Transcription
Over, uh, payrolls, stagflation, uh, we didn't see long-term yields go up, stocks keep going. I'm going to keep harping on the size factor because the only things I get from people are refusals to accept Tesla and refusals to accept small cap. So, we're just going to keep going with it until the sentiment starts to shift, uh, and the data keeps supporting it. Uh, macro players still having trouble adapting, not focusing on AI. Uh, time to focus on national housing emergency. If you only watch one thing today, spend it on national housing emergency. Uh, regime shift signs are intensifying. AI is cannibalizing itself. It's power time and Tesla's $1 trillion man. All right.
Payrolls. Um, 80 economists. Uh, only one, uh, did it surprise on the upside too. Payrolls were bad. Um, the diffusion, six-month diffusion has gone down to 48. Uh, for all the data on this, never not been in a recession. So again, people getting hired, jobs not the same as, uh, total income, but jobs not good. Uh, when you go through the categories again here, you look at the last four months, basically no jobs except for health care. So, X health care, and even though this is health care and education, these are all health care jobs. Uh, you're negative. So I will highlight down here, government negative. Um, I want to expand upon that because it's not a bad thing that we're getting rid of government jobs. Um, everyone believes the government is inefficient, uh, and doesn't need as many people as working it, and making government smaller. That was part of Trump's goal. So, I just wanted to show you this is what's happened so far this year on the government. Obviously, the federal ones are dominating, but you can see negative. And then you can go back and look at the months before, and then you can go back and look at the year and a half before, just massive government jobs. This is not a bad thing. These are inefficient things moving back into the private sector. This is how big the numbers are. So when you're looking at jobs, I bring this up because the amount of jobs that's being hired, whether it's the immigration, whether it's the fact that there's approximately 300 people, 300,000 people in terms of the demographic shift. So we have labor losses that are happening every year regardless. We now have no immigration, which means that's more labor losses. Uh, so when you go through it, look at aggregate weekly payrolls. Um, you know, they're back in here in the levels that they were before, bouncing around. Um, the three-month change, sorry, this is a six-month change, is still right in here at this point. The three-month is weaker, but overall, uh, the aggregate payrolls are fine.
Now, uh, Waller had done an interview on CNBC. He again emphasized that labor is the focus, uh, for him. And obviously after the Jackson Hole speech, we got Powell moving that way. And now with this payroll number, we're over 100% of a cut in September. Uh, getting to, you know, about three cuts by the end of the year being built in now. And rates, two-year rates went down, not surprisingly. Surprisingly to everyone who's been trying to short bonds, though, 10-year rates went down. So that whole argument last year that when the Fed cut here, it led to this big rate. Well, if that's the case here, we've now had three sharp moves lower since Powell. We're now almost 30 basis points lower. So for everyone that thought again that that was the best trade, follow the jobs. And here is something Dennis Busher put out this week with his team, uh, on just what's been driving tenure rates and just showing that it is the Bloomberg US labor surprise index. Should not be a surprise to people, but it was. So, of course, we've had fears over tariffs, we've had fears over recession, we've had fears over inflation, and now we're going to go into stagflation. Uh, so the commentary, of course, uh, maybe someone should tell the stock market, like the stock market should trade off your views. Stagflation has officially arrived. Doesn't even mean anything, guys. It means nothing. Uh, this, the definition for stagflation is not, uh, 3% core inflation, 2.7% headline inflation, and 5% nominal GDP. So go do some work. Stocks should really be in the red. Oh, this is Marco. I don't need to show it. The, the reason I want to show it is he even reposted his posting to prove that he was right. Stocks finished down on the day. Uh, UBS gives America. Oh, now we're going to go into recession. So once you leave stagflation, the next fear for you guys to focus on, uh, is recession. So stupid.
Um, all right. And here's long-term yield. So you got to worry about these because again, there's a big deficit, there's a lot of debt, bonds are collapsing, blah, blah, blah. It's actually fairly boring. But if you guys want to keep doing that, go ahead. What you should be doing instead of stagflation is reading about Solow paradox. That's what I wrote. It's a Solow paradox. Um, you can see the computer age everywhere, but in the productivity statistics. So this is in 1987, and this was really the beginning of a paradox as to why wasn't the digital revolution showing up in the numbers? Not just again, productivity gains. Traditional GDP and productivity metrics miss quality improvements, intangibles. You're, you're following. When you say stagflation, you are basically saying, number one, I believe the CPI numbers, which no one believes them unless they fit their narrative. And number two, I believe the GDP numbers, which cannot possibly deal with artificial intelligence. So you look to profit margins and corporate profits, which, no, we shouldn't talk about those because in stagflation, corporations do phenomenally well, like they're doing right now. Uh, early adoption of computers often slowed workers down while companies learned how to. And this is one of the big things, the early adoption. So we started to see numbers be moved up in the 1990s once businesses adapted. This is what's coming in front of you. The adoption phase for AI is next year. Every time a macro person speaks and doesn't mention AI, turn it off. Every single time. Can't emphasize it enough. Remember this was a report last week from Eric Brin Jolson, who is an economist who is out at Stanford, and they put this piece out on the effects that are having on the labor market from artificial intelligence. This is going to be a theme. I've talked about it for over a year now that we would reach a point that the Fed would be in trouble. That growth would still be okay. You'd have inflation going higher. I thought it would be coming from the power needs, which I still think are going to be there, the buildout for AI. Instead, it's coming a little bit from the tariffs right now. But I don't think inflation is coming down. I think they need inflation to be on the higher side, as they've said, to try and grow their way out.
