Bottlenecks, Momentum Breaks, and the Next Phase of AI

Uh, in Maine for the week. We'll be up here, uh, for at least a couple months, beginning soon. Uh, holiday weekend there. There, I'm going to cover a lot of important stuff today, uh, in terms of some of the things that, again, uh, that I'm going to keep talking about, just because I think we're, we're at and getting close to an inflection point. Uh, and I want to make sure this is clear every single week. Uh, from a trading perspective, we're in the middle of a secular trend in AI, and not for any moment are you going to see me talk about the fact that this is not going to happen. Uh, rather than worry about not having the financing, uh, this is a bubble and once it collapses, the whole economy is going to go down, you're not going to hear any of that. Uh, this is a different time. It's being funded with companies that have the money and the revenues are coming. They're just on delay, but they're contracted and the momentum is, uh, it's never happened before. So, there are no free lunches. Uh, markets are all about risk-reward. We've had a huge run, as I'm going to go through, and I am seeing some global and sector correlation breaks, which usually are warning signs that, uh, you should be careful. We had a very volatile week for momentum. Momentum shifts are usually around regime shifts. Uh, I have a lot of podcast stuff in here to, to listen to, because I think we have to start building in that, uh, it's been now two months of talking about the Strait of Hormuz and how important it is, and there's still no ships going through in any meaningful way, and inventories are being drawn out. So, again, this podcast is about the factual side, and now we have inventories breaking through the bottom end. We did have the Google IO. There's two reasons that this is important. One is to show you what went on last year. Another is to connect it back to crypto, briefly. Um, for the RAAS, uh, I do, and I am going through with Morgan Stanley on index launches. Uh, I'm not planning to do an ETF on the AI thematic portfolio. But, uh, the 100-name portfolio, I am working on getting, uh, two things. One is an index that's tradable, not just for mutual funds, RAAS, but for hedge funds, options, everything along those lines. But most importantly, in my opinion, um, as I go through this, you have to understand that, uh, the outperformance AI is, is driving the entire market. Uh, in my opinion, it's driving the entire economy. I'll debate any economist on that who wants to play the math game. If it wasn't for AI and the buildout, if it wasn't for the profit margins exploding for these companies, uh, and getting the wealth effect, I don't know what would be happening, uh, in the economy at this point that wouldn't involve needing the government to a greater degree. Uh, I also have worked on, and I'll show you the results of, creating a concentrated portfolio, uh, which will go up at some point this week. Uh, I've had a lot of RAAS that have discretion looking for something that's not a hundred names. Well, now you have the index, uh, able to be traded for anyone who has the ability to do ISDAS. But then you're also going to have a concent, two concentrated ones. One's of 10 names, ones of 25. I spent a lot of time to basically put together a list that replicated to the highest degree I could, uh, what the portfolio is.

So, let's just go through this quick. The S&P, uh, up 88 basis points for the week. Rough start, good finish overall, strong week. Qs again. And by the way, in the S&P, I didn't mention it, but you can see it here. Um, we're extended. I mean, this is, this is a lot of weeks in a row. We're going to get a correction at some point. I think it's going to be more than a one-week correction, uh, when we finally do get it. Uh, it wouldn't surprise me if it started next week. As we get into, uh, close to the end of the month, we've had a huge month, and I think mean reversion should be on the top of everyone's mind. Uh, especially since momentum showed weakness last week. Qs up 1.2%, IWM up 2.8%. The hyperscalers, they're basically unchanged now. So, again, this is Meta, Google, Amazon, and Microsoft as an equal-weight basket. Uh, and like I've said, I think these guys are the key to the market. As much as the AI trade is driving the alpha, as long as these names are bouncing, since they are the spenders, everything's okay. Well, they broke below the 20-day moving average. It's unchanged pretty much over the course of the last three weeks. So, I would watch because this is telling me the market's tired. At the same time, Goldman made all-time, made all-time highs this week, uh, with the S&P sitting basically at all-time highs still. Uh, when that's going on, nothing bad is happening. So, you're going to have to watch. I've shown this before. I'm going to show it again because this is the environment that I think we're going to be in. And this is the regime shifts that we're going to go back and forth with. We're not jumping into here. Whether it's positive or negative, it's over here. And if you go through this, we've got rising oil prices. We've got, as I'll show you, high mo, high volatility with inside factors, specifically in momentum, not so much in the VIX and other places, but when things do go, I think it'll be correlated. As we've seen, you've got stock market multiple compression. You've got trapped money and forced selling with inside the credit markets, and now you've got interest rates rising. So, again, this is a version of the 1970s, which I think is going to be a part of what you should expect.

Last week I showed this, which were basically show me, uh, the two times I put a monthly chart and a weekly chart, uh, of RSI. So, monthly RSI, weekly RSI, both of which were above 70 last week. Uh, on Thursday, I ran it. So, Friday is when I did the video, and I just highlighted that historically, this is not a great time in terms of what goes on when you get momentum up there, and that it either means we're at a top or we start to kind of build a top and then we head lower, and it takes a while to get back up to the highs. Well, we did have a sharp move lower in momentum this week. The S&P didn't really follow. Uh, this is on a close basis. We actually got all the way down here on Tuesday morning, which meant we had given up a significant portion of the last month's, uh, momentum rise. And here you can see the rate of change. So, I put these arrows in just because we've been on a trend now, and I think this is going to happen because I believe we're at the beginning stage of a secular trend, which means momentum is going to continue to work, uh, over the course of time here. I don't think the names are going to change that much. I'll go through why that is, uh, because we're at the beginning of a buildout for the agentic world. This is a once-in-a-lifetime opportunity over the course of the next three to five years to play a secular trend, which will remind us somewhat of the commodity bull market that occurred in China when China was building out from 2002 to 2007. Nothing could stop that train. Um, and it ended when the US blew up. Uh, but if you remember everything that was going on, it was a lot of money being spent. So, we are seeing continued unwinds. This is a two-day rate of change that the numbers are just getting bigger on both sides. And I think that's what you're going to have to get used to as this trend continues to extend. Um, we've got a historic breakdown in correlations between software and semi. So, if we're going to get a momentum sustainable move the other direction, it's going to include software that is either not going down, uh, or is going up while semis are coming down. And that's what I'd be looking for. I wrote this for 22V this week. Um, it basically just went through what I see as the issue. And again, I'm going to say it again and again. I don't think this ends because of a bubble. I don't think this ends in the way that most people are writing about. Uh, meaning, I, I've done this for a long time in terms of thinking about the things in go. When we had Amazon and we had the Mag 7 that were going up, it didn't ever end. You just got periods of time where they'd have corrections, but nothing stopped them from demolishing the other companies and creating moats around it. It just continued. So, I think this is going to be in there, but the risk is that at some point we hit bottlenecks. Hoarding happens. We don't have the physical world like we did with software where nothing could stop the growth. You can stop the growth, uh, basically by not being able to build the stuff anymore and by holding all this inventory. So, that is the risk that I think will happen.

