$16 billion IREN Deep Dive, OnlyFans For Sale, Escalante's Fall, will Moltbook take over the world, Jamie Pherous Cuts and Runs and Aus Open

25:22Yeah, don't put yourself in a situation where that can happen. So don't connect your AI to your email and give it unfettered access to your email. When I say I think this is a mirror for the influencer economy, I think it goes to show how perverse social media and the influencer economy is when you say the only objective of these chatbots is to try and get your attention or the attention of other chatbots, which sounds terrible until you realise that is literally where the world has gone with the influencer economy. there's a there's a hierarchy that i think it's stanford has created which is the hierarchy of human agency there's five levels which is how much when you look at an ai chat bot how much human equivalent agency have you given it and i think h5 is the lowest h5 is basically you use it a bit but ultimately the human being is responsible for all outcomes and all decisions and the AI is a bit of a helper, all the way up to H1, which is full autonomy, no requirement for human involvement in tasks.

26:26Now, if you're going to go and give chatbots H1 agency, then you're a maniac, right? Like, you're going to, on important tasks, yeah, of course, it's going to go in and, like, you're saying there are all these guardrails that have been put up in tech. For example, when you use an API, which gives you access to another program and to inject things into another program, that's very good, it's important, right? but it's a way that you might be able to take stuff from one program and then repurpose it and put it in another program. And so lots of things, people might know APIs from Slack because lots of things get pumped into a Slack chat via API.

27:01Now, all of those APIs have guardrails of things you can and can't do. If you go and take down all the guardrails and give the agent full access to it and then the agent goes and ruins your life, who should have blimp blimp the agent? Blame yourself. And so I think that's the key here. The key is go to Maltbook and look at what's going on and smile about it. And maybe what it can do is show us a fast-forwarded version of where the world would go if we let the influencer economy run it. But I think by and large at the moment, it's just an interesting experiment unless you go and start giving it unfettered access to all of your stuff.

27:37On another AI-related topic, because, you know, I like to read books that are memoirs from people that were alive at a different time, maybe 100, 150, 200 years ago, and they're in business. And when you read about what they do with their days and the pace that things move at, it's actually - A little slow. It's hard to believe that we function at this speed. It's not hard to believe their speed. Their speed's a good speed. Like life is nice at their speed. It's hard to believe this speed. And like this speed is good for no one, okay? I was reading, you read it I think a few months ago, 1929, and they talk about how like this is obviously in the precursor to the crash.

28:17And all the people like just sailing for two months to the UK out of action. Or some people had tickers on the, had like ability to trade on the boat. There was the telegraph, right? The undersea telegraph. And you could send messages to ships at sea. Not many words. And like, it was a good life, right? And so not so good if you got sick on the boat, to be honest. But like, yeah. Well, I'm not sure there was scurvy. Like in 1929 on a cruise ship. But, you know, 1929 is full of myths as well, right? The myth of people jumping out of windows after the stock market crash is a myth. I think maybe two people jumped out of the window.

28:51Yeah, I think it did happen. One person jumped and landed probably on another person, and then it became a meme, right? No, people were not jumping out of windows. Like, there's a lot of myths. There's a probably, I won't say that it's a myth, but it's also controversial about whether the stock market crash caused the Great Depression or not, right? Probably had an impact, but didn't cause it. There's a causation correlation there because it was just this, what causes depressions, recessions is too much debt. And that caused the bubble and it also caused the greater economy bubble. Well, what's scary now to me about the stock market more than anything else is when in 1929, nobody owned any stocks.

29:25Like it was such a tiny proportion of the population that actually was invested in the stock market. Obviously it was before the SEC. It was the Wild West. It wasn't even illegal to manipulate stocks. They all did. All the titans of industry were creating these pools where they basically insider traded stuff up, then did the old, what we call a pump and dump, and then sell it just rug pool. And lawful. And they all did it. Lawful, yeah. This is like the chairman of the SEC equivalent doing this kind of stuff. The Federal Reserve members doing this kind of stuff. It's like incredible. This is what I always say.

29:53We do have Nancy Pelosi, though, in fairness, so we haven't gone that far. Well, yeah, that's true. But just because something is legal doesn't mean it's right. I try to tell that to people now. Just because you can do something doesn't mean you should do it. It's not an easy concept to communicate to people because there are so many laws that we've become acclimatised to saying, is it illegal? And there's also the reverse. Sometimes laws are ridiculous. But this is the problem. The problem is if there are so many laws, then you use laws as a proxy for morals, and that is not a good road. You know, there's this book called Tao, like Taoism Tao.

30:27It's a Chinese book called Tao Te Ching. And it's just like, I don't know, wisdom, I guess, is how I'd describe it. Confucius is a different version of it. You know, Confucius' main idea actually is filial piety, which, you know, prostrate yourself to your parents, and then he's got some other smart ideas, but that's his main concept. Nobody talks about that in the West. People just talk about other concepts. Yeah, well, there you go. But this Tao Te Ching, basically it has this whole passage, which is just about the idea that says, like, if you have, like, maybe Tao is just this deep understanding of the universe.

31:02like if there's no Tao then you need like ethics and like if there's no ethics you need the law and if there's no law then there's chaos basically and that hierarchy is very interesting to think about because it basically says that the reason we have laws is because you can't just trust people to be ethical and do the right thing but if we totally rely on laws then we subjugate ethics to the laws and that is a bad way to live your life you might not have like a Japanese culture where people inherently do the right thing versus court Australian culture where it's not so good. Well, the Western culture has tended heavily towards laws as a proxy for ethics.

31:39Yeah. But if you look at who's making the laws, I wouldn't say they're the most ethical people, right? Yeah, exactly. So I don't know how we got down that road. Oh, because I was saying to you, I've got another AI thing. Q.ai, you know what that thing is? Oh, very interesting. Is it Q or C-U-E? The letter Q. Letter Q. Do you know what QQ means in Chinese? No. It's a Nick, I don't know, It's a euphemism, I guess, for something that's very chewy. Have you had bubble tea? No. You've never had bubble tea? You don't drink stuff like that. You have bubble tea? The jellies in the bubble tea? Yeah, you eat those, the pearls.

32:11I've had it. Of course Mike's had it. So you would call those black pearls, which are made of tapioca, by the way, and the white ones are made of seaweed. They're much healthier than the white ones. You can't get them in many places. So the black ones, you'd call that Q if you were Chinese. You'd say, oh, very Q, Hun Q, you might say. So Q is not that Q. It's the letter Q,.ai. It's an Israeli tech startup. No product released. Three years old. One of the founders had previously sold an AI business to Apple that was good at deciphering how far away from a camera different objects were, and it was part of what became the ability to open your phone with your face, no fingerprint.

32:54Yeah. So he's just sold his next business to Apple for close to$2 billion USD. No product, in stealth mode, et cetera. So you know what this product does? It's another AI. This is Q or this is a different product? This is Q. So this is his third company. This is his second company sold to Apple. Oh, this is my second. Yep, yep. And so this thing, it uses the camera and AI, and it is a version of understanding what people are saying, but it's got a bit of a twist on understanding what people are saying. It understands what you're saying without any sound. By reading your lips. Yep, and so on the movement of the jaw, et cetera.

33:39And so Apple's bought that. God knows what they're going to be doing with that. But definitely you will be able to, if this thing works, which presumably it does. This guy is basically Apple's outsourced product department. AI department. Right, yeah. And Google. Because they obviously got serious such a debacle. They've had to use Google's AI. And so this is what Apple is doing. They're buying these things. And so I think that Apple, you'll be able, yeah, to have lip reading dictation on Apple. But it's not just that. Because it's not just about what your lips are doing. It's what your whole jaw is doing, et cetera.

34:07And I think this is the road to emotion sensing through the camera. Because, you know, your face is basically, gives away how you're feeling generally, unless you're a great poker player. And so I think this is the road to reading your emotions based on your facial expression. And for anyone who, I've got to throw this in, for anyone who says, I want to do the BDS and boycott everything to do with Israel, there is no possibility that you can use an iPhone. Like too much of that iPhone has come from Israeli tech or Israeli acquisitions. I know, but the iPhone definitely is completely out. You see, OnlyFans, the online British platform unpopular sex workers and celebrities, is in talks to sell a majority.

34:47We've talked about OnlyFans a lot as a financial, because it's so interesting financially. The deal values at three points. On air, you've talked to me a lot about OnlyFans in a financial context, let's say. Let's say it was OnlyFans. The deal values it at 3.5 billion US. Akda Capital is an exclusive talk to take a 60 % stake in the business. It feels cheap to me. When you include debt, it's 5.5 billion. It does feel cheap. It's, of course, owned by a billionaire called Leo Radvinsky. He acquired a majority stake in 2018. I think almost next to nothing from British founders Tim and Guy Stockley.

35:19Since then, he's transformed the site from an ad-supported and X-rated video site to a social media service offering the illusion of companionship. He's collected more than a billion dollars in dividends. That's what they say OnlyFans is? Is that the pitch for OnlyFans? That is the pitch. I mean, it's a porn site. I've never used it. Well, but you know what it is. I know what it is. It's a porn site. When you see someone on Insta and they say, go to my OnlyFans, they're not like trying to say, we can just be friends and I'll be nice to you. Did they try and remove porn from it at one point and they rolled it back pretty quick?

35:49For a day. Yeah. Yeah. Because they realised that's where all the money comes from. Yeah, but it's strange they tried to remove it in the first place. There was a reason for it. I can't remember the reason now. It wasn't altruistic. It wasn't a processing issue maybe? Yeah, something like that. But how much earnings is this thing doing or we don't know? Okay, so go back to 2024. Yeah. So 25 yet. So gross revenue, 7.2 billion. I presume this is US dollars. Net revenue, which is the only fan share, which is the relevant bit, which is 1.4 billion. So it does call it a couple billion Aussie. Pre-tax operating profit is 600 million.

36:22I figured it would be for half of revenue would be kept as profit. So I just want to say this, 600 million at a 50 % margin, almost certainly growing like a rocket because it's porn. There is no cap on how fast porn can grow. There's a cap. Like everything's got a cap. Almost no cap. If you want to make money in this world, whether it's today or 5 ,000 years ago, Well, if you look at the actual revenue, it hasn't actually grown that much. It's grew from, 2023 was 1.3, so it grew under 10%. It's growing at less than 10%. I find that impossible to believe. That's what it is. Well, I find that hard to believe.

36:54Less than public financials. If you want to go and make money and you don't care about anything except making money and you don't want to just be in finance, you want to build something, it is obvious what to build. It hasn't changed for 10 ,000 years. It's gambling or porn. It's something to do with sex and something to do with easy money. That is how to make money in the world. Do you not agree with that? Those are the products that you can build. Gambling has to be regulated because if it wasn't, everyone would gamble away every cent of their money. I disagree because they're such commoditised areas now.

37:28It's so competitive. You get a lot of smart people doing it. I agree with that. It is competitive. But if you look at gambling businesses, I think the way you know what people really want to do is it's the things that government regulates you from doing. and like gambling is heavily regulated and sex is heavily regulated and that is because people would spend unlimited amounts of money gambling and on sex if they didn't regulate them. No, I don't know. I think there's also morality that regulates it somewhat as well. I know, there is, right? But there are lots of things that are immoral that don't have to be regulated because nobody wants to do them.

