$iren Sunday Night 8.9.26 Vengeance on the Horizon

Hosted by @₿itcoin ₿utcher 🥩 🐑 🐷 · 2026-08-09 · Tags: IREN

TLDR

Hosts discuss IREN's strong relative week, Texas Governor Abbott's data-center audit politics and Batch Zero delays that may actually benefit Iron by tightening supply, and the imminent Horizon One handover to Microsoft as a major de-risking catalyst. Site progress at Childress and Sweetwater looks ahead of schedule with accelerating builds expected on later Horizons, while a guest raises Elon Musk's ambitious multi-gigawatt AI power plans as a potential competitive threat that still leaves IREN undervalued with margin of safety.

IREN had a rare green week and remains viewed as strongly undervalued versus peers
Abbott's midterm-driven data-center audit and ERCOT Batch Zero pause hit developmental power names harder than IREN's already-energized sites
Horizon One cluster delivery and Microsoft acceptance expected imminently as key execution de-risking event
Horizon Two late-September delivery likely; later Horizons should accelerate with standardized designs and parallel builds
Childress mining-hall demolition clearing space for Horizon Five; Sweetwater dry-cooler plants rising fast
Guest flags Elon/SpaceX 7-10 GW claims and Nvidia memory architecture shifts as possible threats to time-to-power advantage
Hosts see demand still outpacing supply and IREN's modest guidance style as preferable to over-promising peers
Earnings expected around August 27; kitchen-sink quarter with impairment charges from mining teardown anticipated
Overall tone bullish on operational cadence and contracted ARR ramp despite macro and political noise

Speakers

₿itcoin ₿utcher — Host who framed the Texas political news, explained Batch Zero and SB6 implications for IREN versus peers, detailed Horizon revenue potential and operational cadence improvements, engaged the guest on Elon and memory topics, and kept the discussion moving toward catalysts.
Frans Bakker — Co-host who opened on the green week, downplayed Abbott's audit as midterm theater already addressed by SB6/Batch Zero, provided detailed proprietary intel on Horizon cluster delivery timing and site progress at Childress/Sweetwater, and closed with skepticism on Elon's reproducibility versus IREN's under-promise style.
Sumit — First-time guest (Bay Area Meta/WhatsApp infra engineer) who raised Elon's 10 GW claims, SemiAnalysis coverage, and Nvidia Rubin/HBM architecture shifts as potential threats to neo-cloud time-to-power advantages, then explained memory types and why he is researching before allocating to Hynix/Micron.

Notable quotes

“I still think iron is strongly undervalued at these levels given, everything we know.” — Frans Bakker
“I don't believe that it impacts Iron right now, given that Sweetwater One will take a few years to build, as well as building out Childress.” — ₿itcoin ₿utcher
“I think it's a, it's a, a, a futile, effort, and it is most likely not going to do much in either direction.” — Frans Bakker
“delivering Horizon One and getting a official acceptance by- Microsoft is a, massive milestone for Iron, and they will announce that accordingly.” — Frans Bakker
“Horizon one out of four Accounts for, you know, almost five hundred million dollars a year of annual recurring revenue.” — ₿itcoin ₿utcher
“I believe that, this week will be the kick-off of the deliveries Season for Iron.” — Frans Bakker
“if Elon comes and wills, you know, five gigawatt into existence, how, how are you guys thinking about the potential, threats?” — Sumit
“in the forties here, like I just I think your, risk reward is in your favor here” — ₿itcoin ₿utcher
“Iron likes to, you know, give a more modest underpromise, kind of- Kind of style, capacity, and then, over deliver.” — Frans Bakker
“Vengeance is on the horizon and, we'll talk soon, everyone.” — ₿itcoin ₿utcher

Transcript

₿itcoin ₿utcher: We got Franz co-hosting, Kama Sutra back from his video games, Kama, good to see you tonight, buddy. And Franz, why don't you lead us off with some opening words? What's going on this evening?

Frans Bakker: Hey, Boucher. thanks for hosting. yeah, opening words, I think we can look back at a very, good week for Iron, especially, given the, the, you know, the, some of the other stocks, in the same, industry didn't, really, perform or impress that much. So I think relatively we can be satisfied. Yeah, of course, I still think iron is strongly undervalued at these levels given, everything we know. But, you know, there is so much happening in the world that, impacts this particular, small stock, small cap, I mean, it, it's, it's being, thrown around like a small cap, like a penny stock almost at some, sometimes. So I think that It's very, you know, happy. I mean, it's, it's, yeah, I feel satisfied, to see this happening, especially given the fact that we don't have Green Fridays very often. So, yeah, I think this is a good, stepping stone, towards, more catalyst, in the second half of this year. I would really hope that, the macro turbulence can, come to a halt at some point and, you know, leading into midterms, I would hope the market would resume, a bit more of a bullish, broader, You know, rather than just the indexes performing well, I would hope the single names that we are invested in, are going to have some, nice, green candles, every now and then rather than just, being slammed down on the daily. So yeah, that's, that's it for the stock from, from me. I, I mean, that, that's as much as I can say. I'm not a TA, or I'm not a TA guy, I'm not a, a day trader that much, especially not for, for For comments. so I guess the most interesting thing to talk about, will still be the fundamentals of the company, but, I'll, let, let, you, bring the, the topics, Butcher. So, as far as, Iron is concerned, I'm, I'm happy with the progress and I'm happy with, a, a green week, for once. So, yeah, that's it for me for, intro.