So, the S&P with all of the shenanigans was basically unchanged for the week. The NDX is up. Um, Russell's up for the fifth week in a row. And see this here, 18 of the last 22 weeks, it's been higher. Small caps are alive and well. And as I go through this, be wary of the small caps. Gold finally broke out of this beautiful channel going higher. Some of the comments people have said, "Why don't I talk about gold more?" I love gold. I own gold some. Nowhere near as much Bitcoin. Uh, but again, it's a narrative that works in the debasing theme. So, anyone that owns gold should own Bitcoin. Anyone who owns Bitcoin should own gold. Um, revisions. Again, I'm talking about earnings in companies. Another massive week for revisions. Here's the latest weekly revisions. This is from the, the City Group for the US. Came in up here. This is the now the 10-week average. So we continue to move higher. This is overlaid with a PMI. When you see these kinds of numbers, as I've talked about in the last few weeks, it's a tide that lifts all boats. PMIs go higher. Here are the earning surprises. Again, if we're in stagflation, why are companies surprising? Why are they making so much money? In this case, Q3 bottoms up estimate. I just want you to read this increase. So the estimates, the bottom up increased by 0.4%. Doesn't sound like a big deal. That's since June 30th. So this is for Q3. This is not the quarter we just came out of. This is going forward. Analysts usually reduce earnings estimates. Over the past five years, 20 quarters, earnings expectations have fallen by 1.4 on average. So we're up 0.4, normally down 1.4. Over the past 10 years, they fall by 3.2. I can go on and on in this. These are facts. Stagflation is crap. It's a word. It means nothing. Every time someone uses it, ask them, "How are you incorporating artificial intelligence and earnings into your thought process of stagflation? Where is this coming from? Some historical thing that doesn't mean anything in a world where we're staring with humanoids coming at us?" Please do the homework. Semiconductors, largest contributor year-over-year growth. Semiconductor industry, large 43% year-over-year growth. Energy power. This is data centers. This is data centers. Make money going with where the big dollars are being put to work. Earnings call citing AI. Thing I want you to focus on here. The jump that we're seeing in financials. I've talked about this. They're breaking out. They have a huge tech budget. You are going to see profit margins in here. You're going to see it across sectors. Only one that cited it a little less was the tech side, uh, interestingly, uh, enough. Uh, and I'll get into more of that later. I'm going to keep emphasizing the point. I, I'll get into robo taxis again later, but regime shift. This is something I went through. If you didn't watch the video two weeks ago, I highly recommend watching it because it's still going to be the same story going forward. Size, the size factor, sharp ratio 2.7. It's working its way lower. It's been the best factor. It's the one you want to watch. Size has worked in a market. You've had large caps outperforming small caps. That is not normal. Here's what the size chart looks like. And again, every week I get pushback on certain themes, and I love to get pushback because that gives me a sentiment gauge. I am, it is whatever number to zero. No one wants to embrace small caps. Everyone wants to fade small caps. So if you want to find something that even from a contrarian basis, forget that I'm going to show you why it should work, even if you just want to do it purely from a positioning basis. I showed you last week that CTA futures positioning massively short, uh, Russell and long NASDAQ. So take your pick in terms of where you want to go. You want to be long S&P, you want to be long NASDAQ, you want to be short Russell. Here's what the size factor is doing. Beautiful chart on this. John Rog, I've highlighted, loves the chart of IWM of Russell. I'll highlight that there. But this is the pure size factor, which has a nice kind of rolling time. Now, we've had other topping formations back in here. This was, uh, uh, during the tariff situation, but this was really during the tech side of, of the Mag 7 going. This is happening in a different situation. This is while the market's going higher. So, I just watch the size factor in terms of the Morgan Stanley size factor, which doesn't make it pure. It doesn't neutralize it. The thing I want to show here is it's now hit higher. And again, there's a cyclical component to this. The Fed is about to cut rates. I'm saying PMIs are going higher. We've already seen that with the S&P. I'll get into the difference between the two of them. Um, but right now you've got the S&P PMI moving higher, and you have the national ISM PMI not, but you have the size factor going up. If you have rates being cut, which helps size, and you have that, it's just another reason. Uh, Neil Sethy put this out from Bespoke. Uh, again, these worked really well in terms of these breadth signals. If you go back to the ones I did back in April and and May coming out of the bottom from, uh, uh, from Liberation Day, when I wrote the paper, uh, back on April 10th about turbulence and the fact that you have to go in because the government is going to support the market. This is a breadth signal on the