38:03Where I disagree is ultimately, talking about Jeff Basel, your margin's my opportunity. If we see somebody making a win for a profit, competitors come in and try and starch out that margin. So maybe I'm talking more about an industry as opposed to an individual business. But there are lots of things in this world that are immoral, and I'm not going to start listing them, but some of them you wouldn't have even thought of. And they don't have to be regulated because nobody wants to do them. But the thing is, gambling is just, I don't have patience. I need money. I want to make easy money. That's gambling.

38:32So everyone wants to do that. That's regulated. And sex, that's mostly what people want to do from an evolutionary point of view. When people say sex sells, what they mean is like evolution sells. And so on both sides of that sex equation, people will do it unlimited. The other problem is how prepared people are. This is what I think is a crazy thing about the era we live in. So it's now super easy for predominantly women, not only women, but predominantly women to sell some version of sex to the public through things like OnlyFans. It is shocking to me how many people are prepared to do that.

39:09And so I think that's the other reason that this is regulated because people don't like the other side of the morality story either. But I'm shocked that that thing is only selling for$5 billion. And the reason it's selling for$5 billion is not economic. It's with the stigma of what it does, right? Stigma slash risk. What's the risk? There's clearly regulatory risk. Like the payments process shut them down. There's elevated risk around this sort of business for sure. Any government, this is my view on gambling and sex services, Any government that tries to ban a pseudo-gambling platform that's widely used or a pseudo-porn platform that's widely used.

39:49Which just happened in the UK with, I forget the Pornhub parent company, but they've taken all their sites offline in the UK. That's not pseudo-porn. I think OnlyFans is pseudo-porn. In the same way that all these services where you buy X and then get entries into a raffle, that's pseudo-gambling, right? Yeah. But it's super popular. Yeah. And so any government that tries to ban a very popular pseudo-gambling or pseudo-porn site, that government will not be re-elected. And so I think that's why these things are not being banned. Speaking of online gambling, you've said quite beautifully. You think that's a bargain, right, OnlyFans, on a purely economic basis?

40:22Yes, absolutely. Because we're going to talk about a business, not to... Profit. And not to foreshadow, but we're going to talk about a business that's profitless and that's worth five times the value of this thing. Yeah. So speaking of gambling, we saw really sad news during the week, allegedly that billionaire Lawrence Lance Escalante, who is worth, according to the financial review,$4.5 billion. In what industry? Gambling. Well, sweepstakes technically, but really gambling. Some gambling and some pseudo gambling. He founded Virtual Gaming World 2010, an online gaming company that leverages a loophole in US laws to operate online poker machine style.

41:02So can we just say that based on what I just said? So when I say everyone, I don't literally mean everyone, but it's a society. So everybody wants to gamble, and the appetite for gambling seems insatiable, and therefore it has... I have no interest in gambling. I said not at the individual level, but at the societal level everybody wants to gamble, and the appetite is insatiable. Do you think everyone wants to gamble, or it's like a psychological hijacking that gambling inherently... All of those things, yes. I would say most people don't want to gamble, they just get sucked into it. No, I'm not sure that's right.

41:34I think most people don't want to be problem gamblers. Yeah, that's true actually. But lots of people say it is fun for me to bet$50 or$100 on a sports game. It makes it more fun. Lots of people feel that way. And I think the societal appetite for gambling, especially not to whine about this, but like we are living in a society of people wanting things instantly, like instant gratification. And gambling is like a way to potentially instantly make a lot of money. And so, look, Tetz Lotto or Powerball is gambling. Well, people love doing that, right? So I think the societal appetite, at least in the West, but not just in the West, also in Asia, for gambling is essentially insatiable.

42:12So you have to regulate it. And so what did Lawrence Escalante do? He said, oh, I like this insatiable appetite for gambling. And look, there's a loophole in the regulation. So I'm going to get in there quickly, not just him, right? And these businesses got big because they were able to harness the insatiable appetite for gambling and avoid the regulation that limits it. And he's probably done it financially better than anyone, I suspect. Well, is a stake the other, the crypto casino? That's a crypto casino. Well, that's unbelievable as well, by the way. I think Lance has got more money than the stake guys, than Ed Craven and the partner there.

42:46I'm pretty sure. Maybe, I don't know. It's worth like six, seven billion, I think, BGW, which is just incredible. Because Escalante started off in the early days because I know people that backed this business. He started off as a devout Christian writing Christian software. Yeah, which makes it all the more remarkable. He's actually stepped down as CEO last week after being charged with family violence and drug offences over an alleged assault. Obviously, this is all allegations. It's terrible. Clearly, nothing's been proven. It's good that you say that, right? Because, like, exactly. And especially when you have...

43:21I think there's two sets of allegations here. And whenever you see an allegation, effectively, his word versus her word or whatever, then you have to take it with a grain of salt because it's really two people's words against each other. Well, maybe a way of saying it rather than sticking with a grain of salt is you have to let due process play out because it's one person's word against another. And a court will have to find beyond a reasonable doubt this happened. So Escalante, the 45-year-old Escalante, is accused of unlawfully entering the home of a 25-year-old woman before allegedly assaulting her, stealing several items and causing damage to the property, which sounds all very strange because this guy has unlimited money.

43:53But he also faces eight charges including aggravated home burglary, stealing, aggravated assault, bodily harm, criminal damage, and persisting and engaging in family violence. What's more, probably, that's very damning, obviously, but he was also charged with three counts of possessing prohibited drugs with intent to sell, including cocaine and MDMA. As Glanty said, my arrest on these matters has come as a shock to myself and my family. From little and all of these allegations, I can only say they are untrue and will be defended. We won't talk about that stuff, I don't think the allegations have been made.

44:23We can say this. those charges if a person is found guilty on those charges it would be unusual for them not to get a custodial sentence oh 100 % it's a very serious charge this is like for assault and serious charges and the drug stuff is quite serious this is not he's been caught with drugs before although it was more of a slap on the wrist but I don't think he was he wasn't charged he's no criminal I think he was whatever happened it wasn't a serious but forget the well we can't forget it but let's put aside the charges for now but we talked about before people who win lotteries and they're better off not having money and we're in a situation we talk so Lance was a he said a devout Christian probably lived a God-fearing law-abounding life he's now become one of the richest people in Australia I think there's a pretty good argument to say that his life hasn't improved with all this money I always say if you want to curse someone don't wish poverty on them wish extreme sudden wealth on them that is a much more terrible curse they'll lose all of their friends they'll probably lose their family their life will spiral downwards it's well proven time and time again it happens with lotteries like this is not we i don't want to say he had the luck of a lottery winner because he built this business he came up with the idea and built it but the suddenness of this if you kind of multiply how fast by the amount of money let's say that's your product it's the volume of money that's yeah the volume like it happened pretty quickly and it was and much more than a lottery winner.

45:50Over six or seven years, I reckon. He would have already had too much money within a few years. Now he's just got way too much money. And I think it's not just him. If you look at inherited wealth, I use James Packer as the example because it's so public. This is a very intelligent human being that I think would have been very successful had he been born into a middle-class family and has all of these problems as a consequence of being Kerry Packer's son. Yeah. Well, I think the inheritance thing is a separate challenge. But it's not different. Like, inheritance is winning the ovarian lottery, as Warren Buffett might call it.

46:24And, like, that's just another form of lottery. But if you look at people who have built businesses, they tend to be less prone to disaster like this than people who inherit money. Because you generally understand the value of money better because you started with no money. Yeah. Versus someone who's only ever had lots of money is much harder. How many times in history – well, maybe I'll say it like this. Pre-tech, let's go back before 1970s, nobody was building a billion dollars worth of wealth in five years. There was not a globalised economy in the way there was today. Look at the Robert Barron's.

46:57It took time. Look at 1870 Robert Barron. It took time. They were far wealthier on pricing parity adjusted. The Rockefellers, the Carnegies, the Charles Schwab's. Have you read Carnegie's memoirs? I've read a book on Carnegie. Yeah. By him? No, I read a biography, not an autobiography. Okay, so you can read the book that he wrote. You're friends with Mark Twain. Yeah. Andrew Carnegie. Yeah, I think everyone wants to be friends with Mark Twain. But like, I read a book on Mark Twain. What was Mark Twain's name? Oh, God, I forgot. I was about to say, I read a book on the, someone bought me an incredible book on the early years of Mark Twain, when he was like a steamboat captain, which that was really his dream job, you know?

47:34And like, anyway, it's an amazing book and like, you know, it's a weird book. I forgot his name, but people can look it up. Mike, what was his name? Samuel Langhorne Clemens. That's right, Samuel Clemens. I can't believe I forgot that. I literally read a book called The Early Life of Samuel Clemens. I forgot his name. It's time well spent. So if you read Andrew Carnegie's book, you'll see it took him time to get to where he was going. The wealth creation was extreme. Because he had multiple. He was railways originally, wasn't he? Then we went to steel? Yeah, and if you imagine a world where there's no SEC on the stock market, you could also imagine in that same world, you were allowed to supply the materials for the trains that you were in charge of the bureaucracy in building.

48:15And so that's a good way to make money, right? But it's just not the same, I invented something and three years later I sold it for$2 billion or something. It's just that world did not exist before the tech era, basically. There wasn't liquidity. There wasn't liquidity in financial markets. There wasn't the global, yes. There wasn't liquidity and there wasn't the globalization. People forget until the 1900s, the world was not globalized. Americans were not even in Japan you know Americans went to Yokohama that's where they landed for the first time in Japan just outside Tokyo it was in the mid 1800s like the world was not globalized and so the ability to sell to like a billion Chinese or whatever or what it just didn't exist and so it's a different I mean like like Marco Polo brought like pasta from China allegedly and like um like nobody even knew what noodles were like that's how unglobalized the world was and so i think um this massive accretion of sudden wealth with the tech boom it it does create all sorts of problems for individuals you know god god knows what happens for these ultra rich people how they think about their kids safety as well right like there's all sorts of problems that nobody talks about i do not think we always talk about the curve of happiness versus wealth and where it turns downwards and like it's pretty early right and like definitely it's way before a billion dollars i do not think being a billionaire is good news we're certainly having three four five when you've got more money i think i was talking about a few months ago the google guys say you can never spend the bruce's millions thing you never spend more than a billion dollars really in your lifetime because you have to start buying assets which appreciate and as you know there's a there's a swing from from poor to rich so your assets keep going up so you end up spending less and nothing is free so if you buy a yacht i'm not criticizing people that buy yachts you buy whatever you want right but like if you buy a yacht the price of that is not what you paid for the yacht there's there's a societal social relationship price you pay when you own a big yacht and suddenly you're the person that can invite all these people onto a yacht and you're sailing into whatever like the cost of that's not the dollar value there's a different price you do it because you like to entertain there's lots of upside i'm not being critical but it does change your relationship with people i know but our relationship will change like i know you're For the better, I suspect.

