Transcription
All right, busy week, lot to go through. Um, all AI again. You guys can read on your own.
Uh, first of all, first time since November 2021 that all four major benchmarks closed together at fresh highs. Rare feat, only occurring 25 other days this century. So, equity markets are screaming. We love Fed cuts. NDX, Russell, S&PX, and the Dow all up for the week, uh, with the Russell 2000, uh, slightly below the NDX for the weekly move.
And again, for everyone focused on policy mistakes, and that was the, that was the word of the week this week, stagflation last week. Um, for all the ex-people posting and everything about policy mistake, uh, and tenure rates, rates are down since Powell did his Jackson Hole, despite the fact that most of the data has suggested things are stronger. Uh, bond yields, the white line here, really falling sharply and getting back down here again. If there was a policy mistake and fears over what the Fed was doing, we wouldn't be seeing bond volatility breakdown.
Uh, at the same time, we've got the triple B spreads down near the all-time lows and junk. This is a monthly chart going back to 1990. We are easily at the tightest levels ever. So, not only is the economy good, the Fed is cutting rates. So, the playbook should be simple. Be bullish. Jump into things.
Uh, momentum has soared higher. I think it was right around here that I highlighted again for about the fourth time that we had MACD sell signals. I think I've done this about six times over the course of this year, looking for some change in momentum on the back of PMIs going higher, and momentum just continues to power, uh, dominated again by the AI, AI theme. And even though the Russell had a good week and outperformed the S&P, the pure size factor, which strips out the sector biases, uh, continued higher as well. So the factors on the pure basis continue to move, and I think one of the dominant reasons is this.
Um, retail had a party last week. So funds and institutions yelling for policy mistake and sell the news and everything I had heard leading into the Fed. Uh, here is retail's favorite index. This is the Goldman Sachs Retail Favorite Index. This is the meme index. This is the most shorted, uh, for, um, for hedge funds index, and this is the non-profitable tech. So again, you can sit there and say this is retail is going crazy, but retail's just following the playbook, which is they're cutting rates near all-time highs. So Powell and them, they, they cut rates and they've shifted to focusing on the job market, which I completely agree with. And there's a variety of reasons. But again, we've got people saying policy mistake. And again, just like I did with stagflation, this word means nothing. It gets thrown around. Policy mistake. I, I just don't even know what it means other than I'm right, you're wrong. This is a mistake. That's not the way you make money in markets.
Um, just something for the future that's been brought out this week in his congressional testimony. Uh, he did mention the third mandate for the Fed, uh, which again, we're starting to get back into this thing, which I think people should again be focused on. And this gets into this concept that the Fed chair will be pushed out and the new Fed chair will come in, and Trump wants long-term rates down. You can argue with whether it's a mistake to do that, but that's what they want.
Um, and just to show how long this question has been on the third part of their congressional mandate, uh, Powell said, "We haven't thought about that for a very long time." But Long View, one of the better macro, um, posts out there, is highlighted three times over the course of the last seven years about this third mandate. So, I think it's just something to keep in the back of your mind, uh, as we go through this.
Now, it's not just one rate cut. Right now, we have a 92% chance of an October rate cut. We have an 88% chance of December. So, again, don't get caught up in this quote unquote policy mistake. And most of the policy, policy mistakes again that I hear are inflation. And I'm just going to continue to say, here's inflation year-over-year, 2.9%, basically within the range of where it is. We have rates higher. I've shown this before. Rates should be coming down. Policy mistake is a bet and belief that inflation is going to go higher. If it wasn't for this during this period in here, people really didn't focus on inflation, even when we had these little up moves in this point.
Now, next year, I do believe we will see close to 4%, but only because I believe oil will surprise and go higher in the back, and PMIs and that gas on the pump will go higher. This is just showing that it's really hard to get headline inflation to come over. If we had gas at the pump up 20 to 30% year-over-year, putting it up basically at $4 a gallon as opposed to $3.20, then you'd see this up near four, if not slightly higher, and I don't think the Fed would be cutting. So, I think that's an important part to look at.
