Transcription
week with Donald Trump as president. Uh, Fed independence growing, uh, your way out of it. Eric Schmidt, the natural limit of AI is electricity. 10,000 times demand versus supply. The real AI race, inflation or growth, and crypto week and all breakouts. Let's get right to it.
So, uh, S&P for the week, uh, is now consolidating, but again, able to make new all-time highs. Chart looks great. Um, still a lot of uh sentiment that I would say is worried. We don't have extreme bullish sentiment. We're still kind of middle of the pack. Uh, and this week we got rate fears again in terms of uh 30-year yields around the globe breaking out. Uh, but 10-year yields still sitting here. So, really, it's been a 10-30 steepener. I'll get through some of that as we go on.
Fed rate cuts, we still have better than 50% chance built into the market for September. Uh, and if you go out to the end of next year in December of '26, we're talking uh significant five rate cuts almost still built in by the end of next year. So, uh, as I go through this and talk about growth, government's focused on growth. Uh, so rate cuts into growth, rate cuts into no recession, just very powerful for stocks and liquidity. Uh, so at this point, as earnings are better, uh, there's really, you know, there's there's just a lot of good stuff going on, and I don't see a recession uh as a problem. So I think it's more likely that inflation is the next scare, but we'll go through that.
So, uh, dollar, it's bounced now for really about two and a half, three weeks, uh, but dramatically oversold and positive liquidity boost, uh, with a delay of about three to six months. So, still coming through. Um, CPI this week. So, I just wanted to show this. These are the last five. These yellow dots are where the actual number came in. These histograms are the economist forecast. So, you can see it's five months in a row. Uh, this one getting a little bit closer to the center, but still, this was supposed to be the tariff fears on this, and there is tariff uh inflation within there. The inflation data has been weaker than expected, and it should be no surprise that Donald Trump went on the war path again, uh, for for the Fed and Jerome Powell, uh, specifically now choking out the housing market.
I think this is interesting. I've talked about the fact that I think they will do something to help the housing situation, but one of the things they want to do is uh get rates down to 1%. Yes, 1%. It wasn't just Trump, though. You've got PY continuing. Uh, they're trying to use the building out of the innovations for two and a half billion as their Rick Scott all over him. Uh, then you had this on Wednesday asking GOP lawyers, this seemed to be a trial balloon to see what the market reaction would be. Trump likely to fire Powell soon. White House official. This reminds me of exactly what happened before the 90-day tariff side. In that case, they were wondering what it would do. The market ripped higher. In this case, I think they wanted to see whether the market would respond better to one, you know, saying moving rates lower while the economy's strong, and obviously the market saw the dollar come down, you saw rates go higher, and you saw stocks come down. So you had bonds, stocks, and the dollar down, and, uh, it seems like, uh, they're backing off, and he said later in the week, they're not going to fire him.
The question is, and where this gets into it, can the Fed stay independent? Trump advisor may put it to the test. Uh, I'd say it's already done. That's my own opinion. Um, fiscal dominance and financial repression ahead. I put this in there because again, I think the goal here is very simple. The mounting interest rate costs associated with serving the massive government debt. It's going to get worse going forward. He already had to put the one big beautiful bill in place. Uh, got a lot of conservatives very angry that the spending was there. So he's going to be focused on growing and somehow another getting the deficit down. Some of that's going to be through the tariffs, which is the aggression on that. Uh, and then the other one is to make sure that rates are coming down, uh, to service the interest costs. He'd like them down now to try and get this going ahead of the midterm elections and have something to show for it, but he's probably going to have to wait. Uh, but regardless, there's a shadow Fed. Now, Wall Street hates the idea of a shadow Fed. It may already have one.
