WARNING: The AI Stock IMPLOSION..

Is the AI stock crash over with or is it just beginning? In this video, we will answer exactly that question and what areas of the AI trade are more vulnerable for additional downside. But first things first, I want to share this image with you. This is Morgan Stanley's momentum ETF. And just so far in the month of July, this is down almost 36% in 18 days. less trading days than that. This is the most violent downside move we have seen ever before. This is more violent than when the.com bubble burst. This is more violent than 0708. This is more violent than the 2020 pandemic crash. This is on a whole new level. Sentiment has completely flip-flopped. But it's not even really just sentiment itself because most investors out there that, let's say, have went on CNBC in the past week or so, they're still just as bullish as they were on AI stocks a month ago. They're reiterating a lot of the same arguments that they've made for a year or so. So, I would actually argue that sentiment has not changed really at all. just the price action has changed which is very unusual because normally price action follows sentiment and really what that highlights to me is there is a heavy confirmation bias in this market right now to the bull side. So I actually wrote this post on X. If you guys are not following me over there, go ahead and uh follow me on X. It it is the real TCI. I wrote this I want to share it with you briefly. It says, "I am neither a perma bear or permab bull. I sift through the noise to the facts and come to my own conclusions, many of which tend to be unpopular until they are proven correct more often than not. This is because I can spot turning moments faster than others. Not because I'm smarter, but because I don't approach the markets with a heavy bias. Humans from a psychological level place heavier importance on things that confirm existing beliefs. This is called a confirmation bias. Often causing perfectly smart investors to miss things. You can manage a hundred billion dollar portfolio and still lean into your confirmation bias and thus completely miss the ball. I've been warning about AI stocks ever since Microsoft, Google, and Amazon begun selling off for spending more. If Wall Street punishes the spenders for spending more, the spending will slow. It's like Isaac Newton's law of gravity. It works every time. When Wall Street says enough is enough, the timer has already begun towards the end or the slowdown of the spending. I think China is showing that alongside Enthropic, you don't have to spend like you're drunk in order to win in AI. You have to spend better. So even though I only expect capex only expect capex to go from 700 billion to 900 billion next year, you might deploy more compute than that. So, am I bearish on the markets? No, I'm really bullish. I'm just not bullish on the stocks everyone is bullish on. Have AI stocks bottomed? Probably not. But confirmation bias is strong and I do expect aggressive bounces and declines over the next two weeks. Sincerely, an AI hyperbol. And I really want to highlight this because have AI stocks bottomed? That is a broad thing to say. Some of them may be in the process of bottoming. Some of them might have more downside to come. And I think we are no longer in an environment where you can just paint all AI stocks with the same brush. I think the semiconductors, the actual like GPU and CPU makers or designers technically you should say are in a completely different category than a memory stock or a data center stock for that matter. They're all in their own lanes. And I want to break them down for you right here and right now. And here's the problem with AI stocks in the simplest form possible. The spenders are being no longer rewarded for spending. Now, I would argue that hyperscalers are already showing you return on investment from AI. Their numbers are at all-time highs. The problem is not their numbers. The problem is the spending. And if Wall Street gets the signal over the next couple of weeks that spending will continue to be out of control, yeah, AI stocks are going to bottom a lot faster potentially. hyperscalers are going to sell off. The only thing that matters here is do hyperscalers sell off on higher spending guidance. If that happens, yeah, sure, you're going to have a short-term rally in AI stocks. But the problem is once the spenders start to get punished for spending, you've started the stop clock towards the end of the spending cycle. And we have had other factors in the last couple of months that have really driven a lot of capital and pushed up expectations in AI stocks. One example is the Iran war. I made this example in one of the last videos, but would you rather own Delta Airlines or Nvidia as oil goes, you know, over $100 a barrel? You'd clearly rather own Nvidia. And there's a lot of funds out there like Kathy Wood for an example that has to be invested at all times. They can't hedge the market. They can't short the market. They have to be invested. So I think part of the reason you've seen this vertical move for really all of AI, all of AI stocks is because partially the macroeconomic conditions supported crowding into the only bull trade on Wall Street. From a tactical perspective, it makes sense why AI stocks have done so well in the last couple of months. But here's the real problem. If