's very rare for Dan to have access to that many high level individuals in close proximity as opposed to having to fly from Australia to them. So being in, I believe it's in France, that offers an opportunity to have a lot of meetings all at once in a very short period of time. Would you agree with that?

Lando: Yeah, I mean, to your point, just to give a little more context, something I thought was unique about this specific conference is it's going to be held at the Louvre in Paris. I thought that was like, it's not like a exposition center or just some random location, you know. And then to your point about the more business to business connections and like potential partnerships. There's a dinner the night before this convention on July 7th at the Chateau de Versailles, which is Palace de Versailles. I thought that was a very premium event. I mean, the tickets to be able to attend that dinner and conference are 6,500 euros, which I think is around like 8,000 USD or something like that. But that's besides the point. It's a very exclusive event that I think the Iron team and Dan will have access to pretty much anyone serious about needing their infrastructure, needing their expertise with compute. I also noticed Mirantis is also a sponsor, so I don't know why. they would have a double sponsored or maybe Mirantis will have a booth there. But currently this is also another announcement for anyone in the space. I already have three people that will be able to go to this event. The one of the directors for this event reached out to me after he saw my article and is willing to give us Six total take us to this event, one being a VIP. I'm not sure how familiar you guys are with Kaizen Invest, but he lives in Belgium, I believe. He's got a relatively large account. I think he has a smaller position to iron than most of us, but he's still an investor and a very prominent name. And he'll be having the VIP ticket. And then I've also given, or I haven't given, but assigned two other tickets to Leo and then one other investor that sent me his stuff on X and DM. So if there's anyone else in here that would like to attend, I'm kind of just going to take the first, the next three people that send me their name and e-mail via DMS. But if you are interested and you live close enough to Paris and would want to go July 8th and 9th, that would be great. I'm sure everyone here would love to kind of get whatever experience you get and whatever notes or anything you take there. But I mean, to your point, Butcher, this this event is a lot different than Dell World and Nvidia GTC. I feel like both of those events are kind of highlighting Nvidia architecture engineering. It's more for developers and kind of server side with Dell and kind of all their technological advancements with Vera Rubin. And this event seems more directed towards like business to business relationships. There's going to be institutions here that have over 600 billion assets under management, which is exactly what the Iron team would want, is people that have the capital ready that need compute as fast as possible for whatever clients they need. So this seems like a really great opportunity for Iron and I'm really excited. I think there'll be a lot of coverage. I know the All-In Pod is also going to be at this event, or at least it's on the website. So there'll be a lot of retail coverage, similar to when Dan went on the All-In Pod to talk about Iron a few months ago. So yeah, this is, I think it's going to be a great event. Iron has their own lounge at this event as well, where there will be talks all day with Iron branded across everything. But yeah, I think I think we'll get some really exciting news either before this is all just speculation on my end now, either before. So it gives Dan and other investors to talk about something or just show them like Because I mean, Microsoft, like in the business world, the Microsoft contract was relatively close. I believe it was last November. But we all know we've been waiting for the next deal, right? We got the NVIDIA cloud contract at Childress for 60 megawatts. I believe it's 3.4 billion, which is great. You know, that's a great first step. It'll help Mirantis get up to speed. But I think we're all we're all waiting for that next like big couple 100 megawatt contract, right? I don't know if that'll be before. If I had to guess if I was going to bet money, it would be after this event at some point. But I know Jim Liu and God particle and a few other guys have been thinking that that. end of June, July, and maybe even August is kind of when they're expecting that next deal for whether it's a McKenzie or another smaller megawatt. I really don't know. I was talking to Marcos as well. I know you guys are familiar with him the other night about this event as he has a little more experience with these kind of events and research in the semis and AI infrastructure world.

Frans Bakker: So Lando, are you still offering these tickets or did you already find enough people?

Lando: Yes, I am. I need three more people to get the Pro Pass, which is just the, it's worth 2,000 EUR. I believe it's just all access to the two days that, I believe it's July 8th and 9th. It's the two days for all of the keynote events at the summit. So if you are interested, just let me know.