Now, I wanted to just highlight that one of the correlation breaks that's happened, there's two parts to the AI trade. Um, when I put together the thematic portfolio of 100 names, it was really broken down into two sectors. Um, and if I said two sectors, it's obviously heavy semiconductors, but then so technology is part of it, but the other side of it is industrials. Um, and even with the power side, the majority of the names, if you guys have gone through the power side, they're industrial names are things like Caterpillar and Eaton. Um, so momentum unwinds and industrials are very important. And this is what happened in the industrial momentum side. This is through Thursday. Some of the charts on this stuff are through Thursday. Friday was not a big day for anything. Uh, so I didn't want to go back and make all these changes for nothing since I put some of these slides together on Thursday. But momentum in particular on industrials, which is still down here. SMH has gone back to the highs. So, this is the correlation between these two, and this is the problem. So, you're getting the long side of momentum in industrials. This is the long side of momentum. The Morgan Stanley Industrials Momentum. This is the actual Morgan Stanley momentum side. So, this is both of them. You can see this head and shoulders pattern. This is what happened to the overall momentum in industrials. It stayed down, and we have a MACD sell signal on this in terms of what's gone on. So, I think momentum, um, is showing signs that maybe we've run too far and we need some kind of profit taking. And whenever we get to the end of the month when you've had a huge month for things, I kind of worry about people taking off positions either in that week or the following week. So, I think the risk is there for, for more momentum, uh, pain. Uh, I just wanted to include this. I did this at the beginning of the year. This came from a Michael Seymble piece, just he brought up the fact that 40 names were driving earnings grow, uh, driving 70 to 80% of earnings growth, not the excess earnings growth, the actual earnings growth, uh, that was coming into the year, and they were related to AI and capex. And I basically did this to say that, um, correlation spikes in stress periods, you're going to get all of these moving together, meaning all of the things creating the alpha will move together, and that's one of the issues that you look for is if that's the case, you have to start looking for breakdowns and correlations. And here's the first chart of a broad-based global breakdown. So, each one of these is a different market around the globe. The S&P is back up near the all-time highs here, but the Russell is, is, is, is still below the highs. It did bounce and it's getting closer to the highs, but you have this situation where the US markets are here. These other markets here, it's European markets with the CAC and the DAX. It's Asian markets with the Hang Seng, the Indian market, and Jakarta and Australia. So, you got Indonesia, Australia, and a lot of these. Now, this is all normalized over the course of, uh, the last year, meaning they're all together. These are big divergences that are going on. If you look at Indonesia, and again, it went down, and then you had, and this is the key point. When you have oil going higher, and now you have rates going higher around the globe, you're starting to get these divergences. So, initially the oil market, it was a fall. We get the bounce in the US market to go up. The Jakarta market goes up. All of these markets go up, and now they're starting to roll over again. And that's the risk that I see. I see contagion amongst the down move. I see everything having a hard time going higher. I talked about the importance of industrials. This is the world industrial index. Again, it peaked and barely bounced this week. So, the industrials around the globe are showing weakness. And it would make more sense that they would be the first group to show the weakness over the semiconductors. So, if you're going to have a problem with rates going higher and with oil going higher, you would expect it to be more of an Asian and European thing than a US thing. We're seeing that. You'd also expect the industrials to be, to be hit hard. They're the manufacturers and very important to the data center side. They're very important to all this. And if they're not getting in what they need on the natural gas and the oil side, or any of the chemical derivatives of that, they're going to start to have issues. And I always believe that the market is going to show where the pain, uh, is going to be beforehand. This is that world, uh, industrial index overlaid with the S&P. So, again, the S&P has gone to new highs, gone higher, while this world industrial is lower than it was basically before the, the war started. So, you had gone up into the war, then you came down, then you went up, and you're below it. So, we never took out the highs, and then we failed there. So, the industrials are being impacted. Here's that now with the Cosby machinery. My, the, the yellow name here, which is up at the highs. This is my thematic portfolio. And the red line here is indicating where we saw the breakdown in the Cosby machinery index and the long side of the industrial momentum side. Basically, the industrial winners are under pressure. The thematic portfolio as a whole is up. A lot of those are, are US names. I don't have any real names in the industrial side from Asia. So, what you're left with is you're getting a breakdown here in these, uh, these names. This correlation has gone down. And to make it look a little bit more, let's say, severe with inside the country. I've shown this in the past couple weeks. This is through Friday. So, this is the, um, the construction side in Korea, the construction sector. This is the machinery sector. They both peaked at different levels, went down, and this is SKHEX. I've talked about, and I did a webinar this week for subscribers where I went through how important Korea is. Korea is an important part of the market, but it's not just them. This is another market that's been important that's also showing signs. This is Japan. Um, Japan is one of the markets that stands out the most to me, uh, on the Asian side. They, or on the, the AI side, they come up in the chemical side, they come up in the semiconductors, they come up in all parts of this. Mitsubishi from the gas turbine side, they are heavily involved in multiple levels. And the Topix basically has rallied, but now all of a sudden, you're getting this divergence again, just like we saw there. It has been very correlated to my thematic portfolio, but beginning towards the middle of April, the thematic portfolio continued to rip higher. A lot of this has to do with what's going on in Korea. A lot of it has to do with SKHEX. A lot of it has to do with Micron. A lot of it has to do with the power semis. All of this stuff has been kind of impacted by what's happening in Korea. And this is when you got the breakaway. So, in Japan, we've kind of broken away. Now, if you go inside the Japan market and look at the machinery side, you have the same chart as you do in Korea. Basically went up and has gone back down. Very correlated to the thematic AI side. Worse than that is the construction side. So, not only does Japan have the issue right now of oil, it also has the issue of long-term rates going higher. And this is another part of the equation that people have to start factoring in. Remember, since the Strait of Hormuz was shut, we've not only had oil prices go up significantly, we've seen a Fed, which was expected to have three rate cuts over the course of the next 12 months, and now we have one rate hike over the next 12 months. So, we've shifted 100 basis points out of that. We've seen 10-year rates basically go up, uh, about 75 basis points from where they were before the war. And so, in countries like the UK and Japan, you want to watch to see how things like the construction market are doing, which are obviously more sensitive to long-term rates. We're breaking below the 200-day moving average. And to put that in perspective, since ChatGPT came out, which was in here, except for this panic during the yen unwind and during Liberation Day, we haven't been below the 200-day moving average. Well, we're below it now as of Friday. Um, and this is the Topix overlaid with those. So, the Topix has hung in. And again, when I start seeing correlation breaks on what had been driving the market, I start paying attention. So, you've got more there. This is the machinery side in Japan relative to the, my thematic portfolio. So, the AI side. Now, if you take it even further and you go in the US and you say, "Okay, here's two-year rates overlaid with the S&P." So, the S&P is ignoring, again, the rate side. It's also ignoring the currency side. The dollar has been strengthening continuously since we took out the rate hikes, the rate cuts. So, we now have the dollar going, which is generally a tightening liquidity situation. You also have oil. Despite everything happening in the Strait, all these promises about the deal, we don't have anything yet. And if you're living in facts, eventually you can't keep this game going on. You start to become an inventory problem. You go from a supply glut to a supply shortage at some point. And based on everything that I'm going to show you and everything I've heard, I listened to probably three or four podcasts this week to listen to four different people on the oil patch that have been around a long time. And the message is pretty similar. And rather than the doomer side which was coming out initially after, which was focused on how bad it'll get, well, now we're in the reality side where the inventories are being drawn, not just in the US but across the globe, and it matters.