Russell. And the interesting thing, based on, uh, a breadth being in the 99th percentile, Bespoke went through six months out, up every single time. And this is over the, you know, there's 15 of these signals back to 2010. The numbers are very high over one year. Uh, I would be very wary of being short the Russell. And then you have this one, and again, this is from September of last year. But it is important to pay attention again. Since 1980, five of the 10 best years for the S&P 500 happened when the Fed was cutting rates without a recession. That's where we are right now. So if you are being bearish at this point when the Fed cuts rates into this, where I believe the PMI is going higher, the Fed has cut rates 12 times when the S&P was within 1%. Again, here we are. The market was higher one year later all 12 times. You're playing against history. You're playing against the numbers. You're also playing against leading indicators. So Russell over the S&P, the orange line barely moved higher. Again, no one supports this. Well, this is micro cap over the RTY. So what hasn't happened yet is the Russell hasn't gone higher, and I believe that is because there are futures on it. There just hasn't been the covering. But regardless of it, it's been there. There's a cyclical nature of this with the dollar. The dollar has moved down historically. That has meant that the size factor, which in this case, obviously, is, uh, is moving down. You've just started the dollar weakness historically has a component which is good for both commodities and good for beta. Well, in this case, small caps have not large outperform, and what benefited again was the Mag 7.
I'm going to go through a couple of macro and these are macro podcasts. Two of the, the few that I, I listen to, I would say at least monthly. I like these guys a lot. Uh, probably because, um, I really don't care that much about the Fed, but I think they're very good at at getting in front of things. Uh, they do a roundup every week. I highly recommend it. They're also good Bitcoin people, so they do a good job of connecting back the Fed's, uh, decision-making, uh, back to crypto. But in this case, every single thing in there, I could care less about. Um, again, gold thesis, not a surprise. Stagflation light. Uh, I think these guys are more realistic. I think they don't get over their skis. I think they do both sides real well. Do they mention AI? They say AI is taking jobs. Again, no offense to people, but don't understand AI. But I think at some point, you have to admit that you don't understand the environment we're in. I had to do that in 2013 when I didn't understand how Amazon could finish at, could trade at infinity forever. Uh, and I wanted to learn why it was happening and why it was justified. I went out to Silicon Valley, spent a bunch of time at Singularity University. It led to me doing these things and being very focused on AI. I highly recommend you do the same thing and open your, uh, your brain to possibilities that the reason you're wrong is not just timing. It could be that you're not paying attention to signals that matter the most. Um, Jim Bianco, who I like again, he was on Macro Voices, one that I like. I'm not going to read through all this. I got bored within two minutes of this. I actually listened to it solely for you guys just to show that nothing in here, in my opinion, matters. He's been saying the same story for a while. I like Jim Bianco. Nothing in here I found to be worth anything. Did he mention AI? No. AI isn't mentioned in this episode. Discussion centers. I just don't care. If you can't even mention it, it's not there.
Now, let's go to my boy Pump. We not only do a weekly thing together. The guy hustles and interviews a bunch of people, and he asks good questions, and more importantly, pays attention to what's happening in the news as opposed to what happened in history. That is one of the things about people in their 30s compared about people in their 50s like me. You have to spend more time on the news that's happening and less time on the history. The history doesn't matter as much because we're not riding horses anymore. And unfortunately, for tariffs and stuff like that, we were riding horses. He did an interview with Darius Dale. I've talked about Darius. He was the only sell-side person that I paid any money for for research. Uh, I've done Luke Groman as well, just because I think both of them get some insights. Uh, he interviewed him for 30 minutes. In this 30 minutes, there is more in here in terms of the conversation than anything I heard in those two podcasts by far. And the reason is because Dale stresses investors don't need to be the smartest, just disciplined and willing to do the work. The administration is saying its policy path, and those who follow the playbook, long risk assets, Bitcoin are outperforming. He goes through these reasons, and I'm not going to go through everything here. His fight with the institutions at the Fed, the housing market, which I'm going to get into, which is the most important thing, and they talk about this 3% inflation is the new percent. The Fed raised their inflation target. I don't know if they would have announced it then, everyone would have said buy stocks, buy gold, whatever. Well, that's what's happening. They just didn't announce it. So, if you go through that, you'll get it.