I want to basically get into people need to think about this not from, uh, a what they think President Trump should do or what they think the Fed should do, but just on the rules of the game. Um, you got one of the potential, one of the four, uh, listed candidates, uh, saying, uh, he might dissent at the next meeting, thinking there should be a cut because the labor market is soft. I'm going to continue to say the labor market is going to be an issue. Uh, we got some more details on this this week in terms of just how long it's taken for people to get a job, particularly for 20 to 24-year-olds. Uh, there's no doubt that we're getting a change in the jobs market from AI. I've talked about this being less about job losses in terms of people getting fired, and more about a combination of less hiring or no hiring from the bigger companies. Uh, startups happening which aren't going to create lots of jobs because they're using AI, and you're going to be left with this, let's say, not not confident, uh, that their jobs are safe. They're going to be living in a world of paranoia, wondering when they're going to lose their job. That just makes the confidence in the job situation where you're getting paid, but you don't feel good. So, I think that's going to be a story, and I think that's one of the reasons why, uh, whether the Fed likes it or not, I actually believe they should be lowering rates at this point because if inflation does get higher, I don't expect it to go too far, but they can always turn around and raise rates. But again, not what they should do, it's what they are doing, and what President Trump is doing, and what Scott Besson is saying. Again, we can grow the economy and control the debt. That is the what they're going with. Um, and whether or not they're able to succeed, that's not our job. Our job is to make money on the markets. Uh, and I think, uh, this podcast this week with Raul Pal and Julianne Pal, uh, is worth listening to. Very few macro people, uh, that I listen to on a regular basis, uh, are able to kind of go through the quote unquote no recession thing. The fact that we've transitioned from a credit recession related time to where the government is just too big, and where innovation is exponentially moving. Both of which have occurred since the great financial crisis. But Raul Pal and Julian and I are all in agreement on this in terms of what's going on. So rather than only listen to me on this, they have a different, um, approach towards this. I think it's worth a listen to. This podcast was very good. You won't hear many macro podcasts, uh, that I think focus on the things that are going to help you make money at this point, as opposed to trying to scare you.
I talked about the PMIs last week. Uh, we got two of the early two manufacturing data points, and I just want to reiterate this again. I believe the most important theme for the next year is going to be that manufacturing is entering a very, very big boom period driven by AI and the infrastructure buildout. Empire Manufacturing came in 5 something. Here are the histograms on the forecast. Way above consensus. Philly Fed came in at 15. Way above consensus. Again, I just think we're going to start, and at some point, whether it's this month, next month, the month after, I do think we're going to see a big jump in the PMI. And it comes at a time when people are worried about the trade war, or continuing to focus on it. This was something people should spend more time on. Wedbush made a commentary on, but this is about the announcement that was made in Pennsylvania for a $90 billion plus focused on the technology and energy sectors. $36 billion for data center development and $56 billion for energy infrastructure projects. Wedbush noted the scaleless public and private investment expected drive a super cycle of growth for companies positioned within the AI energy and digital infrastructure ecosystem, underscored that the buildout will lead to tens of thousands of construction and permanent jobs with multiplier effect expected across local economies of data centers. This is what Donald Trump wants. He wants to see the manufacturing sector pay. We are spending a lot of money. The hyperscalers are spending money. I believe you're going to see something similar in terms of the impact it had to the fracking boom, the various parts of the economy they're going to benefit from this, and then from an overall dollar amount and impact to commodities and impact everything. I'm going to continue to say this is going to be similar to the China urbanization story, at least from how people are going to have to look at it.