hyperscalers are going to be punished for spending more, they will inevitably slow down. Why? Because if Microsoft stock falls 50% from highs, which is already down like 30% from highs, but if it continues to fall, if they continue to spend, it's going to be harder and harder to raise tens of billions of dollars in the credit market. People don't want to lend to stocks in which their stock is in an active freefall. That's pretty common sensical. Nobody gives a [ __ ] about the long-term AI story for hyperscalers if their stock is actively declining because they are spending money. People are going to say, "Yeah, I don't want to be involved in that." That's what happened in 1999 and 2000. Companies, they had to go out and raise capital. But the sentiment shifted, the stocks got sold off. the capital was there, just people didn't want to lend it because their stocks were in freef fall. You're seeing a mini example of this starting to play out today. And the real problem is at the end of the day, we're going to spend $700 billion roughly in capex this year. It's probably going to come in at like 750 billion. Well, Wall Street is like, "Oh my gosh, yeah, the the the the AI capex trade has years left." and they've pushed expectations well over a trillion dollars of spending for next year. But if hyperscalers are getting punished for spending, that's impossible to happen. It's like Isaac Newton's law of gravity. It is impossible. Don't fool yourself. You know, does Microsoft want to spend $500 billion next year on capex? Of course they do. Can they? No. Because simply the capital will not be there. You know, institutional investors, fund managers, they only care about their clients. Not even their clients. They care about beating the markets. They care about not losing their clients. If you're a hedge fund manager and and you love Microsoft, but Microsoft's down 50%. And you're going to lose clients if you give $10 billion to Microsoft. Are you going to give $10 billion to Microsoft? Are you going to personally take a a hit because Microsoft wants money? No. You know, people will start to pull their money out of funds that are giving hyperscalers endless amounts of of their capital, right? It's capital preserv it's self-preservation at its core. It's human nature. It's psychology. So, it's impossible to continue to spend like you're drunk when Wall Street is actively punishing it. So the problem is not that the capex trade is going to end. No, absolutely not. The capex trade will continue to grow, but it's not going to grow at the same speed that Wall Street was expecting a couple of months ago or a month ago for that matter. And simply put, if Wall Street's expecting 1.2 trillion of spending next year and we get 900 billion in spending next year, that's obviously a big disappointment. But you have to change the expectations going out for years, right? If we spend two 300 billion less than Wall Street's expecting for next year on capex, the year after that, Wall Street's at like 1.8 trillion, that's got to come down to like 1.2 trillion. So you you you really mess up the expectation that Wall Street had for a lot of these AI stocks and the spending that they were going to see not just for next year but for years down the line. That changes the valuation calculus for AI stocks at its core. And that is why AI stocks have begun to sell off as much as they have. Now, I actually think it's kind of an interesting scenario because, you know, why is the capex so high? A lot of it has to do with memory cost and server cost and some of these, you know, costs that have went up. If costs come down for something like memory chips, you could actually see more data centers built with less money. So, it's not even that I think there's going to be less data centers built. I just think they're going to be built at a lower cost because I think cost will come down. After all, if we're expecting massive demand, you know, that that's higher than, you know, expected, companies are going to keep prices higher like Micron, you know, they're not going to be able to keep prices as high as they are if capex comes in lower than expected over the next 12 months. So, breaking down each sector of this AI trade, I think you have like lowquality, high quality, midquality. Low quality would be the memory stocks right now because it's not because they're they're selling a lot more, right? They're just raising prices. So, inevitably, that's a supply and demand issue, right? That's a expectations issue. If you run a business and you think you're going to build a 100 homes next year and you have the capacity to build 90 homes, you're going to be able to charge more. If oh [ __ ] next thing you know, 6 months later, you're like, "Oh my gosh, I'm only going to have 70 home sales and I have capacity to build 90 homes in a year." You're going to cut prices to try to get your capacity back up to full capacity. And I know that's kind of complicated. It's simplistic in the example that I gave, but that's partially why prices have went up so much. That is unsustainable. That's not going to last. So for a lot of these memory stocks like a SanDisk, you know, it's up three still like 2,000% in the last 12 months. That's unsustainable. Could SanDisk or Micron or some of these memory companies continue to sell off? Absolutely. And that's why their charts look