The reason the cuts make sense is because, and I'm going to write about this this week, we are at a very, very K-shaped economy, and the Fed's job, if inflation is not soaring higher, is to deal with this situation, which is we're in recessionary territory for job creation. Everyone can go through their explanations and immigration and all of this stuff. It doesn't really matter to me one bit. We have a problem here where there's no job hiring going on. It shows up in all the statistics about hiring. Again, AI is coming. It is only going to worsen in the coming years in terms of the pressure on the labor market, and also the deflationary pressures that will be coming from AI three years from now. They should be cutting rates, and the administration, which is AI-focused and has every AI leader in the world in the White House on a consistent basis, including David Sacks sitting there as the AI, AI Czar. They know what's coming. You guys know what's coming, but we're still focusing too much on what has happened over the last year and what you think may happen with inflation, which is just not going to happen without gas prices going higher. That is the weak link in AI is oil and power needed.
Retail sales came out. This is overlaid with nominal GDP. Retail sales, basically close to the highest level since 2021. So they're cutting. We have strong retail sales year-over-year, which was not what people were expecting. And that's because we do have a K-shaped economy. There's two economies right now. Here's what's happening in terms of after-tax wages and salaries by household income. This is the impact that is going to continue to be pressuring because of technology and because of rates being at higher levels. The stability in the economy just continues to shift.
And again, I wanted to show this just to show, um, the charts of just how much of consumption in this country now is from people above the age of 55. This red line here is the percentage of people above the age of 65. This is a. So when you get through these numbers and you combine them, these are massive numbers of people that are basically leaving the labor force and spending money.
I showed this back in August. This continues to go higher for people who keep sending me things about PMIs. I don't know how a hedge fund person in the world cannot be focused on what is happening with inside the commercial side due to data centers and be constantly pressing back on PMIs that will change going forward because of the demand and the shortages that are showing up. So the Dodge Momentum Index soared again on the commercial side. All sectors sustain momentum over the month, notably led by strength in data centers. If all data center projects between 2023 and 2025 were excluded, planning would still be up 38% year over year ago. The largest commercial projects include $500 million Big Sky Data Center, battery storage, data center, blah blah blah. We've got a tremendous boom going on in AI, and Trump is fixated on the manufacturing and bringing stuff back for security purposes. So, team explores government-backed manufacturing boost housing. I've talked about their desire to declare a national emergency. Expect that to come.
And here's what's happening already. They're bringing down mortgage spreads. So, 30-year fixed rate mortgages, lowest level since late 2022. This is inverted. Here are the mortgage apps. This orange line here is the mortgage applications. You're starting to see it flow through.
The fact that the Fed is about to cut rates while animal spirits are rampant brings back the post-LTCM easing cycle. Um, I only bring this up because cutting when there's not a recession is a big deal. Cutting when you're within 2% of all-time highs to shoot against it the next year, going back the 20 times, 100% were up the following year. Could it not happen this time? Of course. But when you read through, cutting at all-time highs is not something that you would expect to fade if you just talked about it. For some reason, people are focused on the other side.
And it's not just a US thing. MSCI ACWI All Country World Index continues to make new highs with 75, 74% of its constituents trading above the 200-day moving average. So technically, it's there. It's complemented by earnings momentum and breadth. This is not something minor, and revisions ripping higher still. So this 10-week average is now up to 0.22. We still are going to have some other things fall out which are going to continue to keep it. This is a PMI overlay, a tide that lifts all boats.
Now, when you go through FactSet, so this is Trump back in 2018, once we started to see the earnings calls not mention tariffs as much and we saw the second derivative, the rate of change improve. This is when stocks bottom. This is when they were bad during this period. This is when stocks were bad, then they make a bottom, then you go. The rate of change is what matters. People adjust, companies adjust. Uncertainty. Remember how we got all these uncertainty things? How big of a deal it was? It was. And stocks went down. Now the rate of change is coming down sharply on uncertainty. Inflation. Remember policy mistake in the earnings calls. Inflation completely coming down still, despite the tariffs. This is probably the most important one. Everyone talking about tariffs on inflation and everything else. We still have all of this coming down, and here's the way it looks overlaid with CPI. So again, we don't have anything going on, and without the tariffs doing it, which is shocking. You're going to need gas at the pump to go higher. We are going to have inflation for transformers, for turbines, for generators. I'm going to go through this stuff later, but for the most part, don't worry about it.
Everyone who talks about the US market being overvalued, of which it is, a drumbeat of macro people talking about valuations without mentioning margins, without mentioning return on equity. This is how long this has gone on. If you've been bearish US equity market since many, many people said it was overvalued, you've really paid a price watching this go this way. So 91% say it's overvalued. Not the best way to do it.
And here's the sentiment. So, this was at the beginning of the week, uh, September 12th. I'm just shocked that we're not up here like we were when the market peaked. So, again, try to figure out ways that people are not bearish. I wrote a, uh, a Substack and two pieces for 22V this week. And one of the things I highlighted was this AAII bear. This is the five-week average. It's just shocking to me where we are on this. So what I did is I took all the times over the course of the last 23 years. These, these vertical lines here are the points that we hit that level. In every case, the stock market had fallen. This is the only time that you have bearish sentiment like this with stocks ripping higher and at all-time highs. All of these were after either major corrections, major corrections, not small ones, 15 to 20%, or the beginning of something really bad, or some consolidation that had lasted a period of time. So again, that one's there.
And then this one I highlighted just in one of the papers as well. This is where the, um, bullish came out last week. 55.4. There's two points here, two points here, and one point in, in here that we were this level or lower. And in every one of those cases, it was after a 20-plus% fall in stocks. Again, people are not positive. This is not just an equity thing. This is not just an AI bubble thing. This is actually about the state of the economy we're in.
I'm writing a Substack right now on the, the convergence of the Fourth Turning and Joseph Schumpeter's creative destruction. And instead of sitting there predicting when this will go, I am the most bearish person because of AI. But that's just further down the road. For the time being, we're in the game of making money, and sentiment remains way too negative for where we are right now. And the fight, the current fight the government's trying to do in terms of running the economy hot to get us out of this while the acceleration of AI is directly happening now.
So this article came out, and again, all the posts are about how this is a sign that the AI bubble is happening. The Economist is known to write things when a bubble's in, except for the fact that The Economist framed the $3 trillion piece. I asked ChatGPT to go through the article and on a scale of 1 to 10, with one being bearish, 10 being bullish, is this a bullish or a bearish headline? Uh, four, closer to bearish than bullish. Basically, it goes through the fact of all these things.
So, when Oracle came out last week while I was in Vegas, I figured when I came back and I did all my weekly conversations with people, that they'd be shocked at the Oracle numbers. Instead, they're all talking about this stuff. They're never going to, how are their margins going to do? They're in, it's a, it's a bubble. Sign of a bubble. This, we can't trust Larry Ellison. They don't, they're going to have to issue debt, the valuation of the company. I can't believe what I listen to. Guys, this is about the demand side of AI. You can doubt forever and ever, but markets tell you if you're right. If there was really a problem to start with with the debt side, as this article goes through, you're seeing all these massive capex numbers. Okay, you can either invest in it because it's going to happen because we're only here, or you can avoid it. These are numbers that are going to happen, and it's not just the US, it is around the globe. And so what would be happening if people were worried about the debt? The spreads would be widening. Junk spreads would be widening. Something would be widening. Not sitting at all-time tights. So, let the, let the market tell you when it's time to doubt it, and we'll all get bearish together.
Now, I wrote this paper at the beginning of the week because of the disbelief that was out there, and it goes through a lot of different things that I talked about, but it really goes through the numbers and just how important this is from an economic basis.
So, then we get the Nvidia Intel deal, and I write this one, which is all about the new semiconductor inflection. So there's two data points that have come out in a week. One is Oracle reports a number and talks about inference. Then you have this, where again, two large semiconductor companies are signaling what the next stage to invest in is in AI. Now, on the first paper, I wrote about inference back after Q1 earnings reports, and I published the paper on May 15th, that it was the inference moment. In there, I highlighted 22 companies that would benefit from the growth in inference that was starting to accelerate. Now, this was all based on about 13 to 15 companies that I listed where there was direct quotes in their earnings reports on how much inference was starting to pick up dramatically, which again, is a sign of adoption. This is why when these reports come out, they are meant for clickbait. They are meant to go through. If you do your own work and you just believe that the market knows, here's May 15th when I released it. These are the names that make it up. They're up 30%. Since May to September, that's four months. That's an annualized 120%. The S&P is up 13%. If you just follow what is out there, I'm not a bottoms-up analyst. I'm just using AI to help me find things.
On that note, a master prompt. This is only a piece of it to go through the supply chain analysis of the AI inference buildout catalyst. This is for me to come up with 25 to 50 names as to what is going to be part of this. This master prompt that I do ended up being five times the size of what you're seeing here to go through.
So then I wrote after I did that one, now connect the announced Nvidia Intel deal to the statements from the company. So basically, I'm taking the Oracle data point and what they said, what companies across the spectrum should benefit, and then connected it to the Intel Nvidia deal to give me then two data points to come up with this. So I just want to take you through where this all comes out.
So the Nvidia Intel partnership can be matched up with the Tesla Samsung partnership. These deals point towards edge devices. If you don't know what an edge device is, spend time on AI asking it. You have to learn it. This is going to be the next stage. I've talked about this, the phases, but we're going to go through it now. Training was the first one. That's the one where Nvidia really benefited the most. This is the one mainly about GPUs. Then we get into the inference scaling. We're still using GPUs, but now we're broadening it out. You're getting into other components. And this is what you're starting to see now. The next phase, which again, is going to start next year in my opinion. It's one of the reasons I wrote the Tesla piece. It's one of the reasons I'm focused more on the power side now, is the embodiment and edge AI. These need the brain. So all we're doing right now is teaching this thing to learn everything from high school and college. This stage is where it's learning to think. It's taking the knowledge it gained in high school and college and it's making decisions by thinking through. This is the adoption phase. This is where the AI agent side is going to accelerate. This is when you need to have the brain. These are cloud-based. This one cannot be cloud-based. You can't have humanoids going to the cloud to get their, uh, information and then come out and make decisions. So, there's different things that are needed here. But the new term, not GPUs, not CPUs, NPUs. CPUs. You guys all know what those are. Those are very big before we got to the importance of GPUs for the cloud. Well, now we're going to get into NPUs. This is for the thinking on the spot. This is the decision-making. This is the thing I talked about with Tesla. The pushback that came on Tesla. If you don't know what an NPU is, you have no business talking about Tesla. It is not a car company. It is an NPU company. End of story. Humanoids, robo taxis, that's it. So, if you're negative on Tesla, go study NPUs so you can at least have an informed opinion about what the man is trying to do and how important it is for the next five years of their business and why he could get a trillion dollars because of NPUs.
As you go through this, this is really important in terms of the edge devices. And this is not just humanoids and robots. This is robo taxis. Anything that makes a decision. This includes your phone. This includes the computer. The reason we can't have Siri work the way it should, it's really hard to go back to the cloud and come back down in information. You want the thinking to be done on the device, that's what we're getting into as the next phase. And this week, Meta launches smart glasses. This is all part of the same thing. This thing needs to have a brain in it for it to do what it needs to do. It's not there yet, but we're going to get there.
Now, one of the outputs on that, and I did another routine. If you guys want to see the names on this for this next stage, John Ro went through, did his technical scoring. I went through and got the output of all the different names. There's a lot of names on here, and many of them because of the NPUs. So, NPUs need CPUs. CPUs are the things that went in washing machines, in autos. Again, guys, think PMIs, PMIs, PMIs. This is where we translate the next one.
So here's one, ASML, one of the major things they are important in this. They're one of the names on the list. Is this what is critical for the brain and the machine? And this is, yes, EUV lithography, where ASML has a monopoly. I love monopolies when I'm investing. Is absolutely critical for producing the advanced chips that serve as the brain in AI-native devices, robo taxis, and humanoid robots. So it would make sense for ASML to be basically unchanged since 2021, while the PMI has been unchanged. It wasn't needed as much because we were in the cloud. We didn't need ASML chips as much. We didn't need them to be doing what they were doing. So now you're getting more their back order backlog of the minute stood at third multi. You can just read through this stuff on your own. It's also necessary for batteries, which I'm going to get into as well.
So the impact on AI-native devices, robo taxis, and humanoids. If you wait for this stuff, it will already be higher. And the reason is because retail's already involved in this. This is the thing. When I hear and I meet people on the retail, they're really good traders. They're communicating via social media. They're finding these names. They're doing the same thing I'm doing. So, if you're sitting there as a hedge fund, as an analyst, and you're wondering how retail could possibly be right, they're putting in the work and they're doing the homework and they're going through the same thing that I'm going here. So, if you want the information, reach out to the 22V people. I'll go through it. I'll even show you how to do the things that are on there if you want. If not, you can wait for Goldman Sachs and Morgan Stanley to actually produce the research. It's never going to come because there has to be a timing element. And the timing element has to do with the trading side. The trading side is getting away. And it looks like they're expensive, but that's what happened to Bloom Energy. That's what's happening with a lot of the names that I showed last week and things like Babcock and Wilcox. These things just move and they're gone. And you have to be on top of them because they start moving before the actual data comes through.
Can you talk about the importance of the battery for both the inference, scaling, and embodimenting stage? So the next stage that I want people to really understand as I segue, batteries are necessary in the same way the brains are. You need a lot more battery and a lot more inside the devices for them to be able to run properly. Unlike data centers, edge AI runs off local battery packs from smartphones to EV packs powering robo taxis and humanoids. Power per inference watt efficiency. This is critical as you guys think about the next phase.
So Oracle and talks with Meta on $20 billion AI cloud computing deal. So after all the stuff with Oracle last week where people doubted it, there's a conversation about another deal. This one with Meta. At the same time, Meta pushes into power trading as AI sends demand soaring. So the power side, where are you going to get it from? Meta is seeking authorization to participate in the wholesale power trading business to manage its data center electricity needs. Again, the extreme nature of how much of a power shortage we're going to have is why people are doing this.
Now, a report that I went through post, uh, Oracle from myself, surging backlogs and unprecedented forward visibility. Okay, so just the backlogs that are there, a global construction boom for AI factories, massive capex outlays, specific large-scale projects. This rapid expansion is hitting physical limits. Power availability is now the primary criterion for site selection. We are, we have bottlenecks building up everywhere. PMIs are going higher. PMIs are going higher. Next-gen hardware rollout. So the hardware procurement on this, the supply bottlenecks, they're everywhere. Their gas turbines, their generators, their transformers, their cooling systems. And the backlog is going further and further into the future.
We also have DRAM. So, if you've wondered why Micron, which again I referenced a while ago when I first got to it to, uh, 22V and I've talked about on Pomp many times, has reportedly been forced to freeze pricing on DRAM and NAND as AI demand causes shortages. Lead times for high-capacity hard drives have extended to a year. The demand is reflected in the bullish forecast from key suppliers. It just doesn't stop.
Now, most people have heard about this. Gas turbine manufacturers expand capacity, but order backlog could prove stubborn. Gas power boom sparks a turbine supply crunch. So what I did was, I know from everything that I'm reading that the situation on the turbines is taking us out five years. Everyone I talked to who's invested in it, who who's invested in GE Vernova, Siemens Energy, Mitsubishi, they know this inside and out. But the real thing is that they don't know AI demand is going faster than we expected. That's what the Oracle news was. Instead of fading it, the demand is going far faster than we anticipated, and the products keep coming out. So the shortage is getting worse and worse. The demand is growing. When demand is growing faster and the supply not only has a bottleneck, the turbines cannot possibly meet the numbers that people want. So natural gas is not going to be the story. In fact, what I wanted to do is say, okay, putting all together the shortage that we have with the demand, what will gas turbines solve out of the entire equation that's likely necessary from the high case? We are tracking right now above the high case. Roughly one-third to two-thirds of the high case. So, let's assume about half is going to be from natural gas from the turbines. That means the rest of it is coming from these other areas, which means the most scalable near-term fillers. New nuclear is a post-2030 contributor. So there's a tremendous amount of places, and this is why I wrote this report on hybrid solutions.
And just as machines only delivered efficiency once they were plugged into electricity, AI only realizes its potential when plugged into the vast compute and data infrastructure that's needed. So for everyone looking for ideas, just stick with compute and stick with power. It will get you to the endgame. The problem is you need to talk to people that can speak both languages. You need the demand side on the AI side because if you talk to someone on the compressor side, they say, "Well, it all depends on what the AI demand side is." Well, to get the AI demand side, you got to talk to a technology person and actually get the Oracle thing. Since most people are negative on this stuff or don't understand that this is a national security issue that goes far deep and something that they can't turn around on and it's funded with cash, there's no signs that this won't be done.
Now, outside of, outside of shortages and transformers and cooling systems and turbines, there is a shortage of large-scale generators as well. So, if you're wondering where there's many spots, and I've listened to a bunch of podcasts this week on this, it's a perfect storm as is in here. Uh, we need to be able to have backup power, and this is going to mean generators. And so, what you've seen is so far both Caterpillar and Cummins have strongly highlighted surging generator and power systems. Now, Caterpillar is also, um, getting other things from the AI side. And I wanted to show this because Caterpillar is purely an AI company at this point. Uh, their power generation revenue is now 14.5% of total company sales. And if you go on the growth rate, you're dealing with 28% year-over-year. These numbers are only going to grow bigger and bigger. So you have to make sure that you understand the energy and transportation segment where data center generators recorded outpaced construction machinery and revenue for the first time in '24 and '25. This is why the PMIs haven't gone higher yet, but they were starting to go higher. It takes a while for the data center construction numbers to get above commercial real estate, which is what's happened now. Same thing goes for Caterpillar. So at some point here, the sales start mattering for these companies, and they start to be driving, and the same thing will happen with PMI. So if you've missed Cummins, here's the chart of it. It's an AI company. You can go look at where it is now and where it was a year ago, and how these companies are making money in an environment where there is a recession happening in PMIs. This is a scary thing to see that this is happening, and it will spread because it's a tide that lifts all boats because the bottlenecks are growing.
So with Oracle's numbers last week for demand, will this put more pressure on the grid? So I expanded this and said, well, what other generator companies? I did another list. I went through and I looked for small-cap, mid-cap places, and maybe ones that wouldn't be the first place to go. And Generac came up on a bunch of things, uh, mainly because that's not where their traditional business has been. So where you go is then you start going through it. Data centers, obviously, are prime target. Also, our existing customers, they all require larger blocks of backup power as well. So if you've been historically not talking to the hyperscalers, but you were talking to the utilities, you've already seen the demand. So you can go through and read the quotes on your own in terms of what's already happening. The market is in an early cycle of very strong long-term growth driven by AI. The capital expenditure cycle for data centers could extend for many years driven by the AI trends. A robust pipeline and the company's first major international shipments, first, you're getting all the information you need. Um, management repeatedly emphasized that demand for transformers and backup power is a critical concern for data center developers. They cannot have these things have to run 24/7. We have to find ways to use generators to make sure that happens on-site. It has to be from the grid in terms of making sure that peak demand can be extended. They're all types of grid optimization side. So if you look, here's Generac's chart. Good from the lows, but basically it's unchanged since November. So in looking for names, go through and see where you can find them.