Best podcast of the week, if you listen to one of the ones that I give you. I've got a bunch. Moonshots with Eric Schmidt. Definitely worth it. Eric Schmidt is by far my favorite person to listen to on the AI buildout. He's not only articulate and informed. It reminds me of Michael Milken. He's just involved with so many people in conversations on a regular basis. He doesn't run a company at this point. It makes him unique where he's just gathering information. He seems to have the brain to be able to consume tremendous amounts of data, and he's great at connecting dots. In fact, one of the best I've seen. He doesn't have a big side of him. He is the one that has said repeatedly, AI is definitely underhyped. This is the thing that's most important, uh, in terms of, he said this at the very beginning, in the first two minutes of the podcast, AI is a learning machine, and in network effect businesses, when the learning machine learns faster, everything accelerates. It accelerates to its natural limit. The natural limit is electricity, not chips, electricity, really. The bottleneck in AI and acceleration is the available of the availability of energy, particularly electricity. I've said this all year. I'm going to continue to say it. You have to be focused now on power, and not just the things that you've read about. And the reason is, he talks about this explicitly. In my recent testimony, I talked about the current expected need for AI revolution, US. It's 92 gigawatts of more power. One gigawatt is one big nuclear power. There's essentially none being started now, and there have been two in the last 30 years. As important as nuclear, both fission and fusion, they're not going to arrive in time for us to get what we need as a globe to deal with our many problems and opportunities. So, we're not going to be able to do it with the solutions that people seem to be focused on. And that's his point is that this has to be other things. And so he goes on to talk about the energy sources that are going to be there. Without going through this, the computing needs that we have are not going to come from, or are now going to come from, traditional any energy suppliers in place like the USD world. This means fossil fuels, and he specifically said from the fossil fuels, natural gas, hydroelectric, and traditional suppliers. Again, I'm going to say Exxon, Chevron. I'm going to talk about all the buildouts of these things, the things where we already have stuff. You're going to have to see every source of electrons of energy used. Everything is going to be needed because we cannot possibly get the power that we need at this point. He talks about the fact that there may be solutions later to solve this that could speed up, meaning ways to get more energy, but we're not there yet. So, he goes through the arrival of super intelligence, AI's insatiable energy demand, again, goes through that. The hardware and computer arms race. Talked about this, and he went through the software in this. Andy Grove giveth, and Bill Gates taketh away, meaning software always eats new computing capacity. I'm highlighting this because the next time you get to the point where there's a sell-off in AI, which I think is going to be less likely to happen, meaning the deep sea, uh, side here, where there's fears, just remember underneath, you should be staying with these things. The demand side is going to be much, much bigger than people believe, uh, even at this point.
Just going through this, just other things in there. I'm not going to go through the reading of it. You can go listen to the podcast. Uh, another podcast, uh, that kind of reinforced the energy problem. This is from, uh, Andreessen Horowitz's podcast, A16Z, came out this week. America's Energy Problem: The Grid That Built America Can't Power Its Future. Again, the amount of dollars that are going to need to be spent on this. Two more Moonshot podcasts. Believe it or not, Peter Diamandis and Moonshots. I very seldom meet anyone that listens to these. He does at least one a week, and they're all brilliant in terms of the details. These guys all know the AI space well. As much as I like BG2, they don't have their hands in this the way that these guys do. Um, these guys are heavily involved. They're MIT guys. It is just deep and dirty in terms of what they're going through. So, you'll learn a lot if you listen to it every week, and they ended up having three podcasts, or three podcasts this week alone in the last six days. Um, they go through the GPU demand and supply, talk about Grok 4, that's what this one is about. Um, if you haven't spent time on Colossus, I highly recommend that in terms of there's a couple YouTubes out there. They're each about 10 to 15 minutes, just to get the scale and the size of this, because as I go through this, you're going to see, uh, a ton. They go through the bottlenecks again, the GPUs, the interconnect infrastructure, power, cooling, logistics. All of these things are in big problems in terms of supply chain issues, particularly on the GPUs. In two years, robots will be good enough to do 90 to 95% of labor. The only reason it won't flip the global economy is the supply chains. I think that's important for everyone to realize. When you hear, and I said it last week, um, that humanoids will not be dominating, uh, our marketplace for at least the next five years. It's a supply chain issue. It's not the ability to build them. It's not the capabilities. Within two years, they're saying they're all ready. It's just a question of building it out, which means once it stops, the dollars will flow from everybody who wants them. There will be infinite demand. If you've listened to Chris Camillilo talk about his experience with this, there's already the big companies going to get infinite labor. The cost, which will be a rental situation, which will also be the case, uh, long-term with robo-taxis, everything is going to be working for you. The amount of disruption that's coming inside the next three years is dramatic, but the dollars that are flowing will just continue. They'll be coming from other sources. Uh, so I won't go through this whole thing. We're heading toward a world where the most powerful tech in history will cost pennies to use. Don't fade it.