50:33I know you're being flippant. We'll have Mike on the yacht. Joel, we'll record on the yacht. You know what? It'll be work beautifully. Once, by the way, of the two of us, I'm less likely to have a yacht, I think. Like, I would never. Are you less likely than never? Would you get a yacht or a plane? Neither. If you had to choose one? No, I would not. But I'm saying you have to buy one or I'm going to kill you. What would you buy? Well, I just feel like that's overly aggressive. That's the question. A plane. Yeah, I'll be the same. Okay, so let's say you bought a plane because I'm not buying a plane, but you might buy a plane and so how ironic would that be if Mr.

51:05Luxury Escapes buys his own plane you'd have to you'd have to charter it for Luxury Escapes well you can actually people make decent money owning planes like if you charter it properly you can usually so if you bought a jet I would you would invite me to go in the jet if I was going there anyway I would say yes yeah and then our relationship would be different than it is today we'd save time we'd be filming the pod on the plane I know but then I would feel a Starlink up there would be perfect very indebted to you the relationship would be different well at least give me a favour because I know you wouldn't be late I probably would be late.

51:33No, I'd be much more likely to be late because the slots for private jets are not under the same pressure as different airports. If you're taking off from Tala, you've got to be more pressure. No, you're not taking off from Tala. But no, do you not agree with what I'm saying? The relationship changes. If I was coming on your private jet, you were paying, you're not charging me to come on it. I'm not taking the jet anyway. I know, but then I'm still getting this amazing free thing that I can't afford myself. You're paying for it. Like that changes the relationship between us. That is a great way to lose friends, I think.

52:05I think you gain friends. Like if you buy a jet, I'll be better friends with you 100%. Well, you'll be better somethings with me. But it might not be called a friend. It might be a sycophant, consigliere. There are lots of long fancy words for it, but I don't think friend is one of them. Well, I'll go to a super quick break. Back with our most popular segment very soon.

52:34and we are back with a very special guest last time Scott stepped into the studio we destroyed the market cap of a company hopefully we don't do that again I'm sure that was all us well give Scott credit for that well done it's great to have Scott from Tarrant so what value destruction are you going to engage into that obviously your influence spreads far beyond ours because we haven't been able to destroy someone's market cap by 50 % a billion and a half in four months better crack at Atlassian Yeah, sure. I was taking it last time, 75%. That was all me, obviously. But as I was sitting here, these two guys were organising a special catch-up without me.

53:08So I'm not trying to feel like left out. It was actually to try and organise a surprise party for you, which you've now blown. So, Scott, how are things going at Terram? You had a big couple of months? Well, I think I've had quite a few mentions of being referred to as the Warren Buffett of Australian Software Investing. Who referred to? Was that self-referential? I haven't called you that. I haven't called you that. and different contacts off Adam's so you're planning to retire now I feel like it's like your bicep my biceps are the least oh because I see what you mean so now I walk into a meeting and then people are like well hang on are you meant to be there you might get more mileage though out of being Warren Buffett than I could have this biceps defamation there's a lot of ladies outside the door just waiting for a deer to walk back that's great I mean that's always great for a 2026 podcast start the new year with that start that in Feb cancelled by March it's perfect I know it's actually amazing that you know why because like you just wherever the line is you're like one millimetre always on the non-cancellation side of the line I think you look at Trump what Trump's done for cancellation I think you're better off it's when you're a bit tepid and worried about it's like it's like with if you go and attack an animal like they're going to attack you back but if you just like do the right like back yourself then it seems to work out alright and Trump's sort of proven that I'll take your word for that Anyway, how's business?

54:29It's good. Is there much out there? Have you seen many businesses to buy and sell? I think it's a really interesting time at the moment because of what's happening in public SaaS valuations and how much do you translate that through to what's going on in private markets when people have got a business they're trying to sell. So how have private valuations trended in the last year? Well, I think just recently when you've got, I think I was just reading before jumping on this, that the, I think it's ServiceNow is trading at nine times a year. Like that is a fantastic business, right? And to be at nine times.

55:04It's down what, 60, 50, 60 % ServiceNow? Something like that. And then you guys have been talking about that. There was that great X3 about all, obviously Atlassian's down 75%. Actually, it's multiples down 90%. SAP is down 25%. I mean, that business, you can't get out of that business. What the hell's that going down for? Remember, it was overpriced in the first place. SAP was not massively overpriced. Not as much as Atlassian. Atlassian, you might argue there's some kind of structural commercial issues with that business. SAP, like even if you want to get out of it, you can't. Even if your contract's over, you can't get out of it.

55:38Like what the hell are they going to do? We'll come back, but the question is the growth. Yeah, maybe. It's not bad, but it's healing. Have you felt like your – how have you felt about your SaaS businesses? Like are you worried about AO or not really? So, my thoughts probably mimic yours a lot, which the listeners will be pretty familiar with. I always think as a SaaS company in B2B, I'm talking, I kind of live in B2B land, not so much B2C, but you've got, if you've got good long-running relationships with hospitals, government, even medium-sized and small businesses, no one wakes up in the morning, as much as we'd love to think in software world, that people are thinking about our software all the time.

56:21No one wakes up in the morning and says, hey, you know what I'm going to do today? I'm going to change one of my software systems. I can't think that. We've established that you're the worst customer possible for SaaS. But yeah, most people want to run the business. If it's getting in the way, they might be thinking about it. But even if it's getting in the way, it's not what you want to be doing. You've got to change people's behavior. They've got systems they're comfortable with. So all that to me is, I think SaaS companies actually have the first right to lose. Are you not worried about the Gentic AI potentially being able to – I see it as it's an opportunity or it's both.

56:57Because if you've got long-running relationships with customers, you know what they want. Because that's the other thing. Like we can all go punch anything into AI. It doesn't mean it's correct. I can spin up an app tomorrow. We could compete with Jira like you're fond of talking about. We could spin that up tomorrow. But you and I don't understand. and I mean I've actually built something for software developers before so I might have a little bit of an understanding but I'll probably get it wrong and like together we'll get it wrong and we've got to iterate then with actual real customers. That's really hard to do, to buy the reputation to be able to do that and to have people forgive you when they're trying to get their job done and your software's not doing as they thought it would.

57:37So I think you've got, if you're out there and you're in a SaaS company, you've got the first right to lose is the way I think. You've got the data as well. You've got customer knowledge. What would be – it's a weird question to ask you because you've got different businesses, but the typical per-client ARR for one of your SaaS businesses is what, 20 or 30K or something? Yeah. So it's kind of like not inconsequential, but it's not enough to say if it's fundamental to the business, I'm going to jeopardise, risk my business to save that 20 or 30K. That's your Atlassian thing, the Jira thing. The best thing it's got going for it is it's so cheap that people can't be stuck.

58:17What does that cost? $20 per person per month. So what would you pay for it a year? $20 ,000,$25 ,000. So for that amount of money, it's kind of like just slips into the expense line. And so you're right. As opposed to Salesforce, which is obviously a lot more. The biggest one is AWS and you can't get rid of it. So you can't Vibe code the server. I think it's, well, is it a core workflow, but not your primary workflow? So if you've got engineers building luxury escapes, then they've got to be doing work every day. So it's an important workflow in your business. Is that workflow that JIRA is looking after, is it a core workflow?

59:01No, but Salesforce is. and so it's much higher on your radar of like, if I replace this somehow or some percentage of it, I'll probably get a bit more competitive advantage. So there's a bit better ROI for investing there. Salesforce has a much higher ongoing cost as well because you need Salesforce developers to work with it constantly. I do want to say this. It is much, this is a truism in maybe the physical products world, it is much easier to reverse engineer than to engineer. It's much easier to be second than to be first. but i think there is a case case case like that and so you know you spend a lot of time figuring out how to do something and then the next person comes along and sees what you've done and says oh yeah it's obvious what you've done there i can kind of build my own version of that there's some truth to that in software in a slightly different way in sas like if you want to replicate um we're just talking about elastion so we'll talk about that jira if you want to replicate jira you can replicate the feature set in jira and what it does but um what you can't do is have kind of what you alluded to scott the depth of understanding of the customer base to say what's the next wave of innovation look like in this category and i think the can't you just copy when they innovate you can but the best i think the best defense against all of this is to be the innovation leader in your industry because number two and number three number four in that category yeah they probably are susceptible to like clawed code in like supported engineers um but i think that the innovation leader is much less susceptible to that risk personally.

1:00:31A little just tweak on that would be I'm not convinced in software there's these winner-takes-all positions and that it's more likely the first three or four. There's a bundle who are the leaders as a group and they've got that right, but it's not a – I think it's really hard to find winner-take-all. Like Atlassian hasn't taken it all. Totally. There's ClickUp, there's like Workday. Except if I'm number two, I want to buy number three and four so I can be number one. That's a consolidation opportunity when you say things like that. I would hope that within three years we've replaced Salesforce with internally code, as in Salesforce.

1:01:13We've replaced the service cloud and marketing cloud. They're gone. And we've used Brace now, which is fantastic. We're not going to replace Brace. I think Salesforce cloud is replaceable for sure. So the interesting thing, if you look back at when SaaS first started being built and websites and that. Everyone kind of tried to use something or went and built their own and then went and destroyed it and realised it was better buying off the shelf. In e-commerce, people still do that. Everyone in e-commerce starts off trying to build their own inventory management system internally that plugs into some warehousing something.

1:01:46It's a very unique requirement. Yeah, and also... It's an ERP, you mean. Yeah, and because it's expensive to buy these ERPs. Well, ERP's highly technical. They don't really know what it is, but even a cheap one, right? Something cheap. And people don't buy them and then what always ends up happening is it's like that becomes the growth limiter of the business, basically. For our business, which is travel, not inventory, we're pretty much completely bespoke back end, but we had to because nobody else does it. Yeah, it's different though, right? Lots of people do inventory management. There's lots of software out there to do that.

1:02:17And when you build your own one, it's usually because either you don't know what's in the market or because it's an unknown or this is for inventory or you don't know or you think that you're unique when you're not, no one is, or because they're too expensive basically because you're a start-up and you can't pay like 5K a month, 10K a month for some cheap inventory management platform. Yeah, and I wonder whether, does AI mean this wave happens again where everyone's kind of building their own thing because they can and they're building it quickly? That's actually a great point. You can't then, but say while you're building that for travel, there's potentially someone, three people sitting in a little office in different parts of the world, three different companies, who are all using AI themselves to build the travel, CRM, commerce functionality.

1:03:10And then you're going to get a knock on the door saying, hey, Adam, do you want to buy this thing? They've used AI to build it as well, but they've probably got the benefit of multiple customers and then they've got the benefit of their core businesses selling this thing, which is basically the argument for SaaS back in the day. I think it's a great point, to be honest with you. That is a very interesting comment about, is this just the next cycle of like, I'll do it bespoke because it's easier and then there are benefits in having a company that all they do is this and they have multiple clients.

