Transcription
The story of the week is uh, is going to be earnings, uh, along with some other stuff that's mentioned in here, including importantly, the GPT-5 release. Uh, so let's just get right into it and start with a big week for stocks. Um, so S&P up 2 and a half% for the week. I don't need to show you how close to all-time highs. Just continue at this point to consolidate, um, after the strong July. Uh, NDX led the way. Uh, pure momentum has gone up. So two weeks ago, I showed a triple divergence on the MACD's and basically was looking for, uh, momentum to grind lower. It has ripped in the last two weeks to new highs and basically this was the best week for pure momentum. It tied it a couple times, but really of the year.
So the reluctance in earnings season of people to embrace anything related to anything other than AI continues. So AI has continued to rip and led the way. This is the AI pair. Uh, very, very lagging for everything related to the PMI stuff, which I'm looking to go higher. So as of right now, from my, uh, seat of looking at what's going on, earnings have been great. Stocks continue to power higher. AI is still the story, but as of now, we have not seen anything really expand from the AI side. So, I'm going to go through that as, as, uh, as we hit it. Uh, there's no way to look at the earnings any other way. And the way that I want to do this, because I've seen a lot of people say, "Well, they're pretty good reports." I listened to a macro podcast yesterday where they said, "Yeah, it's, they're okay." These are not okay earnings. You're going to have double-digit earnings growth again. And as you read through this FactSet recap, we're now, uh, almost, uh, where 81% of the companies have reported, or 90, I guess, or close to 90. Uh, and basically you have to remember this is for Q2. Q2 is when the tariffs and the uncertainty were at the peak, and yet somehow or another companies have reported fantastic earnings. They've beaten numbers. These are higher. I'm not going to read all the stats on this. Uh, but especially on the revenue side, when you get into revenues above estimates, 81% reported revenues. Revenues directly tie back to nominal GDP. It's above the 5-year average of 70% and the 10-year average of 64%, and they beat by 2.4%. It was the best quarter for beating stuff since Q2 of 2021, which was basically when we had were hiring millions of people.
So, as you go through this, I just want to remember when we came into June 30th, the estimates were for 4.9% earnings growth. Through Friday, we're at 11.9 year, and revenues are 50% higher. Instead of four, they're coming in at six. So, anyone who's still believing that the economy is not good when we just had the most uncertainty, uh, that we could possibly have, everyone was sending around these doom and gloom charts, and we not only beat at the overall level. Here it is by surprises by sector. It's across the board and revenue, every single sector. The only one here, uh, sector-wise, that didn't beat was materials in terms of surprises. Uh, but everything else completely across the board. Profit margins again, if we don't see profit margins come down, which again, with AI coming in, and as I go through this, we'll keep going, nothing is going to happen. Profit margins at this point for the quarter, uh, are coming in at 12.8%. It's above the previous quarter. It's above the 5-year average of 11.8. Profit margins are still the story. Companies continue to benefit, and it's not just a technology thing. You can see the profit margins here. Continue to come through. I'm not going to harp on this. I'm just going to say if you're fading AI, you're making a mistake.
Now, if you want to again, and I'm going to do this and go through these quickly. I'm not going to read every single part of this, but I think what's really important to look at is comments like this, "almost unprecedented interest." This is for Caterpillar. Uh, highlighted major strategic focus on data centers, AI, and energy power infrastructure buildout. How can the economy be bad if Caterpillar is blowing away things? Trading at all-time highs and is talking about this kind of stuff. Power generation sales jumped 28% year-over-year. Demand from hyperscale and data centers both robust and highly visible, uh, years in advance. Taking orders for equipment well into '26 and '27. Uh, despite margin pressure from tariffs and construction softness, this long-term secular growth story, data center is providing stability and back. This is what you have to understand. This is why I keep talking about there are two sets of the economy. Yes, housing sucks. Yes, commercial real estate sucks. Yes. Uh, nothing's being transported. Yes, you can keep going on and on. Auto production sucks. As of now, I would not fade the fact that Donald Trump will find a way to get this going and the fact that people don't have to worry about the world ending the way they've been worrying about it. I got more stories on that this week. People are surprised the world hasn't ended. So, when you're preparing for the world to end, you get nervous. Businesses get nervous. They don't hire. All of this is built in.
Here's more of Caterpillar's stuff. The backlog is at a record level, uh, $37.5 billion. It's just all there. They're, they're pretty extended on the data center side. Uh, Parker Hannifin management stated that total company backlog remains at near record levels. These are not the types of things you see before you go into a recession. They're not the types of things you see when the economy is slow. Parker executives and company materials have directly highlighted the unprecedented demand being generated by generative AI workloads. Cummings, another big industrial, uh, revenues jumped 19% year-over-year, fueled by surging demand for generator sets and power solutions. Uh, power systems and distribution now account for 507, uh, 57% of total revenue, showing that secular data center demands are helping offset soft truck engine demand. This is what you're going to start to see as this buildout becomes far more massive than what's going on. I've talked about this, 92 gigawatts of electricity is necessary. We are finally getting into the building of the power side. So far, we've only had the building of the data centers, and that was the last two years. Now we are getting the massive spend, the, the numbers you've seen that you've heard about from the hyperscalers. This is all now. So when you fill up the capacity of the electricity, which will happen next year, then you have to get into, oh, we filled up the excess, we have to build new. This has a compounding effect along with the rare earth mining side and along with all of the things that will be coming down the road for humanoids, for robo taxis, for everything. So, anything on the housing side, anything on the commercial side is just bonus on top of it because these numbers are growing rapidly.
Emerson Electric, U. I'm not going to read this one for you. You can go through it again. Book-to-bill ratio stood above one, signaling continuing visibility for demand tied to AI, LNG, grid modernization, data centers. Siemens, uh, backlog, our intake remains strong. Long-cycle demand for data center. Uh, management maintained a cautious note on global macro uncertainty but reiterated confidence in secular trends. Again, everyone's seeing that we've got weakness from the tariff side, but on the AI side, Siemens Energy, enormous demand from data centers. The grid technology set new records for both quarterly orders and backlog. This is just starting. Uh, this is from Arista. Uh, AI cloud titans, enterprise company have only strengthened demand. Was not limited to just hyperscale, extended to enterprise edge data centers, campus upgrades. You're starting to get the expansion. That's the key thing. This feels to us like a unique, once-in-a-lifetime opportunity. We recognize the potential to build truly transform. Bloom Energy, a company I've mentioned on this before and one I've talked to accounts about that kept showing up as a company to just focus on because of one part that they mention, the power of, within just 90 days, solving the AI speed. This is in quotes, "power challenge." That's the main thing you have to do. We need so much power and we need it so quickly that you have to focus on companies that can deliver the speed and have the size. This is why for me, if you have an energy problem and you need to solve it, you need to solve it quickly. These small companies, that's great. You're going to need to bring in the big companies. Well, the Chevrons, the Exxons, everyone on the power side. So, just continue to focus on this. Um, management notes, AI companies need power for AI speed. Waiting five to seven years is untenable. And that's the key thing. It's growing at a pace I haven't seen in my 45-year career. This happened on, uh, Thursday or Friday.
Um, Juul Caterpillar, along with Wheeler, announced a landmark partnership to cover high-performance computer data center campus in central, uh, Utah for 4 gigawatts. I'm not going to go through the whole thing, but you're talking about this stuff is expanding and going quickly. This is the have and the have-nots. And this is what I wanted to emphasize. So the S&P Industrials are outperforming the S&P this year, but not by a lot. Not by, and I've heard a lot of people reach out and say, why aren't the Industrials outperforming by more? They're up by, they're up 14%. These are the AI names for the most part. You got G. Vernova, all the ones I mentioned, uh, in there. All well and good. You've also got the fence names. Here is the other half of the story. Basically, homebuilders, anything related to transports, anything related to trucking, anything related to all of these other things right now are not moving. They're down for the year. So far, it's been about overall the AI situation. But again, as I mentioned in the earnings, all of this happened at a time when we had tariffs.
So, just to recap, I heard about this consistently whenever I was bullish about the capex. You can't have anything good go on when capex are down. So, here's capex numbers. This is from April from Peter. And I, I have a lot of things in here from Peter. Peter's a smart guy. I like reading his stuff all the time, but he's been offsides a little bit. And the reason I'm bringing him up here is not to pick on him. I've been wrong about energy so far. I've been wrong about momentum. It's really more to highlight what smart macro people are thinking about and how the AI thing is getting past them. That's what the whole point of this is, is how they're underestimating the AI side. It wasn't just the capex side. It was the uncertainty. We saw these uncertainty charts consistently. Uncertainty is continuing to come down. It had gotten above COVID levels. It's come down. That should mean that things are going to get better. And they are getting better. The revisions continue to go higher. Remember this here in 2020 is coming out of COVID. Since the rate hike started, we've been in negative revisions consistently without any period like this. We are back in this. And to me, this is just like coming out of 2020. You have high uncertainty. Then all of a sudden, you get companies taking revisions higher. And we are at levels now. The most recent week was the highest of these. You have to go back again to the COVID period when nominal GDP was exploding. Uh, Bank of America highlighted this, one-month sales revision was two times four in July, the strongest since 2021. Again, this is a different way of looking at it. All sectors have moved upward, more upward than downward revisions to sales over the last 3 months. How can we be heading into a recession? How can all bad things be happening when that's going on? I'm tying this back to PMIs again because now the 10-week average of revisions is up here. Every time we've been at this level, this has been associated with a PMI above or near 60. PMIs to me are going to go higher, and people are still missing this, but as of now, it's been wrong.
Uh, in terms of current data, the most, you know, accurate as of now, or the things that I'm looking at that are suggesting things are about to get better. You got the Goldman Sachs Current Activity Indicator using 37 indicators. It's up to the highest level of the year now. You've got, and I didn't mention this last week because I didn't get the report in time. I had not seen, and it makes harder on Chicago. Chicago PMI is a very important one. Uh, along with Dallas, it's the middle part of the country. This is the part that I think Trump has emphasized. So when we're, you know, looking at new orders, you want to see the new orders category for the Chicago PMI. This is not public data anymore. So you actually have to go to their site. The ninth largest increase on record, and it now has the highest level since May of '22. That was when the ISM was far higher. And the backlogs rose 12 points as well, unwinding all the declines seen in May and June. In terms of other side, so the white line here is the Johnson Redbook, the weekly Johnson Redbook. So the most timely of the retail sales numbers, it went, it jumped all the way back to the highest level since before the COVID impact. Uh, the green line here is the Dallas weekly, uh, economic release in terms of the indicators. This, the last two weeks have put it at the highest level since again, before liberation day. And then you have the trucking data, which was still heading lower, has finally seen. There's the internet trucking data. And I just wanted to remind people that this was the new orders category of the Philly Fed. So I showed you the Chicago one. The Philly Fed had already gone higher as well. So you've got a lot of the leading indicators that are going higher. And then on the spending side, we got this at the end of last week, which is Visa's spending momentum, uh, making new recent highs. And again, you go through this, it's just continuing to high. And this is the highest level in some time.
And then on the payroll side, I'm going to get into the payroll side because this seems to be what got most people negative from listening to, uh, macro forecasts. I've heard some people all of a sudden that were bullish, turned bearish because of the, uh, the employment numbers. The employment numbers were absolutely fine. I showed that last week. I'll go through the reasons why, but it fits into this category. You've obviously got these massive pay structures happening in at Meta. They're not going to change the entire picture, but this is the most timely thing, the employment tax receipts, and this person puts these out every week. These are big numbers. Um, so again, you had weakness in the March, April, May, June period, and again, we had revisions come down. Then you had a bump in July, which matches up with the jump we saw in the payrolls. And now you've got the August number still coming up at 7%. These are very good real-time numbers. Despite all of the data that's been better, you have basically the bearish sentiment went to a six-week high, the bullish sentiment back to a seven-week low. So, you've got more bears than bulls at this point. The only reason I wanted to bring this up is because when you have any price action that moves down and then you get this exaggerated move in sentiment, that shows you the way people are leaning. People are still negative. They do not believe what's going on. And it shows up in Goldman Sachs' flow indicator in terms of their equity sentiment. It's sitting here. This is a very, very good indicator. In fact, if you want to see when we hit the peaks in stocks, it was up here. When we hit the bottom in stocks, it was down here. We're still not even in positive territory yet, which is with stocks at all-time highs.
