Everyone Is Panic Selling AI

This is a structural bull market, structural demand, structural change in the world that's just not going to stop for years to come. We literally have countries that are in a race to beat each other, companies that are in a race to beat each other. Like this stuff is not stopping because they need to win this. >> What's up everybody? It's LG Du Set here and welcome to Milk Road AI, the daily AI show that knows there's no gain without the pain and man, does it ever hurt right now. Today is July 29th, 2026 recording on July 28th, the day before. By the time you listen to this, you will know whether Kevin Warsh has hiked the rates and if we're entering a new cycle of higher rates. And you'll probably also know how Apple, Amazon, Meta, Microsoft and many others have reported on their earnings for Q2 and how the market's reacted. We're clearly at a contentious point for the AI build out trade as the factors I listed above start to circle and the threat of Chinese open-source models also looms and the market starts to look at actual applications for AI rather than just more memory and energy companies. We'll discuss these themes and more today on our weekly roll up episode with our leading AI analysts and a reminder that if you want to see their positions and catch their next big move after they already called Micron, AMD, Bloom and others earlier this year, all of that is in Milk Road Pro and it's just a dollar to test it out for 7-day trial at the link below. Today's podcast is free and it wouldn't be possible without our partners at Securitize, the regulated rails for tokenization and Bitget, stocks 2.0 with real liquidity and real dividends. Keep an ear out later in the show for a message from them. All right boys, we're recording this a day before people listen to it. So who knows what the market's going to be like by the time they listen to it, but I don't know if it'll be that different. Market is let's say down is probably a way to simplify it today. I want to hear from each of you how you're feeling cuz that's how we just start the episodes. Melvin, how are you feeling? You look like you're actually kind of sweating it a little bit. >> I am a little bit sweating. >> [laughter] >> My I'm having a terrible week. My power's been out for 2 days. Market's market's getting nuked. Micron's getting crushed. Nvidia's is down. Sweating hella hard. Uh but um you know, this this whole market sell-off has been a whole bunch of fud, I feel like lately. We have Citadel came in coming out yesterday saying that they're going to, you know, hike rates to Fed might hike rates. Um then you had the Cosby sell-off um because of CXM MT IPO and um China's DEE uh machines. Um and there's just been a lot of like negative um news behind this sell-off and I feel like um I I think this is not right and I I think market will recover in the next like next week and I think we're just going to go continue to go higher. Um I think this is all just a big fud and uh this sell-off is is a good opportunity for you to buy right now. >> Oh boy, Vincent, how about you, man? How do you feel? Same thing? >> [snorts] >> It's a tough day, for sure, but uh I remain bullish. I think what we're seeing now is that the market is basically shifting from only looking at TAM and demand for companies towards how much of that demand can they actually uh materialize in in in revenue and how much reaches their P&L, right? We saw that with with Corning today, with Amkor as well. All going down on on really good earnings, but guidance was the issue. Guidance came in more or less with what was expected by markets, but um yeah, we're at the point in market where where we're really looking for how much of that future demand can we bring in in revenue early. And this is especially important for suppliers that are tied to kind of future systems. Talking about um uh companies that are supplying into the the Nvidia Rubin Ultra supply chain, into the Kyber rack, the 800 V. This is everything I was covering that is coming in the future, which will be that next step up. Those names are get hit hard, I think, because that revenue still lies in the in in in kind of the future and is not here yet. And yeah, but I remain bullish and thinking about buying some dips here. >> Well, yeah, and Vincent, we'll talk portfolios later, but Vincent has the biggest cash position out of anybody on the show today. Can't even say it. Nobody has any. Kyle, how about you? How do you feel? >> Yeah, I feel I feel good. I think um I think what we're seeing is over the last few weeks, the the chances of a rate hike have just been slowly increasing. And as that's happened, the market has continued to pull back. Nasdaq's now down uh little over 10% after the the dip this morning. Um and I I think the market is a little or was a little spooked anyway that war is going to start raising rates and sort of um start to shrink the balance sheet. Um and when you're shrinking the bal- the balance sheet, then investments years out are less um exciting to investors, right? If they're growing the balance sheet and rates are going to zero and they're printing money, um then you want to invest everything you have today, right? Because you know that your your dollars are inflating. Whereas the market is kind of thinking the opposite's going to happen. Now, we haven't seen that that actually is going to be reality. I don't think that they're going to raise rates tomorrow. Actually, by the time you guys are listening to this, you're already going to have the answer. So, um but I I don't think they will. I think inflation has uh at least the last inflation print we saw sort of uh surprised to the downside. Um oil had a little jump last week and then now has come back down um with, you know, everything going on with Iran and US. So, I think ultimately they're not going to raise rates. They don't really need to. And then I just think on top of that, the market is kind of forgetting that we're in this technological revolution where we now have infinite artificial intelligence and that's literally going to change every company in the way that the economy works and this is a structural bull market, structural demand, structural change in