Google IO. This is a chart I just wanted to show because I want you to pay attention to here. Um, last year at this point, in a span of a year, we had gone from about 9 trillion to 480 trillion. Now we've done seven times that growth. This is in tokens processed. So, just think of this as compute usage. And just look at this parabolic move. So, this was reported this week. The reason I want to bring this up, and this is the other reason why I think people should be wary of just where we are now. I wrote this paper on May 15th a year ago. I could not get a single person to care about this at that time. Not one person. I spent a lot of time with institutional. Everyone was busy debating tariffs. Nobody cared about how inference was made. It was mentioned. And this was from the Q1 earnings reports where I talked about how many companies specifically mentioned what we just saw here. This is May of last year. So, we had seen this dramatic uptick. No one cared a year ago. And I highlighted these are the names if you want to play the inference breakout: Nvidia, AMD, Intel, Qualcomm, Lattis, Micron, Western Digital, Seagate, Sienna, Marll. Every single one of them. Okay, I created an equal-weight basket off today of them. Here are the returns of them since I wrote the paper, and now people are bullish these names. Yes, this is the reason why when you get through everything and you're seeing kind of a bubble formation, this is my inference side. So, if I create an equal-weight basket of these names, these 12 names that I mentioned, this is what the chart looks like. Okay, so a year ago, here's where we are. This is why when I say I've bailed out of my Micron, a lot of that has to do with why wouldn't you get out of a chart that looks like this? Yes, it probably goes higher, and it may go higher because there's no way you can pick. I've been out of this and in this area, but here's the thing. If you want to judge retail into this, this is the Roundhill DRAM, uh, ETF which was created in April, which has $10 billion. I listened to Yan Van Eck speak today or speak yesterday on Compound in France talk about SMH, which has been around forever, has about 60 billion AUM. This thing, I, I heard is up to 10 billion in AUM, uh, at this point in two months. This is retail, and it's very correlated to what I wrote about again back here. So, I think this is an indication that maybe everyone's jumping into something that's already had its move, and everybody's in at this point. And even though I don't think institutions are in a big way, I do think that you got to be careful right now that if it turns the other direction for any reason, and I'll go through some of those possible reasons, it's an issue. Um, here's another problem right now. Is I don't use a lot of different things for sentiment, but I do think this Goldman Sachs Risk Appetite Indicator is one that I look at just as a general gauge. And again, it's at the highest readings since 2021. So, don't, that's no reason to get bearish, but it is a sentiment indicator that I like to look at when they post things in X about it. Um, this is the reason everyone is so positive. And again, I've said it before, I'll say it again. Maybe we just did two quarters of orders, and maybe the next one's going to come in here. When you're up at this high, so this is net income for Micron. Obviously, they've been able to price this at any price. They've had huge orders, but I mean, I don't want to go count the stuff of what you're talking about, but you're talking about basically the last, however many years of total net income. That was for the quarter. So, again, we'll have another reading in, I think it's six weeks. But this is the other issue. Um, profit margins in the S&P are almost entirely seven stocks. And again, profit margins are good by the different sectors, but the majority of the profit margins which are geared have taken this S&P up to these levels. It's being driven basically not just by the MAG 7, and that's where the focus of this is because that's what BFA did. But if you add in Micron and you add, just put Micron in there, and SanDisk and Western Digital, this, this is a lot of them. This, we don't have margins going up for the S&P at this point, and especially with oil going higher. It's the only thing I'm saying is if all of a sudden the margins went down, which wouldn't be bad, but if it, it went down because we had hoarding in the first order, in the first quarter, and there were so many shortages, and then all of a sudden the demand starts to go down because people have either overbought or the costs have gotten too high, you'll see profit margins come down. And if profit margins come down, I just think the quant strategies and everything will see margins come down, which is usually a bad indicator for the market as a whole, and you might start to get some unwinds of positions or a change in certain kind of risk.