₿itcoin ₿utcher: Okay, so here's what I was thinking, Franz, and we'll probably cover iron specifically a little later, but I thought the most notable news event this week is the news out of Texas. And those who aren't familiar, the governor of Texas, Greg Abbott, is up for reelection, and one of the Key issues that his state is facing right now, in the state of Texas, where Iron and a lot of other data center providers call home, is this pushback from local communities, where I believe I read a stat saying at least seventy percent were, it was something that they're, they wanted, not necessarily opposed, but it was just kind of a high topic of mine, given the ramifications, there's some-- in the past, there's been, I believe, tax sale, sales tax reimbursement to incentivize data center providers to call Texas home. Certainly there's Childress, Sweetwater, and so just a high level overview, he had a tweet that came out essentially saying that Echoing kind of this populist sentiment that he didn't want, this being Governor Abbott, didn't want the data center providers to increase the electrical, cost for citizens of Texas, and he didn't want to be a strain on water resources, and I think most people in here have been around long enough to understand why. At a minimum, he's misinformed. He's probably more just playing, with two or three months out from election, more playing a political game to appeal to voters, which is kind of what I think it is. But for the sake of conversation, the most important thing is I don't believe that it impacts Iron right now, given that Sweetwater One will take a few years to build, as well as building out Childress. Which are already energized data center projects that are approved. So the temporary pause, really the impact of pausing those new data centers, projects relates to a process that ERCOT came out with called Batch Zero, which Was supposed to be announced on Friday, August 7th, but that got delayed and that hit, some stocks pretty hard, most notably HUT and Cipher, given a lot of their respective power was, developmental power was in that batch zero process. So that's a long way of saying two things. One, it may or may not slow down the rate of grid approval. I mean, it already has slowed things down, and that might impact some of Iron's Sites, but I was talking with a few people earlier about this today separately in chats, and I think the consensus from most people that study the space is batch zero was a process created to already address the problem that Abbott calls out. So Abbott's Gripe is that there's well over three hundred gigawatts, or maybe even over four hundred gigawatts of power being requested for fulfillment, by ERCOT when the existing grid is somewhere around seventy to eighty, gigawatts, so multiples over what's already in place And the whole point of the batch process that ERCOT came out with was already to address the concerns that Abbott's pointing out. So if anything, he's kind of late to the party, but at a very high level Going forward, it's gonna clear out the, bad actors from the good actors, and when I say a good actor, there were companies such as Cipher or Iron or even Hot or CleanSpark or anyone who has an interest in, producing a data center in Texas, they would go through a process where they would submit a load request to ERCOT and then they would basically put down these large deposits so that the utility could procure long lead items in advance and ultimately provide the power. So the, the point of the deposits is so that the taxpayer, you know, the, your everyday mom and pop who has a house that's connected to ERCOT, that they weren't footing the bill if someone couldn't come up with, you know, if, per- or I should say- A data center request doesn't follow through, then they weren't on the hook via that deposit. So that process appears to be working, it's just, kind of I think politics more than anything, and small, or I should say, X-Cap has a nice post that I put in the nest kind of explaining the, batch zero process. What I also didn't, mention was there's Senate Bill six, which was also from Texas, which is just Legislation that requires data center providers to have the ability to curtail, meaning shut off their power at a moment's notice, which is Wasn't that big of a deal with Bitcoin mining, but it's actually a very big deal with data centers because when you're contracting power to a hyperscaler or a cloud, you're kind of, a lot of times within your, SLA or service level agreement, you're attesting that you'll provide power ninety-nine point nine nine nine percent of the time, so like basically fewer than, a few hours of downtime a whole year, let alone a few minutes, and the workaround that is iron specifically and maybe Franz can speak to it, but other providers will put batteries on site or on-site generation to supplement the power if they can't import it from the grid at a given time when it's, specifically during the summer during peak usage. So that was kind of just some background. Probably too much. I, I bring it up because, no, it doesn't affect iron immediately, but, well, it does in the sense that the available megawatts that they have at Childress and Sweetwater, the price of those just went up because someone like Cipher that I have an interest in, they only have four hundred forty-seven megawatts available to lease out for co-location as of next year at three Sites, they were anticipating having two gigawatts, I should say two point nine gigawatts approved with batch zero that may have not come online necessarily in twenty twenty-seven, but even if they were coming on in twenty twenty-eight, they could have leased that capacity out to a hyper So now the market of Texas has shrunk in the interim, and I think that works to Iron's advantage, and I could see that creating a little more urgency from people that are trying to get compute online as soon as possible, having less leverage and having to deal with Iron. So Fran, that was kind of an introduction, an explanation of what I see. Do you have anything to add to what I, introduced there?

Frans Bakker: Well, I, I've of course also, you know, read up on, the Abbott audit plans and, so far as what I can see is, basically the majority of the comments and, sounds that I'm hearing is that it's for the midterm. So as to why now, you know, on the back of SB6 and Batch Zero, it- It just seems to me that this is a political, political games for the midterms, and it's already showing now that some of the companies that are, involved are, you know, proactively declaring, being conforma- being conformative. Like, I think there was a post where, that he shared himself, where it said Mara and like two other names, I don't know if it was Digital Realty or, or something Like two or three providers already declared that they are, you know, conform the new audit plans or audit ready, or I don't know what it is exactly, but it's just, I guess if the single names are already coming out now, I think it's just a matter of time until other single names will also be mentioned and, I guess that just means that there is already a, a cure to this issue, if you, if you will. So, as to why, I think it is really, yes, a, a, a midterm play, a reelection play, and it doesn't really add much other than, you know, a little bit of, oil on the torches, that, that go with the pitchforks. it, you know, the, the, the comments from the actual people, the residents, et cetera, for those, you are really better off, scrolling through Facebook, and you can see that nobody believes him. Like, this isn't helping him at all, as, as far as it, appears from the comments, you know, people aren't just that stupid, right? They just see this coming out now, and then there is a massive, Commentary about, so, why do you care about our water and you don't, care about, immigration and people start to like mention other stuff like, that there are too many mosques being built in, in Texas or some, something like that, right? People always have something to complain, but it just, it's, it appears to me that, they, they consider it disingenuine and, so people won't be swayed to vote for him just because of this. That's, that's what I'm reading. and it's like, lo- lots of comments, but these are the loud people, there of course are also the people that aren't loud, that just go to the voting booth and think that this guy has it, has the best interest, out for them, but Originally, if you scroll back like a couple of months, like four, four to six months back, Abbott is being mentioned as the face of, of the data center movement, like he's being, mentioned by, local residents as the guy who is, responsible together with Ted Cruz for the whole data center, boom in Texas. So These are the same people that are now being persuaded to vote in favor of him, and that's just not gonna happen, overnight. So I think it's a, it's a, a, a futile, effort, and it is most likely not going to do much in either direction. So, I don't think it's gonna very much impact anything. I think SB six and, Batch zero have already, you know, taken their toll on the, on the interconnection queue, on all the aspirational plans, et cetera, et cetera, and, As, as it goes for, the actual, goal that he wants to achieve, like, no, lower rates, lower rates for ratepayers, I mean Even in the, even in the Bitcoin mining, era when, you know, companies like Iron would go to places in, in the ERCOT market where there was an excess capacity with, overheated transmission lines, and they would come to relieve the grid, To avoid, congestion rents and stuff like that, to lower, you know, the cost of transmission, which ultimately helps to fund new infrastructure and to, lower rates That was like kind of the whole, thesis behind, companies like Iron saying, "Hey, we go where the power is," and then the whole, movement came that, you know, you-- On the one hand, you have the companies like Iron and Cipher, et cetera, who are finding the power and just putting their data center there to mine Bitcoin and be like a battery to the grid, and then there was the other side where people said, "No, you have to be close to metropolitan areas." I think, I mean, it, let's just, let's just remind, you know, look back at that time, which was only like one and a half years ago, and those companies that were like in massive favor of latency, I mean, we had to debate like these guys, like, I see, I see Pro or whatever his name is that said Iron will never be an AI cloud because they aren't in the center of a residential area. Well, guess what? It's twenty twenty-six and residents, hate data centers So incredibly much that you want to be, as far away from these people as you can. So, you know Yeah, I'm, I'm, I'm rambling, but the, the point is that, I think it's, it's not really gonna impact, any, anything like, I think regulatory and, and, these audits, they will, they will pass, you know, this is just a phase. companies like Iron are, you know, very experienced with handling these kind of audits. I mean If you look at Iron, they are extremely over, compliance with basically everything. I mean, the, it's, it's almost, as if they are a European company. I mean, the way they are compliant with, all the little details and, you know If you look at their hiring, the way they structure their, their job applications, it is just extremely, yeah Overly tolerant, overly compliant, et cetera, and it just the, the, the requests and the requirements they have for their hires are also like so extremely, audit and compliance, ish, you know? It is just, my complete belief that if this comes down to an actual audit for, showing your water source, showing your intentions with, you know, being a good grid citizen, a, a, a good, community member, th- these are things that they have dealt with already for years in, in Childress and I, and in the other communities in Canada, obviously, which is not completely irrelevant. So I, I think that they will pass this with, with flying colors. Yes, I do know that there are, people in shoulders that absolutely hate iron because they are, are, filling the, the highway with eighteen-wheelers, every day, all day, every day. But Outside of that, there are also a lot of good things, that happened in a small town like that, so it-- there will always be people that, that hate or, or are against these changes, these companies coming, and, I think at the end of the day, Iron will just have to declare, confirm, the audit, or we have passed the audit, or we will pass the audit, some, some kind of communication around that maybe during earnings, or, or maybe not at all, maybe they just won't mention it. And just assume everything is going according to plan. I think that, that is also a possibility. So I'm not concerned, and I also don't think this really slows down anything. I mean, I know, Agrippa had a couple of, comments in a, in a group chat where he said he thinks this is already impacting Texas now in terms of data center, construction, et cetera. I don't think that's the case, honestly. I think that these- The projects, projects that are currently ongoing will just keep going, so construction isn't gonna stand still just because there is a new data center being built. I think these, these These companies are just gonna keep on building, assuming that they will handle this audit, you know? So I don't think that Greg, Abbott is gonna send like, teams to all these sites to watch for, construction at new data centers to, you know, write a cease and desist or something. I, I just don't see that happening. I think this is a, you know, More of a, you know, just, just, being very vocal about it and showing his, goodwill towards the people, but not actually really doing anything about it. So and I, and I think that, you know, there is a movement in the background that is, very well funded, that's gonna be very strong on, lobbying and, pressuring, and, ultimately there will be a resolution in the coming weeks. this- They drag on until the, the midterms elections, but I don't think it's gonna actually impact, the companies that we invest in here. So, yeah, I'll just leave it at that. And, as for the, for the residential, commentary, the-- I made a post about this yesterday. I don't think that these people will, will be convinced. So you will just have to live in a world where there is a big group of loud people People that will never, ever agree with you, and that's actually how, how you should live your life in general, right? Because if you're, if you're overly, leftist, you will probably never agree with someone who is very right, and that's just never, ever gonna change. And these data center people, I mean, sorry, the residents that are affected by data centers They are saying this is a bipartisan, a bipartisan happening where the left and the right will be united to save their water, yeah, but that's, that's just not, not true, because someone who is right-oriented will know that the, it's, it's in the responsibility of, of the, the companies and the people together to find a solution for water, right? It's not, it's not that we wanna pause this to fi-- to let the government find a solution for the water first before we, now, that, that's just- So I don't think this is really as bipartisan as those, community people want it to be or make it out to be, and I think that ultimately Texas as a, state of, business will prevail. so we just gotta, you know, keep our heads cool and, let the, these companies handle it, and, I think Iron is in an excellent spot to deal with this.

₿itcoin ₿utcher: Yeah, let's switch topics. I think we beat that down pretty well. Irene is positioned well, and a lot of, the last thing I would say to wrap that up is The Public Utility Commission, that is the government branch of Texas that is interacting with ERCOT, which isn't Actually a government entity, it's a private entity that runs the grid. They are, supposed-- or I should say, the Public Utility Commission has a meeting on August 20th that I'll be watching because ERCOT immediately came out with a letter asking for some sort of Good faith exception for batch zero for the reasons we've described, that the studies, the deposits, like a lot of the hard audit that the governor was requesting, has already been performed specifically for those sites. We'll leave it at that, Franz. I think the other, main piece of news this week we can hit it on right now, but we're obviously gonna discuss Horizon one through four if you wanna go into that. Or was there anything else? ancillary in the industry that you wanted to touch on, like specifically this week, I believe as far as decoupling from stocks that we trade with closely, certainly Cipher decoupled from Iron due to their exposure primarily to that batch zero that's been paused for now. I saw CoreWeave was up a decent amount week over week because of a new cloud deal, I forgot where the location was, and then most notably, Nebius has its newest Short in the name of Michael Burry, so that kind of, brought them some attention towards the end of the week. But I would echo your sentiment earlier that, Friday You should be thinking about this as a long-term investment certainly, but after six or seven weeks of getting your ass kicked on Friday, it was kind of refreshing to see. some strength. I wish there was a little more volume, but I do think the way it traded on Friday was, out of characteristic in a good way, and I think- Is, signal of some good things that are anticipated going forward that we're about to talk about. But where do you wanna take it next, Franz?

Frans Bakker: Yeah, so what's on everyone's mind, I hope, if you're in this space, is the Horizon one handover, and, for a good reason, right? The-this isn't just a, you know, a happening that, will add to the revenue for the next earnings in November, it's also a massive de-risking event, and, I can't help but wonder if the market is somehow, s-slightly front-running this or starting to- Price it in. I don't know, which one it is. if we go by commentary of Ken Draper himself, who was at race and said, he believes that, or, or, or let's say he hopes, I don't wanna like quote him directly, but it was basically, along the lines of that he thinks or hopes that delivering Horizon One to Microsoft with an acceptance of Microsoft will be, an, an event where the- Market will start to get it, where the market will start to understand like this is actually going to happen and, it is, you know, I, I will translate that myself to a de-risking event. first and foremost, it's an execution de-risking event These, these data center builds aren't easy. It's also not something that has been done before, to the extent that Iron is currently doing it in this particular design. With a variable rack density, with a very, sort of under allocated, IT load, within a data, data hall that is, you know, Iron's typical metal building, and, you know, so there is a, a, a, a, there is a, a high degree of pioneering going on, in Childress, I can just tell you that. So The market will need to be convinced that a former, slash, current Bitcoin miner will be able to actually build to the standards of Microsoft, who is, a, you know, a hyperscaler that, has very high standards. I mean, it's, it's broadly, understand that, understood that hyperscalers such as Microsoft have immense, term sheets with, requirements that, that are like incredibly, extensive and, not to mention, the actual uptime in the compute. So there is, a lot of execution risk that is, you know, that may or not, may or may not be priced in or, not be-priced in to the stock. And in my personal opinion, I believe that, delivering Horizon One and getting a official acceptance by- Microsoft is a, massive milestone for Iron, and they will announce that accordingly. And, I honestly think that, this is going to happen in the coming week, unless there is a event that prevents this from happening, such as Microsoft not accepting it as it is now. But I would say that that is unlikely, simply for the fact that they have been involved in the whole, Bringing up of these clusters, these buildings, they, they are regularly on site. a-as far as I'm aware, there are very big weeks on site right now, from, at least from Dell and possibly from, from Microsoft as well. But from Dell, I have confirmed sighting of a VP, so I think that, they are really on top of this and, You know, initially we thought this, this may be a building by building, delivery, and we cl- clarified that with IR that it's not the case. It is a single cluster delivery, so a tranche is a fifty megawatt cluster, and it will be delivered as such. But that doesn't mean that, they couldn't validate the building per building already on an operational, status, building by building. So All that happened, after the 19th of July when the building was done, and you re- you may recall this, this post where, Southwest Electric, SWEX, had this dinner in Childress in a restaurant where they had these T-shirts that said, "Uh, Horizon One Completion" or something, and, everyone on X lost their mind. but this was just a, single, subcontractor who was finished with their part of the build. And All that happened, beyond the 19th or the sev- the 17th, I believe, was their dinner and their completion date, and then as far as I'm aware, the 19th, it was like mechanically done and finished. All they had to do was, connect it to the backbone, for, for the network core building, where these three buildings become a cluster, so every building has these, scalable units, which is, eight racks Of GPUs, and then, there are eleven of those in every building. So they need to, first, they need to interconnect the GPUs In a, in a scalable unit, so that's eight, racks being connected together, and then it goes through a, backbone, so it, it, it loops all the, units, all the scalable units in the building together to the network core building, and then you-- then they have to do that two more times with the other two buildings, so that means it becomes one single, cluster. Which is all tied into the network core building where eventually the Horizon Two, GPUs will also, tie in so that it becomes a six-building cluster of, six times eleven, scalable units with eight racks of NVL GB300, sorry, NVL 72 GB300 per rack. So that's how you can, figure out the entire, GPU, quantity, the exact one. I mean, obviously Iron as said is seventy-two thousand or seventy-six thousand. I think it's, somewhere around there, like a tad more, but, so, I don't, I don't, I lost my train of thought. I don't know what, what I, I was, I was just explaining, so, why I think that this is what they're, what they have done in the last couple weeks, and, my intel suggests that this is also officially handed over to Microsoft for acceptance testing, and that means, Microsoft officially has five days to say yes, no, or say Nothing. and, I, I believe we are at the final days of this, consideration. I would personally s- say that if they take the whole five days, I think that a Monday announcement is off, because that, I think that Monday would be the fifth day. so It depends on Iron, and on Microsoft obviously. If, if they accept it in one go, will they do it proactively? Will they just, accept it as it is? Or is there something to change? If there's nothing to change, we still don't know if, if there was an early announcement towards Iron from Microsoft or, or not. And even if there is, it still doesn't mean that Iron is like, forced to announce it immediately. so there are- There are var-variables inside the intel that we have, like the, even if you apply the information we know from the tank queue, you still cannot, you know, prove that, that Iron will announce it on Monday or Tuesday or Wednesday. It's more like a reasonable approach, so I would say, unless something is wrong, I can't imagine they will drag it over the current week, and in, in the com-- sorry, coming, coming week. So in the coming week, we also expect Iron to announce their earnings date, because traditionally, Iron announces their earnings fourteen days in advance, and, the, the best estimate for earnings is August 27, with a number two being August 26, but I, I expect it It's going to be August 27. I don't have any proof of that, I just expect it to be, 27, and that could mean they will announce, their earnings on the 13th. And if they announce-- they already have an announcement planned for the 13th, maybe they will just- Make one big announcement, but yeah, I mean, obviously there are people who have a preference for multiple announcements to keep momentum, but, Iron typically doesn't give a shit about momentum, so it may very well be Wednesday or Thursday just for the sake of, You know, I are not having to put out two announcements in a week, which saves time, which because they're obviously extremely busy. So, that, that's just-- I'm just kidding, but, so, yeah, no, that's what we are looking forward to, and I think, we are, at the final days or, or hours, before that, that announcement is coming. So very excited to see how much of this is priced in and how the market will react, because let's not forget at the same time There are a lot of projects that are delayed, there are lawsuits, there are, you know, new projects facing audits and, and, and all kinds of batch zero troubles, there are all kinds of, reasons why projects, are delayed or may be delayed going forward. I think a, delivery within the announced quarter, I know that originally, Horizon One was supposed to be done in Q2, they didn't, they didn't make that. I, I acknowledge that is a, that is a delay, but it is a, an o-overseeable delay, right? I mean, they announced it, this week, middle of August, we are six weeks beyond the end of the, of the last quarter. I mean, it, it isn't like The, the worst, compared, especially compared to other projects who are going to be, late by multiple months, or, or more than six months, so. I'll just pause there, but I'm, very bullish, for, this week. That said, I don't know how the price is gonna react because first of all, I don't know what's priced in and what's not, and I don't know how bad the macro is gonna develop, and there's supposed-- I think there are also economic data on Wednesday, so, yeah, I'll pause there.

₿itcoin ₿utcher: I think the only thing I would add is by officially handing it over, you address the execution risk, but I also think a lot of the market has been favoring, CoreWeave and Nevea specifically, just given the revenue number that's showing in the financials. So by handing over, then at that point, anyone who's short at a minimum has to rethink that and/or if- If someone was interested in buying, there is a path to the financials, the top line, I should say, improving with the revenue number, given that, Horizon one out of four Accounts for, you know, almost five hundred million dollars a year of annual recurring revenue. one other thing I had posted in the nest prawns that you had included was the graphic that you had created, and I didn't have a chance to ask you, but I think I would feel the same way regardless that you had spoken to delays in the marketplace and if iron is, past the pain and the- growing pains, I should say, of Horizon One and has a new, new and improved Horizon Two template that they're working through right now. I think what I'm looking forward to the most, certainly contracted revenue is always welcome here, but, from an ops standpoint, I think seeing the building cadence specifically of Horizon Two now and two and three, I think you're gonna see, the most in Improvement. Like if I, back when I used to play football, like your, the difference between how you played your week one versus a week two were some of the greatest improvements during a season. I kinda feel that way in parallel about how they're gonna be building Two and three respectively, and I think if they have all the right long lead items secured and it's a matter of just two thousand or three thousand, however many people are on site now working hard and working through it at an improved speed, I think that's when the market you know, once you get that announcement of two being delivered and accepted by Microsoft and then three in the consecutive months or maybe a few months apart, that kind of snowballs and creates a confidence that These guys, this team can start delivering on some of those promises, 'cause right now I, I think just the biggest thing that's on their mind is they said they were gonna deliver the Well over, four billion in annual recurring revenue by the end of the year, and I think we're gonna see some potential upside here from new, you know, most Notably, the rest of Childress getting signed or ideally Sweetwater by the end of the year, but if we don't see some of that soon, at a minimum, you're gonna see operational improvement that gives context to someone that's looking to invest in this, that they went from last year not being able to produce more than Seven million dollars in a quarter to next year, possibly being in excess of six hundred or seven hundred million in a quarter in that, in Q1, I should say, or be closer to, I sh- close to actually, a billion dollars, so year over year, you know, well over a hundred and fifty x from, quarter, which is kind of crazy when you say it out loud. So those were just some of the add-ons. Other than that, I think this upcoming- week, it's hard with the war to kind of take it seriously anymore. I think I read something from Cal State today saying that The, so you don't know whether you can accept it as true or not, so this isn't verified, but the possibility of not having a nuclear agreement and just asking for the strait to be reopened, I saw, post related to that. So I, I think the macro is gonna be what the macro is and All you can do is hope it works in your favor, but talking about it more than just saying that it can impact your pricing day to day, I think, is kind of a fool's errand because the truth is, none of us have any control, unless someone Can get in touch with whoever's leading Iran right now and work something out, but our own president can't seem to do that yet, so we just kinda wait and see. But what else, Franz? I saw you had a Sweetwater photo, I believe, today, or is there anything from Mackenzie that you wanted to report on?