Now, expand on the housing options they mentioned, and this is what I want to get into. Treat housing as a national emergency. That's what Besson talked about this week. Support a broad economic boom. Fixed housing is seen as a lever to lift the middle and lower income, directly addressing the K-shaped economy that Trump often highlighted in his rhetoric. The policy options discussed, all of the things that they say could be done. And what does Darius Dale say? High confidence in the action, repeatedly stressing that the administration says what it will do and then it does it. It says it wants rates to be lowered. It forces the Fed into a corner. People disagree with him. They say Fed independence is going to lead to bond sales. It's going to lead to stock market falls. It's going to lead to Turkey. He compares this approach to China's command-style policymaking. Oh, another reason to be bearish. No. If they declare housing as a priority, even via emergency powers, which I'll get into, investors should assume policy will follow. Dale warns investors it would be a mistake to assume they won't happen. The administration has both the urgency and the political incentive, and I'll even say the political ability. Market implication, investors should expect an eventual housing turn as government policy levers will be pulled. That's the most important thing of the week. And the reason is I showed this two weeks ago. These are the winners in the industrial space. All of them are related to AI and the data center buildout in terms of the first phase. These are related to jobs, in terms of the lower-end consumer, in terms of consumption, in terms of housing. This is what Donald Trump was elected on. If these go higher, even if these continue to go higher, which is normally not the case at the beginning when you have stuff like GE Vernova up 99%. These things have had an easy ride. This has been in play for a while. If he stimulates the housing side, it is a massive shift. And again, it will add to the PMI side. And that's why I mentioned it weeks ago. Bill Pi has been alluding to this. They've talked about getting rid of capital gains tax. They've talked about all types of things. And I mentioned this is a major issue in the midterms. So if it's a major issue in the midterms, the administration keeps finding ways to do it. Whether you agree with it or not, that's not the point.
Do they bring up AI? AI comes up briefly, but it's not the central theme. Oh, okay. So I guess Darius doesn't do it either. But wait, while AI isn't discussed in depth, it is flagged as one of the drivers of structural economic change that makes the Fed's current policy framework misaligned with reality. That is the key thing. You have to mention that AI is screwing with everything, and it is impacting the way things were done in the past because when you had higher growth, you would get job creation. We are not going to see a massive job creation this time, even if nominal GDP stays at fives. So for all of you sitting there saying this is going to end badly, I'm telling you we're going to sit around this level because of AI. Trump weighs, weighs declaring national housing emergency. So they're talking about it. It was something that was mentioned about, it's a critical leg of Republicans' 2026 midterm election platform. What does declaring a national emergency allow him to do? Basically anything he wants. So a national emergency. So you go through it. These are all the things that he can do. You can read them on your own. Expand sanctions authority. Use industrial and emergency powers. Force companies to prioritize contracts like for AI data center buildouts. Control allocation of electricity, oil or gas. If electricity prices go higher, there will be a national emergency declared. Override environmental restrictions. I say that because once you start talking about inflation, just remember they will find ways to contain this. Override environmental restrictions for national security purposes. If a housing emergency is declared, Trump could. These are all ones that were just created by ChatGPT. When I asked the question about a national emergency, which ones has he already declared? Here are the ones he's already declared. National emergencies, they've been happening consistently. This one is in the courts right now in terms of the trade, the the trade side. He just did crime. Trump has done nine plus in one year. Donald Trump let, did seven in four years. So that was the the fastest pace the last time. So he's done these before. This is what he does. Obama did them, but it was 12 over 8 years. Well, uh, he's declared nine national emergencies in less than nine months. So, in effect, these emergency declarations can be forms of government stimulus. Yes, they can be function as powerful forms of targeted government stimulus. So, whether you agree with the way China's doing things, in the case of AI, I've said this before, in the case of all of the things we're talking about, yes. Are, are is he acting like China? I, I don't know other way to say it. He's making sure that he fulfills the campaign promises that got him in. So, you can argue with them. You can be academic and go through and say this is going to be bad. You can be bearish and maybe you'll be right. But the reality is you're doing it at a time. Well, the market's going higher, profit margins are going higher, earnings are going higher, and inflation is not at 9%.