Uh, GPU is being driven by two forces: intelligence and scale, AI agents, and the video world models. They talk a lot about the importance of video. I brought this up with the JibJab stuff for the token usage and then also V3. I bring this up because they go through why video is so important. It is not for movies in terms of the ability. It's the ability of doing things. Robots for bodies, the GPUs for brains. Everything in there in terms of vision becomes critical for using these things. Main thing I want to highlight here, which I highlighted in the in the outlook or in the beginning, demand for compute to support AI is projected to be 10,000 times higher than supply for the foreseeable future. This was definitely somewhat tongue-in-cheek in terms of what was said, but regardless, this is it. It's the amount of power that's needed, the memory, the cooling. It keeps getting more and more as we go forward. No matter how much the efficiencies are, you have to remember this is just going to keep going. Orders of magnitude gap between where compute demand is headed and where supply currently stands. Accelerated computing is not a luxury, it's a necessity. And again, I'm going to go through the reasons why the funding for this will just continue because it is a race for obsolescence. It is a race for military supremacy. I've said it before, I'll continue to say it. You have to spend all your time on this going forward. It's less about the people spending money and what the AI software companies have done. I believe you need to be doing this all on the PMI names and all on the hardware side. So, I'm going to keep pounding the table on this. Uh, if I'm wrong, I I'll be surprised. I've been wrong plenty of times before, but this one, because we've had the PMI sitting below 50 for a long time, I like to look for scenarios where the positioning is off sides. And right now, when I look at Exxon, and I look at Chevron, and I look at Freeport McMoRan, and I look at Caterpillar, and I look at Deere, and I look at Parker Hannifin, as much of those companies have done well, they have still underperformed the other side, and they are trading very, very cheap two years out compared to where a lot of the Mag 7 and a lot of the software names are trading. So, data center, semiconductors, high bandwidth memory, liquid cooling, power utilities, renewable sources, all to benefit. Consider long-term structural plays in energy, real estate, chip manufacturing, not software. Again, these are all from the the insights that came out of that. Goldman Sachs, new stuff on the hyperscaler capex increasing sharply through 2027. Um, almost all places are now taking up their numbers because of what we're seeing. And this is one of the reasons. Meta to spend hundreds of billions to build AI data centers. So, this is out this week. Meta is not only hiring tons of people, they realize to actually get super intelligence, they need to secure more and more places of power and data centers.
So, let's just go through. I'm not going to take the whole thing, but Zuck confirms that they are building multiple gigawatt clusters, uh, large ones, Hyperion, Prometheus, these are the names on. To give you an idea of the size, if you haven't seen this, uh, this is Manhattan. This is the Hyperion data center, basically the design plans in terms of size. Just this is building cities. So when I say the China urbanization, you have to be there. So what we look for is we go in earnings period since this tariff stuff has now gone down, we look for any kind of details that we can find in terms of what's happening on the industrial side and going through it. Uh, uh, had, uh, a call this week with, uh, a data center guru. The rate of orders remain elevated with no anticipated air pockets given the sheer demand strength for capacity augmented by rising AI workloads. Given the massive increase in demand, there are three main issues impending supply from the coming market. Top constraint is energy, again, energy, energy, and supply constraints for mechanical and engineering process may also struggle to keep pace with the buildout as thermal density continues to increase. Supply and demand issues will be out there. And then as you start to look for companies, we had two companies on the industrial side in Europe that reported that had big up moves this week where they came in higher than expected. ABB, a Swiss company. US orders were especially notable, jumping 37% year-over-year in the second quarter, far outpacing the company's overall 14% earning growth, marking the US as ABB's largest and fastest growing market. Booming demand from AI data centers, electrification, infrastructure upgrades. Uh, the trading environment in the US remains strong with demand sustained by investment. The next time you hear about recession, just remember this. This is why I'm bringing this up. The PMIs are going to show you. Look to see what the companies are saying. And again, this is all from Perplexity. Just going through and looking for it.