And that brings us to the next thing, and I showed this last week. So when I say Tesla is working on the brain for humanoids and robo taxis, they're also bringing the Mega Pack. Tesla's new Mega Pack system can power 400,000 homes in under a month. Tesla is not a car company. They are an AI company. Everything that they do is related to the brain and to making sure there's enough power. And batteries, which is a focal point of Tesla, is now starting to become a major trend because we're entering the stage where batteries become most important. Everyone has probably touched batteries before, got invested. They're very speculative. We need batteries now like never before, and it's only going to increase going forward.
China to supercharge energy storage tech with world-leading advancements. I highlighted this yellow stuff there just to make sure that I knew what it was. Says the items circled in yellow in the image are large-scale battery energy storage containers. Think Mega Packs. Think what's going on? China's battery makers get boost on hopes for advancements. We will get advancements in batteries. Batteries have to go through an innovation finally. There's going to be an enormous amount of dollars thrown in the pace, and these names are going to be a place you want to look. So CATL, the biggest battery maker in the world, jumps on JP Morgan upgrade on the China energy storage plan. Again, PMIs peak here, going higher. I can do this all day with the PMIs. Um, here is the Chinese battery company. And now all of a sudden, you've got LIT, which is the ETF for batteries in the US, which has been correlated, still correlated, but now you're starting to see the battery thing show up in the US. So, if you go into X and you go follow retail, what's retail looking at? Well, they're looking at NRGV as one of them, which has ripped higher in the last two days. And again, this will show up as speculative memes. This, that the reality is, if you go do your own homework on these names, you will find they're not. They are doing stuff in AI.
I mentioned this one a couple times, uh, in public speaking engagements, and I talked to someone in August about the fact that this name kept showing up on all of my podcasts for energies and batteries, and who was already working on this, and it was this company, EOS. Uh, EOS was up huge this week. It's up big. I, it could be speculative, or it could be the fact that EOS executives have stated on earnings calls that big hyperscalers and developers are coming to EOS because of its American supply chain, cost-effective relation. It frequently cites the rising demand as a driver of EOS pipeline. And then here's their backlog. They talk about their backlog saying they're seeing strong signals ahead. And then this week, they just announced a new major platform called Dawn OS in September. It's purpose-built for the battery systems designed to enhance advanced control, automation, and system optimization for grid systems. EOS emphasized that Dawn OS was built based on direct customer feedback and is specifically designed for American security, energy security. Um, again, the reason I'm bringing this up now, this was September 11th, before the big rally. I mentioned this stock. We've talked about Cummins. I've talked about Bloom Energy. Here's EOS. This is someone who barely has any views. Had 805 views. Has barely any followers. This is the way the information spreads, guys. These names, they're following what's happening. They're paying attention to the deals that are being done.
Um, oil again, some point next year, I believe oil will be high, especially while everyone is negative on it. Uh, why is China stockpiling so much oil? I think every country is stockpiling BTUs. Um, so energy will be a major theme next year because we have bottlenecks in everything.
China bans its biggest tech companies from acquiring Nvidia chips. This was a big story this week. David Sacks highlighted why this, or how this is a big story, and just showed that, okay, they're not needing to buy our chips. They're not buying our soybeans, and they're still messing around with the world on rare earth. Companies in the EU incurred seven production stoppages in August because of the shortfalls of rare earth. All this feeds into the deal with China. I'm going to reiterate again since you're looking for information on what things will happen between now and the end of the year. China is clearly showing their leverage before the negotiation is done. And Trump is saying he will meet Xi in APEC. Uh, this is what Andy Rothman had talked about, basically, and wrote in on July 9th. The odds of a deal, the world's greatest deal, uh, are more likely to close a deal around the time of the APEC leader summit in Korea. So Andy's been talking about this. Uh, a deal is not a sure thing, but the odds are higher enough that investors should think about how they would take advantage of the potential impact on Chinese equity markets. I find it impossible, given the importance of rare earth, uh, that there isn't a deal. Uh, we don't have enough rare earth. I'll get into that in a second. Uh, if there's a deal, and if you want to start looking for options and things to play as you get closer to that point, ChatGPT 5 Pro is really good at giving you a bunch of places that you can go put some, go look and see where the options are. Go see where things are. Again, you've got all these different places of places that you can play. One of them is semiconductors and equipment. If there's a deal, just, just suggesting US dependence on China for rare earth elements sparks security concerns. So knowing that we don't have the labor, we don't have the expertise in mining, and we have permitting issues along with many environmental delays that would occur, where are we going to get our rare earth from five years from now? So a deal, I'm sure will be that they will supply our rare earth, but this is going to be a deal that obviously is going to end at some point. And so where do we get them? So the countries again, Australia and Canada. Great. I wanted to find countries that had the labor, they had the ability to get this done, and they had the mining experience. The ones I want to focus on are Brazil and Chile. We'll get into them later, but there are a number of US companies that are developing, deploying. All this work again is in AI. Where can I find companies that are in the forefront of rare earth recycling? American Resources Corporation. The company is focused on the critical midstream of the supply chain, separation and purification. It's sub, blah blah blah. American Resources Inc.'s with PASC to boost rare earth supply chain. That was this week. And here's what the stock has done. It's a small-cap stock. Again, you're going to see all of this stuff go on. And if you do your homework, you can find these names. It's the gift that keeps on giving because once they get orders, they continue to go. AI is not stopping for the next four or five years. And there's no surprises that are coming other than the fact that we could run out of power. And so energy plus AI equals a good way to make sure that you're covered.
Now, the metals. John Ro has been highlighting this stuff. Commodities, commodities, commodities, but in particular, we need tons of copper. You've known that. The chart looks great. He's highlighting the fact that that's going higher. The Bloomberg Commodity Index, same thing. So, what I care about. Oh, we're back to Brazil because when commodities are going higher, I want to be long Brazil. Having lived there. But also, I wrote this piece on a lot of different, uh, reasons. A lot of it has to do with their rare earth minerals.
Now, we have the LME metals. Okay, so if copper is going higher, this will go higher. This is the relationship historically between EWZ and this part of my belief was that the dollar would be on a weakening trade. If you believe the dollar is going to be weaker, if you believe American exceptionalism is over, and I do believe the American dominance of coding is over, and the dollar was representative for 30 years as the monopoly on coding, which has ended with artificial intelligence. So I believe we're in a long-term shift here where the democratization of AI means that the Mag 7 don't have that software, doesn't have it, and we get more into the hardware stage. EWZ, and if you don't like EWZ, just look here. It's a great year for the S&P, up 13.3%. The Mag 7 now up 20%. Sorry, Mag 7. This, this is year to date. And here's what we've got up here. These are all up 34 to 44%. This is Brazil. This is China. This is Mexico. Uh, and this is Chile. That's where I'd be focusing my attention.
To finish up, Bitcoin chart's still great. It's still stuck. Um, we're doing the usual routine, and I hate to say it, but the hedging that seems to go on to the end of every quarter until we get out of the quarter, it seems like it's stuck in this 110 to 120 area. Eventually, it will go higher. And part of the reason that it's in a structural bull market is this one. A majority of fund manager survey investors are not allocated to crypto at all, with 60, uh, with August, September. Now, 67% have a 0% weighting. Uh, 8% or more. Next to know it. That's it for this week, guys. Have a good one. Reach out to 22V if you want to talk about any of the work that I showed in here and if you want to start having some, uh, classes on how to do the work that I did on here. See you.