Frans Bakker: no, I would say, just to touch on my post, with the image where I predict a acceleration, from Horizon 3 going forward, because I kinda like, said promise I, I would talk about that But, I still really believe that is the case. So, Horizon two, is in a final delivery stage, but it is, since it will be a single cluster tranche as, as well as, Horizon one was, They will all three need to be ready, all the three buildings, and currently, my intel suggests that they are trying a new, Trying something new in the third data hall for Horizon two and, paired with, the liquid cooling plant, I think that this is going to take a little bit longer, than the Horizon two first two data halls. so, that will lead me to conclude that the, tranche delivery will still likely be happening in September, but because of the third build I think it will be, late September. so, but that is basically just, that only means that the, the subsequent delivery of Horizon 3, on a relative basis will be delivered earlier compared to Horizon 2. I, I mean, not, not in absolute terms, but relative. So, the follow-up delivery of Horizon 3 after the delivery of Horizon 2 will see a- A shorter timeframe versus the delivery of Horizon two compared to the delivery of Horizon one. So, I mean, that's basically obvious from the picture. but that's, you know, that's from, proprietary only France research, and, there are really, there is factual proof for that, which I have shared, so it's not like I'm making this up, or something. so, yeah, the- That, that is, making me very confident that it's gonna happen this year. I don't think there will be any delays. I mean, if you see how far they are already with the Horizon four, buildings plus the whole liquid cooling deployment, it is going really fast, like, you know, blazing fast. And at the same time, the internal outfitting and, the buildings, are very standardized already since Horizon two, so even though on the- On the outside, it may look like, you know, oh, they still need to do, all these piping systems, and, you know, there are no glycol tanks yet. in, inside the buildings, they are just, you know, raging on with these, deployments and, and, and the outfitting. So, yeah, I'm, I'm confident and bullish, and, I, I believe that, this week will be the kick-off of the deliveries Season for Iron. So that's, that's it for, for Horizon. otherwise in Childers, they are still, continuing to take down the mining buildings from Block Five, and the most recent satellite imagery suggests that they have started to take down the first building of Block Four as well. So as I predicted, they are really going to, demolish/slash disassemble because it, the- The truth is, it's disassembly up to the concrete, pad, and they, they only demolish, the, the concrete pad. So, but it's gonna, really cover twelve buildings and, not just the six, I belie- I believe they needs, six for, to make space for three hori-horizon buildings. So, they will be taller and they will be wider apart. Further apart, and, so this is still an ongoing process, but they are really close to concluding the, Block Five, which is six, former mining halls, to conclude those, disassembly slash demolition. So there will be a flat ground, a clean canvas to build Horizon Five. that will be most likely start, in the coming like two, three weeks. So most likely during the earnings. presentation, they are gonna talk about this. I am pretty, sure that they will show some, snapshots from the site. it is obviously a, it will be a kitchen sink quarter. the, you know, demolishing slash disassembling these buildings is, not going to, go unnoticed on the balance list. so we, we know the impairment fee is gonna hit, but I already talked about that, earlier. I don't wanna, you know, go into the earnings, predictions now too, too much, but, yeah. so that's, that's going on. and, the third thing that they are doing is, in block three, on the very far east of the site, they are, figuring out the air cooling for the P3. 300s in the retrofitted building. So, it is a three faceted, site development which is very interesting to follow. we follow it of course through satellites and, other aerial imagery and, yeah, it is, fascinating to see, you know, three d-different, things happening at the same site. Sweetwater, on the other hand, is, just one thing that they're doing, but it's going extremely fast. So there was someone who posted a drive-by video, like, like earlier last week, I believe, it was Mike, one of the Mikes that we, have in the, in the subscriber group, but he, had a, a great drive-by video where it was visible that, Dry cooler, installation is already vertical, and, I am happy to, tell you all that the second dry cooler, plant is also already vertical, so they are, As I have, estimated based on the rendering that Iron shared, with the last earnings call, is that the, the first one is a twenty-five megawatts IT load dry cooler, plant, and the second one is a fifty megawatts, dry cooler plant, and, they've already gone vertical with this, with the fifty watt, fifty megawatt one as well. So I mean, it just shows that they are, going rather fast on a, staggering, like, it's not as if they do like, data hall, liquid cooling plants, dry cooler, plants, and then onto the next one, right? they are building, the entire, dry cooler row out for the whole, two hundred megawatt IT load building, in one- One single row, in one single go. So it's not like they are pausing and then first building the data centers. These are like different teams doing different things. So they ha- it appears to me, at least, they have a dedicated dry cooler team that is, not stopping, that is just keeps on, pumping out these, these dry cooler, buildings. Well, I, it's not really a building, it's more like a frame. and, the satellite imagery also shows That they have already started to put these concrete foundations, well in excess of, what, what is currently vertical. So there is already, one hundred and twenty-five megawatt of IT load worth of dry cooler frames that have a concrete foundation in the ground. I mean, it is, of course, this is just early stage, vertical building, but it just shows that this isn't going to be a you know, a, partial, of, you know, tranche by tranche or, or, or phase by phase deployment, they're gonna do the whole thing in parallel. With the current, scope of, people on site, I think it's impressive that they are going this fast because, from what I can see, I don't, I don't think they have more than a thousand people at Sweetwater right now. most likely it is, somewhere between five hundred and, and, and one thousand. compared to Childers where there is more than four thousand people on site, so that is, there is a big difference there, and it just, gives you an idea that if they can actually, you know, pull a couple of thousand more towards Sweetwater, we are gonna be, very spoiled with progress every, you know, week over week. So, that's, that's all very, bullish. I don't really have any conclusions yet as to when or, or, or, what else, but Sweetwater is just progressing along, very well, in my opinion, ahead of what I would expect it to be right now. So, I'll just leave it at that. And otherwise, I have no real insights on Canada right now because I'm, I don't have any, up-to-date, footage. So that's it, but I can give you, right now, Butcher.

₿itcoin ₿utcher: Let's, change things up and go to Sumit. Sumit, how are you this evening?

Sumit: Hey, Franca, hey, Butcher, I'm good. this is my first time, so thanks a lot for pulling me up to speak. so I just wanted to really cover on the whole Friday weekend news after the SpaceX earnings calls, right? that Elon saying that he'll be up with ten gigawatts next year and, Semi Analysis following up with their piece, you know, saying Saying how it is possible, who might offtake it, and stuff like that, and I was going back and forth with one of the authors on that, and they're saying that they have done some, site inspections with a couple of million square foot sites, and, you know, there's some, you know, seven gigawatts of, turbines available for that. He might get the, permits as well, or he might skip the permissions because of the mobile, generators he has. So are you guys, are you guys thinking about this, if he's able to will at least half of it into existence, does that cause a threat to the whole neo cloud sector because, time to power and time to compute is one of the things that, at least iron on the iron side, time to power was one of the main things, our main, advantages because of the grid connected power, but if Elon comes and wills, you know, five gigawatt into existence, how, how are you guys thinking about the potential, threats?