Marco Papic, my boy, this is really interesting. In my view, the White House is laser-focused on unleashing the massive amount of wealth stored in US housing equity. I fully expect them to find a way to let the floodgates open. Whatever it takes. Marco's in his 30s. It's what it takes, guys. It's what it takes. Household durables. Hm. Here's the chart. Wow. It's been rallying strongly despite a crappy housing market and no PMIs. H I like it when charts are going higher. Sentiment isn't and data isn't. Here is the four-year chart. The John Rog special. He loves housing. He loves small caps. No wonder why. This is Home Depot. The white line about to break out. Overlaid with IWM or Russell. Why would you try to pick a triple top when we're about to cut rates and PMIs are about to go higher and the government is about to help the housing market? Figure it out on your own. I remind you, look at all of these names. Come back in three months and see what's gone on. I guess everything could fall apart. We could have a recession. I'm, I'm gonna go the other direction. Um, this is the PMI overlaid with that chart. Again, as a reminder, I did this two weeks ago. So, this is what happens in PMIs. Is there's always some either move to safety, there's some move to AI, which is safety right now. Wherever the growth is, you move to. So, this is inverted. This is momentum of industrials, which has started to pick up. It was down sharply. Uh, this is inverted. Uh, it was down sharply on Friday on the back of the payroll numbers. Uh, for anyone who hasn't seen this, uh, you know, I, I went to 22V to basically connect AI to macro and crypto. Uh, I didn't, you know, I knew that I'd be there. Colin Fenton and John Ro, I've known for 20 plus years. I love them. They're great. They're both buy-side people. We are going to focus a lot of attention. I think this is one of the most important things for mutual funds in particular. If there is a shift and if these charts are telling us a story, if I'm right on AI, there is going to be a massive shift away from the Mag 7 and into a lot of names that have no liquidity. They are starting to show up. I will do more and more of these in terms of showing you names that are trading at a dollar at $5 that have now suddenly had back orders of things related to AI and data centers that are going to be the buildout.
So the power trade, and this is around energy, but this is really more on power. We're going to define power because it's not just energy. It's batteries, it's storage, it's boilers, it's, uh, transformers, it's, I can go through a list, but instead, if you come, you can see us. We're speaking about this on Tuesday. Um, I wrote this paper this week, efficiency versus exponential demand. If you didn't read it, um, I'll just read you this line. Efficiency is the most overused word in AI. For those fading AI, like most macro people, efficiency is hope. For those riding AI, the momentum crowd, efficiency is fear. We will use up everything for the foreseeable future. Will that change at some point? Yes, but not in the next few years. And that's more than enough time for everyone to move their portfolio and to trade. GPT5, which came out, power appetite consuming energy like two to three nuclear reactors. That's what I highlighted in the, uh, in that article was just to say that we have so much demand coming. It will not stop for the foreseeable future, and any efficiency gains which are happening every day are being gobbled up by the insatiable demand that is still necessary. Data centers that don't exist yet are already haunting the grid. You should go read this article. It was out this week. Um, there's more coming. We did GPT5. Well, now Google's going to try and up them with Gemini 3 coming out. It's shaping up to be an incredibly performant model. Gemini is starting to kick ass. I use it and I'm going to bring it up more going forward. Uh, Musk claims Grok 5 AI could be true AGI by year-end. I bring this up because it doesn't stop. We're just going to keep going and going. Apple demands suppliers this week, switch to robotics for manufacturing. More of the job situation. This has to happen. It's going to happen. Uh, Michael Symbolist put this out today. The reason I wanted to bring it up, he goes through these capex numbers are enormous. Just look at the hundreds of billions that are coming through. All the things that make him up and then being predominantly these guys, the contribution to GDP, how it's growing while the other part is there. This is the macro community down here focused on recession. This is the part they're leaving out, which continues to grow. It surpassed personal consumption expenditures. Trump is focused on this. He wants this to go higher. And then the danger to these, which he brings up as well, is that we are getting into insane territory as capex and R&D is a percentage of revenues. This is the ROC argument. This ROIC, this is where I get worried for the Mag 7. I'll go more in depth soon. Peter Diamandis, moonshots. I say it every week. The only thing not the only, if you're going to watch one AI thing every week, this is it. They go through the entire week's stuff. Lamas colossal data centers. Um, for those of you who've never looked at Colossus or go watch a documentary on on YouTube. Um, the main point is that he's finishing his second one. It goes through powers of bottleneck. They talk about it, all the different things. Energy, AI is energy limited. They go through the Grok stuff. They go through the talent wars. Everything along those lines. Google's nano banana. So, if you haven't heard of nano banana and the importance of it, you're going to get a little bit on this. You got to pay attention to these things. So, Gemini released it. Nano Banana disrupts Photoshop. They talk about it. Google Translate came out as well. Disrupts Duolingo. Go look at what happened in Duolingo and go look at what's happened to Adobe over this. I've talked about software. Long power for me is the same as short software. Same trade. Texas Hedge. This is the trade you want to have on. Long power. You want to be short software. You want to make them equal weight, but you can be short Microsoft as well. I'll go through that too. But you want to basically believe that the transfer that we've seen in market cap away from power and into software, which is what's happened since 2007 and the release of the iPhone, everyone is sitting and overweight these things. They will underperform, underperform going forward, not collapse. The Mag 7 are not going to collapse, but they are going to underperform in my opinion. They go through Nvidia and some of the risk that's rising, and they bring up the Chinese chip makers. I'm going to bring up some other stuff, uh, which I brought up last week as well. Tesla optimist shift to vision only training. They go through this. I've talked about this as the most important thing right now is the next stage of AI. They go through electricity prices spiking. AI's natural limit is power. There was a debate, and one guy said recursive efficiency breakthroughs possibly reversing the demand shock. This is what is 100% true and it's already happening. I've talked about recursive efficiency being the thing that both increases demand, but what it means is that eventually AI will get so good that it'll start fixing the inefficiencies in its code. It'll fix the inefficiencies and come up with inefficiency fixes for everything. It'll actually fix the energy problem, but that's not coming in the near term. Nano Banana. Okay, let's get through this. Is Nano Banana a risk to Adobe's business model? Unlike other MLs that struggle with manufacturing facial features and identity across, Nano Banana excels at preserving character likeness throughout transformations. Adobe's Creative Cloud generates over $13 billion in annual revenue from 37 million subscribers. Nano Banana threatens this foundation. So I'm going to show you some of the work of Nano Banana in a second. Democratizing editing, eliminating learning curves, and it's free. This is from Gemini. This is just one of the tools they have like deep research. Yep, there's me at the US Open. All I did was take a photo of me in one of my grungy clothes and said, "Put me at the US Open hitting a backhand and then show me holding the trophy." It does it in seconds. It's free. Duolingo. Duolingo's business model heavily depends on converting free users to paid subscribers. Key vulnerabilities include only 8.6% of Duolingo's users are paid subscribers, but they generate 80%. This whole model is dead. All of them are dead. That's why this is the, what I want. I want to be long Chevron, short Salesforce.com, which reported earnings this week, missed. Everyone's trying to pick the bottom of software. I read it in exposts all day long. Everyone has a mean reversion bent. Everyone seems to think that they can just jump into something that it's overdone and that AI is a hype. Like I said, the fading of AI will cost you more money. You have to spend the time in terms of thinking about it. And the only way you can do that is you can't read the bank's research. You have to read 22V's and mine. And I'm not doing that to pat myself on the back. I spend hours of time using it and hours of time listening to podcasts every week. And then I give you guys the playbook and I give you the information. Chevron over Salesforce.com. That is the thought process. That is the trade.
So, here's Microsoft relative to the S&P. This is a weekly MACD. We just got a sell signal. It's come down sharply right here. I don't want to show how big of a move this is, but this is a very strong move. Uh, when you do it on a rate of change basis, uh, relative to history, we normally don't get these types of moves. So Microsoft, which is the second biggest company in the world, is underperforming. And the biggest company in the world, which I highlighted last week on their earnings, you not only have declining revenue, you're starting to get worries. This one, you not only have again, a MACD weak sell signal. It's a divergence relative to price. I would be very, very worried if these were my two biggest longs. And every single person and every mutual fund who is benchmarked has those as their two biggest longs. I think their alpha days are going to be challenging. And again, I'll go through one of the things that says that right now it's happening. This white line here is an equal, this is the S&P 500 equal weight semi. I created this in Bloomberg. Uh, we were up again on Friday, but more importantly, Nvidia is underperforming. The S&P semis are still going higher, meaning the equal weight semis, which has a lot of small cap stuff on it. I've highlighted LSCC, C on here, Lattice Semiconductor. The thing is, it's the tide that lifts all boats. And before Nvidia had a, a do, a do, a dominant market chat. They have a monopoly on GPUs. They might be able to extend this. They might be able to do something, but it is not as safe a trade. And I think you have to build that risk into the stock price, and it's still trading at a high enough multiple that people can start to believe that if they can't do that, maybe they should be trading off next year's and the following earnings with more and more competition. And the reason is Nvidia's chip dominance, recent competitive threats. These are recent China's rising chip makers. So, we've seen the whole thing with China and the US. But what probably hasn't happened, and I'll go through more China here, China has made huge, huge advancements on their chip side, and they've made huge adv, huge advancements on figuring out how to be more efficient with what they're doing. You can't not see that they're probably going to be further ahead on this as we go forward, like all the other chip makers are. Google's TPU commercial. Puss Davidson analyst released a report on September 2nd suggesting their TPU business could be worth $900 billion if spun off. You're starting to get more and more messages of them allowing other people to use the TPUs. So, you're getting competition from there. Broadcom was up huge on Friday with a partnership with OpenAI. Again, market disrupt. This partnership represents a direct challenge to Nvidia's inference chip dominance. Tesla announced a $16.5 billion dollar deal. They designed their own chip. Again, these companies are designing the role, and they're the ones with the capex. Is there going to be other places that need Nvidia chips? Of course. But Nvidia is moving on to something related to humanoids. It's moving on to native devices. They're trying to gather that. If they do that, they'll get another wave of this. But that is very different than what they've had before, and they're at such a big level that even if it starts working, if the numbers don't get huge very quickly, it's not going to help them. So, I would just keep it in there. Go through. You can read these things on your own. These four developments collectively signal a fundamental shift in the AI chip ecosystem from Nvidia-dominated supply to diversified custom silicon solutions. Uh, again, use ChatGPT.
Okay, two podcasts again that focused on China. The entire BG2 one was on China. These guys, even when they, I'd say about a quarter to a third of it was on China or indirectly on China. China has made open source an explicit policy vector. You have to understand that what they have done is open source. They are competing with the Mag 7. So the Mag 7, Google, for example, 90% plus dominance in Germany across the globe in search. That is not going to be the case with AI. China has made their stuff cheaper, and their open-source models are close or better than the open-source models that the US is offering. That's why Llama's made this whole big push into ASI. Will all countries use Chinese-made AI? No. But if all of Asia does, well, that hurts the market share. You can only lose when you're on top. You can't take more market share. So, the competition from having cheaper models that are almost as good and continuing to grow fast in China, where they have endless energy. And I will tell you, using ChatGPT all the time, I get very frustrated with how I am now restricted in ways that I never was before. And this is happening on most of the models. I used to get 50-page deep research reports easily. Now Gemini is the only one that I can get them from. That's why I'm saying as all of these models start choosing some kind of vertical that they want to focus on, finance for Perplexity, or, uh, the consumer for OpenAI. This is a, this is a battle that's not as easy as it was. And China has wanted other battles in this. So using the open source, and they go through this, they highlight Ford, and I'll show you some of this. The mismatch, China challengers capture share more quickly, putting pressure on Mag 7 margins and future revenue growth. Uh, there is competition, and it's going to be there. In older industries like cars or steel, dominance played out over decades. The problem is with this one, if you get ahead in the models, or if you're really cheap, while everyone is growing rapidly, this is growing like wildfire, the Chinese are willing to lose money now in order to dominate later. It gives them an advantage, and it's much harder with a private business, even though OpenAI and Anthropic, all those companies are losing money. Are the Mag 7, who are public companies, willing to lose money if that's what starts to happen? And the answer is absolutely not. So I would just keep it in the back of your mind that this is going to be pressure coming from all directions. Um, China's industrial, they have an extremely sophisticated market-driven approach on how to deal. They get obviously the government support. Um, in 2006, Hu Jintao gave in his speech the boxes that were national security issues like semiconductors, rare earths, pharmaceutical APIs. Doesn't it sound like we're doing that now, 20 years later? It's happening right now in terms of global dumping with BYD. It's disrupting American, German, European auto manufacturers' ability to compete. And if you haven't read this story from June, Ford CEO says China's EV progress is the most humbling thing he's ever seen on BYD, on a Xiaomi car. We flew one from Ch, from Shanghai to Chicago, and I've been driving it for six months now, and I don't want to give it up. This is the Ford CEO. It's the most humbling thing I've ever seen. 70% of all EVs in the world are made in China. They have far superior in-vehicle technology, meaning you walk in with your phone, and everything lights up. You don't have to connect via Bluetooth. You don't have to do that. There's a variety of things you can go watch online. And their cost is far superior to what I've seen in the West. That really helps in a world where there's very few wealthy people. And that's the distribution of wealth problem. And that's the plate that Ty is targeting. The economic Fed versus the inflation target of the world. I just want to remind people again, if the Mag 7 is under pressure at the same time that power is where all the money's going, PMI and housing and everything along a sectoral regime shift is more powerful. It usually reflects dramatic changes in monetary policy. The Fed is going to cut rates while the S&P is near 1% while the government is saying they want to lower it more. We have labor weakening and inflation staying high, and they're going to choose labor. Fiscal policy, national emergency measures, the AI action plan, whatever you want, and the targeting of housing at the same time as innovation, AI. I believe we're in the early stages of one right now. Call me, talk to me. This is what's happened over time with momentum. When you get one of those points, and I've shown this before, we didn't fall here. We're beginning to here. I believe we're going to see momentum shift. There'll be new leaders going forward. Um, one of the new innovation leaders going forward over the Mag 7 will be crypto, will be Bitcoin. Uh, I highlighted this in here. Uh, someone highlighted to me and was angry that I haven't disclosed that I've been wrong in Bitcoin. At this point this year, I thought Bitcoin would already be over $150,000. It is currently 110 as we do this. I have been completely wrong on how far I thought it would go this year. Uh, I can give you theories why, which I go through here, but that was the point of this is basically to say I still think it is the, the horse you want to own in this race. Uh, and maybe it starts after the rate cuts. Maybe it starts after people realize what I'm talking about here in terms of PMIs going higher and cutting rates. I don't really care the network effects of stable coins and the size of the asset class. But more importantly, if the Mag 7 aren't working, people need to have innovation beta in their portfolio. The only large cap of over a trillion dollars aside from Tesla that you'll be able to own that can provide you the beta that you would want if the Mag 7 aren't doing it for you is Bitcoin.