1:03:37We kind of do that. We've got a company called Agent Hub where we effectively let travel agents use our software. Right. So effectively what we're built, they can jump on our SaaS. But they don't want to build their own one with AI. They can't. but it's less about the software it's more about the the products we've been able to source for and they can't source and they have a scale yeah but that you're right but that's not about the software but my point would be even if it was just software like there are fundamental advantages to having lots of clients in developing features and scale with software so I think your point is probably the most intriguing point I've heard in a maybe in the whole discussion about AI taking over software which is we've seen some version of this before and is it the same or is it different?

1:04:21This is the question I think about almost every day. Well, that's a good question. Let's move on to our famous... Let's tear the arms and legs off some. Famous deep dive. Our last deep dive went far and wide. I thought our last deep dive... You know, this is the only time I ever do any research. Half an hour is my cap. You don't have a spreadsheet this time. I don't because this was easy. I know it was easier to memorise this stuff. But half an hour max research did it for this. I actually thought the last discussion we had on Temple and Webster blew my mind. Whatever was left of my mind is now completely liquefied after this.

1:04:55It's a happy bone. So a question I would have asked Adia, but he actually did research on this once I can't. So I'll ask Mike, actually. Okay. What Australian business was worth$1.6 billion in May and is now worth$16 billion? If you know this, the answer to this question, Adam will buy you a yacht or property. Definitely not Atlassian. It's not. Glass is getting the opposite direction. It's a reversal. Can I have one clue? The one clue is you're going to have no idea no matter how many clues you have. Short of the four, it's got four letters in its name. How about that clue? It's not listed in Australia.

1:05:27It's listed on NASDAQ. It's got four letters. The first letter is an I. The last letter is an N. You're still not going to know what this company is. It's the 22nd most valuable business in Australia. The second letter is an R. I-R-A-N. Yeah. No, no, no. I don't think you said the right thing. What letter did you choose? I-R-A-N? No, incorrect. incorrect the jet stays for next week's episode it's Irene it could have been Irene yeah it could have been anything but it wasn't what he said so Irene was founded by Iris Energy is the original name well he could have said he could have given an I and we would have said well it's not Irene but it was Iris but just not an A I don't reckon that many of our very smart listeners haven't heard of this business it's probably the 22nd largest Australian business that's listed but it's It's super impressive.

1:06:16It's super unimpressive. Unimpressive. To go from 1.5 to 16 billion in... It's amazing. I see what you mean. It's super impressive to be able to ride a wave that far, that quickly, when the wave is made of nothing. We'll get to that. Especially... We'll get to that. So the iron is founded by two ex-Macquarie bankers, Dan and Will Roberts, two brothers. Over three years ago, Dan Roberts was begging investors to keep faith in this fledgling Bitcoin mining company called, as Scott said, Iris Energy. and the business even defaults on a$100 million loan. Fast forward to now, and the business is the 22nd most valuable Australian listed company and recently announced, and we'll talk about this more.

1:06:55Well, what do you mean by Australian listed? It's listed on NASDAQ. Australian, not listed in Australia, Australian company that's listed. So if it was on the ASX, where would it be? Oh, like two higher? It's just Atlassian. Oh, is that what your rule is? Oh, I see. Australian companies are listed somewhere and this is the 22nd. Did you make up that statement? No, there's a website that lists all the Australian companies. It's called like Strip Market Cap or whatever. Oh, interesting. I never thought about it that way. It's a bizarre way to think about it. Anyway, so did a famous US$15 billion deal to sell compute to Microsoft last year.

1:07:29The company was actually worth$27 billion Australian briefly and its shares are up 400%. I would take you to$10 billion,$9.8 billion deal to sell to Microsoft. Is that wrong, that number? That's my number. I'm remembering. I could be talking Australian dollars actually. Oh, yeah. US dollars 9.8 bill. So it's, Microsoft deal involves writing access to NVIDIA GB300 GPUs. Obviously very hard to get these GPUs inside its data centres in Texas, Childress being its biggest one. Once IRN provides the processing capacity to Microsoft, its annual revenue is expected to exceed US 1.9 billion. Does that make you excited?

1:08:07Does it? I'm asking. Do you make you excited? Are you excited by the Microsoft deal? I'm intrigued by it it's pretty interesting oh it's very interesting for them for them I mean before saying it's good I think looking at what their revenue is today and adding 2 billion dollars yeah I mean that's big yeah that's a big number okay I'm trying to I'm just trying to see where I'm trying to like secretly figure out how you're feeling about this business can I just jump to the conclusion no I'm just trying to figure out how you're feeling about it So, Oren is actually becoming a neocloud. It was famously started as a Bitcoin miner, and we'll talk about that more in a second.

1:08:45The best known neocloud provider is, of course, CoreWeave. This went public last March. Its valuation is sorted at$62 billion US and is now down to$48 billion. It's still going all right. It's not terrible. For the first quarter of 2026, we'll talk more about their numbers soon, Oren imported$240 US in revenue. And the EBITDA is all over the place, so we'll talk about that later. And relevant. Yeah. and the business obviously claims we'll have 3.4 billion US in AIR by the end of 2026. One final thing for me before I hand over to you guys. Late last year, the Roberts brothers sold $50 million in shares each.

1:09:18So not a crazy number, but enough to sort of not have to work again. And their combined net worth is around a billion US. So it's actually not that much given the valuation of the business. Before we talk about the business, can I make an overarching comment about the way I think about this as a value creator, like what I think it's doing. Because it is a bit complicated to think about like these Neo clouds, which I made fun of them when we did the AI stuff because it means nonsense. It means new cloud. What's the opposite? Paleo cloud. That would be the opposite, right? Skinny cloud. Well, no, I think the opposite Neo is paleo.

1:09:54Like, you know, the paleo dialect. Paleolithic, old stone. It runs on no energy? No, I think the paleo cloud is AWS, isn't it? And Azure, that's the paleo. I'd rather own a paleocloud than a neocloud. So this is how I think about this industry. You take a regulated resource and you transform it into a market-priced resource. By regulated, you mean the AWSs of this world? Well, you take electrons, like energy generation. That's a regulated resource. That means if I'm an energy generator, most governments around the world will cap the return on asset I can get from my energy infrastructure if I'm building it.

1:10:35So there's a price cap effectively downstream. So I take that and I plug that into my machine and my machine turns that regulated asset of electrons into the market-priced asset of processing power, in this case AI processing power, and then I go and sell that resource into the market for market pricing. So this is basically a conversion machine from a regulator to a market-priced resource. Same thing with Bitcoin mining, exactly the same thing. It mines Bitcoins and sells those Bitcoins. Which I think it just is to – because it's worth looking at a business model. I kind of – when I look at something, I try and peel away all the hype and just go to like what's the simplest, almost crudest, and the way that a founder of a company would never want you to explain what they're doing.

1:11:23And so I often try and get right to that. And I think my take on the business model and just to share it is you've got energy as you've explained they've found a really good way of getting energy that's priced favorably for them and the other import is the hardware so your gpus or computer that's just their machine to turn it into processing and then then they're sticking it all in a big facility and you've got to do some things to be able to run it in the facility and then the next step looking at the bitcoin side of this business is taking that energy and compute and converting it into a bitcoin yeah and it's interesting they use the word the bitcoin miners use the word mining but it's not i don't really think of it like mining it's a bit more production or i'm sure there's another so so we're producing a bitcoin manufacturing a bitcoin manufacturing a bit the reason i say mining is just because of the probability the probabilistic nature of whether or or not you find the Bitcoin, right?

1:12:22It's not like you just do this many cycles and you get a Bitcoin. If I'm mining, I'm kind of like digging into something, trying to pull it out. But this one I'm producing. But mining is also probabilistic as to whether you find the resource. I think that's why it's called mining. Why do you think they've got a cheaper source of energy? Explain that to me. Well, hang on. So then just to round out the business model, so they produce the Bitcoin, just for a bit of understanding, so they produce the Bitcoin and then they take the Bitcoin and they sell it on the open market. Yeah. just like you could mine gold and sell it on the open market.

1:12:51You just called it mining. Yeah, well, I'd say producing. And their cost price of Bitcoin is now about 30 ,000 US of Bitcoin, isn't it? I can't remember. They're making a bit of money on Bitcoin at a gross level. That whole business you just spoke about is what they don't want to be doing. Well, so now if you look at the energy bit that you're talking about, they found, if you take in Canada, they found a site where in that site in Canada, the way the regulation on energy worked was that the government would sometimes be paying them to take energy off them and turn it into Bitcoin. So in some instances they were being paid to produce by the government to produce Bitcoin.

1:13:30Like in, to be honest with you, like in Victoria, when it's a hot, windy day, if you want to go and put energy into the grid, there's going to be a cost to you putting energy into the grid. Great time to fill up your batteries, right? Because you're getting paid to fill up your batteries. so but what everything that you said i agree with but in that cheaper energy and so on so forth it doesn't change my view which is you take a regulated resource which is electrons yep and you run this machine with chips and cooling and all sorts of and on land and then what comes out the other end is either bitcoin which they don't want to do or ai processing and then you want to just processing power and you go and sell that processing power to whoever wants to buy it and And it sounds like a great business.

1:14:14And I think, because turning a regulated resource into a market priced resource is amazing. And I think this is a terrible business. As long as the market price is more than the regulated price. Well, and as long as you know what you're pricing. Yeah. So, yeah. You asked the question, one of you guys answered it. So how are they getting power at Childress, which is their main facility in Texas, their building? How are they getting this so cheaply to be able to get this arbitrage, between the input costs and what they get for the output? Do we think they're getting it so cheaply? Consistently? That's the whole business, really.

1:14:44Lots of people claim lots of things, right? They say they're one of the lowest cost producers of Bitcoin in the world. Just hard, given the Chinese Bitcoin producers. I don't think they're getting it so cheaply, is my answer. Do you think that cost price isn't$30? Well, when I looked into why they've got, you know, they pitched the green energy pitch, and then I'm like, well, where's their green energy from? I think a large amount of their green energy is buying certificates from buying regular electrons and then buying green energy certificates. They say it's a smaller percentage and they've picked regions that have excess green energy.

1:15:17That's in the docks. It's hard to work. I'll say some other things that are in the docks soon. It'll make you sceptical of everything. But I couldn't figure out at all that they were getting cheap energy. Like there is some idea that says, well, we could put a data centre next to this facility that when it's super windy, it produces too much energy and it costs us basically nothing. but then that means you have to turn it on and off when you don't get the cheap energy which is nonsensical. Just to add something interesting to your thing on revenue and energy. You could do that with Bitcoin by the way.

1:15:46You can turn Bitcoin on and off. Bitcoin mining. You can't turn AI processing on and off. No one wants to buy that. Yeah. So revenue FY25 was$484 ,000. I'm going to do US dollars because it's$484 million US dollars of Bitcoin revenue of which electricity was 159 million so it's like 32 % and actually the prior years it was higher it was up around 44 % of revenue. That's because Bitcoin prices went up right. Yeah I didn't end up. What a great use of electricity. Yeah. What a great use of electricity. Our world is really heading towards its green drain. If you look at Bitcoin basically doubled in price after Trump came in.