₿itcoin ₿utcher: I think the first thing that caught my attention and, to me, if you don't mind going on mute just while so it doesn't echo, I would appreciate it. thanks. I think the first thing that caught my attention from Elon's presentation is if he's able to build that quickly and now he has a strategic relationship with Jensen, how does that impact Iron? And I don't have an answer for that right now, but I think it's something that if we have the ability to talk with Dan after earnings I know that there's, you know, certain number of GPUs, I don't remember it off the top of my head that Nvidia is committed to Iron, but I think it's fair to say that if Elon can build faster, does that- Call into question where iron is in line, and I don't have an answer for that right now, but I think it's-- I think it'd be foolish to not acknowledge it. What I also would say though is I think You would agree to me that the rate of, AI usage, the growth, I think you've seen more of a transition to some open-source models more recently from the frontier models, but I think overall usage Regardless of token costs is still, token consumption is still increasing, and I think it's increasing faster than the infrastructure itself. So I, I still think Iron is in a position to do well, but my question for, Dan, if we ever have the chance to talk with him after earnings, is how does someone like Elon building at that scale, does that interrupt their ability to procure cheap- GPUs and deliver on their promises, so I, I think it's a fair point.

Sumit: Yeah, especially, oh, the way I'm thinking it is a series of events happening, especially this weekend, right? So NVIDIA released their new, Rubin architecture where they moved to a different memory architecture because Jensen is pissed, because, HBM has become the limiting factor. If Jensen is able to take out the HBM limiting factor and move to a different memory architecture, and, now next lines up the Elon, Elon factor where he- Can spin up power out of thin air, sounds like, so if he can spin up power like that, so Jansen would be, if their new architecture scales, Jansen would be easily able to, you know, have Elon consume a lot of Rubens, because he, he's, pissed with the whole memory wall as well. So if, if the series of events line up, then I can, I see a world where, like, you know, Jansen can, you know, keep up, Spin up a bunch of DCs, you know, upwards of like ten mega, ten gigawatts. So, I'm just starting to think and also, Stock Analyst Pro has done some, research on like how Elon can, potentially, power this with one of his acquisitions of APR Energy, if I'm not wrong, they have aeroderivative, turbines up, up there. So, yeah, I'm, I'm just starting to think more on that terms because, if the power advantage is kind of threatened, then Like reevaluate the thesis. But I, I agree with your other point that like demand is, like as soon as one of the bottlenecks are, removed, there is demand coming in from different aspect. Now it's mostly coding and bunch of other use cases, I think demand will sprout up from different use cases. So, yeah, just starting to think on, on that lines. Just so I'm aware of like, what, what are the potential threats?

₿itcoin ₿utcher: No, I, I think we have to be honest with, you have to be honest with yourself in your investment thesis and to not acknowledge someone like Elon coming into the space or even Meta selling compute now and, I, I still don't think that moves the needle enough to lose confidence in what Iron is doing, but to not acknowledge it, I think, is also foolish, and it's definitely something where, all you can do is kind of wait and see. I, I think it would be really knee-jerk to just- Assume like day one, like, "Hey, we're, we're screwed," and it was fun while it lasted. Like, I'm not, I'm clearly not there, but I do appreciate you coming up and speaking with us, and I've enjoyed, interacting with you, outside of this, setting here. But thanks for coming up tonight, Franz. Do you have anything to add?

Frans Bakker: no, not really. I just know that, Elon is a binary take across the board. there are people who say never bet against Elon, and there are people who say Elon never, met the deadline that he, imposed on himself. So, because it's such a binary, personality with takes, you know, that are very wide apart, I just like to, stay away from it.

₿itcoin ₿utcher: I think the only thing I would push back on that, Franz, and I've been one of those people who will jokingly talk about like Elon Cybertruck, Presentation when the glass broke or some of the other ambitious goals he set, but there is precedent with Colossus that they put that up faster than anyone, and even Jensen spoke to that. So I, I think that's what makes this a little different, and what I would also say, I think Sumit had said seven to ten gigawatts. I, I haven't had a chance to listen to the call yet, but I do know that next year, And Franz, you can correct me on this, but I believe this year we're delivering, Weibing Iron is delivering four hundred and eighty megawatts by the end of the year, and then next year that gets up to seven hundred and thirty. So If Elon is saying seven to ten, and even if he comes up short and delivers two or three, like that dramatically changes the equation, and I do think one of the Value props that we have been pushing as a way that Iron can differentiate themselves is getting to market faster and being the escape valve for Jensen, and now there is someone that is able to produce or produce faster that does, I think to not acknowledge it, like I said, would be a mistake, but I think there's still so much under Iran's control that they haven't delivered on yet, and/or if they have, we haven't heard yet, and that's kind of where I'm, especially at this valuation. you know, so me, if, if we were, if Iron, I should say, was trading at, sixty or a hundred billion dollar market cap, I might feel a little differently about that question, but right now, like Just the value of the infrastructure alone and look at the cash balance and the value of the hardware on the balance sheet, like in the forties here, like I just I think your, risk reward is in your favor here, whereas if we went on some run and it's at a hundred, hundred twenty dollars a share, then maybe you would Be looking a little closer at some of the external threats. So I'm not sure how you feel about that, but you got your hand up, what were you gonna say?

Sumit: Yeah, yeah. So I, I, I kind of agree with you guys. Currently, the market cap is giving us a lot of like margin of safety, and also with the developments that, you know, Francis shared through his research and everything is also giving a lot of margin of safety. It's just that I am trying to like, track the leading indicators of Elon more of like how he's built Colossus one Of like got those, one year, one year exemptions with those permits with the mobile units and is kind of moving, eventually to the permanent power generation source, right? So it's just like, I think, I think it's the right time to track all the leading indicators and not really worry about like iron as such, it's, it's just like, you know, something to, you know, put, put on your tracker. And also I, I appreciate you guys, you know, you guys do great work and, I've of course echoed the feelings that I've had, pleasant exchanges with you both.

₿itcoin ₿utcher: We appreciate the compliment and, thanks again for coming up. Are there, let me ask you this since you're working on the Or tell the crowd a little more about yourself. I see you're San Jose based, I see the full house, background with San Francisco. Can you help us better understand what you're doing in the industry right now and what other things besides Elon caught your attention this week?

Sumit: yeah, so a bit about me, I, I come from India, I came here for my masters, and I've been in the area for the last like eight years, as a software engineer at Meta slash WhatsApp. So I am more of like a core infra guy, not to the level of like data centers, but more of like the top level infra. so I've, mostly, experience working on WhatsApp side. I'm not sure how many of you use WhatsApp, big in, Brazil and India and, Yeah. And a lot of my fri- friends work in the industry and also in the semiconductor industry, right? And especially this week, I, with the memory architecture changes and also Elon's news, right, I'm trying to like meet with a bunch of folks from Nvidia to see, what are the memory architecture changes they're coming up with, because I am trying to, you know, buy more of SKHynix and the memory thing, because at least my thinking is that HBM has, isn't six Clinical anymore because of the architectural changes, so I'm trying to meet up with a bunch of folks, here in the Bay Area because literally everything is in and around where I stay, so a lot of action, action happening always here. so yeah, I think, I think this weekend, these are the two things that kind of like caught my attention, and I'm trying to meet folks to understand on the Nvidia side, and also I was going back and forth with one of the authors on the semi-analysis piece to give me more data, but he kept me pointing to the models and To buy models and stuff, I was like, no way, I'm gonna pay for your models. so, yeah, that's, that's pretty much about it. Yeah, been in the Bay Area for almost like eight years now, working at WhatsApp.