Liron, a French company, delivered 20% organic sales growth in North America. Uh, US again was the dominant standout performer. Data center segment now accounts for 20% of total sales. Again, think about that. Data center segment now accounts for 20%. CEO specifically credit US hyperscale projects. The entire organic growth for the quarter came from data center demand. Now they're raising guidance. So you can see everything down here. Despite some ongoing macro uncertainties, tariffs, both companies raised their outlook for the US manufacturing environment through the rest of 2025. Uh, this all comes at a time. Remember this is what we look like back in 1990. Manufacturing and retail dominated the job market. And now the top employer in each market is basically healthcare and professional services. This little area in here is the green. So Trump's focused on a swing state over here. Manufacturing in terms of the 92 billion. Uh, we are going to focus ourself on manufacturing. This buildout is going to happen, and this comes at a time when the BFA fund manager survey number one tail risk. It's still global recession from trade war. People are still worried about trade war. Inflation's number two. Uh, dollar, I didn't understand this one. AI equity bubbles going down. Uh, so we'll we'll see where all this goes. But the given fact people have recession, trade, trade war fears, I'm going to continue to say something I've said, which I very seldom see people say. I'm going to shift my focus to the China-US trade deal. Um, Trump softens tone on China to secure Xi summit and a trade deal. So regardless of what you think is going on, the reality is there seems to be a truce going on between the two countries. And again, you're seeing, oh, China lifts exports of rare earth. China's exports of rare earth products rose 80% in June from a 5-year low in May, according to customs data. Uh, Nvidia got to resume AI chips to China this week as part of rare earth talks, says US. Some articles for you guys to read on why again I highlighted why the hyperscalers are just going to continue to spend, that Trump is going to, uh, put money from a manufacturing-based perspective. But then there's this. Is the US ready for the next war? This was in the New Yorker this week, and you can go through the details, but basically, is the US ready for the next war? Policymakers, as artificial intelligence and drone technology transformed the nature of global war, for recent conflicts in Ukraine and the Middle East have revealed vulnerabilities and sparked urgent moves to modernize the US military. Again, the rare earth situation has scared the hell out of them. They can't do any of this stuff without rare earth. So whenever you get the point that China and the US are going to break off trade deals, I don't see that happening. And so as long as China, where everything on the supply chain is dependent on China, is an issue, I think you're going to see the tariff situation not matter in terms of this. You should go read the article and just see. But again, the Defense Department is overhauling the drone. American-made affordable and AI-enabled systems. Competitive urgency, analysts warned the US could lose its edge if it fails. I'm going to read the whole thing again. The whole point is that military supremacy is critical for this. Uh, you've got quotes here in terms of how important it is. Uh, the outcome will depend on how quickly the US can adapt its industrial base, supply chains, military doctrine. Basically, we've been dependent on China. We need to raise things. Go look at what's happened with MP Materials, uh, a a rare earth place. This is going to continue, guys. This is not going to stop. Why? If we're going to lose military supremacy, you're not going to see it.
So, the real AI race, again, America needs more than innovation to compete with China. There's a fear because of our dependence on their supply chains. So, we're going to have to raise the manufacturing side to compete with them. They realize they're going to have to depend less on us from a consumption basis. And that means that they're going to have to increase their domestic demand, which is what is going to happen. So you have both places spending. And at the same time, in Europe, Europe needs to increase defense spending because the US is now pulling away from NATO and spending in this. Everything fits towards domestic onshoring and getting going, which means they're spending, the fiscal deficits are there. You see the 30-year yields go higher across the globe, not just in the US, yes, in Japan, and yes, in Europe.