Frans Bakker: Anyone who wants to go to RAISE and, you know, see Aaron's lounge in person or connect with the people on the ground, make sure to take it up with Lando, send them a DM and see if you fit the profile. It's in France. So if you're not in Europe, it's probably going to be difficult to get there unless you're willing to fly over. My personal thoughts on this event, last year, I believe, Dennis Skrinikov, our CTO, had a fireside chat with the guy from Poolside, if I'm not mistaken. I think that was also at Raise, and I believe they were sponsored then as well. I'm not sure if they were a headline sponsor, but this year it looks a little different with their own launch. So people were saying, like, why doesn't Iron have a keynote. That's because they own their own lounge. And I think they're going to have like a C-suite speaks to the crowd kind of thing. I don't know if they will have a proper keynote by Kent or Dan in there. But my personal thought is there is a chance that Dan will appear on the all in part on the last day. the last, after the program finishes on the last day, I think the All In pod is going to do a recording or a live version of their podcast there. And it just, it would make a lot of sense as a headline sponsor, particularly Dan, who has sponsored the All In podcast before and who has appeared there once before. Or, I don't know if he appeared on, but it was some kind of like, I don't know what it was, but he showed up in those short interviews where they had the Coreweave CEO and a bunch of others as well. But I mean, that would really fit the marketing MO that they are currently seemingly having. So So anyone who is interested in trying to obtain a pro pass for a raise, connect with Lando. He's done a great job covering the event. You can find the article in the nest. I don't know if this is going to be a when deal moment. I don't think so, honestly. I think there are other signs that are indicating at a deal. of some sorts. Currently, I am prescribing a likeliness or a high probability even of a deal for Horizon 5 and 6 with Microsoft in Childress, possibly before the end of this month. And I'm not going to give a probability number to that, but I think it's it's very reasonable if you consider the following. And sorry for diverting from this topic of race, but I think we've covered most of it. So as it looks now, iron is actively destroying electrical infrastructure for Block 5. And Block 5 is our six mining buildings that they have constructed in the second quarter of 2025. So I don't know if you remember last year that Iron said we're going to have 50x a hash by June 30, 2025. And they announced it literally in the afternoon of the 30th that they reached their goal of 50x a hash. Well, I have seen footage of excavators excavating the ground between the data center and the substation, destroying the high voltage cables that are underground, because these buildings are being disconnected from the substations, because these buildings are going to go. They're going to be either demolished or disassembled. I don't know what's the proper term for it, because I simply don't know what's going to happen, but they are going to move. Obviously, Iron has already explained and disclosed this to the market. They're going to build 2 new liquid cooled data centers there named Horizon 5 and 6. Three buildings each, similar design as Horizon 1 and 4, but with a little twist. The liquid cooling plant is going to look different again. So we're going to get a third iteration of a liquid cooling plant at Childers. where Horizon 1 had a version with fans pointed down at the dry cooler installation. And the second iteration was Horizon 2 to 4, where there were cooling fins on top. And so they have a new design for Horizon 5 and 6. It's going to look a little different again, but the idea behind this is that these will be potentially Vera Rubin and maybe Vera Rubin Ultra compatible in terms of cooling power. So yeah, I'm I'm looking forward to seeing this roll out. But so back to my point, they are currently, you know, breaking the link between the primary substations that will remain and the former mining buildings that will go. And I think they're going to take offline and demolish 6 mining buildings in block 5 and six mining buildings in block 4. That's a total of 12 buildings that they constructed last year that are going to move, that are going to be razed to the ground or somehow salvaged I don't know but uh that's uh so um there this is happening and um we are now at the uh at the periods just before the actual buildings are going to take being taken down and we're at the last two weeks of the fiscal year so my point is iron is making irreparable damages to their electrical infrastructure mining buildings are about to being removed. ASICs are palletized and being moved out of the buildings and possibly out of the site. Iron has announced they're going to take a $540 million hit on accelerated depreciation or displacement or whatever. And so My theory, and this is not financial advice, it's not investment advice, is one, I think Iron's going to use quarter four of this fiscal year, which ends in two weeks as a cleanup quarter. I think they're going to take the entire hit in this quarter, which will be a big strain on the EPS. It will look horrible on an EPS level, but it doesn't really matter because it's a non-cash event. But it's certainly, it's going to look bad if that is the case. Secondly, I think they are going to sign Microsoft into Horizon 5 and 6 as a counter to this. So I think they're going to tell the market in August or earlier, that depends on how they're going to disclose it. But I mean, I think they're going to sign Microsoft into these new Horizon 5 and 6 buildings. in the same period. And why do I think they're going to do that before the end of the fiscal year? It's because I think they're going to use this to somehow end the fiscal year with a higher market cap, which could somehow unlock certain benefits. Or I don't know. There's been people that told me that the market cap RSUs have been reached already. So I wouldn't say that Iron's going to announce a deal to pump their share price to pay themselves, but it does seem to me that it would make the, you know, would you really destroy 12 mining buildings and all the electrical infrastructure connected to that up to the substation just So you can start negotiations with Microsoft. I mean, obviously not. There is a, they are in extremely far advanced negotiations with Microsoft. Possibly they have an LOI already, you know, but can they really just destroy all this CapEx that we paid for with an ATM last year and with our mining cash flows last year? just for a build it and they will come. I think I deem that unlikely. So that's what's going on. Horizon 5 and 6 are going to be developed this year. But first, of course, they're going to have to clear the grounds for that. And the implications of that are, you know, costs. So I see not the raise event, but the raising of the buildings as a more likely reason that they will announce a deal. Pun sort of intended, but as you know, I'm not very funny on my Monday mornings. So what else? Yeah, I wanted to talk briefly about what's going on at Iron. As you know, we follow the sites very carefully with our satellite analysis in only France. We covered, we had a high definition Mackenzie shot just 1 1/2 weeks ago. Here we saw that the site is progressing. We've identified concrete paths that we've also seen in Prince George before. where generators and UPS buildings typically are being deployed. But I haven't seen any generators or UPS buildings as of yet. So there is one row that has these buildings and has generators, but there are three rows where I cannot conclusively say that they have these external components placed outside of the data center. So that would lead me to believe that they are closer to 20 megawatts being retrofitted for actual GPU usage rather than, you know, 50% of the site or more. I don't I think McKenzie is sort of slow, if I'm really honest. But it could be a perception of external of use, maybe inside these buildings, they are very far and they've advanced everything and they are doing a sort of inward out approach where they do everything inside first and then, gather the components. The caveat to that is that in Prince George, we found out that IRA needs permits for these generator buildings. outside of the data centers, which is, I mean, it's almost laughable if you think about it, that irony is a building permit for a small generator building. I mean, no, it has a word building in it, but you see a data hall of 25 megawatts, and then you have these little external, you know, enclosures is maybe a better word than a building. But they literally needed building permits for these enclosures. And I just wonder if this is also the case for Mackenzie, because it's the same district. So the district Mackenzie, Prince George could very likely have the same requirements. So as of now, I haven't really found anything with regards to building permits for Mackenzie, but I'm not extremely satisfied with the progress happening on the outside of the buildings at McKenzie. And that's just my honest take. You know, if I see something bullish, I will tell you. If I see something that I don't like, I will also tell you. That being said, It's only middle of June right now, and we don't know when the GPUs will come. It could be happening that the GPUs will come in December, right? second-half of 2026. So, you know, bear case, the GPUs will just arrive so late that they're taking their sweet time to do it with a couple of people. I don't know. You know, it just doesn't look very like a big undertaking like it does in Childress. So what we've seen in Childress is, first of all, of course, the hash rate is going down. The hash rate is literally plummeting. With that, the electricity costs are also going to plummet. So that bodes very well for the margin profile that Iren will lay down in the August earnings. It's most likely going to be an improvement in terms of operating margin, but it's going to be a