Warren PI's put this out this week. This is another thing that again, um, the forward sales growth is projected at 18% over the next 24 months. Again, I think the orders are going to come in. I just think you're going to have lumpy problems along the way because you're going to have bottlenecks. Um, I don't know how much you can order, just keep ordering, ordering, ordering if you don't take delivery of stuff. We already see that delay in terms of the RPOs for Oracle. We see that for the RPOs for pretty much everyone. I mean, Nvidia used to have a big receivable issue. They still have big receivables, but they're changing it to make sure they get cash. And that's because everything was in the future. Same thing goes with Microsoft, Google, and Amazon cloud, where a total of $1.4 trillion. They're going to get it at some point. They don't know when. They're hoping it comes soon. Wall Street Journal had an article today on how the AI chip mania sows the seeds of its own destruction. Gets into the cyclicality side. But again, it just shows the analyst forecast income for this year. I showed you the 13 billion, and then for, for next year. Again, a risk is that supercharged profits attract new rivals to enter the market. I am not worried about competition by itself. And I want to make this clear. The competition side has a bunch of issues with it, but most of them have to do with the fact that I think the delays that are likely to happen prevent the capacity from coming online quickly enough because the orders are just too big. So, regardless of whether it's a timing issue or it's an actual competition issue, there's a lot of things that can go wrong a year from now. Not that it's not gonna still be a good investment, but you're going at a pace where an eight-bagger in a year at some point, it's going to slow down, and at some point, if it slows down, the question is, do people want to be long a stock that they only think is going to go up 50% when now it's bigger size? Remember, I've shown you that Nvidia's multiple has collapsed over the course of the last two and a half years. Uh, and their earnings have been great. Um, this got everyone excited this week. Um, probably led back to the bounce. Uh, this is the dollars to be spent in the Vera Rubin on memory, 435%. Just to give you an idea of where the change is as you go through and how people get excited on the numbers that are just heavy growth periods in this, in terms of dollar size. And again, that assumes that Vera Rubin and the buildout actually happens in the time. Um, so Zero Hedge kind of went through it. They're focusing on the memory cost in Vera Rubin, but the US power grid can barely keep up with the current setup. How are we going to get them more? And that is the issue. And remember, Nvidia was hunting through Korea to try and find help on the industrial side, and we have the industrial side breaking down right now. So, this is the connectivity between Vera Rubin and the needs of the Korean side for the Vera Rubin side. They're all connected. Um, data center pipeline and gigawatts has tripled during the past year. That's great. We get more and more orders for things, but that just means the bottlenecks continue. Gartner AI spending forecast. Remember, it was only six months ago that people didn't think we'd get any spending, that it would be canceled. And now what we're talking about for 2027, the total AI spending will be 3.5 trillion. Now, again, this is three and a half times the size of what's expected for next year, about a trillion dollars for the, just for the data center side. So, again, the AI infrastructure spending, service spending, software spending, cyber, like these are big numbers, guys. And this is the reason why, even if it comes in at 2.5 trillion, this is, this number is more than 10% of GDP. So, again, these are, these are big numbers and orders.

So, inside the biggest bet in corporate history, I thought this was interesting just because of this chart. So, first of all, the year-over-year growth in capex expected to be here. Now, again, they keep revising these numbers higher. So, who knows how high it'll be, but over the course of the next eight years, the reason people are now fully excited. So, when I wrote the inference piece a year ago, I want you to remember, no one wanted to talk about this chart. Think about it. A year ago today, nobody wanted to talk about this chart. And this chart is now higher than it was. So, last year they doubted that we could do a linear line that looked like this. Now they're doub, they're jumping in because they they accept a higher number. I've just seen this too many times in my career where there's going to, it's just not going to work out that easily, uh, in terms of this. And I think we're starting to see the pain points because rates are moving higher, oil's moving higher. We've already had input costs between micro, memory, CPUs, silver, everything that goes into building. And guess what? The most important thing about this $8 trillion, the most important thing, this is what it looks like in terms of the buildout that needs to happen. It's that complex. We have a ton of things that need to get built. How far into the buildout are we? By the end of this year, we will have only spent about 18% of what's expected. The physical system is already stressed before the largest dollar years even arrive. Is this likely to be smooth? No. The chart is smooth. The buildout will almost certainly not be. And that's because this is not a software capex. It is a physical world capex cycle. It needs HBM chips, racks, liquid cooling, copper, fiber, substations. If any of them aren't around, it is about the sequencing. Meaning you can't add in, and this is what my paper was about. You have all these issues that are happening. You've only done now through Q1 of this year 12% of that spending. We are barely through it, and you have multi-year delays in so many things, it's impossible to name at this point. So, the risk is that the buildout becomes lumpier, more inflationary, more volatile than the market is modeling. That's what I believe is going to happen. So, all I have to do is upload the chart and say, what does this say? What is the risk in it? I didn't even ask anything about it. I just said, how far into it is, and only being this far when we already have bottlenecks. What does it mean? Bottlenecks create cost inflation. Cost inflation creates delays. Delays create revenue recognition risk. We are already seeing all this stuff, guys. And that's the risk is that you're taking Q1, which is only basically two months after people thought it was a bubble and didn't want to be involved in it. The time when I started hearing people accept memory was in October of last year, and that was because the memory prices had gone higher. Before that, everyone was still worried about what was going to happen with tariffs. So, you've had this huge move in things. You've had all this stuff go through for a variety of reasons which I mentioned in the paper. And I just think we're not taking into account what's going on. A must-read interview with a Seaman's employee explaining just how big the demand is for energy equipment right now. And again, 12% of the way into this. Are these decisions speeding up or are they slowing down? Decisions get made very quick. Decisions are quick, but there will be delay in the production and delivery. Is any of this surprising to you? I would say yes. It's not only me. I'm working with some of the people who have been here in the business for the last 40 years, and they say also cannot get a proper explanation of how these things are happening. Just imagine we are getting orders double the size of your whole factory can produce in a year.

I've shown the Dodge Momentum Index. Um, it's, you know, it slowed down obviously in the winter time, but now it's starting to pick up again. Uh, but labor shortages, high material costs, and supply chain disruptions are weighing on owner confidence in the near term. Year-over-year, uh, DMI, the Dodge Momentum Index, was up 14.1% when compared to April 2025. The commercial segment was 37.2, but only 5.8% when data centers are removed. Again, 37 down to six. Data centers are all of it. Um, this is the reason why I think people should now start paying attention to oil. So, Art Burman, uh, I spent a lot of time with Art this week listening to him, but also, you know, I tried not to get into the, the, the oil side in terms of the details other than just to mention that we still haven't gone anywhere, but sentiment just doesn't care about it. And now we're at the point that people should be caring because we're starting to have draws. The shortage hits US inventories. So, we've been releasing things on the SPR. That's part of the draw. Everything is happening. US total crude inventories fell 17.8 million barrels. That's commercial and SPR stocks combined. On that basis, it's the largest weekly fall since data available starting in 1982. Warren pies last month was the largest cumulative draw of total US petroleum inventories ever. Gasoline inventories are now the lowest since 2014. Distillate inventories, the lowest since 2003. Next few weeks is a crucial window for oil fundamentals. Now, you've been hearing this for a while, but again, when you start seeing what is happening, which is the inventories are being drawn down, the inventories are being drawn down, you start running into a situation where this person highlighted that before Ukraine, Russia, we actually had low inventory. So, and that was because of COVID with all the shutdowns. We had a supply glut. So, we had room for this to go on. The question is, is the war actually going to end? We've been in a ceasefire for a while. And yet, you don't have anything changing. Trump went to China. Nothing seems to have changed this week. Uh, there seems to be a couple more tankers going through. They seem to be China. So, maybe China's going to get their oil, which would help. But oil's imports, China's oil imports have plunged 20, 30%. The global onshore crude oil inventories ex-China. So, everything ex-China, I mean, again, you can see the years here. We're really drawing down. So, I highly recommend listening to this, uh, interview with Art Burman. I think he's very, I don't know, honest. He seems to talk about this in a way that is not hyperbolic. He really does go through, like I said, I listened to about four this week, but Art was the one that I think went through why this is so serious, and I think he highlights a lot of points. We get caught in, it's open or it's not open. The reality is it's probably partially open. The question is, it's partially open. We don't know the demand destruction. We don't know if it's going to be easy to build up things. So, the reality is it's probably a bigger problem than people realize. And most importantly, people don't seem to care about it anymore.

I'm not going to read all of this right now, but you can go through this to, um, see it on your own. In terms of, uh, another podcast worth listening to, Luke Goman was on TFTC this week with Marty Bent, who I'll be on with soon. Um, and Luke brought up another point that I just think is worth people thinking about again. Um, people have forgotten about the sovereign debt problem. He says the Iran conflict has pulled forward the bond market problem. Uh, I don't want to talk about the AI buildout, but he warns that current valuations enthusiasm assume the the economics will work smoothly, even though the whole buildout could be vulnerable to competition, debt financing issues, or supply shocks. I, I'm the supply shock is the part that I'm, I'm going to agree with. I'm not worried about the debt financing, uh, unless again, oil starts to go higher, and I think it will impact all of these things. If oil doesn't sit here and we're draining inventories, we should have a spike then. Um, if it's been held in check because we've been drawing or we've been using oil that was in storage, we should be in a situation where if it doesn't get resolved now, it could be an issue. So, think about the risk-reward of things when again, you're seeing signs right now that it's already impacting supply shocks, and we already have supply shocks. This is just an add-on. Um, yeah, I'm not going to go through the rest of this stuff in here, but it was a good interview. Um, CPI, we're now starting to get the Cleveland Fed inflation nowcasting. Uh, and again, CPI is looking at point four or 0.5 again. PCE is looking at point four again. That would put both the CPI and the PCE above four. And I've talked about the historical returns in the equity market. So, is 4% the line in the sand? I don't know. But rates are moving higher. And it just seems like if we get another surprise in CPI and PPI, and you start seeing that flow through, all of a sudden, you end up in a situation where you're starting to get a little bit more antsy in terms of the Fed. And we'll see how the long rates go. Here's the PPI on things that have nothing to do with oil. This is the input cost. So, again, you've already seen this massive rise. If we don't continue to see the rise, is that mean prices are starting to cause demand destruction? Um, demand destruction does happen at some point, but also bottlenecks cause demand destruction. Truckload. I've talked about this. Um, again, people were fighting this back in here, and yet here we go. Uh, I don't know how inflation is going to come down when you've got truckload spot rates heading higher.

Welcome Kevin Worsh sworn in today. The Fed will have to raise interest rates in July, says Edard Denny. I don't believe that's going to happen, but if it does happen, again, you've got the uncertainty at a, at regardless. Um, Donald Trump couldn't go without speaking on the day that Worsh comes out. We're going to grow our way out of it. We're going to grow the national debt so fast. Economic growth doesn't mean inflation. You don't have to stop the world because you're doing well. So, it sounds like we're going to go with the inflation as transitory routine. And more importantly, because the debt problem is so big, we're going to try to focus on nominal GDP. Well, that's good. The issue is, will long-term rates be able to stay in check? And that's where the issue is going to come. And that is where I believe the major inflection point on the regime will be. So, to get a regime shift, you have to get a momentum unwind because the momentum unwind would probably mean the AI names get hit. If the AI name gets hit, all the markets are going to get hit. Whether it's a 10% correction, 20% correction, 30%, 5%, I don't really care. But at some point in there, if rates are going, if rates are going higher, oil's going higher, and you have the market going down, I think people are going to be confused on what to do. And right now, everything's quiet, the VIX is low, but I think people should be worried about the what-ifs more than they are.