So, as you go through this, just go revisions. I talked last week. This is the most important point for me. Here is the PMI, the white line here. Here's what happens to revisions when the PMI goes higher. PMI hasn't gone higher yet, but the revisions are going higher. You've so far seen the revisions go higher because of tech. That is why I talked last week about momentum. At some point here, you are going to have the commodity stuff go higher, too, which has been consolidating. People took down numbers on the tariff situation. Analysts are still worried about the tariff situation. They're still hearing it from some people. That is going to change. So revisions in the US have started to go higher. PMIs will take them higher. The power situation in terms of what we heard from Eric Schmidt, what we've heard from everything. If you just invested in power, here's the white line here. This is the Goldman Sachs power index. This is the PMI that I created, which is the six names that I mentioned: Exxon, Chevron, Freeport, uh, Deere, Caterpillar, Parker Hannifin. It's lagging behind right now. You're going to need these companies as part of this this buildout of the power situation. You have to look to these mega-cap names because the dollars that are flowing are too big for the other comp. All of these companies are going to benefit. Um, terms of the power move. Just look at this weekly performance for this Goldman Sachs power index. This is the realization that has hit in that AI was not slowing down. And so now you're getting all of these companies, the things that I mentioned, and from the earnings reports that are coming through, it's very clear beta working. You're hearing, you're seeing non-profitable tech companies massively outperforming the MAG 7. This is considered a beta chase. Really, what's happening is when PMIs go higher, tide lifts all boats. This is white line here is pure beta, pure versus pure profitability. Uh, again, implying PMIs are going to go higher, but really you have the quant strategies that are moving into things because we're starting to see it. The credit markets are saying the same thing. They've been saying the same thing since 2022. Whenever there's a recession fear, you see the credit. This is inverted. This is junk spreads. Here's the PMI. I fully expect the PMI to be going higher because we have no recession. And we're finally starting to see commodities go.
So, the CRB raw industrial market made a new 2 and a half year high and getting close to a three-year high as it's got this beautiful rounding bottom base which has gone up. Remember, this was the post-COVID period that was as much about, uh, the supply side as anything. Now, we're getting more into the spending and the demand side. It's been on a lag, but now you're starting to see, you've got the Shanghai Composite also correlated to CRB ren making new three-year highs, starting to break out. Usually good for commodities. I fully believe China released data this week. All I heard from people was how they are still weak. I want to highlight to you their AI situation and the way they have to be thought of. So, their IP rose 6.8%, beating expectations and marking the fastest growth since March. The upside surprise was driven by robust performance in several sectors within manufacturing and a broader. All I heard about was the weakness. What is happening though is again, things related to high-tech manufacturing, machinery. They're building out energy as well. So here's what you have. The white line here is IP year-over-year. Again, bottomed in '22, has been going up. This period is back into the levels it was there. Retail sales way below where they were. So retail sales year-over-year are not happening. So we don't have the consumer. That's what people are focused on. And part of the reason is fixed asset investment in real estate down here, but IP's up here. So again, I've talked about this in the US. If you focus on the housing market in the US, it's going to look like this. If you focus on the commercial real estate market in the US, it's going to look like this. If you focus on the data center side, as I highlighted last week, it's a direct line higher. China has the same situation going on. You have to think in modern terms, not in old terms. That is the wrong way to go.
So what was the notable growth product? Forget just going through all this. What really were the outliers for China in terms of the industrial growth? 3D printing up 43%. New energy vehicles up 36%. Uh, industrial robots 36%, lithium batteries 53%. China's upside beaten industrial production was mainly driven by equipment manufacturing, high-tech manufacturing, autos, and electronics with new energy vehicle. Again, I I I don't know how to make this any clearer, guys. You're going to have to get rid of some of the old things you use as signals. China humanoid robot maker Unitree sees significant delivery growth in 2025. And when we say significant, their numbers are getting large. 370,000 units in the first six months of 2025. The first six months, that means they're on pace for over 700,000. Production of service robots rose 25.5% to 8.8 million units. This is only going to get larger. Robots, humanoids are taking over 3D printing. They had they went through this whole big display in Europe last month. Uh, it's a completely different world. And so the commodity side is not the same. But eventually, when this raw industrial go higher, it will be because energy goes higher as well. Oil moves higher, this stuff moves higher because the cost of those things goes higher. We're going to start getting into that again, uh, in my opinion, and the market knows it.