the worst. you know these stocks are down 25 to 40% across the board and I do think there are other reasons why other stocks have done a lot better like in AMD for an example AMD coming after you know Nvidia's lunch per se AMD has held up a lot better than the rest of the AI trade I think in I think AMD is going to do a lot better than the rest of the AI trade I think Nvidia AMD um like a broadcom right there's going to be a lot of demand for GPUs. Still, the demand is going to go up for GPUs as we get to mass enterprise adoption between 2028 and 2030. I'm not as concerned about the semiconductors here, right? Sure. Could they come down a little bit more if capex doesn't go up as much as expected? Sure. But I think, you know, the semiconductors would be a very attractive area to actually be buying the dip. something that's more so based on not necessarily price, but selling more units. You know, Nvidia has been selling more units, they really haven't been raising prices on their existing units. And I think the share of capex, you know, I'm just going to make up make up a number here. If out of the $700 billion spent this year, if 25% of that is directly on GPUs, well, next year if the capex goes to 900 billion, I think the share of capex could be 30% GPUs or 35% GPUs if memory costs come down or construction costs come down a little bit. Some of those other factors that are driving up the capex number. So, I know little complicated to try to understand here. Even as I'm saying it, I'm like, "Yeah, this is a little complicated." Long story short, Nvidia, AMD, Broadcom, those would be like top tier buy the dip opportunities within the AI trade. And even if they do come down a little bit, I think they're going to bottom a lot sooner than others. There's nothing to say that Micron or SanDisk or Western Digital or, you know, some of those stocks have to bottom anytime soon. there there's no indications of that in fact and then we move into like the mid tier right some of these are going to be better than others it's not as uniformal uh Marll you know Marll at the right price phenomenal you know it's come down a lot the I think the data center connectors and all of that mix that with a theme of what I just said that GPUs could actually be in more supply ne or more demand next year I think something like Marll could do really Well, servers, you know, server prices might come under some pressure. Um, Dell has obviously held up a lot better. I think companies will look to build their own servers if prices continue to go up. Dell could be interesting at the right price. You know, 200 bucks for Dell if we do go through a larger downside move could be interesting. The data centers and neoclouds, I think the fact of the matter is companies want compute power wherever they can get it. Whether it's a Neil cloud or a independently built data center of you know meta whoever it is whoever has the compute is going to be in demand and I especially think you know 2028 through 2030 there's going to be a lot of demand there if you know expectations come down if nebius comes down to a hundred bucks that's a great buying opportunity in my opinion do I think it's you must go out and buy the stock to right now not you know you can be patient with these things. But I think like the data centers, the Marll's, the Qualcomms, and I would probably put Qualcomm in like the A tier category. Those would be um attractive as well. So to make this simple, I just put um these groups of stocks in A tier, Btier, and Ctier categories. Yes, I do think there can still be downside for all of these tiers across the board for the AI trade until you reset expectations lower, but again, some of them are going to do much better, bottom faster, and be better buying opportunities. So, within the A tier, you have Nvidia, AMD, Marll, Broadcom, Qualcomm. These are companies that are going to experience more demand even if capex goes down or or or doesn't go as high as expected. I think some of these stocks are more accurately priced and will actually see more demand next year, not less demand. B tier, you have data centers, AI industrials, AI financials. These are companies that make sense to buy at the right price. You have to understand that compute power or compute cost are expected to decline 90% per year for the next three years. So data centers, you really have to see them execute. You have to see them continue to grow. Assuming that happens, they are in the Btier category. They are interesting at the right price and and that's where I'm putting them. AI industrials like a caterpillar, right? Interesting at the right price has a big tailwind. It's kind of an AI stock within industrials. Um, but I think the consumer is going to get stronger. There's other reasons to own the AI industrials outside of AI. Uh, AI financials, you know, like a Morgan Stanley or Goldman Sachs, you know, I think they're going to continue to do well. There's going to be a need to finance debt and new IPOs coming over the next couple of years, but they're also a broader play on the consumer as well. I think they're Btier categories at the right price, they're good buys. Ctier, these would be like don't even bother buying the dip on them in my view. Something like photonix because of expectations. A lot of these photonix companies are up a lot. And yes, there's going to be a need to