Uh, I got to give these guys a shout out. I, I've started, I've, you know, I've, I've given up a lot listening to most macro, um, economic podcasts. I've talked about it. I just find it to be way too doom and gloomish. Not focused on AI, still calling AI a bubble, saying the economy cannot go with AI doing this, focused on the delinquencies and all stuff like that. Compound and Friends does a really good job of just talking about what's going on in the market for the day or for the week. I think they have a good grasp on the general, um, things happening and what's important. And so by listening to them, they also have good guests. And this week, I thought they did, they had two good podcasts. So, I'm gonna show, you know, both of them in terms of this. They talked about Nvidia. They talked about Google IO. They talked about Gavin Baker. Um, Gavin Baker gave an interview. It was a great interview. It was on another one on Patrick O'Shaughnessy, which I'll get into in a little bit. But the point that they made, which I thought was interesting, it's not a bubble, it's a wave. Bubbles imply something that suddenly disappears. See this type of logic to me. Um, and I give them credit. Both Michael and Josh do a good job of being rational. Uh, and the fact that Michael is a big Knicks fan and wears Knicks paraphernalia all the time, uh, means a lot to me as well. But they've done a good job of just going through it. So, I highly recommend if you're looking for people to listen to regularly, uh, to find kind of under the hood things going on, I would give these guys a shout. I like what they do. They also have my boy Adam Parker on there all the time. And he seems to be spending a lot more time. This is something I wrote about in a paper called The Art of Unlearning, but they talked about Adam Parker's idea that investors may need to unlearn the past. I wrote a paper on that. I was going to show you guys some of the details. If you guys want to see it, just let me know. I wrote it in 2016. I was rereading it this weekend. I will highlight that in 2016. Uh, Dr. Michael Bur, uh, I used him in there to say he has to unlearn. It's amazing how many times a guy can call for a bubble. He called for the collapse of markets. It wasn't even just that. That was in 2016. So again, if you're paying for his Substack and you're listening to stuff, just remember he's only interested in one thing, which is the markets don't make sense mathematically. And that was the point of the art of unlearning is that you're going to have to start putting math onto everything when you're dealing with AI and when you're dealing with the debt, the size of the country, and the printing press. Exponential innovation. The printing press has created a different thing. Um, in terms of Adam Parker, software versus semi divergence, that's where that chart was from that I highlighted. Um, I just wanted to show this on the bottleneck side just because, um, I, I, I, I really do think it's important to kind of pay attention and for you subscribers, I have it in there. Um, the grid problem is an issue. So, again, I've shown this before, but if you haven't seen it, it is up on on the website. The data centers and getting them done. This just shows all the different places of of the weakness, but I just want to show you the numbers. The probability of being realized in terms of getting built by 2030. Look at these numbers here. And this is just half of them across the country. And this goes through all of the issues. I weight them and use all of these. We have so many acute issues, high issues, um, to get the buildout out. It's just a lot, guys. And so to think that this is going to happen like without a problem is not there. Um, Adam also kind of went through, um, basically saying that investors are using mental models that were built for the last regime. I couldn't agree more. Um, and he says capex-heavy businesses may deserve a rerating. That's already happened for some of the industrials. I think that's happening for the semis, and I think that's part of the issue, especially for Micron. Um, capex is not wasteful overhead. It is the toll to participate in the new growth cycle. Again, I couldn't say it any better. This is why your capex is my opportunity is a big focus on it. So, again, big shout out to those guys. And here was number two. Uh, Yan Vanek was on. If you guys have never heard Yan Vanek from Vanek, uh, I've now, I was at an Anthony Pompliano event where he spoke.

After I did and I found him to be have great macro comments and he was on the show today or yesterday compound and friends. Their second episode and I thought he did a great job and Van has the SMH. Uh, they acquired it uh, I think after the great financial crisis when Meil Lynch was closing things down or Bank of America, I guess it was Meil Lynch.

Um, he talks about the issue with memory stocks and he basically brought up on the question of whether memory stocks is a bubble, especially compared to Nvidia. Um, memory looks more bubble-like than Nvidia because memory does not have the same moat. Memory, you need my product. I can so I'm I can raise price. Nvidia, I can give you more intelligence. Blah blah blah.

Then he lists, and this is the part I was most interested in, the possibility of things that could be questions over the course of the next year and break the memory trade. Chinese memory supply could eventually become a competitive threat. Model efficiency could reduce memory intensity. We've already seen that with TurboQuant. The market ignored it so far. Third, customers will eventually respond to higher costs. This gets back into your costs and he gives the specific example which I think is important, telling his CTO to stop spending so much on compute. I think the spending on compute is becoming more of a story. The AI stack itself could change and that's what brought memory to what it is. Once we got to inference and we started getting into Vera Rubin, which was a new thing at the beginning of this year, which hasn't happened yet, we started buying memory because it meant the long-term makes it more likely that it's going to be needed. Well, maybe some other change will happen. Memory is the bottleneck trade. Nvidia is the platform trade and I completely agree. And as someone who's been in the trade, I just think the risk-reward has changed and maybe it does double from here. I like Marll to do more than what I think memory will do. I think Marll is more important to the future and more of a moat. I could be wrong on it. Uh, but I leave that up to you guys. This is the cost side to AI and how much it is starting to go higher.