So when people look at the inflation swap market, I hear a lot of people going, well, this is, you know, the Fed's going to cause inflation. The question is, is this inflation more about growth in the PMI, or is it about what we had seen in the past supply and demand problems on other things like core goods, services, and everything else too, going through from money printing? These are the, uh, the swap inflation markets. Every single one of them is going higher, but it's being led in particular by two-year, which I think makes sense. But there's important signals from this. And I've shown this chart before, but I'm going to show it again. So, this is the swap curve now, inflation curve. So, this is 2-year swap inflation swap versus 10-year inflation swap. So, we're seeing 2-year inflation expectations move higher. This other line here is MAG 7 relative to the S&P. It peaked here when we really started to have the transitory move. This is when I think the market accepted that inflation was here and was going to be bad. We bottomed right around the time that we started to see this normalize. We've been going higher in a similar way where they've been going together, but we did peak here. I think the MAG 7 will have trouble because of the spending, but also because I do believe that we're going to have enough inflation fears that the market starts investing differently.
So, what trades work? XLE versus XLK. You do not want to be long energy stocks over technology when you have the swap, uh, curve in negative territory. When it's positive, it's a good time to be in energy. Uh, peaked here just before it went down again into negative, and then you don't want to be in. We were going higher here, and this gets back to the revision side. Energy was outperforming the MAG 7 in the beginning part of the year. But this is when liberation day happened. Analysts took down sharply their revisions on the energy sector. I've shown that chart on Exxon, on on Chevron. You can go see it for every single part of the commodity-based thing. We started building in that the tariff war was going to have a negative shock on on growth, and all this stuff came down. It hasn't reversed yet because the tech AI trade has really dominated. Well, I think this thing is going to go higher, which means you're getting this in at a very good point. So, that's XLE over XLK. I like that one. Uh, another one that historically has done extremely well that doesn't get focused on as much, utilities long over staples. So, if you want to normalize the defensive side so you don't get hurt if there's recession fears, if inflation spikes, spikes higher again, that's usually good for utilities, it's not good for, uh, staples. But most importantly on this particular thing, look at what happened from '02 to '07 during the China urbanization trade. You had a massive move. This looks like a great chart. Uh, they've moved together, but I think you're going to have utilities over staples. I would continue to focus on that space because of the power needs, AI, I, a lot of positive news, new news on AI.
I don't know if you guys keep up on this or if you watch this to go through OpenAI or Anthropic, new competitor for Bloomberg with new tools on the finance side. I've mentioned Perplexity. I'll start using Claude in terms of their finance product. OpenAI agent, uh, sorry, this is, uh, uh, GPT-5, and when it's going to start, you had Grok 4 come out, and then you have this, uh, introducing ChatGPT agent bridging recent. Just continue to have new products coming out on a weekly basis. Uh, Grok 4, I've used it for something which I'll hopefully demo soon and talk about. Uh, amazing what you're able to do with Grok 4 already. I'm paying for the $300 version. I highlighted this on Pomp that I think parents should be giving their kids the ability to build something from scratch for one month, pay the $300 for one month, invest in them, getting used to it. It is a phenomenal tool that allows you to bypass cursor and all these other places to be able to just speak into it. So, um, the banks were started reporting. I said I wanted to go through this in terms of looking for what they're saying on productivity and efficiency. Uh, I'll let you guys kind of read this on on your own, but both companies are talking about how much progress they're already having now and how much is to come. Uh, and that's both Goldman and JP Morgan. Uh, I highlighted this. Goldman claims 90% internal AI penetration, significantly ahead of the industry average. The bank sees this is a paradigm shift with AI tools expected to unlock $340, yes, that's $340 billion plus revenue opportunity by 2025. Uh, both banks stress not just incremental improvements, but a structural transformation of efficiency. It's a massive thing. I told you JP Morgan has approximately an $18 billion cost for technology. I don't know how much of that is people. I don't know how much of that is coding. Uh, but you can see that you will be at a place where they're going to be able to continually get margins when they want them. Uh, Goldman's internal reports warned that up to 200,000 investment banking roles globally could face automation within the next 5 years. Not surprising on that. If anything, uh, I I don't know how many global investment bankers there are, but 200,000 seems like a pretty high percentage of them.
Let's close out with crypto week. Big week, scores major wins after House passes three CR three key bills. Uh, again, market structure, stablecoin bill, uh, and the CBDC, the Fed not doing a, uh, digital currency. This changes everything. For the first time, the US government is not just tolerating crypto, it's codifying it. The world's largest economy is laying down the legal foundations for digital assets to thrive. This sends a very clear message to markets and to other governments: adapt or fall behind. So, I I really want to make sure that, well, I'll wait until I get into the the slide that kind of emphasizes it. But there's a bigger point here for me that fits in with the AI arms race.
So, what has Bitcoin done? It's doing what it seems to always do. Rally up, consolidation. Again, I believe we're in a very, very powerful wave here from an Elliot wave perspective. After we took this high out, five-wave structure, impulsive, new impulsive move, consolidation, uh, I would think we're going to head up to 130, pull back again, then be at 140. I'm I'm assuming based on my own count by the end of the month. Uh, that will be important. If we do, I'll go through why. But this was not about Bitcoin this week. This has been about Ethereum. I talked about Ethereum and I showed a bunch of things for you to read about. We're approaching a very important technical level at 4,000. I fully expect us to get there, uh, soon, and when we break above it, that's when Bitcoin will be heading to 140. Uh, it wasn't just Ethereum this week. You had XRP taking out the highs on the back of the bills that went through, taking the crypto market value past 4 trillion. So, this is now expanding in the ecosystem. I've talked about this being the year of the network effects. They're clearly in play. Ethereum has more than doubled since the liberation day, uh, lows. Trump says Bitcoin could be the greatest financial technology since the internet itself. Using a Mark Andreessen line, uh, Trump said to open 9 trillion US retirement market into crypto investments, approve custody over Bitcoin and crypto.
All right. So, I do want to bring up, uh, that this is another place where it's a race with China, and I think you're going to start seeing more and more focus. I've brought up some other things, uh, as well, but I think now we're getting to a point where the guardrails of the financial system are being set in terms of the rules and benefits by the United States. This is a race for dominance. Again, if they've given up on the global reserve, fiat reserve currency, I don't think they've given up on owning the financial rails and innovation. And I think that's what this is about. So, we are going to continue to power forward with the AI side, and we are going to continue to see innovation boom on the crypto side for right now. Uh, the attention AI is a democratizing event. The crypto side, you want to be invested again in a democratizing, uh, asset, which is Bitcoin. So, I'll keep showing it on here. For those of you who are still not allowed to use AI or invest in crypto, I feel bad for you. But P/E-wise, you can do this.
So, the Chinese have had very large deposit flows into foreign, uh, deposits. Uh, this is the overlay with Bitcoin. So, again, I think the Chinese now are starting to continue to get more involved. Uh, I mentioned that I thought 130, we'd get up to 130. This is 120 right now. This is the open interest by strike in Bitcoin calls on Deribit as of Thursday night. Uh, once we get above here, this is where I think there's going to be a short squeeze. So, if this happens quickly, I think we're going to start to get a rise up, but I think this move in particular is going to cause more and more of a short squeeze. There's been a tremendous amount of buying going on. We've seen a lot of selling. I talked about this being like the dot-com. I really think we've broken through, and we're going to start now seeing the compounding effects where as money is flowing in, it's going to turn into a chase, partly by new investors and partly by short covering.
That's it for this week, guys. Uh, keep enjoying the summer, and I'll see you next.