run data centers more efficiently, but there they would have to fall a lot to actually make them attractive investment opportunities. Other Ctier stocks would be memory. You know, something that is totally based on price and not on supply. That's not an investable strategy in the medium term. big catalyst for AI stocks coming next week is going to be Google earnings and then the following week you'll get your other hyperscalers and again it's not about the capex numbers that they announce next week or the week after it's about what is the general tone look like you could actually have you know really good capex numbers that go up a lot you could have capex guidance for this year that goes up a lot but then you could have on earnings calls Google CEO say yep we do see capex moderating next year. Just that one sentence is going to kill the AI trade. Is it possible that that happens? >> It is. I actually think for advantageous investors that are looking to find that next 5 10x stock, the application layer is really where the value will be created. And I could make a a 30 minute long video on this. You can't vibe code these software companies and make it economically make sense. Like I've shared this example on the channel before, but um Walmart, they don't disclose this, but most analysts think they spend between 10 and $30 million per year on a Service Now subscription. Could Walmart vibe code their own Service Now? Of course they could. It would cost between 40 and 60 million to vibe code that upfront. So if they spend $10 million a year, it would cost three or four years potentially or six years to make your return on that. But it would cost over $und00 million per year to fix bugs, to hire engineers, to maintain it, you know, an actual compute cost. So economically speaking, it makes zero sense to replace a service now or a UiPath or something and vibe code your own solution. It's like if you need to go to San Francisco or to Miami, you don't buy a plane to do that. You buy a plane ticket, right? because economically that makes more sense and that's a similar scenario here. I also think like in the real estate market you don't need to have a real estate agent in order to buy or sell a home but 91% of homes are bought and sold with real estate agents because there's a lot of risk there. There's um a lot of hassle there. there's there's a there's a problem and you know um understanding of the real estate market for people to go out and buy and sell homes themselves, right? There's escrow, there's all kinds of things that go into it. Well, for something like software, I think software is the real estate agent. It's actually the bridge to enterprise AI adoption. A lot of these companies, their demand is going to skyrocket as enterprises mass adopt AI. that happens between 2028 and 2030. That is when you go from, you know, 30% of companies today using AI for minimal tasks to 80% of companies using AI for over 50% of business operations. That's the mass critical mass enterprise adoption of AI moment 2028 through 2030. Wall Street's forward-looking, so they see things before they actually happen. They they position in them before they happen. I'm telling you this before Wall Street is positioning into them way before the numbers actually reflect this. And that is why some of these stocks are 5 10x or more opportunities from here. Zeta, Rubric, UiPath, Service Now Zcaler HubSpot MongoDB Snowflake, Data Dog, Back Blaze. These are companies I think are going to do exceptionally well. Cyclicals. Again, I've shared this before, but I want to share it again so everyone understands this. things that are tied to the consumer. Royal Caribbean, Norwegian Cruise Line, Celsius, Blooming Brands, ELF, Sweet Green Airbnb Hilton Tesla Nike Las Vegas Sands, Uber, right? Financials, Wells Fargo, Robin Hood, Sofi, Fizzer, Root, Lemonade, Oscar. We've seen this rotation take hold in the markets in the last month or so. I think it's going to get a lot more dramatic, especially if AI stocks do sell off. And you really want to be buying AI stocks. I believe when the sentiment shifts, when you're hearing on CNBC almost everyone saying, "Yep, it's not the time to buy AI stocks." That's when it's the time to buy AI stocks. That's when you want to go out and buy those A tier, Btier software or A tier or Btier um AI stocks is when the sentiment just dramatically shifts. And we are not there yet. So, have AI stocks bottomed? Some of them are probably closer to a bottom. Some of them have not even started the process of bottoming in my opinion. Obviously, that's my opinion. Come to your own conclusions. I think we're going to know a lot more in the next two weeks or so once we get hyperscaler earnings. And really, you don't have to make any big decision right now. If you're long or or short in AI stocks, I would just wait to see what hyperscalers tell you, what their commentary on their conference call sounds like around the capex spending. that's gonna that's what you really need to be focusing on right now. So, hit that like button, subscribe to the channel if you guys have not done so already. If you guys want to come trade and invest alongside of us, come beat the markets alongside of us, most importantly, check that link out down below in the description of today's episode. Have a fantastic rest of your day and I will see you in the next