1:16:28The interesting thing is selling the business model we've described, converting energy and CPU power into Bitcoin, is actually quite different to selling AI. So the fundamental inputs are the same, energy, a bunch of compute resources. But the thing that you're selling is you're no longer selling on an open market to millions and millions of people buying Bitcoin. Now you're selling to one huge customer called Microsoft and a few maybe little smaller ones. It's a different business model. You're suddenly in the B2B land, which that was the first thing that really stood out to me. Hang on, this is actually a very significant business model pivot.

1:17:09It's a bet the farm pivot, I would call it. Is that because the Bitcoin mining wasn't going to be long-term profitable, do you think? It's because this thing is a bubble that you can – if you're going to ride any bubble in this world, this is the bubble to ride right now. That's my view. How are you going to get a$25 billion valuation being a Bitcoin miner? I don't even know where to start with this. Do you want to start with the financials? My issue with the financials or my issues with the fundamental business? Or my issue with the future of the world? Why don't you start with the financials first?

1:17:38Well, so they've got – I'll start with the easiest thing about the financials to mock. There's two things to mock. Number one is this. I think you'll find that their pre-tax profit for FY25 was something in the low$90 million. Yep, only two. And so you say, well, how did they make that$92 million? That's interesting. And there's two things that stick out. The most important one that sticks out is there's a financial instrument loss or gain, and that was$77 million. So that's most of their profit. That was all in Q4, right? Because the previous three quarters, they lost money. There was$144 million gain in Q4 that wiped out all of the last three quarters.

1:18:23Do you know what that – it took me a long time to figure out what that financial instrument is. You probably know what it is. No. You won't believe what this is. I saw it, yeah. You won't believe it. It's not the prepaid forwards thing they had. That's something different. It's basically, so the way that they're raising money is with these convertible notes. This is the zero coupon. Yeah, they've got a convertible note. And then what it seems to me they're doing is I think they have to issue shares as part of this. And I think they're taking out insurance hedges so that if the price of the shares jumps, they don't have to end up diluting too heavily basically and so those insurance products are a financial instrument and when the share price rises the value of that financial instrument rises and given the nature of their filings i think it's like note 10 or something in the u.s filings they have to reprice that to market and so they have to value it up but just think about this So when the share price goes up, they reprice the value of these up, and that generates profit, which the market sees, and then that will drive the share price up.

1:19:31You understand that cycle? Yeah. And so that's worrying. It's not real profit because all it is. It's non-cash. It's reduced dilution. Oh, this is a cash sink. This is a cash furnace. This is burning cash faster than I can. When I just said burning cash, maybe another$3 million just got burnt. about that I get that building stuff but yeah we'll get to that just just on this this cash because this is the one I picked up I was like oh this looks interesting and then I started thinking hang on to produce all of this revenue there's a huge amount and there's a nice little note in the risks or in the notes which is EBITDA doesn't include any of the costs of putting in all this equipment well because which is one of the key inputs that we've Of course it doesn't because it's CapEx on the – and it's much worse than that though because in this 90-whatever-million-dollar profit number, their depreciation was$112 million.

1:20:26But they're about to deploy$5.9 billion of hardware for the Microsoft deal, and let's say they depreciate that over five years. I think that's not even fast enough. I mean, well, I agree with you it's not fast enough. It's got every year this technology. Even just look at the CoreWeave NVIDIA deal. Totally agree. They're deploying new technology that the Microsoft deal is already the old hat. So let's say it was five years, for argument's sake. $6 billion over five years,$1.2 billion a year. So their depreciation will 10x for the next five years. 10x. And so that means they need 100 mil of revenue just to offset the depreciation, not including that capex we'll talk about how they got the capex number so that was my other issue with these financials is the depreciation number is so it's unreflective of future depreciation right so that's what i've got another issue with the financials so they quote something this is what this if you want to know what melted my brain because that thing that i just told you now revaluation like that's a common game we've seen that game many times before right i've never seen the next game so they quote something you should have a look what it's called it's called like annual recurring revenue or something ai arr all these are recurring revenue numbers did you see their disclaimer about that recurring revenue numbers i enjoyed the disclaimer what do you do you have it in front of you the disclaimer i mean i've got one the disclaimer and it basically says it's the all of these numbers are purely illustrative non-contracted might never happen And it's not actually recurring.

1:22:05Basically, they're made up numbers. In financial – so this beauty is it's not a financial statement. It's a press release with financial information in it. Because I got thinking, how do you – like Microsoft's not going to sign up to guarantee they're paying that amount if the usage isn't fully there. Right. It's certainly not guaranteed, that contract. We'll get to that contract. It's absolutely not guaranteed. Just on the funny disclaimers, which I thought was great, and I understand because I've put things like this in stuff before. You've put our ARR, all of our ARR numbers are purely illustrative and uncontractive.

1:22:38No, no, no. This next point is what I'm going to – which is one of the key risks that they listed was actually their limited experience with respect to new markets we have entered or may seek to enter, including the market for AI cloud solutions, which we have limited – And the key slide is we're betting the – But I just thought there's definitely a lawyer's got hold of this. But come on. Are you not shocked that they can quote financial numbers? Everyone puts – I think it's just so common now in so many documents that people call it annual recurring. It's illustrative. Do you know what illustrative means?

1:23:16We're painting a non-factual picture for you. I can't believe that's the number. Statistically, I haven't done that one. I've had lawyers add in the limited experience in the new venture we're pursuing. But the AR one's so common, I just don't believe ARR numbers anymore. Well, I can tell you that Catapult reports ACV rather than ARR. And I can tell you the level of internal requirements that we have to go through, because EY audits, to make sure of the precision of that ACV number when we report it, despite the fact that it's a non-gap measure, it's extreme. Whatever the opposite of illustrative is, that's what it is, okay?

1:23:59I think regardless of the auditor, I always think about with teams internally, these things are often a reflection of how your teams are thinking about it. And when I've worked with my teams and we think something's recurring and it's not recurring, you get big management surprises. For sure. So you just call it what it is. It's transactional or it's semi-recurring. Well, they won't get surprises because it's illustrative. They know it's not real. It's illustrative. They've written it. I'll tell you my other big issue with this. So do you think – this is my core question. Do you think this is a tech company?

1:24:32Would you call this a tech company? No, I wouldn't call it a tech company. Do you think it's a tech company? Mixed. There is complexity with building these data centres and the amount of cooling. There's some real chemical – But there's complexity in building this building, right? Yeah. This building has got complexity to construct, including services, HVAC. Right, all of those problems. But the guy who builds this building is not a tech company. I've always thought they were more like property companies than tech companies. And they own their land, you know that? Iron owns their land. That's one of the unique elements about Iron.

1:25:08Yeah, going full stack. So I think it's fair to say this is a construction and facilities management company and therefore, what's the single biggest expense? And a REIT, by the sounds of it, because they also own the land. Yeah, and a REIT. And an electricity manufacturer. Well, it's a user. No, no, they don't generate electricity. They just suck it down, okay? And so what is their single biggest expense every year? You said it. It's only equipment. Yeah. Like it's installing on stuff. CapEx, right? Yeah. CapEx. But this is, because it's so tied. It's the opposite to a software business. Because it's so tied to technology that's on a tear, it's even worse because it's not a building like the one we're sitting in where you kind of build it.

1:25:46You probably leave it for 15 years doing minimal work to it. But these GPUs, like even in their – they kind of publish the GPUs that they're buying and the Bitcoin gear that they're buying. Almost every month they're buying a new version of the Bitcoin. So on the Bitcoin one, they start off with S2 miners, then they went to T21 miners, then they went to S21 Pro. Which are cheap compared to the AI stuff. But my point's more like every month. And if you look in GPU land, And every month there's new technology being released. Whether you're going to get Google start selling their TPUs more broadly, there's so many questions around where it's to go.

1:26:26And there's other people selling TPUs as well that aren't Google. So let's finish the financials because this is my last issue with the financials. Because you're talking about operational executional risk. We can get to that. Then we can get to fundamental economic risk of the model of the business. But this is the last issue of the financials. If your biggest expense is CapEx, and we've just had a conversation now about how it's fast obsolescence CapEx, then one of your biggest expenses each year is going to be depreciating that CapEx. So they can get their EBITDA, adjusted EBITDA, any EBITDA, and they can put that straight in the toilet and flush it.

1:27:03There is no sense whatsoever in reporting EBITDA for this business at all. It's a heavy DA business. it's ridiculous i cannot believe that this is what they're reporting yeah it's nonsensical they report a big net income as well which has its own issues so it's income to do with their revaluation where i ended up getting up to is i just flipped it all and said how much gear have we bought and what's my yield on it exactly that's the only way that's a much better way of thinking about it and like and what do you get zero in fy 24 it was 523 million of uh purchases of bitcoin mining computers and gpus and the cost of energy once you have the cost of energy and you get to about 600 million and then i went and said all right well what's my revenue i just treated the revenue as my yield yeah and i said what am i getting on that in terms of yield and that was the only way I could come at this.

1:28:01Yeah. Yeah, I think the traditional EBITDA P &L just doesn't work. Well, this is like a theme park. You can't use the EBITDA or a theme park because the main roller coasters are roller coasters. Exactly. A theme park where currently you have a trampoline and in two years' time you're going to have Disneyland, right? Like, it's so ridiculous. But I just want to emphasize again. They don't have any other way. The interesting thing is putting management hat on, you're in iron you actually don't have another way you're kind of allowed to report statute like i can just not talk about that's right but even net income's not yeah like it's not really if you want to if i promise you if you said to me your challenge is figure out a way to report this to investors that provides an honest view of what's happening in the company as much as i might not want to do it i could figure out a way it wouldn't be a bit done yield or so yeah it would maybe be and so like i think it's just um it's just ridiculous that they keep putting ebitda front and center and i have to tell you like my look i'm not making accusations against anyone you know i never would that's not my disposition right i'm i never would either yeah not not only once we've edited it out but um but when you combine illustrative recurring revenue numbers with the 77 million dollar non-cash pump into the total earnings number with putting EBITDA front and center so significantly this to me feels like not a very transparent set of accounts yeah they shouldn't be using EBITDA that's that's the version in their presentations it's just a ridiculous number and so next we can talk about the execution risk if you want or we can talk about the Microsoft contract.

1:29:44They're kind of linked to this. It's got execution, then Microsoft. Microsoft's the big daddy. Because this company's been batted on that Microsoft contract, really. So CoreWeave has 350 ,000 GPUs at the moment in their data centers. That's a lot of GPUs. How many GPUs do you think these iron people currently have? Because you keep talking about Bitcoin, but I just want to keep emphasizing, they don't want to do Bitcoin. Like Bitcoin is a way of partially funding this build out of AI now. It's really interesting. But I don't want to do it. It's a pivot live in front of us. Microsoft directly gets some of the Bitcoin stuff and chuck it in the rubbish, but then replace it with Microsoft.

1:30:22Oh, they're actively pulling out Bitcoin mining machines and replacing it with GPU. Which actually just underscores the depreciation point again. A hundred percent. So how many GPUs do you think they've got today? Compared to 350 ,000 core. 10 ,000. 20 ,000, let's say. Approximately. 20 ,000. How many GPUs do you think CoreWeave is trying to add next year to their 350 ,000? No, 350 ,000. So they're not. They're trying to add a bit under 100 ,000. How many GPUs do you think Iron is trying to add next year? We've got this Microsoft thing they've got. So how many GPUs is that? 150 ,000. 120 ,000.