Again, you can see chat GPT5 uh up here. Um, again, I'm going to say that from my perspective in terms of the best model, I have not used Claude more than a couple times this week as I end the week. I have used ChateBT and Codeex the entire week. I have used Gemini occasionally. I have used perplexity occasionally. I have uh become very, very uh sick of of the compute situation in Claude and have spent most of my time on on chat GBT this week. We'll see if I go back to it next week. Here is the token cost. So Bloomberg has this index. I think most importantly you see that this started to raise in February, which means again, the compute needs, the shortages are there, the costs are going higher.

Gavin Baker again, whenever he speaks on any of these things, I think he's worth listening to. He said it again, which he said multiple times and I'll just say this, one of the things he said is it is amazing that you get to listen to, let's just take Jensen Yuang, but all of the leaders in AI speaking multiple times a week. I completely agree. I've created notebook LMS of Jensen Yuang, so many different people to have knowledge brains of this. They're handing out free alpha. They are telling you how things are going on these podcasts, what they need for the future. He talks about that and then he basically goes through the investor questions. If AI can materially improve efficiency faster than demand, then the market may be overestimating long-term demand for memory, GPUs, and power. So again, the efficiency side is going to happen. That I'm sure of. Um, and the reason I say I'm sure of it is because really smart people that have been on interviews, particularly Sergey Brin, have effectively guaranteed it. And I think Google, which had the Transformer paper, probably knows more than anyone that they're they're getting into this. But also, I think you're hearing more and more about if there's continual learning, does continual learning arrive soon? He treats this as potentially explosive. I think we've learned our lesson with how Opus 4.5 came out of nowhere and then the agentic world flew in. Eventually, we're going to have a change in one year is a long time in AI. to bet on next year's earnings for memory. Guys, sorry, you're living in a linear world where continually and uh recursive self-improvement is going to change things. And even if it doesn't change things into next year, if all of a sudden there's a cliff because they figured out some way to use one-tenth the memory, it's going to be an issue. So again, these are all would a could a should a's, but I've been living in the AI world, it seems, a lot longer than most people. These models have moved faster. They're changing unbelievably. I've gone from not using Claude to using it all the time and now I'm back to chat GPT and I love 5.5. By far my favorite experience in terms of AI and I really view Claude as being something a little bit older at this point because it's slower and I haven't seen the improvements and the model is not as smart as 5.5. It's just not.

Um, the profits are still going to energy, data centers, chips and models. The app layer is not dead, but it's much more selective than many hoped. And that gets into the software side where the value is accruing. This is not, you know, surprising, but this is basically the five layer cake. Google IO, the keynote speech. Uh, by the way, all these podcasts, you know, I'll have a summary up on the subscriber website with the link, but if you just want to go through and read the summary of each one of them, and that way you can upload them into, uh, your LLM of choice, uh, just so you guys hear this, I've done a lot of things that I've gone through on the video side. I will release them on the paywall soon. So, the videos that I'm doing to show you how to use the the tools that I've put up there, the model portfolio to use the uh actual um uh transcripts that I'm putting up there uh or at least the links and showing you how to do the transcripts. All that stuff will start going up next week and the week after so you guys can start using it. You have to be using it regularly. Don't just watch this video and take what comes out of it. Upload it to notebook LM first of all, if you guys haven't done that and I will show you how to do that as well. Uh, but I would absolutely highly recommend just taking some of the stuff and even if you take a screenshot of this and upload it into chatgpt 5.5, you'll have a uh a good experience.

Uh, I just wanted to bring up in the Google IO that if you followed what they said last year about the token side, which was really the focal point. The models were just getting smarter, but the focal focal point was token usage was going up, which was again big for their cloud business. Everything now is about agents and as someone who has his own business, his own LLC um and it's on Google Workspace, I I mean I am happy as can be that I'm going to be able to use Google in the way that they're saying, basically as long as your work lives in in Gmail, drive, docs, sheets, calendar and chat, this is going to be different than what a normal chatbot can do. So, I was listening to it because this is all about agent. It's going to be a better experience. They also talked about the consumer agent part. This is the part that I'm telling you, in the same way that Inference last year, if you would have followed it, a year from now, you're going to look back and go, "Oh my gosh, between the Stripe annual newsletter, the Coinbase earnings report and Google AI agent or IO agent for both the consumer and the business. How did we not buy Circle? How did we not buy Ethereum? How did we not buy Bitcoin?" That will be in May of next year, guys. And whether it starts in a week, a month, three months, or six months. Remember the memory trade took a long time. And I saw Micron bang up against 110 for weeks. Back down to 60, back up. I think cryptos in the same place. I don't think people have connected yet that the AI infrastructure is to trade right now, but starting at some point, the handoff is going to the redesigning of the global financial system. For those of you who got to watch the webinar this week that I did with uh the Kraken uh team and uh and uh the other groups that are involved with the spa. Uh, again, I think this is part of what's coming very, very soon and I think you guys should be doing your homework on the space. So the Google AI IO side for the agentic side was very important. And again, I'm not going to read all of it, but you just have to go through and just see what we're talking about in terms of doing things across the ecosystem and what the agentic world is going to do, especially when combined with search. For those of you Bloomberg users, I highly recommend going through and watching um Sundar Pachai and at least learning about Google Finance and what you're going to be able to do because we're getting into the point where with agents you're going to be able to have 20, 30, 40, 50, 60, 70, 80, 90 agents working on a problem for you, all with different specialties doing different things. And that's what leads to the tokens going exponential. And here is that page from the Coinbase side. And whether it's tokenization, whether it's stable coins, whether it's just the onchain economy. Again, guys, I think you're missing the parabola that's happening.