So, now we started with the, the wor, the bad payroll numbers. We got precipice of recession. Three Fed officials voice market concerns over labor market. So now we've taken the Fed cuts, uh, probability of September up to 90%. This reminds me a little bit of last summer with Sam Rule combined with the situation with, uh, Japan and dollar yen. And so now we've got a little over two cuts built in by the end of the year and into January, we're more close to three cuts by, uh, by January 28th. So, the market is building in cuts, and I believe we're about to see GDP move higher because of AI combined with some of the consumer side starting to feel a little bit more relaxed and getting clarity and the uncertainty dropping off because now we know what the tariff situation is going to be, particularly the big ones. Wall Street is warning investors to get ready for stocks to drop. Again, this a Bloomberg article, so maybe it's just them, you know, hyping things up a little bit, but you did have people get negative. And to highlight this, this was a good thing by Parker Ross, which I didn't, uh, I failed to put in last week. He does really good work on on the analytics and kind of picking things up, but what he highlighted here is the labor force participation rate has fallen since April. If it wasn't for that, we'd be at 4.9%. This is allowing, and again, I believe this is related to the immigration situation, uh, and at this point, uh, in terms of, uh, the AI situation kind of forcing people out, and I just think we're going to have this problem that's going to go on where you're going to have a weak labor market on the hiring side, meaning that unemployment rate is about people working. I'll get into why I'm saying working, hiring, hiring is going to remain bad for a long time. Uh, uh, he's going through more weaknesses in, in there, and I think this is worth looking at just in terms of, you know, where we're seeing it's mainly part-time work that's showing any kind of, uh, strength.
So I want to get back into this because I showed this. I got a lot of people reaching out saying, what is this aggregate weekly payroll? So I'm going to say this again, the number of jobs is not as important as the number of jobs with the hourly earnings with the weekly hours. That is a true component of income. So, whenever you hear macro people go on and say, "Income's going to fall off because we're not hiring enough jobs, population is not growing, this is a problem." It's not a problem. First of all, I only hear about the savings rate. I never hear about the net worth. The stock market alone is up 10% this year. That means there's an extra almost $6 trillion of net worth being increased. It's a massive number, but the aggregate payrolls are the most important part of this. And if you don't believe me, then just go ask ChatGPT-5. The aggregate payrolls measure for the private sector is generally a better reflection of labor markets' total income generation than headline non-farm payroll job growth alone. Bottom line, if your goal is for hiring, focus on the non-farm payroll. If you're assessing labor market contribution, economic growth, and spending, aggregate payrolls is more comprehensive and more reliable. For you baseball fans there, that is like judging a batter on batting average as opposed to OPS. You want to have the thing that combines everything. So, use the better statistic and don't follow the people who are out there trying to scare you.
Want data to prove it? Okay, here's the year-over-year aggregate payrolls. That's the white line. This is the PCE year-over-year. This is spending numbers year-over-year. So, they move together. If you don't want to see them in that form, then just see what they are for nominal GDP. This is PCE to nominal GDP. So, nominal GDP is still at 4 and a half percent year-over-year. PCE year-over-year is still 4.7%. We are still way above levels that would be indicative of a recession. Now, you could sit there and say you think inflation is going to go higher, but that has not been the case. And that is not a, a core belief that I have at this point, other than the fact that I think inflation is going to start to go higher, uh, as on the back of the utility side and some of the other things with AI, but not in a manner that changes the nominal GDP or the real GDP side. Here is retail sales related to PCE. So, and then finally, the Johnson Redbook, the one that comes out weekly, again, to PCE spending. So, spending is fine. We're still spending at 4 and a half percent. I don't see that changing. One thing I just want to bring up only because it got a lot of, uh, exposés. This is the Cleveland Fed new rent number. It collapsed. No idea what this is about. No one seems to know what it is, but this is it overlaid with the owner's equivalent rent of the CPI. So, just worth watching. Um, we're going to have a lot of noise, I think, in owner's equivalent rent because of the utilities going higher as well.
Uh, this is where I want to change the subject into again, AI, and how much I don't think the macro side can possibly spend the time enough on AI. So this is something that came out this week in terms of people focusing on the contribution because the US economy has been weak, ex-AI, which again is kind of what Trump is saying. He doesn't want as much consumption as has been in the past, and he doesn't want tech companies benefiting the most. Well, he doesn't have that right now. But GDP, the contribution for the only growth we got was mainly from AI in the beginning. That being said, you're getting a lot of people sending charts out like this. Uh, Bob Elliot, another smart guy, puts out a lot of good stuff, but this to me is ridiculous and stupid. So if people are doing that, I think they're doing it for their argument. I mentioned Peter was going to be brought up again. All of the stuff that he has here is to minimize AI. All leading into the fact that we're, we're, we're not in a recession yet, but, but it's a close call. I mean, for Christ's sake, this is just like ridiculous for people to just stay a certain way. You have to be both ways. You have to admit when you've been wrong. And if you're going to talk about AI and minimize it in the way that they do to try to find anything about it, then at least go through and show that, hey, office construction is about to cross data center construction. This seems pretty important. We all look at commercial real estate. Ed Heyman used to put out every time I'm seeing buildings go up everywhere in cities. You're not going to see data centers built in cities. You're going to have to drive out to the countryside and go see them. The data center build is going to explode, and office construction is not going to explode. The economy is changing. It is an AI economy.
But most importantly, if you want to focus on GDP and you want to focus on the economy, you have to remember that GDP does not have any measurement for productivity. It related to AI, software, everything. They have tried to keep up with this and change it. There is no way to do it. And the way that you can see it is the S&P 500, 62% of the growth in earnings is coming from the Magnificent 8. This is the economy now. Now, back in 2014, it's not. I went to Silicon Valley back here to understand why my bearishness on things like Amazon was not working. Why my bearishness things on the economy was not working. That's when I started writing papers saying "adapt or die" to economists. You cannot look at the economy using statistics that were created and indices that were created like GDP in the 1930s made for the industrial revolution. They have nothing to do for today. And the S&P 500 shows you where the earnings are coming from. And this is their market cap. It's actually lower than where it is, meaning it should be higher at this point based on the earnings growth. And it just continues to move that way. The profit margins, this is insanity. And the reason it's insanity, go back to here. Their market cap was next to nothing. Their profit margins are growing. So for everyone worried about the non-tech side, the non-Mag 8, this isn't even tech. This is the non-Mag 8. You're dealing with profit margins that are not expanding at all. And that's why there's a bearishness that's out there because most of the economy is not moving. So by not focusing on the side that is and trying to use data that says the economy is bad is just wrong.
And now we're at the most scary time, which is their profit margins were already rising. Now they're not even hiring people. So the world has changed dramatically since ChatGPT. We're not hiring people at these companies. And so here's the GDP of the digital economy. Let's just use these numbers as the GDP of the digital economy. Let's use this as the GDP of the non, uh, uh, Mag 8. You can see there is a recession here. What, whatever you want to call it. There's no growth in most of the companies outside of this. But those companies matter more than the other ones because their numbers are massive in terms of the earnings growth. Can GDP measure properly? Can the GDP measure properly calculate the productivity of software and AI? Go do this in GPT-5 on your own. No, it cannot do it. It's not made for it. It doesn't go through. It's not made to capture that. Here's where you can see what has happened to it. This is the benefit that has come. The productivity has come in profit margins, which has come been directed back into the economy through net worth. This is the way that it shows up. So in the net worth, this is net worth in the US relative to GDP. So I've shown this chart before. This was the beginning of the personal computer. This was the beginning of Netscape in the internet. This was the beginning of the iPhone. This is the beginning of ChatGPT. So people in the country, and again, yes, it is isolated to the top 20%, but the net worth has gone up significantly to where it is now 5.6 times GDP as opposed to three and a half, two turns of GDP. There's an extra $50 trillion to spend. Now, it isn't at the bottom level of things, but that's where transfer payments have come to fill in the gap. That's why I talk about Bitcoin because the government comes in to fill in the people that are not benefiting from this, and that keeps the economy going. People at the bottom can get a job whenever they want. They can, they can spend what the government's giving them. They're just not able to go out and get the loans and do things. And that's why the government is going to figure a way to to to to move interest rates lower because it is helping the Mag 8. It is not helping the rest of the country.