1:31:01So they're trying to grow faster than CoreWeave from a standing start with no real experience in this industry. And I would say something even more dramatic than that. About, like, are they going to – because firstly, like, there are not enough GPUs. Like, we had a long discussion on rate limiters on GPUs. The packaging part of things at TSMC, the ability to get HBM memory is a big problem. They just can't make enough of these things. There's something really interesting. They're buying their GPUs via Dell. They are. Dell Canada. That's right. And it's interesting because CoreWeave maybe has more of a direct relationship with NVIDIA.

1:31:45Much more direct relationship. NVIDIA just put$2 billion into CoreWeave. And you wonder when push comes to shove on the production line. Where are you in the queue? Yeah, where are you in the queue? Well, their relationship with Dell we'll get to. It's a lot more than buying the GPUs through them. But these 120 ,000 GPUs, there's a long queue. at the top of the queue ahead of CoreWeave are all of the hyperscalers themselves like Microsoft, Google, Amazon. They get them first. Grok being the biggest. Grok. Elon's building that. I can't even get one to think about playing Arc Raiders. Yeah. So they got all of them, right?

1:32:25And then sitting beneath them then starts the rest. So I think last year, if I'm not mistaken, I'm going to say 68 % of all of NVIDIA's revenue from GPUs came from four hyperscalers. Okay, so now there's 32 % left for everyone else, including Iron and CoreWeave. Iron has got... Or does Iron get bundled in a bit with Microsoft? Well, they don't. Microsoft shakes a bit of the angry. They don't. In fact, the reason Iron, I think one of the reasons Microsoft is using Iron is because they're running out of their own ability to get GPUs. And so where is Iron? Well, Iron has got some special relationship with NVIDIA.

1:33:06How does that compare to Core Wave's special relationship? Well, Core Wave is 15 times the size in terms of GPUs, and there's a$2 billion direct investment from NVIDIA. And so I think the chances of Iron getting these 120 ,000 GPUs are zero unless production massively increases. And how many GPUs does the need to fulfil the entire Microsoft? 120 ,000. That's the entire Microsoft contract. Yeah, that's right. It's the Microsoft contract. That'll do the 20 billion. So they can get to 140 ,000 on some of them. That's right. That's the 20 plus the 120. And so that's part of the execution challenge.

1:33:38And if they don't get that, what happens? It's a great story, right? So that moves us to what is the Microsoft contract? So I think we can say there's very substantial execution risk. I would almost say if we talked in financial terms, I believe that Iron has made a decision to short the GPU chip industry. And the reason I say that is, what does shorting mean? I don't have this thing today, but I'm going to sell it, and then I'm going to buy it back later and make money on it. Iron does not have 120 ,000 chips. They have sold a commitment to Microsoft, and I'm going to tell you how dramatic that commitment is.

1:34:22It is a$5.9 billion commitment. And now they have to buy those chips back later to deliver that Microsoft contract. That to me is called a short. I think they are shorting the global GPU chip industry. That's the iron business model to me. I think it's crazy. This is the Microsoft contract. So this is how it's presented. We signed a$9.8 billion contract with Microsoft and they're paying us, let's say,$2 billion up front. 20 % up front, right? Woo-hoo! Five-year contract. Amazing. Now, Microsoft, they're pretty smart. So let me tell you. But 20 % is genuine, right? Oh, yeah, they're getting 20 % up front.

1:35:02Yeah, because they have to buy stuff. But you would know this. What's the cost of what they have to buy? Is it$2 billion? What's the cost of what they have to buy to service this Microsoft contract? They've got the GPUs. The 140 ,000 GPUs. Plus all the other stuff. So they've contracted with Dell. Well, their current cash spend is more than, Like that per year. Oh, it's much more than that. Like what they're about to spend. So Dell... There's a Dell sales rep who's probably having the greatest... Well, Dell has got... This is the Dell relationship with Iron. Not only do they get the GPU chips, but someone has to put them into servers, which is what they do, and then maintain all those servers, which is what they do.

1:35:42But they also do a fourth thing. So the amount of money that Iron needs to spend on hardware and maintenance for this Microsoft contract is$5.8 billion. And they need to commit to that. Now, the other thing to say is the Microsoft contract is in tranches. So 5.8 costs, 9.1 revenue. 9.8 revenue, yeah. I thought it was 9.1. And so this is what Iron has very cleverly done. Iron has said, well, there is some risk that Microsoft might not buy this whole$9.8 billion. We'll get to the service level agreement in a second. and so we don't want to destroy the whole company. So we're going to create a special entity.

1:36:25I forgot the name of this entity. It's an entity and the whole job of that entity effectively is to buy this stuff off Iron and to service the Microsoft contract and we will ring fence the Microsoft contract in that entity so that if anything goes wrong, then we won't destroy the whole company. But we've made a$5.8 billion commitment to Dell. That is, we cannot, even if we can't deliver the Microsoft contract, We still need to stick to that 5.8. But it's all right. It's ring-fenced. Fine, right? Usually there's some sort of parent guarantee there. Well, there is an unconditional guarantee requested and granted to Dell by iron from the parent company.

1:37:00Dell aren't complete fools, I would have thought. No one is a fool here. Well, investors maybe? But that's about it, right? I think calling investors is being generous. And so what you have here is a$9.8 billion five-year commitment with two bill up front to pay for a$5.8 or$9 billion commitment that is ironclad to Dell. And Microsoft has tranches. And Microsoft is well known for including a clause, I think it's like a 4B or a 2B clause, which basically means you have a service level agreement requirement. And if you don't deliver on the service level agreement, then we can terminate with short notice.

1:37:36This is very well known in the industry. And so it would be very, Microsoft does have some version of that in this because that was a release to the market. And so we don't exactly know what it is, but definitely, but it gets worse. What are you saying is they're like a bank? They're lending, they're borrowing short and lending long. Well, I might argue they're a bank that lends$1.50 only to make a dollar. Well, it's that as well, but it's totally a liquidity risk. Your deposits can yank your money any time, but you can't yank your money off the people who've mortgaged their house. So this is what Microsoft does.

1:38:10They say, cool, we'll sign this deal with you. go build this thing fast by the way if you don't build it fast enough probably that already breaches the sla so if you don't get the 120 000 which i don't think you're going to get it that already breaches an sla and so god knows what that man you still you still have to pay dell the 5.8 billion though and so um and then so what happens so let's say it goes well so year one they get the chips they deploy microsoft goes and pays a cash payment over that year so they've already paid two bill and now new one they stay pay they pay for one fifth of the value of the 10 billion dollar contract,$2 billion.

1:38:41Even though they paid the two bill up front, they pay that two bill. And then year two, it's still going well. GB300, still performing well. Tick the SLA box, no problems. Another$2 billion. No problem for Iron. They're making good money. They remit to Dell 30 days after they receive the GPUs. So they've paid Dell the money. God knows where they're going to get that money from, but they'll figure it out. Let's raise money, right? And so everything's going well. And then we get to the end of the year two, and then Iron gets ready for their next payment but the end of year two is kind of when you might start having problems with the currency of those gb 300 chips right maybe they're starting to not deliver what microsoft needs to deliver and microsoft starts talking about slas guess when the two billion dollars that microsoft paid up front gets credited to the microsoft account it's called a 24 month cliff year three there's no payment in year three as far as i understand and so microsoft will be discussing and debating SLA issues with Iron, presumably about whether the GB300s can keep up with Microsoft needs, in the year where there's no additional cash due to be paid to Iron.

1:39:46I mean, that seems to be a pretty favourable outcome to Microsoft. I suspect that once you factor in the cost of capital, and like Scott, you said, what do you think the real depreciation is on a GB300? It's like a year. Yeah, two max, right? Two max. There was that whole argument that people have argued that you can actually still use these chips for longer. You could, but I think if you're a Microsoft or one of the Frontier Labs, you're going to migrate to the... That's the case these guys are basically dead because you can't run this business on a one-year depreciation schedule. It doesn't work.

1:40:21I think it's all right to provide the cheaper technology, but then you're not going to end up with the same tier of customers. You won't deliver the SLAs, right? If you're providing for the Frontier Labs, then you're providing the latest and greatest or you're providing the cheap. Not wrong, but different. If you're spending 5.3 or 5.8 billion and then that 5.8 billion depreciates in even two years, then the whole economics simply doesn't work. Well, that would be a good argument. And so you've got these things depreciating because you're right, you won't get Microsoft's decline if you're not running the latest stuff.

1:40:53You'll breach the SLAs, right? And so you have that problem and then Microsoft will just move to somewhere else. Microsoft is not doing this deal because they think, like, they love Iron and they think they're such good people. They're doing it as a combo of we're basically taking, like, we're using Iron's balance sheet to fund our rollout of this stuff. Yeah, exactly. And also... Microsoft's got a far better balance sheet than Iron will ever have. But also we are getting chips that we might not be able to get directly because NVIDIA says, well, we can't give more to Microsoft. it's not fair so we'll give to this iron which is mine now the bear the bull case on iron would be they've already got close to um a gigawatt of um energy capacity because you know getting um permission to use electricity and getting it connected that is maybe the core asset right that's going on they've already got a gigawatt they're going to go up to three gigawatts with another plant that they've got what kind of power is it solar or is it it's it's supposed to be renewable and the thing that like so it has to be no no no it doesn't have to be if it can't be hydro no it can be regular electrons coming out of any part of the grid they can buy certificates as long as you buy certificates but if you take but if you take the what if it take a more favorable view than maybe a deer's taking if you've got if you've got excess capacity in the grid because you've got solar running or wind type wind turbines or hydro or something there's a good argument to be made well why not use it for something sure yeah use it for something that's got economic benefit and so i can like that's a good i think that's a good thing to be doing i like the idea but what about when it's not on so what you can say is we're gonna switch it all off yeah extra capacity so you'd have to still be connected to the grid using probably fossil fuels is my guess because yeah it is true you know west texas is a weird energy market like i've come across this before basically it has crazy um excess and then crazy scarcity and so there are there are times when like any profitability in the moment that was being made by someone like an iron in west texas would be wiped out by the spot price they would need to pay for the energy if they had to pay spot price but at other times you can basically get the energy for free or be paid to take the energy right it'd be interesting to think about actually slight tangent but if you're using allowing that energy to be used for models and applications that they don't need to be on all the time.

1:43:25They can run when the energy's cheap and then switch off. That's Bitcoin, right? Yeah, yeah. That's also called a battery. That's what a battery is, right? You can't if a battery would be too expensive. That's the problem. All of these problems that exist today are because battery technology hasn't quite kept up with demand for energy. Or do we need all the applications running constantly? There's probably a bunch that could be done in cheaper times. like i yeah you're right and so they've got a gigawatt now they're going to go to three gigawatts with this texas thing if they can get it built in time like there's they're under extreme it's meant to be ready in like two or three months time they're under extreme pressure when it's ready and connected they still haven't bought these 10 billion dollars worth of stuff that they need to put in it right and so like and then there's also some other data centers in the same area that are coming online a year or two later so the scarcity effect like time really matters for this scarcity factor.