Uh, one group I've had a lot of people ask about um why it's not participating. I want to just highlight. So, this is Vistra, which is basically unchanged since the end of 2024. Uh, its multiples were too high back then. There was less names to invest in on the AI trade. When we hit October right here, um we had a lot of names and this is when the broadening out happened. This is it relative to Nvidia. These two names to me are think of them as the defensive side of AI since I think we're very close to some kind of a momentum unwind where the names are running out of steam. This is kind of like the inverse of Micron. Now it doesn't have the parabolic downside, but I think people were looking for beta. So with inside AI, these are the defensive areas. So if you want to play something for the next whatever months, I would be increasing size into some of those names like Vistra here is a chart with an IP. I created uh an index of 10 names that are IPs. So the power independent power producers are at least close enough to the IPs, including Vistra and nine others, and you get this thing here which is consolidated. I think uh for all you cyclical defense watchers, I think if defense starts working in AI, you're going to see Vistra and some of these IPs. This is where I'd be moving some of my money into to reduce the beta.

Um, for those of you who keep asking, if you go to my video and you go down here and you look, it will tell you where you can go. Subscribing to my Substack for free is not getting you the subscriber paywall. So, I'm only bringing this up as I start to go through the conversations for the RAAS, for the FAS. I've been getting so much progress. People are asking for things. I'm spending a lot more time. I'll be down in the New York Stock Exchange in a couple weeks on tokenization and also talking to them about doing some stuff with them. But I'm really trying to help people with the benchmark arbitrage. So for any of the FAS and RAAS that want to have a conversation, that want to see if there's ways we can work together. Like I said, I've got Morgan Stanley creating uh the ability of of doing this from an ISTA basis. Uh, I've had ETF uh reachouts, but this stuff is moving so fast that I don't want to wait six months on something.

Here's this thematic portfolio made new all-time highs in this again. So you got the momentum trade. The good thing is not all of the thematic portions, the themes made new highs. But again, here is the alpha that's been created since the launch of Opus 4.5, which happened in here. So again, you've had basically five down weeks. Only one has been greater than 2% and you can see how many have been above. Again, it's a hundred names. It's across multiple sectors. This is not just semiconductors. You've got power, you've got the whole rack, you've got chemicals, you've got optical, and you've got package, plus five macro satellite names, which just means five other names that I've written about. So, not all of them made new highs. Most of the ones that went up to the highs were related to the non-memory semiconductor names. So, Marll made new highs. Um, Intel had fallen. Uh, I mentioned on the the subscriber webinar that when Intel went under 110 this week, I started buying some because I believe CPUs and optical are still good trades over memory. That makes people crazy because Intel's up five times, but I'm doing this based on the demand surprise versus the supply.

Um, this is where we get into the concentrated. So, I've had a lot of people saying, "Hey, I can't trade 100 names. I'm a subscriber. How can I do something better?" So, I created a 25 name ticker. This is basically out of the five themes uh plus one name from the uh macro satellites of the five took 25 names and this is the overlay between the thematic portfolio and that 25 name. So you can see it's a direct overlay uh in terms of correlation. It has more beta, so it has outperformed uh by a decent amount and that's because it has 25 names as opposed to 100 names. Uh, but it's got stuff from every single grouping. Uh, this is the performance on it in terms of the spread to the S&P. The reason I wanted to show this is this basket is up 68% year to date. The S&P is up 9% year to date. If you had as an FA or sorry, as as someone who has discretion or someone in your portfolio, whatever you have in the S&P 500 in spies, if you had 10% 90% in this, you got 8% and then if you have 10% in this, you have another close to seven. That's the way that I think about these things is I think you need to create the alpha in your portfolio by overweighting some percentage of this. That is the way also with Bitcoin and things like that. In this case though, I think that's where the conversations I've been having is the benchmarks are wrong. I don't think the hyperscaler spenders, which are huge companies, or the software companies, which are huge companies, deserve to have the same waiting that they had in the past because they are being disrupted by AI or they are the spending side. You want to be long the receivers, not the spenders.

Because 25 names I knew would be too much. I created a 10 name portfolio as well. Again, taking them across each of the themes basically to try and get to the point where you have at least one name in every theme to get it correlated. And this is what I came up with. So, this is the three of them overlaid. Now, uh again, the green one, which is the one with the least names, is going to have the highest beta. Uh I just want to remind you guys this goes up every week. The exhaustion side, the technical side, and the fundamental sheet. It helped out dramatically over the course of the last week from a trading perspective. Remember, we had a significant momentum unwind. This is where we stand on the exhaustion side now. We've reset most of these. So, if you were trading, you got in at good entry points because we had a huge amount of exhaustion names and now they've all run off. This is how you can use these things to at least have a sense. The technical pattern on all of these are are are still strong and I think it's going to remain that way. The only thing that's weak is the macro satellites and that's because Bitcoin's below the 200 day moving average. Silver not trading well. Um you have Eli Liy in there. You have EWZ in there. I forget the fifth one. Oh, Palanteer, which is not trading well. So those five names, you know, they're mixed. The rest of them, they're well above the 200 day moving average. This is the percentage of names in the thing that are above the 200 day, above the 50-day. The average RSI is now at least reasonable. The relative strength is fine. Again, I think there's a risk for all of these and a correction. But from a trading perspective where you guys don't care about the next three months, you care about the next day. At this point, the trend is still very, very strong and you should be trading exhaustion to get out or at least reduce and rotate and then when they reset, be in a position to buy things. Believe it or not, Micron went from above 800 down to about 650 and then closed the week somewhere around above 700. You can make money trading stuff that that wide. So, that's the way I'd use it.

I'm going to show you guys how to do this, but I showed on the webinar webinar side. If you upload all of those Excel sheets into one LLM and then you go go through the three files and find me the names that meet this criteria and you pick the criteria, I gave an example. It must meet all three. I want only ones with PEG ratios below one. So they have some kind of a valuation, fundamental valuation put in there. Technical names with a score of 85. The charts are strong. I only want good technical scores, but I want cheap names that are in there and I want names that have low exhaustion. If you would have done that this week, you would have done well in terms of names that have gone. In terms of at that time, there were five names that met the criterion. Optical and interconnects, four in semiconductors, four in the whole rack, one in chemicals in this. There were 15 overall. That's the way I'd be using this stuff if I were you guys.

Um, that's it for this week. I'm going to leave Chile Maine and head back uh on Monday. I'll see you guys next week from uh