One final one for Peter. Again, I like him. I think he's smart. He called everything right in '22 and '23 when people were bearish. So, I give him his props. But right now, he's just making a big mistake on this. Uh, if LLMs turn out to be a stepping stone to super intelligence, that will all change. He's minimizing AI. Like, I can't even think of anything worse to do as someone who uses it all day long, uh, for everything you can possibly imagine. Uh, it's a strong assumption on super intelligence, of which debates today's equity valuations. Now, he does this when an art, just after an article comes out from The Economist, "What if AI Made the World's Economic Growth Explode?" What is in there, and whether or not you you buy into it, they're talking about AGI, not ASI. So he said super intelligence, AGI is happening now. It's already going on. It just hasn't reached a point. The reason I say it is happening is because we've already gotten into the, into each of the places with math, with physics. We're starting to pass. We basically have AGI already. We just don't have one computer doing all of it. We have specialized computers that can do each of it, and they're winning awards on this. This article posted, "AGI could dramatically shift this paradigm, delivering 20 to 30% annual GDP growth." Again, I've said it before, I don't think this is going to happen. I do think GDP will be higher than what people think, but I don't think the calculation of GDP will allow it. I also don't think the benefits that are going to come from people not getting the wage growth that they should, they're going to remain scared. And if they remain scared, they don't go crazy in terms of spending money. So, I don't know how it'll do. The physical limits, energy, infrastructure, data capacity could constrain scale. I agree with that. Um, it seems like for the 20 to 30% GDP to it would take massive productivity boost. Would this begin with digital employees like AI agents and followed by, yes.
So again, I believe the reason everything is so important now and why it is incredibly bad to be bearish about growth going forward and profit margins on the S&P 500 is we are at the beginning stages of AI agents. As AI agents come, dramatically reduces the cost of information work. Early adoption in white-collar sectors like law, finance, design, marketing, software, and R&D could multiply productivity without hiring more humans. Now, it happens at a time when we're going to have a shrinking of the labor force due to immigration and due to continued demographics. So this timing works in well from the labor market not going crazy, and by having more retirees, they have money to spend, increases output per worker, and allows smaller firms punch scale embodied AI in robo taxis and drones, which is beginning to get rolled out. It will cut transportation costs, one of the largest input costs in the global economy. So you'll get a reduction in transportation costs because you'll won't have human beings driving the trucks. Um, embodied AI in transportation improves supply chain reliability, reduces downtime, there's an efficiency basis. Now, when humanoids come, all of these are happening over the next four or five years. So you have AI agents that are coming now, then you have robo taxis over the next couple years, then you have humanoids after that. So to talk about GDP numbers where we're going to have robots and humanoids doing the work, productivity by definition has to go up. Societal pressures will probably increase, but that's a different story. GDP has to go up. You can go through the compounding effects. In economic history, this would be close to an industrial revolution plus an internet revolution plus electrification, all hitting within one or two decades. So again, this is all what's coming up. The stock market is a discounting vehicle.