the world that's just not going to stop for years to come. We literally have countries that are in a race to beat each other, companies that are in a race to beat each other. Like this stuff is not stopping because they need to win this. Countries companies people everyone needs to win this. They need to do more of this. Um and the ROI is already there in many cases. We saw it with Google. Their earnings was absolutely unbelievable last week. Uh cloud was up was 82 or 87% whatever the number was year over year. Their backlog has grown 50 billion in the last quarter. Um and so they're spending money because they know they're going to make the money when they spend it. Um and then it's not just them but you saw like ServiceNow's earnings last week. Um you know, everyone's saying SaaS is dead but their business is clearly being accelerated by AI. Salesforce, I think we're going to see the exact same when theirs comes through. Eli Lilly, Apple, like all these companies are continuing to see their earnings grow and AI is a big part of it. Um and this is true both on the bottom neck side and the application layer. So I just think the market has just had a lot of things all at once. Fear of rate hikes, right? Which I don't think is going to end up coming through. We went on a crazy run since April, right? Since the Iran war started and so sure we need a bit of a pullback. Korea got over leveraged and so that's kind of falling apart right now. Uh that's why the the Korean stock market's down. And then yeah, we have some fun out of the Chinese market in terms of you know, open source models which I think is a net good for the entire market and actually is probably going to make a broader bull market rather than a more um kind of isolated one. Um and then you have, you know, some innovations around chips and manufacturing and stuff. Uh but ultimately I just think there's still so much demand left and I just think during pullbacks like this the market forgets that. Uh and so I think this is a great time to be buying both the application layer and the infrastructure layer. Uh I think they're both going to go into a a a bull market together closing into the to the rest of this year. And then also don't forget we're in the middle of summer. This is the worst time for markets usually. You know, you've heard the quote sell in May and go away. That's literally just what's happening right now. So you kind of got all these things kind of combined into one. The market Nasdaq's now down 10% or 11%. So that's typically a proper correction. So is it over right now? I don't know. Maybe it's still a few more weeks. Maybe we need to wait till the summer's done. But otherwise I think these are incredible buying opportunities on either side. >> Just going to pause there for a second to point out that the market is showing signs of something kind of different happening. And our analysts at Milk Road Pro are all over it. They spent the last couple weeks making a lot of trades getting out of some positions and then getting into a lot of new ones getting ready for the next wave of robotics, space, or even kind of picking some different AI winners. If you want to see what they have in their portfolios, what positions they're opening, it's just a dollar in Milk Road Pro at the link below. >> And uh the odds on Polymarket for the for the rate hikes went from like 98% a couple days ago to now 78% and in terms of no cut, which is not it's still 78% no cut. So it's not this huge huge odds, but it did change and I think maybe that's also kind of what's what's spooking the market or or the market spook is is influencing Polymarket one way or the other. Kyle, I want to dig into something that you asked or that you mentioned. You said that um we would be in for a broader bull market because of the Chinese open source models. What do you mean by that? >> Yeah, so the the open source models provides much cheaper intelligence. And ultimately this is what a lot of companies need to unlock the next use case, right? We're not really going to get a massive demand in agents which use 10 to 100x more tokens than just like typing to a chat if we don't have cheaper tokens, right? Cheaper models. And so, open source models are allowing companies to do a lot more with AI, which is going to unlock a whole new kind of realm of use cases and efficiencies within companies. So, I think that's a good thing. Now, I don't think that kills OpenAI and Anthropic, the frontier models, which is a lot of what the sell-off is, is, you know, Oracle's down however much, I don't even know, like 50% or something, because a lot of their backlog comes from OpenAI. And so, what people are thinking is, well, if open source models are cheaper, then who's going to use OpenAI anymore? And everyone's just going to use open source. And so, um you know, Oracle and everyone that's related to OpenAI in some way is cooked. I think what's going to happen is you're going to continue to use frontier models, whether it's because um of the models themselves, cuz they're smarter, or because they're in the US and that's just the way that your company works and your regulation works, uh or the way your data and privacy work, um or because the UX is better, right? Like, go try and use Kimiko 3 right now. You can't. Like, 90% of people listening to this podcast cannot and will not use it ever in their lifetime, right? So, I think what's going to happen is with open source, this is great because you're unlocking a bunch of new use cases, and we're just accelerating the use of AI in general. And so, that means we're going to have more use out of OpenAI and Anthropic, and more use out of open source. Now, OpenAI and Anthropic aren't going to do a 50x again like they did in the last 6 months. But, that was never going to happen. No one even expected them to do a 50x in the first place, right? Like, those that growth they had was insane. So, the market thinks it's black and white, and it's it's both. Both are going to grow together. It's a grayer area. And so, I think ultimately what that means is the application layer is going to do extremely well, because they now have a choice of of intelligence at different rates, at different costs, right? At different