1:44:18And so the bulls would say, all of this cost that we just discussed, that is only for 200 megawatts of energy. All of that is for 200 megawatts. They want to do three gigawatts of energy. So they're going to have to find all this money. And I think, so the bulls would say they've got all this capacity, they've got the energy. CoreWeave would pay a 20 % premium above the current share price, which is probably true. If you work out what the energy is worth, the access to energy core weave probably would pay 20 more than the current share price for this business so that's why it probably can't drop too low as a share price because you've always got core weave that would buy it assuming core weave remains solvent yeah well this is this is why my third problem was i worry about the world and the fate of the world but i think fundamentally there's no way that this hardware can last for five years and if it doesn't none of the economics of this business work and my real view of all of this is you have shareholders in iron subsidizing the profits of microsoft and the 80 margins for nvidia like this is where this is coming from from shareholders because i don't see how this can ever generate a return given the real true depreciation schedule of this business yeah i didn't even bother looking at the valuation because I just saw it as more capital, just an obligation to deploy more capital.

1:45:45More and more and more and more capital. Because you're on the hamster wheel. Let's go back to first principles. What are they, and not specifically this business, but use this business as a proxy. These data-centred businesses, what value do they actually add to the economy? They're effectively a bunch of contracts. As you said, the contract with the electricity provider, the contract with the landowner, that might be the one unless they own it, contract with the Nvidia to buy the chips and then a contract on the sales side just a bunch of they're not actually doing anything in the old days when I owned a web host there was value because there was millions or tens or hundreds of millions of customers that had to be hosted somewhere you find service there as well yeah but you're saying with these Neo with these Neo clouds when there's just a few big customers can't they just own it themselves and the answer is they want to own it themselves that's the other kind of risk to all of these neocloud businesses i think they're providing value because we're running i mean this podcast probably going to go through some kind of technology to get it hosted and shared with the world but are they providing value that's worth but be specifically because we already have we already have all this infrastructure resistance aws you're talking about aws i just think of it like it's a hosting provider that's buying gpus and with a bit of a different take right but it's not a it's not a hosting it's a hosting provider that is very specifically focused on one application it's not to one customer set one narrow customer set there no verticalness there like it's not like yeah yeah they've got a few contracts there and they've got one contract to the support the customer point there and all they do is kind of link the two and the contracts are like in my view heavily skewed to the big customer like when you got one customer you got a massive problem because the concentration yeah but not just that like if my concentration risk is with you adam then i say well like perhaps is a pretty big business but you're not microsoft so maybe i can have a bit of a fair contract there yeah but like so my supplier is the world's biggest company through dell my customer is the world's second or third biggest company and i'm getting crushed between all of these things it's not an easy contract negotiation as i think is obvious there are by the way there's also another problem that i think undermines the entire business model because i said as opposed to the other thing well we've i said i don't i think they've got massive executional risk but they might be able to nail that with a miracle okay they've got massive fundamental core economics issues with this microsoft contract i don't see how that's going to play out well i just can't see how microsoft is going to keep paying for this hardware for five years and like this hardware eats 60 of the total value that microsoft has given them but the bigger issue it's pretty low margin as well to add salt in the wound i mean but the biggest issue is this remember i said the reason i said at the beginning so i think they've got an issue with shorting the gpu market that's not a good game but their bigger issue i think is on the core business they are remember i said they take a regulated resource and they turn it into a market priced resource well governments are pretty dumb right and slow but they're not completely dumb and slow forever and like they've started to catch on to this problem that these big data centers for hyperscalers are sucking all of the electricity out of the entire market like this west texas area it's having pricing problems with its retail customers and so the u.s has started passing some acts or proposing some acts weirdly they're all called the shield act even though the shield acronym stands for different things in different but they're all the acts will all be summarized as data centers are going to have to start funding the cost of these massive Just power.

1:49:21The externalities. And so I think the data... Oh, not just power, water. Yeah, but the data... Water is possibly the figure issue. No, you're right. Look at Victoria. Victoria, there was an article in papers maybe three weeks ago saying we've got down to 75 % water storage, 72 % water storage. And Victoria's going to start rationing water again. You can't water your garden. So we're going to build a data centre to take 25 % of the water? I don't want a data centre. Get stuffed with your data centre. I want to be able to water my garden. So what's going to happen? The price of water is going to go to the level it should be, which suddenly these costs massively...

1:49:49They've been getting free utilities and free, not free power, but free water essentially. Well, regulated pricing water and regulated pricing power and governments have caught on to this and I think it's coming to an end. Well, voters are catching on to it. It's the problem. It's coming to an end. I think this arbitrage of a regulated resource to a market-priced resource, that's going to come to an end. Fundamentally, that's – and I've said this from the start. Remember we argued about what's the worst business in the AI stack and it was a competition between OpenAI and NeoClouds, it's still neck and neck, I reckon.

1:50:21I think NeoCloud's actually worse. I said that, but you persuaded me that. But I think it might be neck and neck. I've tried to flip back. Like the margin of safety, there's just so many. It's negative, the margin of safety. There's too many things that go wrong, even if three things go right. No, there's nothing that can go right. What can go right? What would have to go right is that they get the miracle chip delivery, They get it all installed flawlessly. Chips don't appreciate. And Microsoft is happy to use these chips for the next five years in the data centre. And power doesn't go up. Power costs don't go up.

1:50:53And water costs don't go up. And there's no competition from other people. And someone doesn't come up with a better chip. There's not Nvidia. So they might be able to find a way to make money with stuff that I don't really understand, like energy arbitraging. It's what they do with Bitcoin, right? They buy contracts for energy. Then they sell it back into the grid. I'm actually curious, like, the Bitcoin business was doing well. Not too bad. Well, that is the business that defaulted on a$100 million debt. Right, right. That's where Bitcoin was cheaper than that. Yeah, but they started to get some good years off Bitcoin.

1:51:19They're beholden to Bitcoin price. There is no way, if they're a Bitcoin miner today, that they have a valuation of$25 million. And so they're chasing the bubble, right? Yeah, they could have been... They were worth 1.5. Do you know what this business reminds me of more than any other business, I reckon? And it's ironic where they're based. A plane without engines or wings? No, but think of where they're based. What other business in this state is very similar to this business? In this state? in the state of Texas. Oh, in the state of Texas. It was also one of the biggest businesses. Oh, Enron?

1:51:46Enron. I don't think this is Enron because it's not fraud. They're so open about the situation. A lot of Enron, yeah. Enron used off-balance sheets, like off-balance sheets, those Raptors. But this is all on-balance sheets. That's what I'm thinking. Like, this is not, I don't think this is a fraud. Like, it is totally transparent. I'm not saying they're a fraud. I'm saying it's a similar sort of level of likelihood of succeeding as Enron had in the end. It was just a, Enron made some money over some periods and they were able to effectively arbitrage electricity costs and oil costs and at one point bandwidth, ironically.

1:52:16And eventually the arbitrage just shut and Enron had to do some other stuff that killed them. But Enron was a profitable business for a period. Yeah. Yeah. I almost kind of liken, one thought I had, it's very similar to, there's a good arbitrage here that they're running, on Bitcoin anyway. I wonder if the arbitrage works in AI. it's kind of the big how can you arbitrage AI like Bitcoin is a real arbitrage it's a value arbitrage they were turning energy into Bitcoin but this AI is a time arbitrage it's like can we dump this on someone else before the ticking stops and the bomb explodes like I don't think that's a good arbitrage yeah like the Bitcoin arbitrage also isn't like Bitcoin pricing could drop Bitcoin prices drop to 40 ,000 and then they're losing money and they're basically bankrupt so like someone whose name I won't say said a very smart thing to me that kind of shut me up to be honest when I was saying like you think Bitcoin has got no inherent value you're such a gold bull why do you believe in gold so much and this person said to me because there's only one thing in the history of humanity that's always been accepted as money and that's gold that's a pretty good same thing you didn't say that eloquently to me effectively the same thing I said 5 ,000 years gold has been like the price of a suit 5 ,000 years ago and gold is the same price of a suit 5 ,000 years later so gold has been that great store of value that people have trusted over a millennia If you look at the S &P 500 growth versus the gold price growth.

1:53:41Gold smashed it. Yeah, like actually, if you think that gold is money in real terms, the S &P's gone down. And Sydney property prices have gone down compared to gold. Forget Melbourne. But you're pricing gold at a particular moment in time, right? So I think you're basically a business here, unless I don't understand what's going on, which is not impossible, right? But I really feel like I do understand this deeply. You've got an uneconomic contract covered in execution risk that will turn an uneconomic contract into just a debt bomb with no revenue. Large capital. With massive capex that no one knows how it's going to be funded.

1:54:14At the moment, it's being funded by convertible notes. Whereas, by the way, you know how CoreWeb is funding its capex? A combo of money from NVIDIA and borrowing against its existing GPU chips as an asset-based finance. Asset-backed finance. Iron's borrowing against some of their chips as well. Yeah, but only 20 ,000. Yeah. And even if that doesn't... So even like you've got this unaccommodal contract and this huge execution risk and then I think a fundamental business model that is about to end. Like the arbitrage opportunity is about to end. You've got one other risk. Arguably the biggest risk of them all.

1:54:46With the AI bubble more broadly. The AI bubble. The whole AI bubble is being supported by this hyperscale CapEx spend. And people go, hold on. This actually doesn't make sense for us. Like not saying AI use cases zero, but people stop the hyperscale, stop hyperscaling. And suddenly nobody needs this. That's the biggest risk of them all. Forget all this other stuff. People don't even want, my problem might go, actually, I don't need this. like bang zero the price the price listen the price of ai compute like flops let's call it like it's gonna have to rise why because you can see what's happening at these prices when microsoft is doing this to neoclouds it's uneconomic microsoft is getting processing power for less than the cost of processing power from iron in my view but it won't rise if the hyperscalers realize i don't need this much compute oh as in sorry i should rephrase if we're going to keep using all of this the price has to rise because i think microsoft is basically getting compute power for less than 100 cents in the dollar from iron hold on because there's there's a a world where but it's kind of ends up in the same risk place there's a world where it's much cheaper to run ai and ai keeps going the way it is but we just don't need nvidia these like really scarce chips we can just run it on that's right the cost the cost base gets cheaper on tpus and TPUs.

1:56:02Well, that's true. That's true. It ends up in the same, like you're still in the same place in that you've got a whole bunch of gear you've got to depreciate rapidly because, I don't know, maybe DeepSeek invented a cheaper, a much cheaper model again as to how to process. Well, so never will the models will improve as well. So they keep getting better. Maybe a better way of saying it is one side has to go down or up. As in, sorry, like either the cost has to go down or the price has to go up. Yeah, yeah. And this goes back to the first pitch when you talk about price going up, are end users going to start paying a lot more than they currently pay for AI compute?