₿itcoin ₿utcher: Let me ask you this as a follow-up question, Sumit, with-- so I think I read either a piece by Marcos or someone else within that group text that we're in together. Regarding the memory types you'd mentioned, just for people that are newer to it, I believe there's three kinds. There's the HBM, the high bandwidth memory that the actual GPU is using that is specialized and more important for higher intensity computation. Then there, I believe number two is DRAM, which is kind of more for the CPU. And please correct me if I'm wrong, but I just, I think someone with your experience and background is a great asset to the community to explain. Then there was a third type, so if you don't mind giving like a very high level review of those just for context for people, and then I had a follow-up question as it relates to the, just the memory industry and how it relates to GPU pricing going forward, if you don't mind.

Sumit: Yeah, I just want to, I just want to also highlight that I am no expert in this, I'm just learning all the hardware side of it as my mo-- expert is in more on the software side, but just to, just at a very high level, right? so I think you're right with respect to the at least two types of the memory, right? DRAM is more of, you know, like the commodity memory used in, you know, traditional servers, and as the context window length grew, the more amount of input that we give to the model, That and have that context ready to spin up the tokens, right? So that's why it stores everything in the high bandwidth memory, which is stack of DRAMs, for, like faster access. So, as the context window length grew and as the inference models grew, right? it had to access that memory very fast and the bandwidth needed to be very fast, so that's why they kept stacking the DRAMs and, you know, once they kept stacking the DRAMs to make this high bandwidth memory, you know, HBM went into And it is kind of creating a problem, with NVIDIA, right? Because they are short on memory, they aren't able to ship as many chips they would like to be, because that is the limiting factor right now. And what NVIDIA is trying to, do is architect around that and see if they can change, if they can place the memory module somewhere else, closer, I think it is called, system on chip memory module, if I'm not wrong, and also Marcus wrote a article on like how even changing the memory architecture will not really change the HBM Demand because there is new use case that will come up. So NVIDIA is trying to move around different memory architecture to see if they can, like, you know, push more chips without, with less dependency on HBM. so yeah, that's, that's, that's my kind of high-level understanding. So, to just to summarize, DRAM more, most of like commodity and, what Micron, you know, Samsung, Hynix, they all have fought, share for, but with this HBM, Kind of memory with stacking and hybrid bonding, which is, which is architecturally, architecturally required as the context window grows. So the cyclicality factor has been taken away, but if Nvidia comes with a different memory architecture, then it comes up to the question and it goes back to the same question of like, hey, do we really understand, the architecture, the technical nuances to make a fair judgment and industry, and invest in these memory companies? So that's where I'm digging, With a bunch of my connections here in the Bay Area. I'm not sure if that answered your like--

₿itcoin ₿utcher: No, it helps, and I think there was a third kind that we can look up, but just like focusing on the first two, I think you would agree with this statement or if you don't push back on it, traditionally in the past, the memory business has been more cyclical and, at risk to the ebbs and flows of the marketplace and I think the dynamic that's flipped in the last few months or few months or years here, I should say, someone like Apple used to walk all over these memory providers, but now Apple is having to raise prices because those memory providers have, Switched to focusing more on HBM because it's more lucrative with respect to AI, and now that they have such command due to this HBM shortage, they're able to secure these long-term contracts With someone like Nvidia or whoever else is consuming that HBM, and why that's important is there's more, certainty on their end as to their revenue and their ability to make money, and if it's a less cyclical, cyclical business and more predictable, then they're willing to make more CapEx investments that expand production and ultimately help lead to more production of GPUs, which allows- Whether it's Elon or Iron to purchase more GPUs via, whether, I'm not sure how, if I'm not sure how Elon's doing it specifically, but with Iron specifically, Iron is buying their servers from Dell that bought the GPU from Nvidia, but that was kind of trying to- Talk to something that we might not normally talk to with the hardware, but also bring it back full circle and say my high level butcher, Meat cutter analysis of it is there might be some relief in GPU prices in the future because these memory providers are gonna be able to, maybe not next year, but in twenty twenty-eight, twenty twenty-nine, their CapEx expansion, their factories will be able to produce more units, and at that point, the price of memory will come down, which will impact the price of GPUs, which ultimately is less CapEx Back strain and iron. So if you have any comments on that, feel free, but that's kind of what I see. Yeah,

Sumit: yeah, no, your, your line of reasoning is absolutely correct, and also the whole, history with memory players is they have PTSD from these, up and down cycles, right? That's why they-- now that they have recognized this is more of a structural change, they're trying to, lock up all-- with all these like long-term, arrangements or LTAs, right? Just so that they can even out even out there, you know, revenue streams and the margins, but it's still, again, you know, once the supply catches up, you got, you got to, you got to really, revisit like how will the LTAs, stand out to be? And, recently Gavin Baker also mentioned that, you know, it's very hard to break these LTAs, but, you know, once, once all these players sniff that there is some weakness on the memory side because supply comes up or something like that, they will try their hand do some like, exchanges or favors to, you know, really, really break through those LTAs or work around those, but yeah, at least, at least for the next year and a half until the supply comes up, you know, there is, there is no relief as long as Nvidia, with their new architecture, I think in Marcus's article, he mentioned that they cut down the memory, at least, requirement to one fifth, so I'm still trying to study around that. So, I think overall two things, one Architecture really changes where HBM isn't really required. So these are the two leading indicators, at least I'm tracking, and that is what is keeping me away from investing in SK Hynix or, Micron, but, I, I have missed, I missed the easy money right? So now to, for me to come in, with allocation, I have to be really sure that's where I'm talking, trying to talk to my friends who work on the memory side on Nvidia side or go camp, to the, SK Hynix office next to That's where I'm at.

₿itcoin ₿utcher: In your job capacity at WhatsApp, I've used it in the past for, like, how much of that, I saw an article today saying the, use of Basically, what I'm asking you is, from WhatsApp's perspective, like, are you seeing, do you see any use of agents or is your company, via Meta, like, is your team responsible for any agentic use? Because I, what I also think people are failing to take into consideration, every time I hear like a quote from Mark Cuban talking about pickleball courts and the- The less need for power, I think people are underestimating just how exponential the growth of compute will get regardless of these HBM efficiencies or, token use efficiencies from Vera Rubin, because right now, like, primary usage is long, LLMs where there's the opportunity in the future where there'll be more visual and more audio and those will consume more tokens. So just from someone that is practicing in the industry, or it looks like Sumit switched to listener here, but what I was gonna ask him is How WhatsApp is using AI, but looks like he dropped off stage. But appreciate his time, Sumit. Franz, you got anything else that you wanna talk about tonight? Otherwise, we can just kinda wrap things up 'cause, or now he's coming up Back up. Hold on. Okay, Sumit, we got you back. I, I was asking if you, how are you using AI at WhatsApp? So you were

Sumit: asking about the AI thing, right? I think it's more of, what, Zuckerberg has mentioned in the call, right? But my team isn't really working on the agents and stuff. My team is, mostly on core functionalities on the business side, so don't have a lot of like, info on the agents and stuff.