accelerated deep, an accelerated downturn in Bitcoin revenue. And analysts have not forecasted that. So I would, you know, consider that when you see that the Bitcoin revenue is going to make the absolute revenue also go down. But hopefully the market will not really care too much about that. On the other side of the site, Block 3 is being retrofitted for the B300s that are divided between McKenzie and Childress. But interestingly, we have seen here that Iron has announced they're going to retrofit 2 buildings. But so far, what I've seen is they're retrofitting 4. So my first stake was are they going to divert GPUs from McKenzie to Childress because McKenzie is too slow? I don't know if this is true, could be wrong, but they are certainly retrofitting more than what they have said they would do in 2026 in Childress. So unless these are the two data centers that are going to be used for the NVIDIA contract, If that's not the case, then I think they are going ahead and retrofitting more data centers in 2026, possibly for rerouting B300s to Childress partially instead of everything, you know, instead of 34,000 to McKenzie and 17,000 to Childress, it could be the other way around. But that's just, it's just a guess, right? I don't know if that's true. But if they are only running 17,000 in Childress, why would they retrofit 4 mining buildings to the extent that they're doing? Right? Unless there is a new GPU purchase also arriving in 2026 that they're going to announce soon, I don't see any reason why they are doing this. And on the horizon front, Horizon One is getting along very nicely. Iron has sold the market they're going to deliver this first 50 megawatt cluster to Microsoft in quarter three. But my Intel says that they are going to deliver their first building in June. So initially I was sort of excited to find out it would be delivered last week or this week. I'm not entirely sure if that is really the case, but it looks like they have found the leak. Basically, the issue that the long issue that they had with the first building of Horizon One seems to be fixed. So that's good news, of course. And it's it means that Iron is going to be able to bring these buildings online one by one. And the first one is like it either happened already or it's about to happen, but it's certainly going to happen in June. And well, Horizon 2 and three are chock full, sorry, building two and three of Horizon one are also full with GPUs, but they're in different stages of of how far they are along with energizing. Both buildings, all the three buildings are energized, but not all the racks are energized. As far as I can tell you, it looks to me like building 2 is in like a energized, rack energized process. So all the components are inside, including the rear door heat exchangers, but not everything is is working yet, like energized yet. So I expect the second building to be delivered in the first half of July. And I honestly expect the third building to trail that with maybe 3 weeks. So by my estimates, and I'll give you 3 cases, My bull case says iron can energize horizon one before the end of July. My base case is iron can energize horizon one at or before earnings. And the bear case is iron energizes horizon one before the end of August. So in none of my considerations will there be a version where Horizon 1 will not be energized by the latest by the end of August. So when Aaron said we're going to deliver Horizon 1 in Q3, in my opinion, that is limited to July or August. I don't see them run this all the way into September. So I hope there will be some revenue from the first building in June. And I just hope that they will be able to, hit it hard and finish off the building before earnings. Sorry, finish Horizon One before earnings, possibly even a little bit before that. And given that time frame, you can conclude what that means, because this serves as a precedent for all their sites and all their capacity. If they can point at Horizon One, during negotiations and say, look, we delivered this to Microsoft within this time frame. We have now found a way to do this twice as fast in Sweetwater. How much compute do you want? By when? We can do it. So they bring this to the table now. So I think we are in the final weeks before Iron is at full negotiation leverage. That's basically how I look at this. So Good news from Childress, in my opinion. I mean, the retrofitting is, they're retrofitting more. They are accelerating, taking down the mining buildings and the hash, and they are finalizing the completion of Horizon One. So across the board, Childress is doing phenomenal. This is just, I know I was sort of negative on Mackenzie, But the counter to that is that they are doing everything well in Childress. So, as someone, I mean, I can dream this site from the left to the right. I know every little component we have the satellite images of 30 centimeter pixels. It's like extremely dense. And to see the site coming along, as it is, as a shareholder, I feel kind of proud. So I'm very happy with the progress in Childress. Yeah, and I'll just conclude with Sweetwater. Sweetwater is a bit of a, oh, look, it's Mike Alfred. Is it a real one or a fake one? I just press accept, but so Sweetwater is is showing early signs of foundational works. I shared a few pictures of that to the public before, where you can see some of, some shapes in the ground at the data center footprint that Iron has talked about in their previous earnings presentation. But the The thing that they haven't talked about is what's most interesting to me. On the site, on the tracks where they have the bulk substation, they also made a primary substation that is currently energized. And to my estimates and calculations, yeah, okay, it's the real Michael Alfred. Why does it not work to approve people to come up and speak? It's very annoying. This could potentially be a 300 megawatt substation, which is next to a lot that is roughly half the size of the footprint that they have made on the other side of the road for the first 200 megawatt IT building. So what that means, Iron has a plot of land. that is half the size of a 300 megawatt gross, 200 megawatt IT load building that they plan to build next year under the DSX AI factory reference design for NVIDIA. So half of the size of that still has a substation that is equal in size to the 300 megawatt data center on the other side of the road. So why would they have such a small data center construction site with a 300 megawatt primary substation next to it? And why did they not talk about this particular part of the site in their 2027 plans, despite it being the first primary substation that they developed and energized at Sweetwater? So You know, that's food for thought. I'm going to look up my post that I made where I refer to this picture. So bear with me. My scrolling is okay. So it's this site. Sorry, it's this post. It's a post where I was kind of mocking Anibius. The only reason I mentioned Nibius here is because Nibius is making a site in Finland and the developer there posted a video where they where they were blowing up some rocks and the guy was asked to remove that video because apparently some of the rocks were didn't stay under the tarp but were like blown away to neighboring plots of land. So it's, I mean, that's literally a bit amateuristic, right? So I mean, I mean, it doesn't really matter anything. It's not done by Nibius, but it's just to show the difference between having a plot of land like this with a primary substation, which is as flat as it can be, which has like full irrigation measures under and next to the site. and it's currently being used as a parking lot I mean it's almost hilarious why is Iron not uh doing anything with this right I mean this is the first power that they can literally use at Sweetwater and they decided to uh tell tell us about the other side of the road where they have 2027 plans well why don't you tell us about your 2026 plans uh and don't lie to me, it's not a parking lot. I mean, obviously they're not going to use this 300 megawatt substation for a parking lot with floodlights. I mean, it's not going to be that. So this is, in my opinion, the surprise that they are going to tell the market. I had one theory, but that was basically, it's probably not going to happen, but I've had one theory that they could salvage the mining buildings in Childress and palletize it and ship it to here and then rebuild it. But I don't think it really works like that. So it's probably not going to happen. But what I do want to say is my current working theory for this plot of land next to the first substation that they built at Sweetwater One, Because of its size, which is half the size they projected to need for a liquid cooled data center on the other side of the road, makes me 99% sure that they're going to build air cooled data centers here. So my current theory is Iron's going to source GPUs from somewhere and they're going to announce to the market that they're going to build a 200 megawatts or 150 megawatt IT load, air cooled data center, maybe in the same data hall design that they typically run, like these long data halls, the typical high run buildings, because they don't have the space for a liquid cooling plant here. I mean, this is a big piece of land, but it's not big enough for liquid cooling. So just has to be echoed. That's just, there's no way around this in my opinion. So this could be a very interesting development in terms of, you know, ARR and guidance. Maybe Iron will tell the market in August earnings, oh, we, surprise, we have another 200 megawatts coming online. You know, will it be 2027 or will it be 2026? I mean, it just looks to me that if these are air-cooled data centers, then how long would it really take to build them, right? But bear case or base case, it's going to be first half of 2027. So that would mean, you know, more ARR next year and obviously something to monetize. So that's what I've seen so far across the sites. And I think with that I have touched on all the topics. I tried to approve Mike Alfred to come and speak two times, but it just doesn't seem to work. Or is he here?