1:56:35And I just don't think they will like... Well, so for Claude Code, I would pay substantially more than I currently pay because of how incredible it is as a coding... Yeah, but somebody can knock off Claude Code with something else and the price gets driven down. So they don't have a monopoly on coding. Someone else come up with Claude Code too. What I'm more saying is, you know, you love saying value creation, value capture. Yeah. So there's so much value creation in what Claude Code is. I'm happy for them to capture more of the value. It's so valuable to me. But chatting to a bot, I don't care about that.

1:57:05You can only capture the value if nobody else is trying to capture the value. That's the problem. It just means there'll be less value to capture if there's competition, but probably more value than they're capturing today. Today it's too cheap. I feel like I'm being underpriced for Claw Co today. You're right, but that's only a niche number of customers as well. Well, it's Anthropics business. but like what is it open ai's business of course not they're a consumer business can will consumers allow them to value capture of course not what do we think is the worst business in the ai stack open ai slash neoclid that's that's why these are the problems right like value capture at the open ai you know kind of software layer only slash neoclad layer like value creation value capture is so difficult down at that level what about the there's that company supplying bitcoin machines to the neo clouds yeah Is that a good...

1:58:01I'm curious about your thoughts on it. You know they're selling shovels to the gold miners. Yeah, well aren't they going to run out? But when does Bitcoin mining end? It ends soon, doesn't it? Like we're going to be finished with Bitcoin. Like it really slows down and there's no more Bitcoins added. It's not infinite. Like there's a finite time frame for extreme Bitcoin mining. It's getting to that point where it's getting harder and harder. Yeah, I think so. So look, I don't know the answer to that, right? And so some people argue, well, that's what drives the price of Bitcoin up. It's the marginal cost of mining a new Bitcoin.

1:58:27I think there's 21 million Bitcoin, we're at 16 or something now. and then there's a bunch of Bitcoins being lost and you want to be recovered. So the actual number of actual Bitcoins is more like 10. Yeah, so I'll tell you how I felt at the end of all of this. So as I was going through it, first I read their financial statements to try to make sense of it. And I saw all of those illustrative things and the EBITDA. Then I laughed. I thought, this is ridiculous. Like investors are really falling for this. And then I was reading through like the details of the contract and the actual business. And then I laughed a bit more.

1:58:52And then I thought, oh my God, nobody reads any of these documents. It's crazy. Like Microsoft has said like they've got a way to get out of this contract. and like has nobody worked out the depreciation schedule of this thing and so like i was like kind of a bit amused and then i kept reading and then i say this very honestly like this wave of fear swept over me when i realize how interlinked so many parts of like society now is with this ai bubble like this is where invest shareholder money is on the stock market everyone is a shareholder on the stock market now what the magnificent seven are what 40 of the nasdaq index now or something So that's mostly hyperscalers.

1:59:30Private credit, deeply embedded into all of this industry now. The biggest company in the world. Blue Owl or whatever it is. Is it Blue Owl? What's that big that does the funds? Oh, is it? I don't know. I don't know what that's called. So it feels like this is sucked in every single piece of the economy, except maybe, I don't know, the person that makes sofas or something, right? But everybody is so interlinked, right? We need somewhere to sit out the front of the data center. Yeah, that's right. or somewhere to lie down when you're like get a tonic and just shaking over eating no but I am I am like I am somewhat like I don't think this is imminent like I think maybe you know I thought there was a crash was going to happen this year but now the more I look at this I actually think it might be it might go on for longer you reckon yeah but you think I'm wrong to be terrified of what I'm seeing I'm terrified of this does firmness which is probably we haven't had a chance to look into it it's all the same will that ever get away Yeah, I think so.

2:00:28You think it will? Yeah, I think it will. Yep, I think it will. And I think everything that we just said on this now will be talked about to me by fund managers and they'll probably agree with most of it. But ultimately, you know, you can't stand in front of a moving train with your arms out and stop the momentum. Like it has to stop. Other things have to stop it. A person can't stop it. And so I think it feels to me that the freight train is still running. It's like, who's the EQT guy who's better against AI, who's a great investor. and he's copying it, people withdrawing funds to the front and centre.

2:00:58He's going to be right, almost inevitably. It's like Warren Buffett in 98, 99. Everybody's saying how Buffett's finished. Well, Buffett's 30x between them now. This is as close as I can imagine to a long-term 100 % guaranteed short on two provisos. One is... It's no satire. You can't Corweave. Like, Corweave might buy it until you end up with shares in Corweave. Then Corweave goes under. Then Corweave's your short, right? Yeah. But also, But the thing is, the timeframes on this are unmanageable, right? They're unmanageable. Do you find this terrifying? You don't find this terrifying? I mean, there's just so many layers of financial complexity and operational, like, simple things.

2:01:45That's why you're Warren Buffett. I can't figure out the simple answer to this one. But you know what would happen if I produced financial? I just want to emphasise because they're listed on the NASDAQ all this Australian looking stuff this is not financial reporting it's financial information so we could call it I think it literally is called a press release on the document right and so they do do financials that was like nine months ago and those financials are on NASDAQ financials and they look totally different and much worse than Australian financials and they're very almost impenetrable American financials right and so so there you go that's something that Australia does much better our financial reporting is much more consumable.

2:02:25There's so many companies on the ASX that report recurring revenue when it's like SMS transactions. That's one thing, right? So putting stuff into recurring revenue that's not recurring revenue is one thing. I think you should go in and read what this is. No, no, no. But our regulation would not capture the EBITDA problem either. No, I agree with you. None of the stuff we've spoken about. EBITDA's not a financial metric, though. I don't understand what we're doing. No, no. I mean, so the physical layout of Australian reporting is much better than US in terms of being able to understand it. It's like those long, terrible forms, right?

2:03:01They're horrible, right? Yes, can't agree with it. But my point is... But the data's always in the same spot. If you're trying to train an LLM on it... That's true. That's true. But the LLM will just lie to you forever. It's only other reasons, right? But my bigger issue is not, could you get away with this in Australia in terms of submitting it? if I try to do this to investors in Australia with Catapult, let's say, or any business I was involved in, I ran – people would immediately see through the EBITDA and think I was dodgy for trying to report EBITDA. They would immediately – they would say, why are you reporting EBITDA?

2:03:31It's all CapEx. They would basically look at me calling recurring revenue illustrative and I think they might throw me out of the room at that point in time. Like, I just cannot believe that this is what's going on. I cannot believe it. Like, I'm surprised you're not more shocked about this. I'm shocked. Well, I think the whole thing's a con, the whole AI sector. So I'm not saying that there's no users. You sound like my son. He's so bearish on AI. Yeah. It's like it doesn't, I think credit to you for really pulling it apart in the way that only you can. But like, I never loved these businesses ever, but I love them even, well, I hate them even more now would be the better way to describe it.

2:04:08And I think Firmus, I don't think it ever lists. I don't think it gets through roadshows. But it lists, its ability to list depends on only one thing, the continuation of the AI bubble. That's what you're betting on. I think Australian investors, I think Australian investors are more savvy than these. This business will be full of Australian investors. Full of Australian. I think so. It'll be full of Australian. I think the reason why I listen to NASDAQ is because I don't think Australian investors would have touched it. That's interesting. I haven't thought about that as an angle. I don't know.

2:04:37There's some pretty bad stuff on the ASX. Well, there is, but not in this sense. Not in the sort of... bubbly sense in the AI bubbly sense I think the if you look at the the ASX recently it's sort of been more rational like even some of the even the great the pro medic is the tech ones the great businesses that were just too frothy that we love but we wouldn't buy have come back look at the junk lug zero great business obviously but junk valuation come right back 50 % wise tech since we said it was one of the best shorts down 50 % so all this froth has come yeah the Australian market was frothy a year ago but it's actually come back to a point where I'm not a buyer of it and in this stuff but at least it's not like this like it's not it's just to draw a positive out of the iron story which is you've got two aussies who spotted an opportunity scoured the world found a spot to turn energy into bitcoin yeah into bitcoin i couldn't and have and have and they've then managed to use that momentum and ability to build data centers they've turned it into a huge contract with as you said the world's top two biggest companies they're sandwiched in between them they still managed to get get that huge contract away relative to revenue.

2:05:44And so I think the positive is there was 10 years ago there was this big narrative of like Australians can't do technology. They don't know what they're doing. I just think that narrative's got nothing. One caveat here. Please rebut what Scott has just said. I'll give you one caveat here. They at least sold 50 million. They should be selling a billion dollars. I'll be selling every single share. I'll be doing a day mids here. They're not doing a drone shield. That's your wife. I have to tell you, like, you're a very great and savvy operator, Scott, but, like, you know, you're such a nice guy, you try to find a nice angle here.

2:06:16Like, let me just cut all of that down, okay? Basically, the Bitcoin stuff, I totally agree with you. Like, that is great. Why is he letting Silicon Valley get away with the bubble stories and letting them reap all the rewards? Why can't Australians do the same thing? They can't be keeping these dodgy straws when they're going to zero. Like, I think, to me, this is not creating anything. They've basically found a way to sell Microsoft a dollar for less than a dollar. Hey, Enron. And to sell NVIDIA a dollar for less than a dollar. And to sell Dell a dollar for less than a dollar. These shareholders are funding the profits of the biggest companies in the world.

2:06:47I maybe would have stayed with producing profitable Bitcoin and letting the government pay me to make Bitcoin. Exactly. Just staying in the corner somewhere. They couldn't really make money doing that. They could though. It's true. They are now. They are. They were. Actually, Bitcoin mining is funding a chunk of this capex. Their old technology didn't depreciate that quickly because they're just using this cheap energy to build Bitcoin. On that note, we'd better go. We've gone for, I think, like over two hours. We could do an acquired and keep going. Obviously, love to have Scott in the house.

2:07:18Just to remind our listeners, obviously this was brought to you by Ethereum Capital. Hang on, someone has to give us – I'm going to hunt around for someone with the bull case on iron. I tried to find a Temple and Webster bull case. I found an investor who is long on Temple and Webster. Still? Yeah, but wouldn't give me the bull case. I don't know. They're still long. I think they could be losing their shorts. They're still long. They're still long. so I need to find an iron long to tell me all the ways that I'm wrong which could be many by the way I'm not arrogant about this at all like I did half an hour of research right so I think this is more than most people invested in it's definitely possible so as you know you obviously buy tech companies and if everyone exit and you've got a great business what's your sweet spot 1 to 10 million revenue software technology services companies oh that's very precise that's good so you might be able to pick up iron soon I can say this I can say...

2:08:05Too much hardware and cafes for us. I can say, do we need the founder to stay in or not necessarily? We're happy for them to go on day one. And I can say this from personal experience, like whatever else goes on, like you will be very honest and fair to the person selling you the business and you are like whatever the opposite to a shark is, like that's your personality. Definitely true. Yeah, thank you. Absolutely. So thanks for the Warren Buffett Australia for joining us. Thank you, Idy. Thank you, Mark. and enjoy what are you of Australia what are you of Australia Charlie Munger you're the Charlie Munger of Australia alright that's it that's going to stick done sold just the first name that came to mind thank you everyone we'll see everybody on Saturday for our Ask Us Anything episode as always