₿itcoin ₿utcher: Well, I appreciate it, anyways, and thanks again for coming up tonight.

Sumit: yeah. Same here, thank you guys.

₿itcoin ₿utcher: Yep, take care, man.

Sumit: Yeah.

₿itcoin ₿utcher: Franz, anything else you wanted to discuss tonight? Otherwise, probably wrap this thing up.

Frans Bakker: no, I, I did, feel, like I wanted to say a little bit about the, Elon Musk plans. I think this is, it is good to consider both sides of the story. i-it's undeniable, that he has delivered, Colossus one and two in a very short period of time, which is, commendable, and impressive. But the- The big question is, is this reproducible into a seven to ten gigawatt story next year? And, there is the team that says, "Uh, never, bet against Elon," and it's Elon, he can do it, yada yada. And then there is the more, in my opinion, more realistic, point of view where that, that will require a lot of, you know, a lot of effort. And maybe the, the capital isn't the issue, but- I sincerely wonder, if he can pull off such a, you know, the, the whole strength of Colossus one and two, as far as I'm aware, is that it was a, brownfield deployment, and, how many of those, I mean, for example, Niveus is doing one in France, you know, we would need to see a little bit more of these, than, than just going off on one, you know? Elon Musk deployment. In my opinion, it is questionable to say that, you know, it just, you know, he did it once, so he will do it ten, ten times, again next year there are certain, fi- real world constraints that he will also run into, and I believe that that's not really factored in. And I think the whole, situation with Elon Musk in particular is that the real world constraints aren't ever factored in, sufficiently, and this goes from, traveling to, society on Mars, a-and, you know, of course, he did manage to, re, recycle these rockets with the, you know, the, the reusable rockets, but I mean, there is a, a lot of other stories that he didn't, achieve within a given timeframe, and I think that, particularly on the, in twenty twenty-seven story, I mean, I think that he will most likely have the resources and, capital and whatever, procurement, parts of the chain to, actually get seven to ten gigawatt. I just think it's questionable that he will actually do that in such a short period of time. and by that same token, I've al-also always questioned Niveus going from, I don't know, how many hundreds of megawatts they have now to, gigawatts, in the coming, what? Twelve to eighteen months. that will m-mean that, instead of leaning on a already existing data center, which I consider low-hanging fruit, it to becoming a full-scale, multi-site gigawatts, developer over self-built and, colocation sites. and that is a very different story. So the question, for Elon as well is, is-- are these seven to ten gigawatts next? Next year going to be literal copy-pasting of the Colossus sites or are they going to be like new developments? so, I know that semi-analysis says a lot of things and, they have, their own agenda as well, let's not forget that. I think Sumit also- mentioned that he was sort of, persu- they tried to persuade him to sign up first before they would give more info. I mean, it is a very commercial, for-profit company that really wants to sell subscriptions. I don't know why, maybe it's really great information, right? But I know people that are subscribed to the absolute highest tier that they have, and the information coming from there hasn't always been accurate. So I would just take that with a grain of salt as well, and it's just a lot better to do your own research and, you know, I mean Let's, I don't wanna, I don't wanna mention any specific, person or, or topic right now, but, there have all-- I have seen satellite imagery coming from semi-analysis, that mentioned the wolf Abernathy, which was, wrong intel. so, you know, the, the, just because you have a satellite image doesn't mean that you can understand what's going on. And, don't, don't wanna completely throw shade on semi-analysis, but they aren't the god or anything, so I would, take them as, just another source. And, it's always good to do your own research, and I think the same Goes for, you know, a lot of things, the same way that, just taking people's words, about Elon Musk isn't, doesn't mean that it's going to keep on happening, into, you know, in, indefinite, so that's what I wanted to say about that. and I think, compared to Iron, of course, Iron has a very modest twenty twenty-seven build-out plan, but I think that is because they, are very realistic about their, ab-- about their ability to build out, within a twelve-month period of time. And, I think the- the Iron style is to later come and say, "Well, actually, we're gonna build the next Sweetwater one building as well in twenty twenty-seven, something like that." That is, it's just the other way around, right? Nibius and Elon Musk, they like to promise a lot and then, fill it in later, maybe move it up a little bit, or maybe, delay it a little bit, and Iron likes to, you know, give a more modest underpromise, kind of- Kind of style, capacity, and then, over deliver. And I know that the, the revenue that they didn't meet in December and, the first half of this year, or the first quarter of this year, they, they didn't meet their own guidance there. I, I acknowledge that, that was, probably a very steep learning curve with, some Forced, decisions, into, an NVIDIA partnership, on the back of that, but I think they're in a very good spot now, and I believe if they are increasing their ARR from three point seven billion to more than four billion dollars I don't think that they are, you know, I think four billion dollars is really the, the, the bare case almost. it, it will most likely ex-exceed that, but I think that we are getting to the time, in the cycle for Iron where they are actually going to overdeliver. and I would rather be on the safe side, than, you know, now factor in all these massive promises From companies such as Nubes and, and, and, and SpaceX, just because they say they will, so, you know, that's just my, my two cents on that, and, I don't have anything else outside of the fact that I've been, browsing through some of these, more of these community comments about data centers and, I literally found a topic in, it's a, it's a Facebook group called, Texans Against Data Centers, where they were in outrage about seven hundred sixty-five, sixty-five kilovolt transmission lines being installed in, in, you know, through the middle of Texas, because those, big transmission poles were only for data centers and, the rich people. I mean, and this is just showing the absolute, uninformedness of these, of these, people. I mean, the, the, they are upgrading the grid to transport more of the, Permian-based power to the rest of Texas to lower the rates. I mean, if there is more natural gas, cheap natural gas in the mix, what do you think is gonna happen to the wholesale pricing? I mean, obviously with a seven hundred and sixty-five kilo-volt network, you are, more economically transporting this power, all over- State. I mean, it's, this is gonna strengthen your infrastructure and it's gonna, you know, add cheaper power into the mix. And, everyone knows how cheap the, natural gas is in West Texas, in the far West, in the Permian Basin Based on, it is just, these people don't, they, they don't do any homework. So, I, I really hope, they're not going to actually, make some changes because they are, with many, you know? I'm sorry, but they just need to let the adults to handle these things and not, I don't wanna have all these parents, come up and scream about, transmission lines that are actually helping them out. I mean They want a reliable power, that's exactly what Abbott's been, advocating for, right? Then how you gonna oppose seven hundred and sixty-five kilo volt? I mean, it's just absolutely stupid, but, yeah, so that's my final, words on that topic.

₿itcoin ₿utcher: Alright, well, we should help Karen's out as Franz's, exiting, comment of the night. We'll try to do that. Appreciate everyone's time this week, and we'll see if, we get that desired announcement, whether it's tomorrow or Tuesday or whenever, but Vengeance is on the horizon and, we'll talk soon, everyone. Thank you.