Mike Alfred: Can you hear me?

Frans Bakker: Okay, yeah. Okay, Mike, go ahead.

Mike Alfred: Yeah, you can hear me. Thanks. Thanks again for the coverage of the space. And The last time I think I was here was maybe the week of the earnings, last earnings, maybe the week before. I don't remember exactly. I've been pretty slammed personally. But what I tried to convey at that time, to the extent possible, is just how much bigger I thought what was happening in the spaces than any of the dialogue that I was hearing around the space. And obviously, I knew at the time we were going to acquire Mirantis. I knew at the time that we were gonna acquire Nordstrom and go into Spain and the sort of the broader European market. I knew about the 800 megawatts in Australia. I know about a bunch of other sites, right, that haven't been announced. As I've said many times before, like this is not a company that likes to over-promise and under-deliver and constantly be announcing things that don't happen. And so it's a give and take, and it's something we talk about in the boardroom a lot. And I'm obviously an advocate for providing as much transparency as possible, but not to the extent in which it impairs our long-term fundamental value creation, which is the real focus here. As I said on that last time I was in the space, it may be annoying to not have all the granularity, but remember, our competitors don't have all the granularity either. There are other folks in the market who would like to emulate various things. They'd like to be competing for the same resources, the same sites, the same capital, etc.. And I'm not necessarily sure that's a good idea. And so anyway, what I said then was that, look, I don't know what's going to happen to the share price. I heard some comments earlier about, you know, tanking for the share price or whatever. I'm not really sure I understood the comments, but I can assure you, like everybody wants the share price to be higher. It's good for the cost of capital. It's good to be able to raise more money. It gives you more headroom. It gives you more momentum. And as I've said many times, like when I sit in the boardroom and I'm literally looking at the whole strategic map, I can see everything, like the dashboard of everything that's happening and the share price movements on a day-to-day and week-to-week basis, like literally don't make any sense. Typically when things are accelerating, the share price in the business, the share price is going down. And then when Things are flattening out a little bit. For some period of time, the share price is going up. And so I've come to the conclusion that a lot of the share price movements are largely decoupled from anything actually happening fundamentally. And especially with small and even mid-cap companies now, a lot of the daily trading action is technical. And I've seen a lot of people be confused by that over the last three, four years. And no matter how many times I say it, still seems like every few weeks or a few months I see somebody else capitulate. make a major mistake because they believe that the movement in the share price actually signifies something bigger than is actually there. So I can assure you that I'm an advocate every day for doing things that would cause the share price to go up. There's no reason for the share price not to go up. My belief is, though, no matter how much I advocate for that and no matter what we do technically, in order to manage the capital stack and to raise money and to structure things correctly in order to sequence the growth, the timing in which the share price will go up is completely out of anyone's control. And I jokingly refer to Jane Street some of the time, but I do think Jane Street has more control over the daily price of an equity like Iron than the fundamentals do. But over a two or three-year period, the fundamentals absolutely outweigh the impact of Citadel and Jane Street, et cetera. And so, you know, I look at the field today, like I think there's a sort of a systematic relative undervaluation still between the folks that were viewed as pivots into this space versus people who are viewed as more of a pure play. Remembering that, like, in effect, even Core Weave and Nebius were actually pivots, right? But people don't think of them as being in crypto mining or think of them as being in a Russian search engine operator or anything like that because they've been around saying they're doing what they're doing now long enough that they've convinced a whole crew of kid analysts that they somehow were always doing these things. Within Iron, there's no negativity whatsoever towards those other neo clouds. In fact, like I'd say, they're largely not even really discussed, because at least internally, there's a feeling and a belief that what Iron's doing with vertical integration is just different. And so there's an awareness that those other companies are there, and they certainly come up and... conversations, but there's certainly no negativity internally whatsoever. In fact, I even heard positive words about Nebius internally at several points over the last three to six months. And so while I have some skepticism about some of the other companies' ability to execute in the coming years, it's sort of neither here nor there. It doesn't really matter. The only thing that matters is whether Iron executes, because that's the company that we're talking about. So the main focus, I mean, obviously there's a lot of day-to-day operational stuff that needs to happen. Like you're constantly acquiring new sites. You're constantly pushing those sites forward. You're constantly building stuff. You've got thousands of people now working in the company. But as you scale, you know, there are a lot of scaling challenges around hiring because you need more and more senior talent across the organization globally in order to scale up. a global organization. So there's a fair bit of time being spent involving the board as well and making sure that that's done correctly. And that's harder to do now because you've got these AI labs spending, inclusive of the hyperscalers, spending incredible amounts of money relative to any benchmark historically of what would be normal on some of the senior talent in the AI space. And so it's sort of inflating the cost of labor across the whole space. Thankfully, Iron is an employer that a lot of people apparently want to work at. So there are a number of candidates that a company like Iron will see that won't be talking to competitors necessarily because they've decided already that they want to work at Iron, but they certainly aren't going to take a senior job at Iron and be paid significantly below market. And so you've got to rationalize all that stuff, and that involves some complexity because you've got a matrix. right? You've got levels and roles, and you've got existing employees who you value greatly and don't want to leave. You need to push the business forward, but you also don't want to rock the boat so much that you cause people to start jumping off the boat as you bring in more senior people. But I assure you, like the work that's being done there is at a very high level, and the people that are being signed and the people that are being attracted is a very high level. and the changes we're making to the structure of the business to make sure when you scale from 20 billion to 50 billion, and then 50 billion to 100 billion, and 100 billion to 250 billion, that at each one of those steps, you're gonna need, in some cases, better and better people, and you're gonna need different structures in the organization. You're gonna need more of a global structure, right? Where you have people that are in charge of functional areas, and you have people that are in charge of geographic regions. And you're gonna have to feather that all together so it makes sense and it's rational. And so it doesn't slow the organization down. That it's not too complex and hierarchical, but it's also not too flat, right? Because you don't necessarily want Will doing everything every day. Will's MacGyver, he can do literally everything, but if you try to have Will do everything, then eventually the organization will break because it won't scale. But these are all what I consider to be high grade issues, right? Like if you're having these issues, it's because the market opportunity is so large and growing still so quickly that you're needing to do far more than you might have expected if you were sort of wargaming this two or three years ago. And so as I'm watching the evolution of like the tenor and the tone and the substance of these conversations in the boardroom, what I'd say is it's reflecting the increasing scale and the increasing opportunity that I see. So there's going to be a lot of new hires, there's going to be a lot of marketing that feels bigger than competitors because we think the opportunity is just so much bigger than some of the other companies in the space. If they don't want to do any marketing or they don't want to do big marketing, that's fine. But Iron's not going to be quiet like a mouse and hope people understand that it's there. Iron will be globally recognized in the sector and eventually with almost universal name recognition. It's not obvious to most people why. that would be valuable today. But again, for the same reasons why it was non-obvious three or four years ago, how big Iron was going to be today, it'll be more obvious in a couple of years. And again, I don't necessarily think it's a great thing to try to over-educate everyone in the market about it. People who are really forward thinking will be able to visualize some of the things that are likely coming. And some of those things will actually be accelerated by sort of universal global awareness. Remember, this is a company that had basically zero brand awareness a few years ago, and we'd like it to trend towards basically 90% or something, right, is sort of my read on the situation. So when you're going from zero to 90%, first you start with smaller stuff, then you go to bigger stuff, and then you go to bigger and bigger stuff. And hopefully you sequence that along with the actual commercial developments and the increase in the market cap and the availability of capital, et cetera. And I think that's all sort of being done correctly. Every year we've got the compensation conversations, and so we're in that cycle now. There's a whole bunch of things that are, every single year you've got to do, right, to run a public company. And not all of them are as fun and sexy as what Franz gets to do, which is pull up satellite imagery and look at really cool stuff and predict the future. And as I've said many times, I think that's all really cool. And I actually learned quite a bit from listening to Franz, not only from the substance of what he's saying, but also from gathering evidence of what market expectations might be from the best analysts in the world, which are not sell side analysts at big banks, but are actually folks like the folks in this room who really do granular work on it. And so as I listen to it, I think about it, it helps reset and recalibrate my own expectations. And that's cool and fun and fun to think about, but candidly, a lot of the stuff you got to do every day to make sure that we capture the value is we need to make sure the audits are good every quarter, right? We've got to make sure the compensation philosophy really makes sense for attracting and retaining talent. We want to not get on the wrong side of the proxy agencies unnecessarily, although if you're growing and you're doing the right things, shareholders will generally be supportive either way. I feel like to some degree, the proxy agencies are actually losing clout over time as people realize that they're basically like toll booth collectors. They don't really add that much value. And as more and more investors want to be more active in thinking about how the companies they invest in should be managed and not so passive. But we're going through all that stuff just as we do every cycle. There's the annual cycle stuff, there's quarterly cycle stuff, and then there's just the day-to-day updates. And, you know, I heard some commentary on liquid cooled versus air cooled. And those conversations are constantly ongoing. And to some degree, the variance and the volatility that you sense around some of these issues is actually more to do with customer volatility than it has to do with internal volatility. So it largely reflects the fact that the customers don't even know what they want at most parts along the curve, especially in an industry that's growing this fast and changing this fast. So you have customers that will literally tell you, One week, we have a mandate from literally the CEO level to do XYZ thing. And then one week later, they disappear for three days and then come back one week later and say, never mind, we have a mandate to do the exact opposite thing. And so then you go back and you say, okay, we'll reframe this and take another run at it. And then one week later, they never mind, we change it again and we want to do a third thing. And this time it's really the real thing, right? And I know that every single player in this ecosystem has experienced this, so this is not unique at all. But the thing that gives me confidence is just seeing how every single GPU basically that's plugged in is contracted right away. And increasingly, pretty much every GPU that's available in the near future is contracted in advance. And in some cases, you can contract them well in advance. The other thing I'd point to is just the SpaceX IPO. I did some crude Analysis, it's not like deep analysis. I didn't put it into a spreadsheet or anything like that, but just thinking about the fact that in the near term, probably the best immediate opportunity that's not pie in the sky rocket stuff for SpaceX is just to do more of what they did at the Colossus location in Memphis, where they sort of overbuilt for their own purposes. They overestimated their competitiveness in building foundation models and running an LLM, and so they just flipped this resource that they had developed organically from scratch to Anthropic and Google effectively, and are now gonna be generating like almost as much revenue from that from the rest of the business or more. Starlink is profitable, but launching rockets hasn't really been. This business model could actually be quite profitable for them. But it raises the question of like, if that's true, and as you look at it a couple years from now, if this sort of data center business or NeoCloud business for SpaceX were to grow, and the valuation were still to stay in the $2 trillion range, then clearly there's been some mispricings of anyone else in the ecosystem who's able to deliver similar infrastructure on a pure play basis. Certainly you don't have rocket launches and Starlink and some of these other things. And you're not talking about doing data centers in space and some of these magical unicorn fairies type things. But the reality is, If if the terrestrial data center business for SpaceX End up really being a thing then it would trigger in my mind a pretty dramatic repricing I mean I can imagine a world where literally you the same company like for example cipher Or or or even like a core weave or something core weave has the terrible cap structure It's unclear to me that they'll ever be profitable on a gap basis. They're probably just going to consistently the executives are going to sell stock and they're going to keep issuing convertibles and GPU financing, and they're going to hope to show a profit in like 14 years from now or something. They could probably, I mean, absent the dirty cap structure, they could probably double or triple their equity value by just having it be owned by SpaceX. So it might be totally accretive to just flip it and basically let SpaceX take out the cap structure and just take the assets because the assets owned by SpaceX is going to be imputed a much higher value by these retail investors who just love Elon Musk and love rockets and thinks it's sexy and all this, that they would be willing to mark up the value of the assets if it lived inside of SpaceX. And so similar to a company like Cipher or a company like TerraWolf, yeah, they're not neoclouds, but if you look at their site portfolios and what they're doing from a development standpoint, if SpaceX were to acquire company like that now, they could acquire a company for 10 to 15 billion, that the moment they integrate it in and they take that pipeline and point it at the future SpaceX customers, maybe the value is immediately like $100 billion or something like that. And I don't think that's actually an exaggeration. It might actually be more. And I don't think enough thought has gone into that because I don't, remember, this was sort of like a shotgun wedding between Anthropic and Google. And in SpaceX. Obviously, Google's a shareholder in all those companies, and there's a lot of like cross-pollination, but so they know each other, but like doing these data center deals right at the last minute to sort of bolster the IPO is kind of an interesting strategy. And obviously, at the same time, you know, Elon's litigation against OpenAI failed and he's really mad at Sam and hates Sam and Sam's the Antichrist. So what better thing to do to try to kneecap your competitor than to like shift away some of your computing resources to one of their biggest competitors? It sort of all makes sense. Like if you think of this as an existential fight to the death for AI supremacy globally with all of these firms being part of it. But because it all happened so fast just before the IPO, and again, they shoved XAI in there, and now they're talking about shoving Tesla together. There's a lot of like noise in Elon's company. He's a genius, but a lot of this stuff really wouldn't pass the governance sniff test if it was done by anyone else, but he's sort of afforded a lot of latitude. But because of that, I don't think a lot of institutional investors have really had the time to really digest like what the implications are. And I think once these folks do that math, like basically over the next few weeks or over the summer, I think when you come back in the fall, there's possibly, again, assuming SpaceX holds up, and I'm not saying that's a guarantee. I think it's unlikely that it collapses before Anthropic, candidly, because there's still gonna be too much buzz about AI during the buildup over the next two, three, four, five months, however long it takes to get both Anthropic and OpenAI out. But assuming it holds up, then at some point, people are gonna start to do the math and figure out the implications for what it means for the rest of the data center industry. And I think that could trigger like quite a violent rerating. So look, I'm continuing to be pretty bullish and constructive on the macro backdrop on some of these idiosyncratic factors like these IPOs. It still seems like the demand is extremely high. And I see no reason to turn bearish in the short term on any of this stuff. But I do think that the mania about SpaceX-- well, you could look at it on one hand and say, OK, this is a clear signal of like possible like 1999. dot com behavior, the other side of the coin is it could be quite good for those of us who've positioned in equities that are still, at least on a relative basis, like fairly undervalued objectively relative to how the market is wanting to value at least the data center component of SpaceX's business, which I personally believe is the most realistic business to provide near-term growth and profitability. Starlink is a great business, I just don't know how big the growth's going to be relative to the AI data center business. And the rest of the business is basically losing money. So it's unclear to me what people are buying there other than like they love Elon Musk and they love Tesla and they made a lot of money on other stuff with him. So that's kind of some of my thoughts, but thanks guys for holding the space as usual and looking forward to seeing the company share more stuff in the coming months.