Um, now on to Elon Musk, real quick. Trillion dollars, just like a a football player. Here's, here are the milestones he needs to hit in there. You can love him, you can hate him, but do not bet against him. He has hated. I would say it's 20 to one in terms of people that have reached out to me to say you're wrong about Tesla, you're wrong about Elon Musk. I, I don't know how people can fade the guy, but I'll leave it at that. Um, if you haven't listened to Andrej Karpathy, um, this interview, which is almost four hours, um, he worked at Tesla. He was part of the vision thing. I've said this before. I wrote a paper on it. I'm going to say it again. What Tesla is trying to do with vision only for the robo taxis, the success of it is a major game changer and has huge implications for everything on the AI side. It is the next AI part. It is the advancement of humanoids. If it works again, you've got humanoids that can be just like a sector specialist. They can be trained to do one thing, but then if you ask an industrial person to go into biotech, they can't do it. In a great generalist humanoid, they can do anything. What these guys are trying to do is not be a sector specialist like Waymo, which can only do it on a grid and can only do it in a place where they have GPS. What they're trying to do is be able to do it on Mars. I'm going to keep saying again and again, but if you listen to him, not just now, Karpathy is worth listening to and following in tweets on all this. Uh, okay, real quick. Apps, Tesla Robo Taxi, head of Uber. Tesla Robo Taxi app downloads outpaced Uber across all rolling 30-day periods by 40% and bested Waymo's download day ever. Robo Taxi is live in San Jose. Yes, that's California, where they restrict everything. $23 versus Uber's $37 for a ride downtown. Again, this is just starting. It's official. No one saw it coming. Tesla gets secret permission from California DMV to test autonomous cars. For those of you who will be out in Vegas, uh, to see him, please stop by. I'll be there only briefly. Uh, but I will be on stage actually with John Ro, um, at the event. Uh, I'll be there Wednesday and then I'll come back immediately and then I'll be with my boy Pomp on, uh, on Friday at his investor event. If you're going to be at that one, uh, or if you want tickets, you can reach out to those guys. I'll see you all there.
Final thing for the parents out there. Um, and again, I think I've highlighted this a couple times now. This is a really good podcast if you were going to listen to two AI podcasts as a macro person to be on top of things and then to use AI to make sure that you understood everything coming out of it. This would be the second one behind Moonshots, the Artificial Intelligence Show. Um, this is important for parents. First of all, they went through the Stanford report. That's one of the reasons why you should listen. They basically say it's a good report. Uh, it is true. They go through and say it's not just, um, in their opinion, these are AI guys that are in the business of this. Uh, it's not just the early, uh, the college kids that are getting hurt. Middle management will be getting hurt this year. They go through nano banana. They point out that the buzz and how easily non-experts can access advanced editing, which pressures incumbents' creative tools to evolve. They built a tool, the Jobs GPT tool, which you guys, like I did, can put on your own ChatGPT. What this is, is it goes in, you launch it. So you go to this website, you launch it. Once you launch it, it's part of your own ChatGPT on your own box. Then you go in and you type in the job that you want to check. So in this case, I put in coder. It shows you the exposure level you have to AI. Is your job worth having? It allows you to go in and help your kids figure out whether they're choosing a major, choosing a job, which is going to be basically obsolete. How much of it's going to be obsolete. This one, a lawyer, and just to finish off the episode, since I know we have a lot of these hedge fund analysts, difficult job to have. All right. Uh, that's it. I'm glad to be back. Uh, look forward to hearing the comments. And you guys, uh, hit the subscribe button, get it to your friends, reach out to 22V. I'll see you guys soon.