UX. And so, that's going to be really good for the application layer. And then, the infrastructure build-out diversifies their client base. It's no longer just OpenAI or Anthropic. It's now can be thousands of companies that all want their own servers and their own GPUs, their own memory, etc. So, I think this is a really good thing for the market and why I think we're going to go into an everything bubble, not bubble, everything bull market uh to close out this year. >> So, on the open models, I I fully agree that is bullish for the broader AI infra build-out and I actually came to the conclusion for myself that I was getting this wrong on the hyperscalers. I was getting right that their multiples will be compressed because of free cash flow, but what I was also assuming is that the margins for the cloud business that they're offering will will not decline, but kind of stagnate over over the next couple of months, but that's not happening and that's not happening because of the open weight models because, as you said, it diversifies their uh their their client base basically and it gives them pricing power. And we saw the I think it was 36% operating margin of Google for their cloud business, which is a massive growth compared to to to Q1 also. And um we also saw the post of actually Jensen Huang has now an X account and his very first post was on open weight models and he argued how important they are for competitors and and and and for the Americans, blah blah blah. He's basically speaking his own book uh to to to sell more GPUs, but the core takeaway for investors is that it broadens the AI infrastructure build-out and is also very strong argument of why buying the dip on such a bad day as today is could be attractive. >> Okay. Guys, can you just Melvin, as you answer, can you just remind people listening what the open weight models are compared to just the regular models? >> Yeah, you can think of open weight models as basically a downloadable software, right? You download it, you can plug into anything and just run it. Right, sure an average person is not going to be able to do that because you need GPUs, you need cooling systems and everything and it costs like 20 30 grand to run that. So, you use inference like companies like Nevias and Coreweave actually Communicate 3 is actually on Nevias as of yesterday. So, extremely bullish for you know Nevias cuz there's going to be a ton of demand for Communicate 3 through that. So, that's what basically openly model is. You can download it and plug it anywhere. Um and back to the conversation we had is who benefits from this? I we always said this Nevias, Coreweave, the Neo clouds of the world will benefit greatly greatly from this. Memory will benefit greatly because now inference is you know before inference was centralized at like three labs, right? You have Open AI, Anthropic, another one. And now that we move into the open source open source models, you're going to need massive amounts of more inference. Therefore, you need more memory than ever before. So, this is a great great thing for memory as well. This is also really good for AMD as well. Open source models run on whatever hardware is the cheapest, right? And most available and AMD's power with that. They had their big release or big release last week. Um AMD's looking really bullish right now right now. If I if I had some cash left over, I would be buying AMD right now, but unfortunately I don't. Um and also another interesting thing is like if you look at like companies like Broadcom or Marvell who you know designs these custom AI chips, right? Like companies start like these hyperscalers and all these companies eventually start on GPU hours, right? Then they then they make their own infrastructure and eventually they want to make their own chip designed for their own workload because they want more margins, right? Like I I see a world where as companies get bigger and bigger, maybe they want more custom chips to control their margins. So some companies like Broadcom and Marvell could benefit greatly from this as well. >> So one foot up why why bullish AMD? >> Why am I bullish on AMD? Because there's because everybody wants this this conversation we had before. Everybody wants way out of Nvidia. They want they don't want to be dependent on Jensen. Jensen has huge huge pricing power. If you're I mean sure Jensen has the best best like GPUs, but AMD's pretty comparable with their software layer as well. And and Lisa Su is a excellent excellent CEO. I think she's executing at a really good level. So yeah, I remain extremely bullish on AMD because of all that reasons. Cuz people want not just dependent on Nvidia. They want other other GPUs. And yeah, so AMD is the second beneficiary of that. >> Okay. So I want to have your guys' opinion on the market saying that we're selling off the hyperscalers and the others because of CapEx. Like I found this surprisingly stupid >> [gasps] >> by the market to sell off Google because of the fact that this was the first quarter I think ever or at least the last couple of years where they had negative free cash flow. But at the same time they proved that they're investing this money into something that yields great return with massive backlog. And the way we're arguing this is that the computer that is being built is not slowing down. So how do you view this and is there something that I'm missing on that? >> I also think it's a mistake by the market. I think again, I think it's my My is it's the market is concerned that we're going to go into a rate hike cycle. >> Mhm. >> And you don't want to be investing in companies with no free cash flow during a rate hike cycle, especially if they're starting to take out leverage or debt in terms of what they're trying to finance. I don't think we're going to a rate hike cycle. I My hope is the market's going to figure that out at some point. Maybe it's going to be Wednesday, so by the time everyone's listening to this, but maybe not because also Warsh is not really going to give a lot of guidance. This whole kind of thing that he said is he's not going to give any guidance. He's just going to give his reason for doing what he does and then not talk about the future at