Frans Bakker: Thanks, Mike. so you could look at if you have 3 degrees of AI cloud providers and the top tier is SpaceX now and Nebius is the middle tier and Iron is the lower tier and I'm talking in terms of valuation per megawatt or valuation per ARR, then certainly maybe SpaceX is overvalued from a Elon Musk ingredient perspective, but iron certainly is undervalued from whatever you want to put as a reason behind that. Could be Bitcoin, could be, people have jokingly said Australia discount. There are all kinds of reasons to make up, but it doesn't really matter what it is. It just shows that objectively there are two valuations that are higher than Iron's currently. And I don't think the differentiator is software or space. I think this is just, the market is not making a proper differentiation on a financial level as of now. I think it is a lot of things, but not that. And So I support the idea that there will be a re-rating for companies like Iron that are not getting the benefit of the doubt in terms of future projects, expirations or plans, but more being only valued on past performance and getting discounted on their future plans. So yeah, I mean, this is my personal view on that. I do agree with you to that extent, but I just have a personal view on it. What does make me wonder with all this buzz around the Anthropic wanting to limit the most advanced models that they build to the world, Doesn't it mean that Vera Rubin, which is not currently in use and being used to train models, will see less of an appetite from, you know, across the board? Because from what I understand, for inference, you don't need the newest models. But if training on Blackwell is already accidentally creating weapons, to a certain degree, software weapons or whatever, a threat to humanity. What happens when Vera Rubin is going to train LLM models, right? I mean, I know it's not only about the model, also about the cluster size, and I'm sure that Anthropic has a special way to go around, but at the same time, would Anthropic really want to run Vera Rubin's? you can just ask yourself straight up this question. They created a nightmare on Blackwell. Would they really want to have a Mythos 666 built on Vera Rubin is going to literally destroy the world, according to Dario, of course. I mean, is this not something that is sort of bearish future iterations of GPUs and to an extent sort of bearish Nvidia's roadmap? If unless there will be a consensus amongst governments how to handle powerful LLMs, will there even be powerful LLMs that need this much compute in, you know, the future iterations of GPUs? I'm just curious if you have any thoughts on that.

Mike Alfred: Well, look, I I've been saying for a while that I thought this whole Michael Burry conversation around the GPU depreciation was out of line. I mean, I said it from the very beginning because from what I could tell, there was not enough good infrastructure to even plug in all the Blackwells, right? And so we're behind like a generation on infrastructure already. I don't think the chips are the constraining factor. I think the infrastructure is going to continue to be the constraining factor. And I think Jensen Wong has actually come to that conclusion himself. And that's one of the reasons why they're more open to hosting these conversations and having partnerships with the people that are more involved in the physical side. I think there's an awareness that you can't just be a NeoCloud and an asset, a fully asset light way and hope to be able to deliver the right outcomes for your customers long term. I think it was a growing awareness that that's true. That took a long time in the Bitcoin mining space, right? There were still people as late as 2025 who hadn't figured out yet that Marathon was going to underperform the rest of the like true like operators in the space that were able to build their own kind of greenfield infrastructure from scratch and run it at scale, that that was going to be a better economic model than hosting with a bunch of other folks and paying a hosting cost and having a lot of operational issues and maintenance issues and et cetera. The same thing is going to play out in the neocloud space. It's just so early that people are still trading on headlines and deals, not on actual execution. And I think You look at Coreweave year over year, I mean, Coreweave is down, I think, over the last one year. So people focus on what stocks did in sort of arbitrary periods. And you have to choose a period to do an analysis, so you have to. But I think if you don't look at it clearly, you might miss this, because you might say, well, it's up a bit over X period. You know, people looking at Nebby's up year to date a lot saying, oh, well, Nebby's must be better up year to date. Well, look back even just a year and it's not necessarily beating some of those sort of Bitcoin miners that pivoted. I think HUD8 actually passed a bunch of the others recently, right? It went on a huge run. And so HUD8's done a really nice job of marketing itself. It positions itself as having full control over sites where there's no guarantee there's going to be power. And this is one of those things, another example of something where, yeah, maybe we could do that too. Maybe a lot of companies could do that too. But if you go around telling people you have 40 gigawatts of power, then they're going to discount it anyway. And then if you don't actually deliver that, then people say, look, these guys overpromised. And so I think I personally am more comfortable with the underpromising at least a little bit. versus versus overpromising. But I think a lot of the neo clouds have sort of overpromised in mass, like across their whole business, they've overpromised along a number of dimensions. And we'll have to see like, like what actually happens. But to the to the direct answer to the question, like I'm not an expert on the latest, latest generation chips, I think what you'll see it iron is like a you know, a desire to basically serve the customer where they are. So if the customer says they want to use a specific chip and that makes sense to us, then I'm sure, you know, a deal can be done. It's less about trying to dictate, hey, we think you should be using these chips, all right, at any given point in time. Like if a good customer that's creditworthy wants to do a five-year contract, I'd say, with a specific model, then that's probably going to work. Right. I think the only real strategic consideration is when you get into using something outside of NVIDIA. So the moment you go into using AMD chips or the moment you go into using Amazon or Google chips or doing some of these other things, then if you're going to own your infrastructure forever, you have the potential situation where the primary customer walks away, but there isn't anyone else who wants to use those particular chips at the end, whereas NVIDIA chips are so universally respected and desired that you can pretty much always find somebody to use NVIDIA chips of any generation. And it's unclear at this early stage whether people feel the same way about chips that are sort of made or designed or marketed by anyone else. And so I think that's a real consideration. As I said, I'm not an expert on all the details of that. I, of course, participate in all the conversations around it. But I think if you really think about where Iron's going to win, it's starting at the base layer and knowing that you have a huge advantage there. Like you're just, you have more gigawatts, you have more development capacity, you have more construction expertise, you have more ability to deliver the physical components at a global scale. And yes, those other components are important, inclusive of which servers you use and chips and racks and networking and orchestration and all this stuff with the layers on top. And you certainly don't want to punt on that, just like you wouldn't want to punt like Coreweave and sort of outsource all of your physical infrastructure to someone else, because that would lead me to bad. You wouldn't want to necessarily outsource all of your software, all of your networking design or something. You'd want to control as much of it as you can, but I don't think the mix of which chips you have, from which generation or model, is necessarily going to be the determinant. Because what we're seeing is pretty much all the chips, particularly the NVIDIA chips, pretty much all of them are contracted once they find a home to actually be plugged in. And outside of that, I don't know. But so far, that seems to be the demand environment. And I expect that demand environment to be similar for at least the next year or two.

Frans Bakker: Butcher, before I give you, I just want to say one more thing about this. This is like what you said, Mike, about the customers that are dictating or flip-flopping with their demands that one week they want this and then they come back after three days and want something else. This is exactly what my point was. Basically, when you're the designer of the data center and your customer is asking for air-cooled GPUs, As an iron, in this point in time, you need to decide, am I going to build an air cool data center or am I going to retrofit an air cool data center, which is probably going to be looked at possibly, and that's why this consideration is here now with the overreach of the governments, which is potentially starting to happen. Is my air-cooled data center in five years when this contract rolls off not an opportunity cost, even though these GPUs could possibly still be used at some low-level inference or just a token generator? But on a megawatt basis, could these air-cooled GPUs not be an opportunity cost? So when a customer asks for a chip, as an iron, you need to decide, do I want to build a data center that is able to run in the future, run liquid-cooled data center, liquid-cooled chips. so, basically, what it means is you make a air-cooled data center that is upgradable to liquid in the future, because you know you're choosing a path to go with. when you go with air-cooled chips according to your customer's request, but the customer in five years can just go to a data center operator that has liquid-cooled data centers, you know. So you have to make this decision based on your customer, but you also have to take into consideration what happens after these five years. So the optionality.

Mike Alfred: And then there's, it's not just the optionality of the specific site, though it's also. the optionality of that specific customer. Because you may be able to use your willingness, because if you have a lot of infrastructure, you can use your willingness to do a specific design for that customer to get a new customer that otherwise wouldn't sign with you right then. And then that customer feathers into gigawatts of future business down the line because you've already got that relationship in place and you sort of won that goodwill and you figured out how to work with them. So there are a number of other strategic dimensions even beyond what we've discussed for why you might make certain decisions, some of which I've never heard discussed in these spaces. And for that reason, I'm not going to say any more on this particular topic, but I'll just tell you that there are like dozens of reasons why you might do something. To some degree, it depends on your risk tolerance. To some degree, it depends on what your time horizon is. Because if you're playing 10-year, 20-year games, or longer, right, where you're building multi-generational infrastructure potentially on sites you plan to own forever. It's a very different game than if you've given an option to your customer, like some of these co-location deals have done, where like you're not going to own that site. So the economics you get in that 10 or 15 year lease is all of your economics potentially for that site. So it's a different conversation than if you're going to own, let's say, 30 or 50 or 100 gigawatts of infrastructure long term. And you're trying to figure out the customer mix over 50 years. And so you've got to find ways to break the logjam with some of these major players, especially the ones who are able to do a lot more business in the future, because maybe that's more valuable than somebody who may give you the perfect deal today that you can put into a press release, like a lot of these other neoclouds have done. You can put in a press release that pumps the stock in the short term, but doesn't actually give you the leverage and the optionality and the sort of long-term opportunities that you would have if you had done a deal that the sort of kid analysts on paper are going to look at today and say, Oh, that isn't as good as XYZ deal, because they're only thinking about it in terms of the stock price impact in the next two months. They're not building 20-year infrastructure. I mean, they're 22 years old, so they're not thinking about what they're going to feel like when they're 42. They don't even think they're ever going to turn 40, but yet they have an opinion on every single deal that every NeoCloud signs. And again, the implications could be 20 years long for a serious infrastructure developer, but for other players, they may not because maybe they won't even own that site at some point. So I think Butcher wanted to talk, but I just wanted to throw that in there.

₿itcoin ₿utcher: Mike, I had two questions. Given the sensitivity, you obviously can respectfully decline, but... Two things came to mind as you were speaking. The first was you had referenced the share price as it pertains to the cost of equity. Can you comment on any conversations high level with institutions on what's preventing more long only funds from investing in iron? Because, I mean, Franz had mentioned earlier, I think the only pushback I would give to you and Franz on our undervaluation at this point in time is as of last quarter, our quarterly revenue is $33 million. And it's something I believe that's being corrected in real time. And I believe that time is coming. But my suspicion is that's part of the hesitation with long only funds and having a stock that trades at a four beta. But if you had any comments on that, and then the second piece, whatever you can comment on, we certainly appreciate. You mentioned the Mirantis acquisition earlier, and one of the things I'm most bullish on with Mirantis is they had a post. with Saturn Cloud on how they can use Cordant to essentially replace the capabilities of hyperscalers. Yet, if I go to Iron's website, I don't see any mention of Mirantis. If you can speak at a very high level, if there's going to be any additional efforts to co-brand in the future or how you can communicate to the marketplace those additional menu options that you have at Iron Cloud going forward, I appreciate it. Thank you.

Mike Alfred: Yeah, on the Mirantis thing, I mean, it's so early. I've done a lot of M&A across my career, and sometimes it takes a while. to integrate and to the extent you're trying to integrate. And this is a situation where maybe I would characterize it maybe as a partial integration, where like you kind of want some of the things to continue to operate like somewhat independently, but you also want to integrate some of the other things. And so some of those things kind of play themselves out. I think, like I looked at the deal before we did it and I came to the conclusion that it was highly asymmetric where like, If it ended up not being a great acquisition, it wasn't going to cost us very much. If it ended up being a great acquisition, it could add significant value, like orders of magnitude more than the potential risk to the downside. And to some degree, it's almost like an insurance policy. And so what I hear internally is that the value that the company is hoping to get from that is happening in an accelerated pace. But remember, there's dozens and dozens of different vectors for that. Some of them are internal, like, hey, we have to shore up X, Y, Z component of how we run the chips in the data center just so we can make sure we're maximizing the performance of the clusters that we're running, right? So like that might be something that's almost invisible to the market, but you'll see play out in various metrics over time and should feather all the way through, right? the operating profitability and things like that over time, but of course, aren't going to be visible out of the gate. It's very early. There's really, I'm sure the company will provide more color on this over time, but I remain confident that that was a very good decision. And not all of that is going to be because of like very immediate co-branded opportunities or anything like that, but more because of what it enables across time for the whole company. And again, there's only-- I just want to be careful. There's only so much I want to say on that right now. And then in terms of investors, I'm not so worried about it. I hear the concerns people have. And Mike Power, I know, works day and night to try to address any and all concerns. I mean, the practical matter here is that it's very hard for most funds to hold stocks that are as volatile as a lot of these Bitcoin miners who became AI stocks. These stocks go down 50% two times a year, but then they end up higher every single year and significantly higher over several years. That's not the return profile, even though it sounds great if you're an individual investor and you just want to make a lot of money, and you can if you just hold them. A lot of institutional investors have all these artificial constraints that have to do with things that aren't going to make you or I any money, but they have to do because they don't want to get fired by their clients. And so some of that is going on broadly in the sector. Like it's much easier for them to own NVIDIA than it is for them to own a small cap data center developer. And then there are other things like index inclusion and geographic stuff that are just like checkbox things that just need to be resolved. And that just takes it just takes time to do that. But I don't think there's any real like meat on the bone there. Like there's not there's nothing. deep in material that needs to be fixed. I suspect what will happen is the next time that we go into a strongly risk-on environment, the stock just starts going up. And when it goes up, it becomes very reflexive. And once it goes up enough and the volume's up high enough, it's safer for some of these bigger investors to come in. And then once the bigger investors come in, it's kind of like a flywheel. And so we haven't had that in a sustained way. Like every run we've had since 2013, sorry, 2023 in Iron has come right before like a major sell off. So there's never been like a period of time where we just went higher and then sustained at those levels. And some of that speaks to market cap, right? Like the market cap is still relatively small. Some of it speaks to what you said before, which is like if you look at the financials as a standalone, you don't understand the business, then you'll say, well, wow, someone said this to me on a podcast the other day, like, well, it's trading at 100 PE. I said, well, explain to me why the earnings of a company that's like monetized 3% of its power matters. Like the value still is heavily on the balance sheet side until that the balance sheet becomes the income statement, right? Because you leverage this massive and growing balance sheet to generate net income. over time. So really, what you're betting on, how big can we get the balance sheet before we focus on net income? If you go back to the Amazon model, Amazon refused to allow huge amounts of taxable income to flow to the bottom line for a long time, and instead, just poured more and more money into CapEx to continue the growth flywheel. And a lot of traditional investors are like, I can't own Amazon because it never makes any money. I'm going to own FedEx instead. People were saying this 10, 15 years ago. And the people who bought FedEx, because it looked better as on the PE, got smoked by people who bought Amazon because they like that there were so many investable opportunities in the business, there were no reason to show profits. So yeah, like, look, you've got to convert the opportunity set into revenue, and then you've got to convert the revenue into net income. You either believe that that's going to happen or you can't, but it's so early in that process, and there's still more like building the frame first, right? Like if If you wanted to stop and put all effort on just extracting maximum revenue from the asset bases that exist today, and then from there, maximize the profitability, you would just cut down on all new efforts. You would stop doing as much new stuff, and you would just focus on making the numbers look great so that you could convince these so-called great institutional investors to believe in you. And again, if Amazon had done that, it would have been a wonderfully profitable business 20 years ago, and it never would have gotten much bigger than maybe $50 billion in market cap or something, right? If they had listened to that type of feedback early in the trajectory, you either believe that your terminal market size and market opportunity is much bigger or you don't. But if you do, then yes, there are going to be some growing pains and some hiccups, and maybe not every forecast on the financial side is hit at the exact cadence that you'd like initially. But if you know what you're building and you know the ultimate terminal value of what you're building, then that's probably still the right thing to do it like that. And so, yeah, at some point, if you can't deliver, right, if at some point you just, for years, you've been saying you're going to grow and you're not generating the revenue you said you're going to generate, then that's going to be a problem. But at this stage, I personally don't lose any sleep over that at all. And of course, I look at this as a board member and as just an independent investor. right? It's the second largest position in my fund. I haven't traded it. I haven't sold a share. As I've said many times, I have no intention of selling a share. I don't really think we're anywhere near what I think the long-term value is. And so I candidly barely even look at the revenue in the very short term because it doesn't mean as much to me as what the company is doing strategically to make sure that we capture the maximum share of this sector globally. And I think this is gonna be one of the biggest sectors in the world over the next 10 or 20 years. And so I personally would prefer that we just keep going bigger in terms of looking for ways that we're gonna be bigger two or three or four or five years out and less on trying to maximize the metrics in the short term. Even if there are some investors who that means they won't be able to buy the stock until it's 300 or something. If they wanna wait till 300 because they have some arbitrary rule in their investment committee that requires them to wait for that, then that's fine. Like they can buy it at 300. And they'll probably still have a decent return if this industry is going to grow for 20 or 30 years. But it's not like the company's job to convince every conservative investor. Again, these are the same people who passed on Amazon. They finally bought Amazon when Warren Buffett did, or when Warren Buffett's deputy did. But that was a decade or two after most serious technology investors liked Amazon. So you just have to ask yourself, which investors are we trying to cater to? Are you trying to cater to people who didn't get Amazon until three years ago? Or are you trying to cater to people who got Amazon 20 years ago? And personally, I'd rather just build the business and let the right investors find you. Even though, again, that may be frustrating to people who would prefer that things are much more explicit and obvious. But I think the returns are asymmetric here because it's not so obvious still. Even now, even after all the progress that's been made and all the proof that there's real, there's something hard and real here. There's still a lot of people that are like, well, I'm still not sure because the revenue for last quarter underperformed expectation by X percent. And it's like, okay, I'm not going to argue with you. That may very well be true in your spreadsheet. I'm just, it's unclear to me how that's going to change kind of the three or five or seven-year trajectory of the value of the company, which is what I'm really tracking over the long run, because that's where all the big money comes from, right? Like if I'm going to make I don't know, $300 million from being involved in IRIN. It's not going to happen based on the revenue for this quarter and whether we hit that or not, right? It's just completely irrelevant. It will happen, though, if we track along these really important high-level metrics over five or 10 years. Then it becomes much more possible, right? And so that's really where my head's at, and I think that's where a lot of the folks are internally in IRIN.