GPT-5 was released. I want you to read this. Capable of generating software on demand. You guys can all build your own software. Just go tell GPT-5 what you want. Alan just replied to my post. Everything you will need to be generated on demand in real time by AI. Operating systems and apps will be things of the past. Think about how much of our market, how much of the global MSCI World is based on operating systems and apps. OpenAI is going to eat Microsoft alive. So you see what Elon Musk is is posting. People have been trying for 50 years, and that's the fun of it each day, blah, blah, blah. Again, this is a major, major release. This is what, and I mean, Elon Musk is fighting with Sam Altman, but yet he's saying they're going to win this. This is about the end of apps and operating systems. That is what super intelligence, at least from what Peter Barerisen was was posting, it's really AGI, and we're at that point now of this. Now, I want to remind you, I did this a year ago. It's like the one-year anniversary of me doing these videos. I posted on this video and said it was ever something everyone had to go. This is the band interview that was done at Stanford by Eric Brinson in, him talking about a bunch of things. You have no idea what's coming. Now, that was a, that was a year ago. I this video came out. I heard it was done. The, this was in Stanford, so this was during May or even earlier, but it was last year. Schmidt's comments on lowering software creation barriers center on how a combination of massive context winners will collapse both the technical and organizational friction that separates an idea from a working piece of software. So think about what Elon Musk put back to what Eric Schmidt said. He's basically saying the idea to prototype cycle from weeks or months into minutes or hours. His example, "Build me a TikTok clone with my preferred features, migrate my followers, and test viral content strategies." The AI could not only generate the code, but deploy, test, and iterate automatically, affecting asking as a fully developed shop. So, all apps would be under pressure.
Here's another quote that, uh, went into it, but here's what I propose each and every one of you students do. Say to your LLM the following: "Make me a copy of TikTok. Steal all the users. Steal all the music. Put my preferences in it. Produce this program in the next 30 seconds. Release it. And in one hour, if it's not viral, do something different along the same lines." That's the command. Boom. Boom. Boom. You understand how powerful that is? If you can go from arbitrary language to arbitrary digital command. Imagine that each and every human on the planet has their own programming that actually does what they want. That's what GPT-5 release was. Here's Eric Schmidt on, uh, Moonshots three weeks ago, which I referenced already talking about that it exists. Schmidt discusses Google Cloud Platform. So a nice protected environment with inside Google MCP, which is something Anthropic released last year, which allows you to connect in here. So Google Cloud Platform using all of the AI LLMs and data analysis. Use this, think of it as a tunnel into your software. Google Cloud now offers a fully served AI enterprise solution that enables businesses to automate internal functions without traditional middleware. With the MCP, an enterprise can write the task it wants an AI to perform, connect internal databases directly, allow the LLM to generate the necessary code, eliminates the need for legacy middlemen, enterprise disruption, legacy providers, SAP, Oracle. Basically, he's naming the biggest software companies. This allows you to do it. It's only going to be a matter of time before all start. I'm going to do this for startup hedge funds. You don't want to use, uh, uh, a CRM system, just keep the data inside your own server on Google Cloud, and then you can go out and connect to MCP. You can do all of this in a matter of minutes, and then you can ask the LLM questions on the data. You don't have to do this in the AI agent world that was said by Salesforce. They can do most of the work that way, and that's what this all goes through. No more middle error.
Schmidt says, "There are hundreds of thousands of enterprise software and middleware companies built over the last 30 years whose entire business was stitching together databases, APIs, and workflows. With MCP and LLM code generation, that layer is no longer needed. Junior programmers are the first to go. We are on the cusp of replacing most programming and mathematical work in the next one to two years." Again, he said this the other day. This is before ChatGPT came out. I'm not going to read all of this, but I do want to real centralize has dominated the CRM space. So, I asked it to talk about Salesforce. Are they up for disruption? And it goes through all of the details on it. And again, do that in GPT-5. Is this easy enough for someone with zero coding skills to do at a company? The whole point is that you don't need to be a programmer to get useful software built. That's the whole thing. Here's why even a zero coding skills person could use it in a company. You literally describe the task in plain, plain English. "Pull all customers from the last 90 days. Sort by revenue and generate a chart of the top 20 customers." The LLM interprets that, writes the necessary queries and code, and runs it. I won't go through the rest, but then you can have it send out the emails. You can do whatever you want. Again, it all exists right now. It's just a question of people doing it. And it all exists for no more than $20 a month. So, does this effectively make any CRM like Salesforce worthless? So, let's getting into the pricing side. Yes, I don't have to go through it. Um, I'm not picking on Salesforce. I'm just picking on software in general. Uh, with text to action and agents, the cost of building software drops near zero. Other than compute, Mo said the barrier to entry for launching competitive software shrinks, and Schmidt warns could make incumbents like Google more vulnerable to attack from small teams or individuals. Time to market collapses, increasing competitive pressure. This is what I've talked about in terms of companies in the 2030s not surviving. If you're based on code, you are going to be disrupted. This is all going exponentially. So, this is why you have to focus in the same way that Amazon destroyed retailers. They didn't have the time to catch up. Now, we have something that's even worse. This stuff is moving faster, and you don't need to hire the programmers, which were the moat. So, coders were the moat. Allowed these companies to have a moat. Now there are no more moats. Industries that are disrupted map to which industries will be disrupted first. Many SaaS offerings, AI can recreate 80 to 90% of functionality quickly. CRM replacements for Salesforce, project management tools. You know, you can read all this stuff. Slack, Zoom, Notion, go for it. Um, this is, this is why these names are down 50%.
So, what I'm highlighting now is that nothing I'm saying, the market's been discounting it, which is both a good thing and a bad thing. Um, I've been following this person for a while, Jared Kubin. Um, and he highlights the fact that long software versus short semis. Basically, semis have massively outperformed software. So, the market's discounting this already. So, can there be a bounce? Yes. But I believe this is the beginning of a long-term trend in software. And the way that I want to play it on that same interview that Eric Schmidt did, he said, "We will run out of power as this accelerates." What AI agents and all of this coding by individuals does. By allowing them to speak their coding, and the code is massive. It will get more people building stuff. It takes more and more power, and so software companies should underperform. Here is the trend of long Chevron, short salesforce.com. This is my favorite proxy for this turning point. This, I didn't bring it back here. It was much higher back here, but this is when the software boom started, meaning this is when, uh, the iPhone came out. And what you had is Salesforce.com outperformed Chevron dramatically. They're both right around $300 billion. You have this nice five-way move up here. Then we've had a correction that went here. This was all based on ChatGPT's launch. Now we've got it turning back up. So to strip out and to look for long energy, I want to be short software names, and you can go pick the ones you want.
Finish up, this Bitcoin still consolidating but starting to move higher. I talked on pump on Friday about the fact that Ethereum, once it got above 4,000, I thought it would start to go. Um, we're above there as of this morning. We're about 4,200 as I do this. So Ethereum has broken out. I expect all of crypto to continue to go. Uh, Trump signs executive order allowing cryptocurrencies, private equity, and 401ks. All this does is just continue, like it did with the ETF, to create bids underneath. Uh, there's now three cuts expected by the end of the year. If I'm right about PMIs, and even if I'm not right, if the market is getting better, meaning stocks are going higher and they're cutting rates, then financial conditions will continue to ease. The surprise here now, having heard, I don't know, 50 different podcasts with macro people over the last, let's say, three months, anytime you hear someone say rate cut, they all say that means the long end's going to go higher. I'm starting to believe that the contrarian trade might be that they cut rates and that long-term rates actually go down for the time being. I still think they're going to go higher, but not until the inflation starts to come through. I don't think the inflation comes through in a meaningful way until the PMIs go up. So, you have to watch these things together. But maybe the surprise is they cut rates in the expected way because last year when 10-year rates went higher, remember there was only 37 approximately basis points built in, which means 50, which is what they did, was a surprise. Maybe we get rate cuts and you don't get the long end to go up immediately. It goes up later. Uh, I, I think that's the one of the, the surprises that may come out.
US does their stablecoin side, and China's basically, as far as I can tell, throwing in the towel. They're worried, uh, more. In fact, if you come down here, authorities remain wary of capital flight and money laundering, slowing broad adoption. Um, stablecoins, they have to have their own stablecoin. So they're talking about issuing it. Um, and I believe this means that they are continuing to focus on providing liquidity to locals. And finally, Sam Callahan posted this, and it's just something that you can go read about. This is about Bolivia and the land of 25% inflation. Crypto is starting to replace money. This is the future, guys. It's coming this way. We're at the critical point of AI. And as I highlighted, macroeconomists are not focused on what's happening on AI. They're still looking backwards at GDP and at measurements that are no longer relevant. Uh, keep enjoying the summer. I'll see you next week.