all for the most part. So, um it's kind of up to the market to decide here and I think the market is for some reason thinking that. Um and and I think that's been the big change here. To me, it doesn't make any sense cuz I think there's still, you know, there's obviously so much demand. Um I guess the other part is I do think the market thinks the demand is isolated to kind of the two big players, Anthropic and OpenAI, and I think the market believes that their revenues are I don't know if the market thinks the revenues are going to zero or these companies are dead or what, but it kind of feels like that and I just think that's so stupid because people forget that OpenAI and Anthropic are the two most successful companies literally in the history, right? They're the fastest-growing companies ever. And this year, this last 6 months was the moment that they became the fastest-growing companies ever, and especially Anthropic. And like those companies aren't just going to go away tomorrow because open source. You know what I mean? Like they they have hundreds of millions to billions of users. They're already at What is Anthropic like 77 billion in ARR? OpenAI's even more than OpenAI's growing a ton over the last month. So, like these companies are good companies and they're going to have the ability to either move up the stack to the application layer or even down to the infrastructure layer. Um and and create good UX around whatever the thing is they're building. So, if models start to get eaten away, number one, I think those revenues still grow. They just don't grow as fast as they've been growing, but they can turn on ads. They can launch other products. Like there's so many things that they can do to build a good successful business. So, I don't think the concern around OpenAI and Anthropic makes sense. So, those are the two reasons I think that the market is is just so scared and why they're kind of hurting any CapEx, which is why like I mean, Tesla sold off and, you know, Oracle and all these companies and if you are spending this is why I think neo clouds are getting as as crushed as they are. They're great businesses. They're obviously what everyone's going to need, but again, they've got to finance them. So, the data center companies, Galaxy, same thing. You know, this thing is obviously going to spit revenues, but the market doesn't care cuz they have to spend money to get there. And the market is hating that right now. >> But, so the question is, when is this going to change? And if we assume that all of us are right that Google and all the other hyperscalers are are correct in spending that CapEx and the market was wrong, when is the market going to recognize that? And is it the right time now to buy that? Or do we need to hold off a couple of quarters, maybe until even 2028, when kind of the inflection point is in the in the forecasted free cash flow of these companies, when the free cash flow will actually grow again? >> Real-world assets like funds, treasuries, and private credit are still running on rails built decades ago. Gated, paperwork-heavy, slow to settle. Everyone's talking about tokenizing them, but far fewer can actually do it and do it without cutting regulatory corners. Securitize can. It's the SEC-regulated infrastructure bringing real-world assets on chain. Nine years in, native tokenization, not wrapped, backed by BlackRock, Morgan Stanley, and Cathie Wood's Ark Invest, and chosen by the New York Stock Exchange, VanEck, BNY, and Apollo to do it at scale. It's the regulated bridge between traditional finance and crypto. Tokenize the world at milkroad.com/securitize. >> Let me actually give you a bearish outlook on Google. Uh >> because the think the main reason Google sold off is because, first of all, they spent $70 billion on R&D, um, and they're lagging behind behind open source models when it comes to their own models. So, like how how does that work? I mean, I can see a world where like I if I was an investor, I would not be not happy to see like Google not being top three. Not they're not even top three, you know, it's uh OpenAI, it's Anthropic, and now it's Claude 3, right? How do you as a company spend that much money and you're lagging behind? I think that's like the bearish case I can give on Google, to be honest with you. And I I And I I don't I don't think they're going to catch up. I I don't think Google has a shot, um, at catching up anymore. >> The thing is that I don't even care about the model. >> It doesn't matter. >> Yeah. Who cares if it's a little dumber than the other model? It really does not matter at all. >> using Gemini? >> Yes. >> No. >> You You never use Gemini when you've used Google Search or in your Google Docs? Like you all use Gemini. You ever got a transcript from YouTube? I use Gemini so There's a billion people that use Gemini. It's the most used AI in the world. More than OpenAI. >> why search grew. Google search grew mainly because of the built-in AI. >> They don't They don't need the number one model. They They're fine with number four and number five or number six because they have the distribution. They have every single app that we all use. They know how to crawl the internet better, so they're going to create better agents than anyone. They have everything they need to dominate AI, and they do not need to be the number one to dominate. I think that's the market kind of missing that. >> Well, you still want the ROI to come in, right? Like from these models, right? Like ChatGPT makes mainly their money from the subscriptions, right? Like if you're not paying for Gemini as much, >> Yeah, but Google doesn't need that, right? Like Google can make it in other ways. They can make it from ads, they can make it from cloud, they can make it from I don't know what else they'll do with Gemini. Some sort of agents, they can make it through their phones, through their browser. Like they have so many ways to make money from this. >> So Kyle, how do you think about the timing issue of with with this the cup expanders? >> That I don't know the answer to. I would say historically the hyperscalers are cheaper than they've been in, I don't know what the number is, like 10 years or something. So if you just look at a valuation standpoint, I don't really care about the timing that much. I just continue to dollar cost average in them because it's, you know, whether it's this quarter or next quarter or even the quarter after, you know that these companies are going to dominate the world in the coming years with AI. Like they're they're they're literally a part of all of it, whether it's like Google is part of the infrastructure layer, the model layer, and the application layer. You know what I mean? So it's going to dominate at some point. You've seen the chart of their free cash flow is going to go through the roof. >> Yeah. >> I think it was starting in 2028, 2029, whatever. So like at some point they're going to grow like crazy. And so these current valuations of them are are multiples are are are cheap. So I think what the market needs to see though is earnings from companies that are growing because of AI that are not part of the infrastructure build out. So we've seen earnings of like Micron and Nvidias and all these like just rip all year. >> Yeah. >> But what they need is non-infrastructure companies to show that their earnings are growing. And ServiceNow was the first one last week that I think that was a good tell that like okay, their business is growing because of AI. The market got that wrong over the last year. ServiceNow sold off 50% and they've been saying guys, we're growing because of AI, not the other way around. And the market didn't believe it. And now, and I don't think the market still fully got it cuz it I think it was down on the day that it did its earnings, it's now up over the last week. But the market's going okay, are we really going to get earnings from AI now? And I I don't know if we'll see it this earnings this quarter, but probably next quarter we'll see it. And that's the market goes okay, wait, there is real demand here for all models because everyone's growing. And I think that's going to happen at some point this year. >> Yeah. Yeah, sorry, Go ahead, Mel. >> No, I was just going to say there's some interesting data that came out saying that S&P's 500's net profit margin just hit 15.7 for a quarter two 2026. That's the highest reading going back to 2009. And this is the 10th straight quarterly earnings or quarterly increase in a row. So, like you have profits coming in, but you know, this is largely from the Mac 7. Like Kyle said, we do need to see a larger larger increase in the overall S&P 500. So. >> Yeah, but so if you look at the earnings of Corning and and Amkor today, for instance, they're both down 15% because guidance was lacking. But the one point that stood out to me was the margin and the earnings per share that they were actually growing massively. And that to me is coming mainly from they have better pricing power. So, everything that they're selling into the AI space is just more more expensive. And we saw the same with with ServiceNow. In in in the earnings call, they were saying the customers that are upgrading to the kind of a gentle AI package are paying between 20 to 30% more because the value that they're getting is so big. So, I think the volume is rather the issue and pricing and earning is really really bullish and and positive. And if you if you like look out the next couple of years for a name like Corning, for instance, if we if we fast forward to 2027, 2028 when when we have Rubin ultra, there will be much more volume coming in for fiber for Corning, then they can keep this price or even further increase price while meeting higher kind of volume that they're selling. So, this actually made me really bullish through on and this is also tying into what you were saying, Melvin, showing to be that higher earnings across the board. >> Yeah, this chart shows it here, um which is, you know, the best quarter that we've had since 2021. >> Yeah, exactly. That's just the argument. >> Right, 15.7. And I think this continues. Like I I I really think that companies that use AI are going to grow their earnings and their profit margins and all this uh for I don't know how much longer, but for a very long time to come. That is what this this technical technological revolution is going to do for companies that we've been saying. And look, Q1 was mainly, you know, my it was the bottleneck stuff. That's why their earnings jumped so much, but I think it's going to continue to go because it's not just the infrastructure anymore that is having their earnings that are growing. It is other companies. And I think we're going to the market's going to start to realize that. And that's when I assume the infrastructure trade, the hyperscalers will all grow again. Uh that's why I think we're going to go to the everything bull market. >> So is that Do you So do you see a temporary rotation then, Kyle? Like to just to the Appler, which you've been talking about for a long time, but is that Are we Is that like Do you see a a pause in the ripping of the bottlenecks and a move to the companies that are using AI to actually make money, actually using this stuff to make money now, and showing that? >> That's already happening. Like if you look over the last, what has it been now, month and a half that the infra stocks have gone down? You know, Apple's up like 20%. You've got ServiceNow and and Salesforce up a bunch. You've got Eli Lilly up like a bunch of these like non-high-cap-ex companies that are using AI that are improving their businesses. Um they have been growing during this whole market sell-off. So this has not been a like broad sell-off. It's been a an infrastructure sell-off, right? Um and a hyperscaler sell-off. So I think that's already kind of playing out, but I I don't think that it's just like infrastructure stays down and these just go up and that's it. I think they're both going to go up together. At some point. Again, I don't know exactly when the infrastructure trade comes back to life. I think they're both going to go up together though going into the end of this year. >> Mhm. >> So I think you you almost want like a barbell, right? You want to own the infrastructure, you want to own the application layer, you just don't want to own the models. >> Yeah. Yeah, >> [laughter] >> I'll die to agree. The models I would not own. >> So is that So is that is that super bearish for for the potential IPOs this year, right? Cuz this We haven't talked about that in so long since SpaceX happened, right? But this is supposed to be the big the three big IPO year. Are they going to delay? >> They are delaying constantly, right? They were supposed to come out already at least opening eye, right? So I I think that they're realizing that market sentiment at the moment is that the models are commodity and that's not a good good time to come out, especially when Jensen Huang and the others are signing letters that, "Hey, we need more open weight models and to to protect data sovereignty, etc." So >> Yeah. >> Actually, a really hot take is I don't think that any of the models will IPO in 2026. >> I don't think that's a hot take. >> Yeah, I could I could agree with that. I could see that. Um but I guess Okay, maybe this is a hot take. I don't So I just think that the the reason they're not IPOing is cuz the valuation was already getting too high. It was like What was it? $2 trillion or whatever for these companies. I don't think that they're dead companies or they're not going to be good companies. I'm actually still bullish on these companies, but not because of them selling tokens from a model perspective, it's cuz of them doing something different, being an application layer of AI, right? So um do I think they're worth $2 trillion? Absolutely not. Do I think their revenues will keep growing from here and not go backwards? 100%. I think their revenues will keep growing. I think the market thinks the opposite. Um so I'm still bullish on them. I'm just not bullish at a $2 trillion valuation, which thankfully they're not going to IPO at, it appears. >> I think this is this is what what's likely going to happen. I think Sam Altman, um several others have been teasing us for the last like four or five months that there's a some kind of architectural breakthrough that happened in the model layer. But I think what what what they will likely do is release that, get a big boost from that, and then IPO afterwards. And I I agree with Kyle with what he said about, you know, Anthropic and OpenAI. I don't think they're cooked. At some point all these models, you know, whether you look at any model, they're going to, you know, come together, right? They're all going to catch up eventually. But with with going to make them different is the compute. You know, who has the compute to scale up, right? That's OpenAI and Anthropic. And who has the money and the funding? That's those two model those two companies. You can't really You can't really, you know, at a certain point you can't really break that, you know, if you're a smaller company. Yeah, sure can we get three is going to catch up, but can they catch up when OpenAI goes to couple layers, you know, couple higher up in the algorithm algorithmic changes, you know? >> And and don't forget they're now making what? 60, 70 billion dollars a year. That's a lot of money, right? They're now well-positioned to continue to get a moat in in compute. And it was someone from Was it Microsoft? I was reading this out of my live show that I did yesterday. I think it was a guy from Microsoft talking about how the biggest moat in AI now is compute. And the companies that can finance this and can get the most compute, they're the ones that have the the moat because they can deploy the most intelligence. It's all it is, right? Like the whole If you were to sum up what we're doing here over the next five years, 10 years, even longer probably, it's we're just converting energy into intelligence. And the way you do that is GPUs and compute, right? And now obviously you need data centers and you need energy and you need all the little things that we talk about to make it all happen, but like ultimately it's whoever can be the best in the world at converting energy into intelligence and then like those are the ones that are going to win. And no one has done better than that right now than OpenAI and Anthropic. So these businesses are not dead. They [ __ ] know what they're doing, you know? >> Speaking [snorts] of not being dead, Kyle, >> [laughter] >> give me your view on the Tesla Q2 earnings. >> Yeah, jeez, we're going to do Tesla. >> my god. >> Yeah, um so it was basically everything I expected uh except for that big of a sell-off as a result. I did say it was probably going to sell off. It was going to be its best Q2 ever. Uh it's one of its best quarters ever in terms of revenue. All those things happened. Uh but the market wasn't going to love it because they're raising CapEx. Uh they don't have a ton of robotaxis yet. Um and the valuation is obviously high. So, all that played out. I I The thing I didn't love from it is they didn't give a ton of guidance around robotaxis. I mean, they did, um but it wasn't clear enough. Um they basically said it's going to be slow because they want to make sure it's safe. Uh and they don't want to have any accidents, which is like is the right move. Just kind of sucks cuz, you know, you'd love to see it just deploy faster. Now, are they doing that on purpose to try to like, you know, bring expectations down uh and then they can finally start to, you know, produce more than what they say? I I've no idea. Um the one thing that was like somewhat exciting that I didn't expect from it was um Elon said it and someone else, I think their CFO or something, was people are coming in and buying FSD, not a car. >> Yeah. >> And they're just they're they're getting a car with FSD. And this is something I've been saying for a long time because that's why I bought Tesla. I did not care about the car at all. I cared about not having to drive. And since I've used it, everyone I've talked to, multiple of my friends have bought a Tesla since driving in my car. And they'd never like Tesla and they don't like Elon. But they did it because they're like, holy [ __ ] I don't have to drive my car. So, I was always like, why don't they just run a Super Bowl ad to show people that you don't have to drive your car anymore? And they don't do it for whatever reason, but according to this earnings, and I don't know how much like they didn't quantify it, they just said it, so I don't know how much weight to put behind it. But if that's true, this whole like um device upgrade cycle that I talk about with Apple all the time, that could happen with cars through Tesla. And if that happens, that's actually extremely bullish. And I've talked about how I don't care about the deliveries, this would actually be the one thing that changes that because Tesla could go on a whole 'nother level of the amount of deliveries it has because people just want to buy an FSD, they don't even care about the car itself. And so they did talk about their ramping up all their production. So I don't know. We'll we'll see if next earnings they see a huge jump in deliveries again, and again in those being because of FSD, I think that's a huge huge signal for Tesla. If that doesn't happen, then you know, we're just going to wait for robot taxis and that's probably going to be a next year story. Late next year. So unfortunately. >> I agree that it was really bullish that they were saying people are buying cars because of FSD. Though I think the majority of the cars that were sold was just mainly because of lower prices that they were charging, right? And you saw that in the margin and the lower the lower EPS that came in, right? So to me there was no sign at all of when robot taxis will be really launching. I I kind of viewed it viewed it as more of kind of from a from a bearish perspective because why would you send back? Doesn't make sense in their position. I did not sell my position. I just think that if you're an investor in Tesla, you need to have kind of a long breath here until the the the the physically our revenue really materializes. One thing and that kind of does the the the segue into your world, Melvin, was that Elon actually kind of stopped in between the presentation saying, "Yeah, we're thanking Micron for an allocation in memory." >> [snorts] >> I was like, "What?" >> Yeah, I was like, "I got to go buy more. I got [laughter] to go buy more." >> And so Melvin, I wanted to ask you maybe you can kind of talk me through what is happening in the memory world in China and what that company is, how big they are, and kind of what that means for for for the whole market. >> Everyone's tokenizing stocks these days, but almost nobody's doing it right. Thin liquidity, prices that drift from the real thing, dividends that just vanish. Bitget Stocks 2.0 is different. Real Nasdaq and New York Stock Exchange depth through licensed brokers. Prices mapped one-to-one, dividends paid to your account in real time, plus you get the lowest fees in the market at just .04%, and you can trade them like any other crypto as margin, in earn, in grid trading. Tokenized stocks, finally done right. Head to millroad.com/bitget to get started. >> Perfect. I got you. I've been waiting for this. So, >> [laughter] >> So, CXMT actually IPO'd yesterday. If you don't know what that is, it's a Chinese memory maker, and the stock actually surged 466% on its first day of trading. And at one point, it became the most valuable company in China at around 550 billion in market cap. And to funny enough, it was 212 over subscribed. And I have never seen an IPO like over subscribed this much. This is by far the highest over subscribed IPO there is. So, now yesterday, you see the reactions on X, oh no, the Chinese company goes public with a massive pop, that means over supply is coming. This must be bad for Micron and SK Hynix. Oh, no. And this is completely stupid stupid because CXMT is the fourth largest DRAM producer in the world with about 7.7 market share. The the gap between those three is not just market share, it's their fundamental technology. Now, what do I mean by this? So, we all know about ASML, right? The the company that makes EUV, um which stands for extreme ultraviolet lithography tool. Um it's essentially a machine that prints uh the most advanced circuit patterns um into chips. And uh and Samsung uses like for example, Samsung uses up to seven layers of EUV in their um in their edge memory and SK Hynix uses five. CXMT has zero. So, without EUV, uh you have to do something called multi-patterning, which is essentially you split one print um into four, three, four, five different steps. Um it technically works, but it's slower, more expensive, um and it introduces a whole bunch of errors. Um so, the the what I'm trying to get at is is CXMT is not going to be able to compete at their memory is going to cost higher per gigabyte compared to like Micron or SK Hynix and Samsung. And and on HBM, which is high bandwidth memory, um that goes inside like Nvidia GPUs, right? Um and CXMT can't even compete because HPM actually requires a logical die logical die manufacturer like extreme like advanced nodes. And Samsung makes theirs at 4 nanometers. SK Hynix uses TSMC and CM- CXMT does not have any of that. So, this is this is stupid. They cannot even make HBM and somehow people think they're going to oversupply the market with cheap memory and that's all going to, you know, kill SK Hynix and um you know, Micron, which they don't even care about the market that CXMT sells to, which is like the your chips and the iPhone chips and all that, which is like lower margins. And CXMT doesn't even have enough to capac- to fulfill their capacity needs for China. So, this this this makes no difference for Micron or SK Hynix. So, um I I'm extremely extremely bullish on this. >> What What this is, though, is China building their sovereign AI from end to end. They have not Yes, it's technology-wise not there, low market share, quality-wise not blah blah blah. They have the energy, they have memory now, they have something that they can use to produce the chips. They have models, and they have the state financing everything. Um and this goes back to the entire sovereign AI thesis. We also saw a lot of deals announced this and last week around South Korea, with SK Hynix, with Nvidia building stuff there. I think this sovereign AI thesis is really bullish again for the broader infrastructure builder, as all of that all of the bigger kind of regions, US, China, UAE or Middle East, and then eventually Europe, want to be independent. And that just means more infrastructure. And that's I