Transcription
All right. Up in Maine this week. Uh, doing this just before the close on Friday. So, if some of the prices don't seem to match up with the closing prices, uh, just because uh, I'm doing this so I can get out of Maine and get back to Brooklyn.
Uh, lot to go through. It's amazing how many things happen in a week. Uh, obviously, this is uh, definitely taco week, but I'm going to go through and and just again, get everyone to stay away from the S&P 500 as a gauge. Uh, 95% of things that I listen to and watch in podcasts related to the market, it's mainly about whether the market's going to go higher or lower. And I think at the end of this year, you're going to destroy yourself playing that game.
Uh, I'm going to stick with the same theory, which is that this is going to feel like the 1970s. Uh, we're going to have inflation, which is why there'll be a big part of inflation this week. Uh, how you invest in inflation is the opposite of the way uh, people have been. This is a scarcity trade. The abundance trade got destroyed again this week. We'll go through that.
Uh, but, you're going to start hearing me talk a lot more about Bitcoin. I think Bitcoin is entering a period where every investor is this is the beginning of your journey of understanding it. Uh, growth as an asset in terms of stocks has just been destroyed. And I think this is going to continue. Uh, the terminal value story is only going to worsen. We're going to go through that with Mythos. I wrote a piece on this. Uh, I've I've done a lot of writing over the last couple weeks. Thanks to all the subscribers reaching out looking for stuff. I'm going to go through my thematic basket today in terms of just some of the things that I've talked about. Uh, there's been some big moves on the upside. There's always going to be some losers in there, too. Uh, we'll talk about Brazil, which has been one of the winners. And we'll go through silver, which I did. And also, I'll start to bring in some new things.
Uh, all right. I I'm just going to go through these first three images cuz I've been showing this since I think the first time I did it was somewhere in October, November. Uh, this was all related to Opus 4.5, the acceleration, the destruction of software, uh, and the need to have a portfolio which was based on rising PMIs, rising commodity bull market, and anything that was scarce with compute being completely insatiable from the demand side. The physical world upgrade, which was the theme for this year. Again, moving out of anything built on code, moving into anything AI in the real world as we shove artificial intelligence into everything. This is the first inning of this.
So, for those of you still trying to short SMH, still trying to follow Michael Burry shorting Nvidia, uh, anything on the software side can be shorted. Even Palantir, which is a name that I still like, uh, but fully understand that no matter what you like in software right now, uh, it doesn't really matter. AI is really accelerating at a pace that it makes any certainty on the future impossible for commodities. And for semiconductors or any kind of hardware, that is not the case. And again, the speed of this is the damage. So, I've shown the supersonic tsunami. Uh, these are all photos of what's going on. And that's why when you read stuff like this is a big week for stocks.
All right. The S&P 500 finished slightly down for the year as of this week. Uh, it was up at the beginning, then it was down. Now, it's back to unchanged. And within the sectors, you've got a lot of things up and down. So again, we were down five weeks in a row. We got a bounce. I don't think at this point uh, you should be focused on it being up. I think we're going to see a lot of green and a lot of red. I don't think this is going to be a repeat of last year. Like I said, this was liberation uh, week day week. And then most of them the weeks were up. I do not think we're going to see this. I think we're you're going to see a pendulum back and forth, and I'll go through the reasons why.
Q's after five weeks in a row down, made back most of it up 8% over two weeks. IWM, same thing. Here we are year-to-date. Uh, energy, materials, industrials, utilities. This is basically your scarcity trade. And here's your abundance trade down here. Uh, financials are down here because again, they are for the most part something disrupted by code, but also they've got a lot of connection in the debt market to the software side, which obviously has been the private credit side.
Uh, here's the best way to look at it. Here's the bounce in the S&P this week. Taco time. We get a big bounce. The white line here, IGV making new lows. Not only making new lows, making sharp new lows. And at the same time, private equity continues there. No bounce in the deflationary side of the market. Zero. This is why this is not last year. This was not a rally in getting rid of the problems. And the reason this line is here is this line was when the war started. So, the S&P rallied almost back to where it was. This stuff not only didn't budge, it's lower than it was at the start of the war. It had fallen before the war. Do not forget this cuz this story is still playing out, and it does not have a floor.
So, I'm going to say it again. I've said it in software all year. You can try to pick bottoms in something, you are wasting your time. It's like playing trying to pick directions in the S&P this year. You've got better things to do with your money than worry about whether the bearish or bullish sentiment is high. You've got a job to do, and the job to do is to follow inflation on the upside in the shortage theme.
Here's the S&P chart. What I want to show you is this is the way people are going to look at This was last year. Now, this day here, we had a gap back above the 200-day moving average. Very similar to what we did here. So, here is Monday in terms of the reversal of Trump uh, not obliterating a nation. Uh, we get the gap higher. You've got a bunch of short covering. You've got CTAs uh, at very large short positions. They're systematic strategies that are going to have to cover. And by the way, the reason these positions are so big is cuz we never saw the vol expansion that we saw down here. So, when you look last year, we didn't really see the buy flow from CTAs. It was spread all out of here. I think this one's going to be much faster just because we never saw the vol jump. So, they're going to have to switch back and forth. I think this is going to be a painful year for systematic strategies uh, as opposed to last year where we got this nice smooth uptrend. I think this is going to be a volatile, uncertain year.
Now, that gap back after liberation day, that came from, believe it or not, the US-China agreement on in uh, in Switzerland last year on May 12th. That's what the gap up was that time. There was this statement on the White House. Guess what happened again? Monday. Another release. Peace through strength. Operation Fury. Again, this is we know the weak point with the administration. At some point with gas at the pump higher, the midterms in check, stocks down, you got the flip. Everyone uh, had to cover shorts. And now we're back to a more neutral stance.
The difference is, and this is what we're going to see. Last year, after we got the bounce out of liberation day, here are the estimate revisions for the end of the year. This week, we have a big down move. Before then, we had nothing. Analysts did not turn negative. Why did they turn negative here on tariffs? I don't really know in hindsight. They had to revise everything back up. This time, the bar is far different. And that's what you have to remember. We kept getting earnings revisions last year because people expected inflation to be higher and expected there to be a recession. Now, with inflation certainly to be higher, it never budged here. We've already seen it go higher. Analysts are not turning negative. I think there is negative news to come both from AI and the software side. Uh, everyone keeps hoping that there'll be good news on software. At this point, with good news, software has gone down. Can you imagine what'll happen if we start getting bad news in software where you can extrapolate it out to three years? Uh, that will come at some point, mark my words, because this is a true disruption.
Last year, we still had rate cuts going on around the globe. So, during this whole time period last year in 2025, while we still had into September, we still had rate cuts. Well, now we've got rate hikes around the globe that are going to start. And last year, we never got a budge in inflation. In fact, here were the inflation numbers. And today, we got a bump up of 0.9 on headline CPI, the highest monthly number since the peak in year-over-year CPI back in June of '22. So, we already are seeing inflation go higher. Now, core was lower. I'll get into that later. Uh, this will be the year that uh, you'll learn the difference between core and headline. But, it's going to come at a time when affordability is an issue and when there's a distribution of wealth problem. Uh, higher gas prices, higher plastic prices, higher diesel prices, this is not going to do well for either the administration or it's not going to do well for the market in terms of a lot of things in the past through.
Here's gas prices. This is the 40-day rate of change. The reason I want to bring this up, this is the biggest move in the last 22 years in terms of gas over a 40-day period. That's we have a lot of inflation still to come. Diesel, even more dramatic. Diesel, every truck in the country, everything related to refined products is going to continue to see. And right now, we haven't seen any pullback in this. So, even though we have a ceasefire, we've barely had a budge in even crude futures. We had one up day on Friday before the weekend of 11 bucks. Now, we're back down and we basically settled where we were before the Friday. So, as of the Thursday, we're down about four or five bucks. But, diesel prices and all these refined products, we got a lot of damage there. So, I think the big difference is how long is it going to take for this stuff to normalize? Took a long time in 2022. Uh, we'll see how long this takes.
This is uh, basically plastic uh, again, straight up. So, this is the plastic price, which is going to flow through to everything. Uh, this is for fertilizer. We haven't seen food prices up yet. We haven't seen the the plastic prices up yet. These are all prices that are going to flow through. So, that CPI number you saw 0.9, you can expect at least another two months of numbers that are going to be somewhat similar uh, unless we see a quick reversal, which just seems unlikely.
Today, China moved to ban sulfuric acid. That's a key component in a lot of different things. But, in particular, it's part of the fertilizer uh, side. It's also part of the copper side. So, them banning that, again, when the Chinese ban something, that highlights that the shortage is there. It means they're focused on that. It means someone is suffering from that cuz China is a a big exporter. So, they're banning that. You still have helium prices that have spiked. That goes into a lot of different things, but obviously semiconductors. You think about the breadth of the problems that are coming from the commodity shortage side, which is not going away.
Last year, we had rate hikes or rate cuts still built into the market the entire year, and we actually ended up getting rate cuts. This time, we had rate cuts before the war as of here. We had 2 and 1/2 in the market. This is for the December or the January Fed funds. We're zero now. We've taken out the cuts.
Here's the services PMI. We got the PMI this week. This is prices paid overlaid with the CPI. