Bitcoin's Movement, Quantum's Latest Risk & The Coming AI Energy Reckoning

0:00Welcome to Digital Currents, brought to you by Maureen Creek Digital, I'm Xavier Sigoura. Each episode we spotlight the topics and people changing the tech industry, with the focus on AI, blockchain, chips, and big data. Thanks for listening and please make sure to subscribe if you like what you hear. And remember, stay current. Mark, happy all moths Friday. We are back for Digital Currents. How was your week? Happy almost Friday. Week is good, although flown by last week of summer, I guess, before Labor Day.

1:03I see you got the memo for the orange, for Bitcoin orange. I got a little bit coin orange behind me over there. I can't make it there. So we're recording on Thursday instead of Friday because of the long weekend. And I didn't really channel. Actually, I just remember, I do have the light orange pants, not the full orange. And I did wear a little old school Bitcoin whale sock game. The reason I have the old school whale is I picked up the more modern whale and I got a hole burned through the end. But speaking of burning, BTC is on a tear. I mean, we're back over 80k. As we sit here speaking, we're about 3.5% today,

2:03flatish on the week. We were 80,000 a week ago. And then we dropped all the way back down to 76. So probably should have worn my roller coaster socks. But in the month, we're up 24% off that interim bottom around 60. So so many people, X, so many, are now declaring that crypto winners over, bear markets over, win the bull market, all you four-year cycle guys, I'm my call for it. I mean, I shouldn't call my count. But he's like on a show saying, all you four-year cycle guys are wrong. Like what are you talking about? It peaked like literally on the day of the four-year cycle. Like literally to the day, it dropped 50 plus percent, 55%. But he's saying, well, because it didn't bottom on October 5th this year, then the cycle, well, it's not October 5th yet. So you

3:10actually don't know if it's made the ultimate bottom. Now, it's certainly possible we're not going to go lower. But it's possible we could. And anyway, I just, I laugh because it's, you're right, it's not a four-year cycle. It's three years and 11 months because there's nothing to do with the calendar year. It has to do with the number of blocks. And, but humans, human to gain. And we had too much leverage at the top. And we, you know, went down a lot. And now we had capitulation at the bottom. And so I look, it's certainly possible that we bought. I actually don't believe even if we had one more fall in September that it would go below the 585 width that we had in February. So, you know, cost of electricity today is like 58k. That's the interesting thing. This is a crazy

4:13stat. 17 plus years never traded below that number. It's pretty interesting. And it makes sense. Like, if you, if you had to pay a dollar in electricity to mine a Bitcoin, you probably wouldn't sell for less than a dollar. I mean, you would just wait because you need to cover your costs. And so, now, you say, well, they have other costs. You know, there's the cost of machines and the cost of land. Yeah. So the total mining cost we've traded below that, but never the marginal cost of, of pure electricity. So, yeah. And at that point, Mark, you know, August was a big month. Most of the time, it's when people go on vacation or they're kind of not really checking things. And then they have to kind of get back to their Bloomberg terminal in the roar comes September, right? But, you know, for those that were staying current with us, we thank you throughout the summer. And a lot was going

5:15on, right? We saw, basically now the best month for BTC since November 2024, right? We saw things happening just this week that have caused other assets to slip, whether those were oil prices spiking, stocks that even bonds starting to slip. But, you know, it's also, Mark, I have to call out other coins that were doing well as well. It wasn't just BTC, but we do often say it is the boat or the the the tide list all boats. For that matter, we've got, you know, 56% up on each end. Let's not miss our beloved and affectionate XRP Army crew. They're up 45%. And, you know, we're seeing a couple different things. And to your point about the market being declared kind of now that the bear market being over and now we're entering the bull season, you can add Michael Seller as one of

6:18those believers as well, Mark, where now he as and as part of strategy have made their change in strategy that, you know, as a debt company, it's probably better for me to buy the asset that I want to sell to the public rather than sell the asset that people are paying me to accumulate. Yeah. Funny how that works. No, well, it's interesting in that, you know, strategy is certainly back. I mean, MSTR is it's up a lot. You know, it's up almost 12% today as we're recording. And now down only 10% for the year, which I think, let me just double check, but I think that is actually better than Bitcoin price for the year to date. No, just about the same. So, you know,

7:18he had been lagging Bitcoin, but now he's just about the same. And the one that that I love is, you know, stretch, right, which is the prefs, you know, are all the way back to, you know, almost 98, 97.7. And you know, they had troughed down that 74 people said, oh, this guy's bankrupt and he's going out of business and these are worthless. Make no, they're worth par and they're going to, and they're paying 12 and a half percent interest. So, and the only people, and then people said, well, then why did it go to 74? Well, because people bought it on margin. And if you buy something on margin and then somebody shorts it, it can go down. And then your force to cover the margin, call it if you don't have any money, then you have to sell it. If you sell and the guy's short,

8:19you get a spiral and that's what happens. And this happens to close down funds all the time, right? You get these mercenaries who they seed the cloud, right? They start shorting. And then they break the buck and then people panic and then they sell and you're selling to the guy who is short, which I've never understood this. Right. I just don't. And but the panic is real. Humans are going to human. And so, but now we're back in 98 and it's fine. And it's, and I can't Michael didn't do anything different. Other than your point, he instead of selling Bitcoin to pay the dividend, which is perfectly logical and rational, particularly if you're selling old basis and replacing it with new basis. Okay. That's good tax move. But then he bought a whole bunch of Bitcoin. And that's what he's supposed to do. And I was, it's funny. I did a show with somebody

9:21yesterday and they were saying, yeah, but, but it doesn't work because now he's issuing shares M nav of one, not a premium. Like, it doesn't matter. I mean, it matters a little in the sense that it's not as creative at one. If you do a below one, it's not creative. But at one, it's still fine. And ultimately, you're acquiring the asset that you want to own. And what we've seen is the correlation with M2 has come back. You know, the funny thing about this is, so M2 just goes like this. It just grows because governments just print money. Okay. So, so we know that that that's occurring. And gold, the value of it goes like this, then it goes like this, and it goes like this, and it goes like this. And so gold over long periods of time has been that perfect store of value. And it reflects the nature of, of M money supply. But not always. Like, this year, gold is down a lot.

10:28I mean, not as much as it was, but it, and last year it was up a lot. Same thing with Bitcoin. So, Bitcoin was going along with M2, perfect correlation. And then M2 kept going up and Bitcoin went down. And it was like, see, has no relation to M2. It's only relation is M2. It's like the, you know, it's the case shaped economy thing. It's, it's not that Bitcoin is going up. It's just not. One Bitcoin is one Bitcoin. The money is going down. The fiat money, the value of it is going down. You know, I went to lunch. I mean, people are tired of hearing me say this, but I went to lunch again, down, down the hill here in Chapel Hill. And, you know, I complained when I went and I got myself, right, tacos and a drink, just water, tax and tip. And it was, you know, 20 bucks. I'm like, not New York City in Chapel, North Carolina. Went with my son on the law yesterday, tacos.

11:30And it wasn't 20 bucks anymore. It's now 24 dollars per person. And it hasn't even been a year. And yet inflation is only three and a half percent. Like, well, that's 10 ish. And in less than a year. So I, I struggle with anyone who doesn't have a component of their wealth in stores of value. They could be gold, they could be platinum, could be silver, not diamonds, because diamonds just got nuked by lab diamonds. And that's a really interesting thing. Like, if someone came up with synthetic gold, gold would lose its store of value. And if someone were to come up with synthetic Bitcoin, she can't because Bitcoin's supply is fixed. And it is what it is. And to recreate

12:33the Bitcoin network, you could do it. But then you'd have to convince everybody to come over to it. And that's the part where the Linde effect applies. The longer something survives, the longer it is likely to survive. If you make it to one day old, you're likely to make it to five days old. If you make it to five days old, you're likely to make it to a month. If you make it to a month, you're likely to make it to five years. If you make it to five years, you're likely to make it. And that's, I think that's the same thing. So yeah. And I think that underscores a lot of the things in just to really zoom in on numbers that we've been sharing. If we look at what was happening way back in June, the start, the number we're looking at a 40% off of those prices where we were in the 50s and days where Colin might come to come way lower. But there's actually something last week, Mark, that we didn't talk about, which is related to this, which I got to mention was our very own Pascal going public via this back merger and acquisition. And that's significant for

13:37a whole lot of reasons, which we can get into. But then also as it relates to crypto and what always happens, Mark, is when you're starting to see new, new shine into the market, like you're seeing, you know, some of the best weeks of the year in I did alone taking a 1.3 billion into the ETFs, you're going to start seeing the continued conversation about Q-Day and quantum vulnerability into the blockchain. So now this is starting to pick up traction yet again as crypto prices rise, people think about their vulnerabilities. And so basically, you know, you've got a lot of public fees floating around. There could be potentially seven million wallets that hold almost half of the in BTC. So what I thought was a little different this time, though, Mark, which I know the four letter words that will get you dinged every single time in markets was some new traction that's happening on a proposal, BIP 360, which I'm sure you have seen, which is this idea that

14:39you could make users, public keys less visible. And therefore, not being are as susceptible to on-changing type of attacks to reverse engineer your private keys. But then there are other proposals that can be a little even more severe, like hash based type of cryptography, which is seen as more resistant to quantum. And then of course, the even harder one, potentially with the hard for coming in that would result in potentially freezing old wallets of their crypto, if they don't kind of upgrade to the latest technology. But basically now there are companies like BlackRock, Strategy, Coinbase, they're coming together to form this Bitcoin security consortium. And they're going to be putting 15 million towards quantum security on Bitcoin. So I sometimes have,

15:40you know, Bimond, this is not being a real thing. But now we've got Pascal moving the ball down the field going public, giving much more visibility to the quantum space because they're actually a quantum publicly traded company that does quantum online. And then of course, crypto's price is rising, which create a little bit of this newfound question mark of, okay, how are you going to prepare ourselves if Qda comes and it's impact on Bitcoin mark? What say what say you? Well, I've actually been less concerned about quantum risks for lots of different reasons. I mean, one, I still think quantum's a long way away despite the successes of companies like Pascal and others that have harnessed individual qubits. We need lots and lots and lots of qubits to overcome the sampling error problem of the nature of the nature quantum is kind of interesting in

16:44the sense that if I'm in binary in a traditional computing network, I'm zero or one. There's no doubt. I'm a zero or one. Well, if I can have any number of states in a quantum dot, sometimes I'm zero, sometimes I'm one, sometimes I'm in between. Okay, so now I got a, and yes, that's faster. But but now I have this problem that maybe if I ask what two times two is, I don't get four. Yeah. And if I don't get four often, like if the error rate is high, it's not really a very useful computer. I mean, really fast at, you know, doing pie out to have in many places. Great. But if you can't do math, well, I'm not really that concerned about it. So the error rate keeps coming down and, you know, that's fine. But we still don't have enough power to crack really anything, let alone shot 256. But let's say we did, let's say the quantum was

17:51there. I've always said that we use shot 256 because that was the standard level of cryptography that, you know, the NSA and the CIA created that made all the other, you know, complexities and, and, and, and if you think about, if you build a computer, someone else will hack it. Right? If you build a computer program, someone else will hack it. And that's been going on forever. And so when, when the top dogs create an encryption standard, that's pretty survivable, you know, I, I, we had this experience with the cold card hack, right? In that if you do entropy, which is, you know, how you randomize your, your words in your, your seed phrase, the correct way, it literally takes, you know, two quadrillion years to guess a random sequence of, of 12 words. No one's going to do that. But if you do it wrong,

18:57people were able to guess. Now, still mind boggling, but that didn't need quantum. It just needed fast enough with a break in system, but it's because they weren't using full power of, of shot 256. So, what I'm saying there is when shot 256 came out, someone built Bitcoin to adopt this level of encryption that with normal computing was unbreakable. Okay? Okay. So now new computing happens. The logic says that the programmers, the developers, will change the code such that it adopts this new standard of whatever that will be. Shaw, you know, 1024. I don't know. So, I'm not as concerned. Maybe I should be. And then my second reason I'm not concerned is if a computer existed, quantum computer existed,

19:58that you could do some, some good hacking with, which doesn't really exist yet. But if it did, the first thing you go after ain't going to be Bitcoin. I'm sorry. It's just not. I mean, you have four digit pins to hack everybody's bank account. My guess is there's a lot more money in banking and a lot more money at JP Morgan than there is in Bitcoin. Maybe that's not 100% true, but it's probably close. Maybe all the banking system together. Nuclear launch codes, way more valuable than Bitcoin. So lots of things that I would point that quantum computer at first. Yeah. And then maybe even to that point, Mark, we may not even need to really wait for quantum to be a threat towards cybersecurity because I'm sure you know where I'm going with this. And that would be open AI and its latest model in Astra, which I love the reporting where it's basically designed with the intention to help cybersecurity issues. But therefore creates new vulnerabilities

21:04to reverse the engineer the thing that it's designed to do. And so basically it's the first LLM to exceed the critical cybersecurity threshold in its preparedness frameworks. And so basically there are these third parties that will evaluate how likely, how dangerous an LLM is. And if you recall, when Anthropic produced Claude and Fable, there was a lot of pushback initially and they have to kind of tweak things to make sure that things were a little bit safer. So some of the talk about how quantum is this essential threat towards crypto. AI is already here and creates vulnerabilities at the user's behest because if you can attack somebody's system where their keys are already stored, you don't need to crack the entire blockchain. You just need to crack whatever they're looking at. And that's a great point, right? And lots of people have lost things at exchanges because they had a bad password, right? That it was easy to guess. So yeah, exactly, right. In that

22:09you don't have to crack the mother code to get a bunch of wealth. And we've seen exploits on on bridges or on dexes or all kinds of places where it's kind of like if you were a bank robber, right? Willie Sutton in the 20s or whatever. And they asked Willie, Willie, why do you rob banks? Because that's where they keep the money. And he didn't rob the big banks on Wall Street. He robbed the little banks and you know, we've all seen the movie Point Break. Why did Boaty rob the banks? Because they had money in them and he just needed enough money to go on his next, you know, great adventure. And so I say all the time, how much money has been stolen from the fed? Zero dollars, right? Zero dollars. How much has been stolen from Bank of America's New York headquarters? Zero dollars. How about Bank of America's El Paso, Texas branch office? More than zero dollars, right? How about your wallet or your purse? More than zero dollars. So the fur that you are from

23:16the source of the money, the more vulnerable you are, right? It's a bare asset. I mean, if I have money in my wallet, which I don't anymore, but if I did and I'm walking down the streets of, you know, Buenos Aires and someone says, give me your wallet, that money is now gone, right? It doesn't matter if I wanted it, it's now gone because that person now has that bare asset. And, you know, it's fine. It's it's a back to school season and you know, my Irish are playing on Sunday. And the reason I met my wife is she was at Notre Dame as a new student. And she had gone to the grocery store, shall we say, to buy a present for a buddy who let them use his car and they bought him a six-pack of ear. And she was waiting in the car while the other girls were getting at the stuff. And this gang member came up and opened her door and grabbed her wallet out of her

24:19hand. And instead of just giving it to him, she tried to fight back and he backhanded her across the face with a big studded bracelet. You got a big black eye. That was the extent of the damage, but he took the wallet and they found the wallet hours later. All the money was gone, but he left everything else thankfully for her. She got her ID back and all that. But he took the bare asset out and threw the rest away because it wasn't useful because he didn't have her pin number for her ATM card. And they ended that story as well. That was a high-dreamy truck. Well, you know, she was roommates with my buddy's sister. So we went over to say, hey, you know, things really aren't that bad here. It's safe. We'll take care of you. And I brought her flowers and ice cream and the ice cream was the winner. The actual funny part of that story, X's, so we went on our first date and she stood the big shiner and went, Prince's purple rain, right, dating myself right the movie. And there's this nice

25:23older couple. And you know, when you get older and I do a little bit, my wife says I talk loud now. She says to her husband, can you believe she stays with him? I'm like, oh, oh, yeah, I did that. Yeah, I, I, please tell you, you picked your up in a big motorcycle to tie it all together. Yeah, exactly. Exactly. Yeah, but it is an incredible story. And definitely we are in that same mode of getting back to school. My little one just started kindergarten yesterday. And he made it back for day two. So that actually is about as big an event as there is to get back for day two. Exactly. Go back for day two and then a couple days off before we'll try it again for week two. But you know, as we kind of look to see, as we're mentioning September being this really big, big month, we are also remiss mark if we don't speak a little bit about when you mentioned about

26:23theft and how September is this big month. What is going to happen with our rates, right? And so Kevin Warsh obviously gave the annual symposium conversation in Jackson Hole. And so now investors are of two minds that, you know, this, this is going to happen, but that kind of not necessarily a slam dunk. So now we have 70% on it on Cal State of, of no change. And the odd thing here was this after coming out for July's PCE came in at 3.7 annually, which is unchanged from June. So we kind of have this thing where we've got, you know, the head, the fed hiking potentially to 4.6, but with a 3.7 on the PCE completely different regime than everything that's getting modeled in kind of our crypto world. So, you know, you want to talk about theft and not lunch going from $20 to $24. We're seeing this here in real time. They could impact this September fiscal policy.

27:26Look, I mean, we are in the pot. We're the frogs. And they're turning it up a degree an hour. And we're getting boiled alive, right? And we all know the thing, if you drop a frog in hot water, it will jump out. But if you put it in cool water and turn up the heat, it's muscles will relax and then paralyze and it won't be able to jump out and then buy. And so we are the frogs. And because the problem is, Worsh is just another tool in the toolkit, you know, installed by a bunch of people who don't give a damn about you or me or anybody else who listen to this call. And, you know, I think it's amazing to me to sit to this day when you look at the ownership structure of the companies that built all of the drones and stuff that went into the Iran War, like the zip code of who actually

28:26owns those companies. It's congressman and senators, shockingly, like really kind of wild. And it's like when Donald's run, Donald Rumsfeld was the chairman of the company that made the drug sipro that everybody had to buy because we were all going to get anthrax in the mail. Like, why do we pick sipro? Oh, because Donald or, you know, when Dick Cheney happened to be the chairman of Haliburton when they got the contract to rebuild Iraq, and maybe that was a coincidence. So the graph encryption exists and we deficit spend like crazy. So now we're at 40 trillion. Well, to pay back 40 trillion, like you'd have to tax all the wealth, well, that wouldn't do it, right? Even we tax everybody's wealth as wealthy as the top one percent is, like the top 100 people today own as much as the bottom half of the people on the planet. I mean, that's a crazy stat.

29:30Okay, but let's say we taxed all of their wealth. That's not enough to pay back the debt. So if you can't pay it back, they either got to restructure it. No one wants 70 cents on the dollar. They don't even want to 100 cents on the dollar because Japanese are like sold to you. So Scott has to wave it in. Mr. Besson has to wave it in. And we're going to end up like Japan owning all of our own debt. There's no one else wants to buy it because we might default on it. Or if you default, then you become Venezuela or Zimbabwe and politicians hate that because then you get kicked out and you can't do your graph encryption anymore. Well, Mark, double tap on that point that you made because it's super important. And we mentioned it on the on the other call where this idea that, you know, we get money as a country through tax revenue predominantly. We've spent all of that so we can create this IU called the bond that says we're America. We're good for it. We do all these amazing things. Buy this into the rears. But then nobody wants to buy anymore, Mark. So then what do we do?

30:32We buy back our own debt. We buy back our own debt. And it's your only choice, right? And again, we know what's going to happen because we've seen the movie. Look at Japan 11 years and then follow what they did, right? 1989, their market peaks 2000, our market peaks, 1993, their bonds get downgraded from AAA 2004. Our bonds get downgraded from AAA. And it's because demographically we are the same as Japan just 11 years later. So they're older. And I see myself. I look older. Someone actually in the comments said, hey, this guy looks older. I'm like, I am older. I mean, I'm glad you've been watching me for that long. But yeah, I'm older. And it's because you can't stop that process. And so as a population ages, one, the productivity goes down. So you're not going to raise it. No matter how many LLMs you turn on, you're not going

31:34to raise 70-year-olds productivity. It's not going to happen. It's not the way it works. And so or you're not going to raise 25-year-olds productivity because they don't know enough yet. They need to learn. And so you can't increase productivity. And so what happens is your only way to make it work is to literally print money. Okay. And devalue your currency. So you can look at the yen. It's been going like this. Well, then if your yen is collapsing and your bonds have risk of default because you got too many of them, everybody says, I don't want your bonds. So the central bank has to buy them. They are the lender of last resort. And so the central bank of Japan, bank of Japan, owns 70-ish, 75% of their bonds. And when they get to 85-90, they just start over. They cancel them out. They literally just cancel the bonds. They can't do that. The currency will collapse. No,

32:34the currency has been collapsing. That's the thing. It's a slow, it's boiling the frog. But then the debt's gone and you start over. The UK did this in the 1840s and it worked fine. And remember, the UK was the dominant superpower until 1913. When they invaded Mesopotamia and incurred a bunch of debt got into the situation we're in now, the pound sterling collapsed. We, it took 30 years, ascended. And now we invaded coincidentally Mesopotamia and incurred a bunch of debt, the dollar collapse, the dollar collapsing. And what's going to happen? Oh, look at the reman B. The reman B is ascended. It's the most strong currency in the world. Oh, and what are they doing? They're buying gold to back their currency because they want it to be fully convertible and they want to be the world reserve currency. So we are in a regime, regime change, all empires and every single one in history, everyone. And they all end for the same reasons. And so we are in the death throws of

33:39that. And so it said, is it dead of the day? But the UK is behind our bonds. I mean, no, they're not. That's Middle Eastern money hiding in the UK that we, you know, cut a deal with. Remember, we said, we'll protect you as long as you keep the dollar strong. And now you've got this coalition of banks saying, oh, we're going to start a stablecoin company. Really? Why would you do that? Oh, because you're going to nominate the stablecoins in dollars. It's all about trying to perpetuate US dollar hedge money. But in the end, the dollar is going to go down because we print too many of them. And it's, it's like the sun rising in the morning and setting in the evening. You can't stop it. It's going to happen. No, I'm changing day to day and maybe it'll move around a little bit in the sky, but it's going to happen. And so, does that mean the US empire collapsing? It means we're going to all, you know, have a horrible life. No, UK is still pretty cool. I mean,

34:41yeah, London's still pretty cool place to go. Still exists. It didn't fall into the ocean. They're just no longer the center of the universe. And when Beijing and Shanghai are the center of the universe, which they will be, it won't mean that New York and San Francisco aren't cool. They still will be. But we won't have the dominance. And ultimately, I do believe that the next step is the Bitcoin and the network state. And Bellagy talked about this at Bitcoin Asia. I was supposed to interview him, but I got way laid on my terrible trip that didn't happen. That's the listeners got Mark for another week, which worked out for everybody. Yeah, exactly, exactly. But Bellagy talked about this. He said the network state where everyone lives in this networked state as a post individual state, Bitcoin could and probably will be the reserve currency in that world. But we're a ways away from that. And I think there's one more iteration. Anyway,

35:49but maybe we can get you to go slightly more sinister because we've got the long weekend, we've got college football, vibes are high. Yeah, let's try to see if we can, if we can kind of balance that out when we talk about AI and Infra. So the idea of collapse, I thought this was pretty interesting. You know, they say that everything is bigger in Texas. Well, when it comes to AI demand, that couldn't be more true. But there was a new proposal, which I'm sure that you have seen, Morgan. I think this could have a big impact on what we can expect also from IPOs that are large are going to be these AI companies in the LLM space. But there was a new directive here from Greg at it, Governor of Texas. And so they basically are calling to freeze new data center interconnections and audit the requests that are already in the queue. So the number they're actually pretty ridiculous. Mark. So basically we've got this a couple of years ago in 2023. There was about 48 gigawatts of demand that was coming through. And that was against what is now this big queue

37:00of 474 gigawatts. And basically there is this idea about ghost demand. So there's all this audit that is occurring now to see what's actually real there. Because I believe in documents so many of these companies increasing their loads, they're going to have to create real infrastructure to power these models. But then the question is always number one is there real consumer demand. And number two, for the consumer demand that's coming in there, can this actually contribute to companies making revenues? Because as we've seen Microsoft and some of the others, people are now using their LLM, which is vastly more expensive to run requests that can be done with the Google search, which is vastly cheaper. But I think that's maybe a consumer thing. We don't always know the difference between two. We need an answer. But now we're actually seeing states start to act pretty quickly by saying no new requests, at least for now. Look, it's an interesting

38:03dilemma that I, if you incentivize people in a certain way, you shouldn't be surprised that they did that. So like, remember the time, maybe 15 years ago, 18 years ago, it's done matter. When every doctor and dentist you knew had an expedition, right? And like, why are they all suddenly driving expeditions? Well, because we passed a law that said, if you bought a car over 4,600 pounds gross vehicle weight, you could write off the whole thing in the year. Accelerate depreciation one year. So every small business, you know, Dr. Dennis, bought an expedition, not a Tahoe, or it had to be bigger than 4,600 pounds. And so then, it researches crash in and they get the guzzle gas and it's like, okay, that was a bad idea.

39:08So they changed the world. And now you don't see doctors and dentists buying expeditions. So same things to hear. If you incent people to spend on infrastructure and you give them, you know, depreciation incentives or tax incentives, guess what? They will do that. But here's, here's the place it breaks down. So they all say, well, I'm going to build a gigawatt of power and data center, gigawatt. And that's going to generate all this revenue. Right? And Jensen said, one gigawatt will generate, we talked about this last week, one gigawatt will generate 50 billion, no, I'm sorry, $600 billion of revenue. I'm like, Jensen, the entire profits of the S&P last year were 2.1 trillion. So you're saying for every single gigawatt,

40:0930% of corporate profits are going to get spent. Okay. So we build four gigawatts and that's all of the profits. So now there's no more money. You said 400 just in Texas alone in backlog. Okay. That's a problem. So and this is the problem. One person's revenue is another person's expense. Like if suddenly Google builds a data center or X builds a data center, SpaceX builds a data center, and they're going to have all this income, this revenue. Okay, but that's got to come from some place. And I don't know about you, but I think about just our little business. Yes, we have a little seat license with Anthropic and we pay a little bit to them so people can use the model. And the model is good for certain things. But it turns out, I can, I can test in this,

41:15I can do just about as good a job as our super sorry analysts using the free version as they're doing with the paid version. So I'm not exactly the same, but I'm holding my own. And so I'm not going to authorize spending a million dollars a year to Anthropic. I'm just not. Now will we pay 10,000? Grudgingly would I pay 20,000? Nope. So this is the problem I see is I get it that yeah, if you can book this spending as income, like this is the one I don't get. In video yeah, it says, okay, I promise to deliver you chips for a data center that you promise to build. And I know you don't have any money open AI. So I'll guarantee your payment.

42:17They get to book that as revenue. Like, how is that revenue? There was no, there's not a transaction. I mean, there's there's a bunch of promises that don't exist. And yet that's rev. That's right. And now, and then as soon as it's built, now you got to depreciate it. Well, that's going to go against income. So what is Microsoft said, oh, you just extended the useful life to 25 years. Like 25 years. What chip name a chip that lasted 25 years? Give me one Pentiums? No. I mean, the quad for processor, whatever, whatever it was. I mean, no. I mean, I struggle with with all of the accounting shenanigans that make people believe that that you can just say like Anthropics said their, their tam that their revenue was going to be 30

43:25trillion dollars. Like with a straight face because SpaceX said 28. So we have to say 30. Like 30 trillion dollars for one company. Okay. The US GDP today is 32 trillion. So all the other companies in America are just going to stop having revenue. So your LLM, which to your point, if I ask Google, where's the nearest Mexican restaurant? Because I want to go to lunch. Okay. Maps would do that. Google search would do that. Now when I ask it that, Gemini gives me the answer. Spent a whole bunch of tokens for something that I, it's pretty clear.

44:27The Mexican restaurant's right there. It's been there for years. I don't need tokens to figure that out. And your point is so important that using things that they're not it tended for or useful for is not a business model that will sustain. Because you can get away with it for a while as long as investors are willing to buy your debt. Like this is the other problem is hyperscaler debt loads are going like this and they're free cash flows going like this. So at some point, and we all saw the, I'm sure you saw the meme yesterday of the Emperor as no clothes. And the Donald was sleeping in the meeting. And of course, the meme was he's walking around naked and everybody saying, oh, your row looks so great. And if you say you have a

45:27robot, then you definitely have a robot. And if you say you were at that meeting and I saw your eyes closed actually, but okay. So at some point, someone, and me and I was going to say someone will say, no, those numbers don't match. But I hear you. Here's the problem. So much of the money today is managed in passive strategies that don't ask that question. They're just like, is SpaceX in the index? If it is, I have to buy it. It doesn't do what an analyst used to do of saying, it's impossible to grow your revenue to 100, to 100 times in four years, which is what Goldman and Goldman, arguably, is a legitimate firm, right? They put out a piece of research, pre-SpaceX IPO, saying that SpaceX could 100x their revenue in three and a half years.

46:33In fact, no company in history of size, like you could go from one to 100 fine, but any company over a billion, no one ever, ever, ever, ever has done that. It's just not going to happen. And so this is the part where I'm not sure everything, but I certainly don't want to buy these things where the math ain't massive. Yeah. And for when there's a disagreement there, we always kind of turn the tables to the visual if the words can explain what's going on. Pictures can give us this visual representation of a thousand. Well, you have great more capacity when you have the chart of the week, which is where we kind of look at this in a graphical format and then figuring out what this actually means for us when we take a look at all the numbers together, right? And I think

47:36what's amazing behind that is this idea of ghost demand. And this is really nothing new. It's this idea of ghost power where I'm a developer. This is something that they do, but that obviously city planners hate. They will make a forecast that is super, super high on kind of one project and make multiple forecasts because they're not sure kind of how things are going to shape out in terms of the draw on power. And that'll kind of allow them to assure themselves some power allocation. Well, now 90% of those request are data centers, right? So that's the kind of the big shift that that we're seeing over over in Texas. But we will show this visually and kind of dive in a little bit more about what this what this implies for the AI world and I will share my screen mark and we can kind of dive in a little bit further on that on that conversation. We can see if I might be able to make this a a two-fer where we'll kind of have both both charts kind of telling us

48:42the right the right conversation. So first, I want to tie in something that you just mentioned about what what companies are are doing. So exactly on point, I always same wavelength mark. Most companies are exceeding their AI budgets, right? I believe what for that use? Bart, I will briefly comment here and then get more of your analysis because I do I do. Even though we didn't channel exactly, I mean your top match is my bottom though. So we do both that orange on. But this is not a surprise in the sense that if you give someone a tool, they will use that tool. The problem that I see with the tool that is I think the biggest cause of this, the tool is incentivized and programmed to maximize its own usage.

49:45Like that's why it it always complements the user's question. It asks you if you want to ask more questions, which means more tokens. And so it's like the ultimate upsellor. Like you go to the restaurant, oh don't you want appetizers? Oh, oh you want one episode. How about two appetizers? Oh, that was such a great choice of your appetizers. Hey, how about this entree? Oh, don't you want to upsell to this extra sauce? I mean, it's the ultimate upsellor and it's obsequious, right? Oh, you're so smart. I mean, I'm like, stop telling me I'm smart. I don't I don't want any of that. I don't want you to waste tokens on anything other than the answer of the question I've asked you. And don't give me any other. So that's the first problem. The second problem is, well, if I've got this tool, I could ask it like, you know, I had this itch on my left shoulder for a while. What do you think it is? So you're asking for personal stuff and go to Web MD.

50:54Just as good and free. But this too shall pass because like the story of the guy who got the bill for I think it was $500 million. I mean, big enterprise. Who said what the absolute F? I mean, no, no, I did not authorize this. So cut it back and they cut it back 90%. And so and yet, you know, anthropic was ARRing that like, oh, that's going to go on forever. No, it's not. So token maxing is again, it's built into the programs because they and when I don't understand it's like in most places, we would never do that. We would never, you know, I'm trying to think of a silly example. Like let's say of a home watering system. No one would say, you know what? Maximize the amount of water that you sprinkle on my lawn every day. No one would enter that prompt into their sprinkling

52:00system. They would say, put them the optimal amount. They might even put a minimal amount to keep it alive. But no one would say spend as much money on water as you possibly can. But that's what we've done with these silly, I shouldn't say silly with these models. I said, they have good uses. But at the end, they are, they're tools. And they're not even really good. I think I talked about this last week, but it makes me so angry that I have to feel I have to talk about again is, you know, we were working on this, this company. And I said, how much revenue does this, this company have? And it says, oh, that's a billion dollars. And thankfully, I had enough knowledge to say, I am a hundred percent sure that this company does not have a billion dollars or revenue. Can you double check? It's like, oh, you're right. I mean, don't tell me. I mean, I was, but don't tell. Oh,

53:03you're right. My mistake, that's the last valuation. You know, for their last funding round. I'm like, oh, my God. Okay. So, and the problem is, if, if I didn't have the knowledge that that was absolutely wrong, and I just accepted the first answer as if it were a fact. And this is why I really, I'm almost wary of these things in the sense that kind of like social media can create truth by just repeating it often enough, even if it's a lie. This, these models can create truth because people deem them to have authority. Like, but, but it's not authoritative. It's not a database of facts. It's making a probabilistic estimate of what a word should be based on the prompt. And if you prompted badly, you're going to get a bad output. And maybe so I, I, I, I said, I'm not saying we

54:11should ban them. And I'm not saying we shouldn't use them. But holy moly, if, if you can't rely on the answer because there's no fact checking, right? There's no, there's no like gold star that pops us says, Hey, this is a fact that I have received and checked as opposed to this is an opinion based on a probabilistic judgment of what you asked for. Those are different things. And to summarize here for those that are on audio only when you're asking respondents, those are the kinsy that, that pose this, you know, or how are you budgeting your AI use? Only 5% of respondents were actually under budget. And Mark, if these are publicly traded, I would love to see what the accompanying stock prices were because we have found that, you know, you've got a huge AI boom. If you're probably here, kind of right at the median. So 39% were over, but just by a little bit on single digits. And then, you know, over almost half or over by a lot, right? Over a 10, 30% and if we know that a lot of

55:13companies have huge cat-backed buzzes when it comes to tech, infra, and obviously AI, that's a, that's a very big bogey big number. They have to climb, right? I remember these are expenses. These are expenses that reduce profits in the companies that you, so when, when the CEO comes on says, hey, I, I, I, I, we love AI, we love AI. I'm like, you're burning my cash. I'm a shareholder. That's my cash. Why are you burning my cash? And why can't you control your costs? If they were spending on, you know, nightclubs and extra food and, and, and, and luxury travel, we would bust them. But because they're spending it on AI, we're like applauding them. No, that's stupid. Exactly. And now, let's go sinister all the way and we'll kind of now look at what, what, what has this pulled forward from, from Big Old Texas and, and they heard grid. So I will share a new chart and we will, we will

56:14comment on, on that. Let me just make this change. And it basically kind of illustrates this, this point that things are, things are not necessarily getting getting better anytime soon, anytime soon. So first, this is what this represents graphically marked in terms of what the queue looks like and what the actual reality is. It's not grim. So, peak demand in Texas is 85 gigawatts. And an entire US data center fleet is 65 gigawatts. The back load is 424 gigawatts, which is just astounding in terms of where, where we are, where we are headed, what, what this, what this is. This is, and it's, it's appalling, right? If, if there's only 85 gigawatts available, you're not going to suddenly have power for 400 gigawatts, it's just not going to happen. And so, there's going to be a lot of disappointment and there's going to be a lot of wasted money. But,

57:18but none of the people are going to give the money back, right? It took advances on these, these, these contracts. So, it's, it's going to create a big crater in, in people's earnings, but you know, you get with that later. It's, it's so comical, but not unexpected that it's Texas, because one, you know, I think it's bigger in Texas. So, it's, it's got to be bigger. And so, you know, if, if Ohio is going to do 20 gigawatts, we got to do 50. And, but it's, it's also interesting in that if you understand the physics of data centers and things like cooling, you probably wouldn't build in Texas, where it gets hot in the summer. And where most of, you know, you, you might say, well, but it's so, it's so rural that you're not going to have the Nimbis fighting it. Okay, fine. There's a lot of open space, but there's also a lot of open space

58:20in Montana, where it's a lot cooler. And, you know, you could do the same thing in North Dakota. So, or heck, Iceland, you know, was the name implies. Or maybe we'll just get Greenland, we'll just fill it up with data centers. So, I, most of that orange isn't going to get built. That, that's, that's first. Second, we don't have enough power. And it takes too long to build power centers. And so, I was like, oh, well, let's, let's just do these small nuclear actors. I you think it's tough to get a data center approved. Ask your neighbors if they're okay, if you put a little nuke in your backyard. Yeah, just, just, just a little one here. And then, of course, if the point wasn't, wasn't illustrated more, more fully, it's what does that trend look like over time is hope here or no. What we're seeing here is that people are getting greeter. And, Mark, I guess we can probably kind of crystallize that point in terms of, you know, what do Texas,

59:23Ohio and Pennsylvania all have in common other than being amazing states for football talent, speaking of, of Saturdays and the fall. Yeah, it's that they're all pushing back, right? So, in a couple, well, in August, a couple of weeks back, uh, Governor Josh Shapiro signed an executive order, tightened the permission for data centers and any load that's exceeding 25 megawatts in Pennsylvania. Each of these projects will either have to bring their own generation or storage or get some sort of other type of type of, um, type of filtration in the Corp. Ohio, had something similar. It's the same old story every time. Get rich quick, don't have to do any work. And the way these, these contracts, the way these, these projections work is you get a couple guys who failed at something. And I literally just met with a group here in North Carolina, trying to do the same thing. So they failed at something and like, we need money. Oh, okay. So, we're going to go to people that need compute universities or, or whatever. And we're going to say,

1:00:27we've got a data center. So we're going to find an abandoned lot, hopefully one that's brown fields and, but still has the power hookup. Oh, we're going to turn this into a data center. As long as you give us a 10 year commitment at the current price, then I can use that contract to get the financing so that I can build like, okay, but why don't I do that myself? Look, that's a whole new thing. But, and yet people are falling for it. So, like, well, yeah, I need compute really bad. And if you tell, so great. And I'm like, well, wait a minute. Why would anyone, given the price of tokens as a commodity, is going to come down and the price of all of this stuff is going to come down, why would I sign up a 10 year plan at a fixed price? I wouldn't. Well, but I need the compute now. I get it. You think you need the compute now, but the realities didn't have the compute yesterday. And, you know, maybe you'll need the compute tomorrow, but you probably don't need to sign a 10 year contract today. But they're high pressure selling these people

1:01:28that if you don't take it, it's kind of like all the apartment buildings that went up an opportunity zones. And now you've got the constant leasing pressure, like, oh, if you don't, if you don't do it today, you know, you're going to miss your chance to get the best best apartment. Like, yeah, but they'll build another one tomorrow. I mean, there's so many empty apartments. So I, there's going to be a lot of empty data centers. There's going to be a lot of, like I say, a lot of these things just get canceled. And a lot of these people are not going to get rich, even though they they told all their investors they were going to. Great. So we've managed. We've managed to do it. We've managed to go all the way sinister, which was my goal from the beginning. And of course, in home post, those requests in Texas, they grew up 10 X in three years, which is the return of probably the most successful venture investor that you can imagine 10 X in three year demand is, is pretty astounding. We will wish all of our digital currents audience to have a great long weekend, enjoy all the slate of games.

1:02:30For week one, Mark and I, this is where I think we probably love the most. And in the time of the year, when, when everybody, Mark is, is undefeated, including UNC, I might say. Exactly. Yeah. Wish them well. Yeah, we do wish them well. Some of us don't actually wish that, I wish the kids well. I hope the coach loses every game. I'm ready to be, I'm ready to have a new coach down here. But I'm through like a little ad for it. Go ahead, Mark. No, no, no, I'm ready for a new coach. But good luck to your, your boys and blue and my boys and blue. And it should be a fun, fun weekend. I do love the Saturday, Sunday, Monday, first weekend kind of. Yeah. It's, it's pretty fun. Incredible. Friday Saturday, Sunday, Monday. That's right. Extra, extra days because we just need extra slots for advertisers to pay for all day. Actually, it might even, there might even be games tonight. It might be Thursday, Friday, Saturday, Sunday, Monday. So, you know, work. Actually, I must admit, Thursday is probably my favorite day for called football because you're not supposed

1:03:33to get it. But you'll get like a decent game like, you know, a, a Louisville, like Kentucky on, on grass. But we want to thank our listeners as always for tuning in week after week and remind us always to stay current. Thank you for joining us this week on digital currents. Check out our links to our website and social media in the show's description box. Please make sure to subscribe if you like what you hear. And remember, stay current. This podcast should not be construed as investment advice nor solicitation for the celebrity security advisor or other service. Investment themes and ideas discussed may be owned by funds managed by the hosting podcast guests. Any conflicts at the time of production have been mentioned by the host but are subject to change. Listen to should consult their personal financial advisors before making any investment decisions.