think the that the real important kind of point in the whole China China story. >> What do you think is going to happen first? We will have a colony on Mars, or Europe will have their sovereign AI? >> [laughter] >> We go up by SpaceX before buying your >> SpaceX is worth a trillion dollars. >> [laughter] >> Oh my god. >> Actually, there's I mean, there is one big thing in in in Europe. >> Well, big in for European kind of stakes, I'd say. >> Wine and olive oil? >> Yeah. >> [laughter] >> Yeah, that's big as well. No, but there's there's Schwarz Digits, and they're building for a couple of billions of euros their their own data center in Germany. And then you have Mistral in France. But the the Schwarz Digits news is is just coming out and they're they're actually a massive player kind of connected with all governments and so on in Europe. So, there's something going on. Space colony on Mars maybe maybe earlier though. >> Yeah, that project will be done in 100 years. >> [laughter] >> What are you guys buying? We got to wrap up and I know LG's about to ask so I'm just jumping in for you. >> I was I was I going to ask you if you're afraid of the earnings coming up in the next couple days but also what are you buying is also good cuz it's not it's not just potential rates tomorrow. It's all earnings like all the all the hyperscalers have earnings the next couple days or most of them. >> That's a good question. >> you can answer either question you guys want. >> I mean I I think we're going to have an incredible earnings. I don't know if the market's going to care about that but I think you know, earnings are going to go through the roof on all most companies and obviously not all there's a bunch of dead companies but any company that's part of the AI development or building with AI or building some sort of application for AI like I think they're all growing and so that's the good news. I do think CapEx is going to increase for Microsoft, for Amazon and for Meta. I'm assuming the market's going to punish that unless we lower rates tomorrow which would be insane not going to happen but like imagine that happened. So, I imagine they're not going to like that. I don't know what's going to change that story but I think it's going to be great earnings. >> It's a crazy earnings week. I'm really looking for actually tomorrow Wednesday because I think Meta is there. Is Amazon there as well? Not sure. >> Yeah. >> Um and then Hood as well but especially for the hyperscalers I'm looking for are they also like Google increasing the margins on the cloud business? That would make me and increasing CapEx and if the market sells them off, I might be buying Google here. Uh so to answer both questions actually. Doing also deeper dive on Google this week on on on the pod and today's Bloom Energy as well, which obviously is is important for me after the bell. I will cover that as well on the on the platform. >> And Teradyne, I think as well, Vincent. That's today. >> yeah. >> Teradyne also the subject of a recent Vincent Vincent analyst show. Uh Melvin, how about you, man? >> I'm just going to double I I have, you know, so little cash at the moment, >> [laughter] >> but I'm going to double down on um on my core positions, which are Navios, Corus Weave, Marvell. Um Uh uh analyst payday is coming up for uh Milgrom, so we should get, you know, 500 or hopefully 1,000 bucks this time around to buy buy more. So, I'm looking forward to buying some more then, too. >> Wow. >> So, you're not diversifying into any kind of adoption? Interesting. >> I am sticking with my semiconductor trades. >> Digital assets are performing well, by the way, versus semi. >> [laughter] >> Yeah, John was making a fuss about that this morning. >> Oh, god. Oh, god. That's John Gillan's music. Here he comes. Actually, it's going to be a lot >> It's going to be a lot more fun when we have a broader bull market rather than just everything bull market or everything like in one thing bull market. So, I think uh we're going to have a great Q3 Q4 uh and uh make sure everyone come check out our portfolio so you can see what everyone's buying. >> Here's Here's Melvin this week. He's sweating. [laughter] He's sweating sweating bullets. Wait, hold on. I found I found one of Kyle's, too. This is Kyle in the future. >> [laughter] >> Wait, drinking his Jack Daniel's driving his open top >> What do you mean? >> Tesla. Well, well, that This is Kyle right now. This is Kyle. >> [laughter] >> That's Kyle. >> I love it, LG. Love it. >> A lot of semiconductors needed for that world to uh >> [laughter] >> Power semis developed by Infineon. >> Plus Navios. >> All right, boys. Oh, wait, one last one. Tesla. Tesla kissed like 107 yesterday or today. Anybody interested? No, Tesla sorry, SpaceX SpaceX sorry. SpaceX SpaceX SpaceX kissed like 106 107 this morning. Back up little surge back up. Anybody interested? Any buyers? Anybody? >> Don't remind me bro. >> Don't remind me. >> I'm down. >> It's It is interesting cuz it's down like whatever it is 50% or something for at least from its highs. The problem is is that it fits right in the entire thing that's getting sold off which is anything to do with CapEx. And SpaceX is definitely going to continue to raise CapEx cuz they have to do it for everything they're building. So, I'm extremely bullish. I said I was going to buy at 115 which is exactly where it is right now, but I just think I don't need to rush into that. So, I haven't bought it yet. >> Well, if you want to see if Cal ever does buy that or any of the moves these guys just discussed including Vincent potentially scooping some Google or you make sure you guys are in Milk Road Pro just a dollar at the link below to test it out. See their reports. Guys, otherwise great episode. Thank you for all the thoughts and we'll see you next week. >> Want to stay ahead of the biggest technological [music] shift in history? Subscribe now to get insights straight from the sharpest minds in tech and finance. Quickly you'll note this show's for educational purposes [music] only. Nothing here is financial advice. Investing always carries risk. Never invest more than you can afford to lose. >> [music] >> Thanks for tuning in. See you in the next one.