₿itcoin ₿utcher: Appreciate that, Mike. And while we have you up here, Franz, I don't know if you had any other questions for Mike. The only other thing I can specifically think of right now as it relates to, if you can speak to There's the RAISE summit coming up. If there's any commentary you can give on that or just specifically kind of touched on it, thinking bigger, but Iron as a global player with the recent announcement that you referenced in the APAC region as well as Nostrom, if there's anything you can talk about with that, that we possibly haven't covered, we would enjoy hearing your view on that.

Mike Alfred: Look, I used to go to conferences on Iron's behalf because Iron was so small from a headcount standpoint in the US, right? When we went public in the US, nobody knew who Iron was. We had like almost no staff operating in the US, right? Even though we had these sites in British Columbia and Childress, there was very little name recognition. And so I would go to some of these things, like on behalf of Iron as a board member, which is very unusual. Right, if you go to most of these conferences, it's like the CEO and CFO, and that's basically it at a lot of these investment banking conferences and a lot of the industry conferences, you get more of the technical folks and the sales folks and things like that. So I think they did actually call me. I think Dan did call me. It was the last year, the year before, and said, hey, like very last minute, he's like, hey, do you want to go to this thing in Paris? And I turn to my wife and I'm like, I don't really want to go to Paris, right? Like it's just, it seems unnecessary. Like it seems like a lot of work to go for like floating down the river or something and going to dinner. So I didn't end up going and I've been going to less and less stuff, but I think it's great that iron is increasingly going to be at everything. And that may annoy some competitors and make some people feel uncomfortable or whatever, but I think it's awesome. Like as a shareholder, As I said earlier, I want iron to trend towards like 90 plus percent global brand awareness with anybody who's involved in the industry. And so that's why you do the things that they're doing right now. So I don't have any intel on that. I oftentimes don't even like see where the company's gonna be that week until I see it on Twitter myself, right? 'Cause I don't have time to follow that. It's just not, it's not germane. to what I do, right? And I've got a lot of other companies and sectors to follow, and I've got four other board seats, including one other public one. And some of those companies are different parts of their trajectory where they need more attention. Iron is a very well-run company that is on a track where, candidly, it's gonna be harder to screw it up than to make it work, right? Because of the hard work we did over the last four plus four or five years now. So all the things you need are basically there, and now it's just guardrails to make sure that the value accrues to the shareholders over time, and that shareholder value is created consistently over time, and it becomes kind of a flywheel. When you're involved in a turnaround, like I am with Bakkt, it's a little different where there's heavy lifting that somebody like me can still do during that process that can change the slope of the curve. I can literally help inflect the company myself through various efforts. because it's the company that needed so much work to turn around. I mean, if you look at what we did in the last eight months alone, like we basically turned over the whole company, like swapped out almost the entire board, went from long term debt and like no cash, now 80 plus million dollars of cash and no debt. Got rid of the super voting shares, right? Completed an acquisition that that sort of ice had left out there that like fell as a burden onto the board, but this board is not going to just punt like the previous board on everything. This board just said we need to do this, and so we got it done. There's a very different company now, but that kind of work requires more tactile involvement day-to-day. Iron just needs classic governance at this point. It doesn't need anyone to do anything magical. You don't need to come in with a power move because it's not a company that needs to be fixed. It's just a company that you need to let run. So I watch all that stuff too. I think it's great. I'm aware, like the chief, the new chief marketing officer is great. There's going to be, of course, a number of other big announcements this year that nobody's predicted, which is always fun for me. I listen to these spaces to see what people predict, and then I try to remember who said what, and if anybody got close to being right, I try to call that out later. But like some of the stuff that Iron's going to do is just so big that it would just be hard for for anyone to necessarily be able to predict it until they see it, unless it leaked or something. So like you're just not going to know until you see the announcement. But there's a lot of stuff coming. There's always going to be stuff going on. I'm not going to be going to very many things anymore on behalf of anyone, candidly, just because my kids are at the age now or like it's just magical being with them when they're three in one and I want to be with them like all the time. So it takes a lot to get me out of the house candidly these days. So I'll be watching from afar. when the teams go out to these various different events. And I applaud the work that they're doing in terms of getting to and sponsoring and sort of being in the conversation at every event. And I think that's only going to accelerate. I don't think you should expect that to slow down. They're not going to like tap the brakes on that stuff because from the perspective of where I sit, it's working, right? So like anytime they say they want to accelerate, I say, why not go faster, right? And that's basically working. So I think we should keep going.

Frans Bakker: Thanks for that, Mike. I guess we can start to look at closing the space after 2 1/2 hours. I just, one last thing I wanted to ask you. talked about some of these things that Aaron's talking about and planning with customers and relationships with, you know, some of these very big players that it's completely unclear to the general market what the long-term plans are. So we are talking about plans that extend beyond five years from now, for example, and you're talking about building a decade-long business. At what point in time, you know, are investors really going to be, you know, somehow informed of the long-term plans of Iron and their partners. Is that something that will always be shoved forward into the future, like kick the can down the road? Like, oh, we just have another contract and we have another five-year deal and we now have this client. Or will there be a point in time where something that is now being developed with Nvidia and Dell really becomes like a global strategy where Iron will say, look, you guys, now we have all the puzzle pieces together. This is at the real plan, actually. We keep guessing and guessing. And I know that for the share price, it doesn't really matter what's going to happen in seven or eight years from now. But I think it would help. with building a narrative around a company that has a bigger plan than just build data centers fast. If you catch my drift, I don't really know how to convey this thing, but it just seems to me that the iron story is potentially much bigger than anyone can consider right now. At what point in time is it the right moment to start talking about this? Will that moment ever come or is it always something that is going to be something we can talk about, but we will never know for sure until it is that time? You catch my drift.

Mike Alfred: Yeah, I mean, look, wherever you go, that's where you are. And I say that jokingly. It's one of my favorite things from the past. But the reality is like the forcing function that requires people to be on a short leash is usually a lack of performance. right? So if the company's stock price was down 50% over the last two years or something, or three years, then you would start to have activist investors and folks in there and the board would turn over and you would have a lot more mandates from the investors, like, here's what you need to do, otherwise we're gonna do X, Y, or Z thing. The reality, you know, iron's up almost 500% over the last 12 months. to this moment. And some of that is, again, because we came off a macro low in April, some of it is just good execution, some of it is AI. But the reality is in an environment where the space is actually literally just getting bigger and the opportunity is getting bigger, the question of how big can it be is still an open question even internally. So why would you try to attempt to have that conversation publicly when you're still having that conversation internally? You know, I've been on the board since October of 2021 and even though I've seen everything, and I've been more bullish the whole way than pretty much anyone else in the public market globally over that time period and more consistently, I have still underestimated the potential. And so that tells you the kind of conundrum you might have where even if you knew everything, you're still going to-- if you look back every year or two, you're still going to underestimate then it doesn't really matter what you do, because if what happens in reality actually happens, then the market cap and the share price are gonna go there anyway. I think we're back to this fundamental tension that I've heard from you, and probably other investors share, which is like, we'd like to know more about what the company's gonna do. And I'm like, well, if the company just does what it's gonna do and it generates substantial value, will it matter to you? And the answer is always going to be, well, yeah, of course, all things being equal, I'd like the company to generate substantial value and also tell me everything it's going to do in advance the whole way. And I said, well, what if the process of telling you everything that could be known inside the whole way actually made it harder to achieve the maximum amount of shareholder return over those five years? And some people would take that trade. They'd say, look, I would take 50% less return over those five years to have 100% more certainty that my return would be lower, versus having a higher return but having more uncertainty over that period. And that's basically the tension that always exists in public markets. To the extent to which something is more uncertain and it ends up being correct, then it tends to have a higher return than something that is basically certain that ends up being correct. It's just probabilistically and mathematically how it plays out. And so I still think part of the reason why it's going to be unclear is because it's not necessarily going to be crystal clear, even internally, how this is going to play out. We've all had to learn a lot over the last five years, internally, externally, the market. I think there's a ton of large institutional investors who don't even know the basics that are known by this room. If they just understood the basic construct here that you've laid out, Franz, in your spaces, of what you think the company's doing, the stock could trade at two or three times the value, possibly, right? And so I hear what you're saying. You can keep bringing it up and you can keep saying, and again, there's only so much I can do to even advocate. I do advocate on behalf of all the investors that I talk to and that I hear from, and I try to aggregate all of the feedback and opinions and try to distill that into something that I can then share, and of course, in the board meetings leading up to to quarterly earnings reports where we're talking about what's going to go on the release and how it's going to be framed and what the talking points are going to be. That's where I've added the most input relative to what we talk about in the spaces because that's where I can say, hey, I think the market would like X, Y, and Z thing and to the extent at which that makes sense to the team, then we make those adjustments. We made a lot of adjustments actually over time to try to make sure we explain things that folks in these spaces and other places have said would be helpful. So there will continue to be that open dialogue and that open listening. But I wouldn't expect a major shift in forecasting a longer term period until the company feels internally that we fully understand how big it is. The issue for me at least is that it keeps getting bigger. Every quarter or two, when you look at it, you're like, we underestimated again, and I hate that. I wish that we ourselves can understand fully how big it is, because we would have gone faster. Remember when people were criticizing the billion dollar ATM, which by the way, I was partially the architect of, because in some of those board calls, it was like, oh, the expectation was 200 to 300 million. And by the time we left the call, I had convinced everyone we should do a billion. Well, that billion was way too small, right? So the next one was 6 billion. And then everyone was, oh my God, the $6 billion ATM. It's too big. There's no way the stock price can go up when it's 6 billion. I'm like, we're going to need like. $100 billion of capital. You might need $500 billion of capital if we're going to get to like a $2 trillion or $3 trillion valuation over a decade or two. So like a billion, basically the way I think about it now is if somebody is too worried that they don't have enough visibility and they don't understand why we're going bigger, then they don't understand the business fully yet. They don't understand the demand environment. They don't understand how big AI possibly is. And so they want us to stop and stand in place and stand still long enough so that they feel comfortable with the pace and the size and all that. And that's just not really going to be the attitude of a winning company in a space that's this big with an opportunity set this large where there could be multiple trillion dollar companies that come out of the NeoCloud space right over the next, call it, 10 to 15 years. So in an environment like that, do we want to be standing still long enough so we can give a crystal clear picture to the whole market of what to expect so that everybody can do the same thing and possibly outcompete us? Or do we want to just constantly be trying to evaluate and reevaluate the opportunities and make sure we're going big enough? And then probably consistently have people say, well, we think you're going too big, right? Which is fine. I don't want to be accused of going too small. Like it would be such a shame if I saw clearly that there was a trillion dollar market opportunity. We didn't get it because we didn't have the courage. right? We let the market convince us that we should have done a $200 million ATM that we're still working on. You can imagine if when the people were complaining a year or two ago, we did a $200 million ATM and we'd only use 100 million of it and our market cap was 4 billion or something. That's a plausible outcome with the wrong management team, right? Or somebody who's too easily manipulated by institutional investors. I can tell you from experience the last three or four years on a regular cycle every year, institutional investors tell us, Every year, every time we're in a drawdown, every time the market's in a drawdown, that they need the company to do something, otherwise they just can't buy the stock. And then you know what happens every single year, every single year for the last four years, is that the moment the stock goes up for completely unrelated reasons, they change their mind and buy the stock anyway because the stock went up. They're all full of **** all of them, including most retail investors. Like as much as I love all retail investors, like almost everybody sells the stock when it goes down too much, and they chase it when it goes up too much. And people should be doing the opposite. And institutional investors are not immune to that. They'll tell you, oh, this billion dollar ATM is just going to block us from buying it. And they say that when the stock's 12, and then when the stock's 40, they pile in. So look, it's the same people, by the way. We've literally had this conversation at the board meeting. Dan will say, hey, XYZ investor said XYZ thing, and I said, I don't believe them. And they seem pretty convinced. And then we look back a year later and literally, they were totally bluffing. They just were saying it because they were trying to influence us to do something that would make the share price do something very important to them in the very short term. Because maybe they have quarterly mark to market coming up and so they don't want to lose a customer. They don't want to show volatility that's too high. They don't want to show, and then either disclosures, they owned a stock that went down too much or something. And again, all that stuff is noise. When you're building the company for 10 years, you basically have to ignore that stuff. So again, there's two sides to this. One is always listening, always care, particularly the smartest, most clued in analysts, people like you, Franz and Butcher, of course, are always, I'm always listening, the company's always listening, et cetera. The other side of it is a lot of people say similar things, but they'll literally be gone in two weeks, depending on what the stock price does, and that's not our target audience. They're not going to be holding spaces in a month or two. They're just going to be gone based on the share price movement. So why would we change something strategic that's powered by a 10-year view of the future? Why would we make a major change like that to satisfy a large institutional shareholder who's literally not going to be in the stock in two months based on the short-term performance? You can see the conundrum you might get into, and you're going to come back and say, well, that's not really what I'm saying, whatever, when I'm telling you, that's what happens. all the times with these investors. So I don't want to get too much further into it because it's really not really an argument at all. It's more just pointing out that a lot of these folks say these things, but they're not here a year later because they're not really long-term investors. And so we're trying to cater to the people who plan to be around for five or 10 years because those are the people who are going to make the highest returns with us if we're successful.

Frans Bakker: and actually I was trying to talk to those people. the story that you're talking about, briefly is those decade-long plans that internally are very excited with partners that may scale into the future portfolios. I guess my point is just If you could just talk globally, not crystal clear, you don't have to spell out your, give away your competitive edge, but just compare it to, for example, with SpaceX, who's going to say, we're going to launch rockets, we're going to do data centers in space, and in the long run, we're going to build a society on Mars. Like that is such a decade far away plan, but it gives inspiration to the long-term shareholders that are excited to be on board with this plan. You know what I'm saying? It's just like, these are actually the stories that you and Dan and Will and some of these Max 7 or whatever companies they are talking about in 10 years from now, you plan to have this kind of empire of data centers. For example, I'm just making something up. This is the story that We are trying to tell here, like Iron's going to have 10 gigawatts by 2034, but that's just, it doesn't really mean anything if it's coming from me. And of course, Iron doesn't have to tell us it's going to be this much by then, but like a target, like a, like a, like a society on Mars, you know what I mean?

Mike Alfred: I think that would just really help. Let me break, let me, I shouldn't do this, but let me break some news. We've been talking internally, I think we're almost ready to announce, but we plan to colonize Uranus.

Frans Bakker: All right, I guess we have reached the end of this.

Mike Alfred: And I appreciate what you're saying in the colonized Martin. And that's Elon Musk's whole shtick. And he's proven, the reason why the valuation is 2 trillion, he's proven over time, he can say pretty much anything. And a lot of it, people forget that he never does, but the stuff he does do is so magical that like he's imputed this like amazing valuation on something that hasn't happened yet. And with all respect to that, Iran's view is much more, it's maniacally oriented around terrestrial data centers and using, you know, power and marshaling power to generate compute and to do that in a global way to serve the market globally for all of these exponential technologies. And that's not as sexy as colonizing Mars. It's not as sexy as launching rockets that you can bring back into the Earth's atmosphere and land upright and catch. That's cool. That's not making any money, though, right? So from the investment standpoint, any investor who owns SpaceX but then says they're not comfortable with iron, it's like, I don't know what to say, right? Clearly, your investment committee rules are the things you said. you believed in or thought were good for risk management, you were willing to violate because of the emotions of it. And you saw the indexes, the, sorry, the exchanges change a lot of their rules. Like everyone's changing their rules to allow that to go on. And that is the current paradigm. We have a president in the White House who doesn't behave like a president either a lot of the time, right? And whether you like him or not, like it's pretty clear that he's operating under sort of totally different rules. than everybody else, and that's fine. But Iron, I think, has a view that it can just build exactly what you think they're going to build, but probably just a lot bigger than you think they're going to go. And if they do that over time, they think there's going to be more value created than most people can understand. It's really that simple. You already know what they're going to do, in a sense. You just don't know how they're going to do it. And some of that is because to iterate to where they ultimately are going to end up, you can't possibly know all the answers in advance anyway. And during the time you're working through those answers, you don't necessarily want to leak any of that to anyone else who's competing with you for the same resources at the same time, or who may come to the same conclusions that you come to along that curve. So yeah, obviously, there won't be any iron activity on Uranus, at least not this decade. But I would be a proponent of having some activity on Mars, Mercury, possibly Saturn or Jupiter, if we could figure out how to deal with all the heat and the winds and the pesky gases and the lack of oxygen, et cetera. But that might be part of a future strategic plan, maybe by 2050 or 2060.

Frans Bakker: Thanks, Mike. Well, on that note, one final thing, wake up early tomorrow, yes or no?

Mike Alfred: I wake up at 5.30 every single day without an alarm clock local time, so I'm not a good person to ask.

Frans Bakker: Sounds early to me. Well, I guess we've almost reached 3 hours now, and I don't think we have much more to talk about after discussing colonizations outside of Earth. I want to thank the speakers. It was a great space. It started off a bit before, but... Thank you, Mike. Thanks for speaking and providing a great inspiration to a lot of people on this platform. And thanks for all your work as a board member of Iron. And yeah, we hope to see you in one of our next spaces. Butcher, thanks for being my co-host. Lando, thanks for speaking. I want to close the space here. Unless anyone has anything very pressuring, I'm just going to close it here. So thanks for attending and see you the next time.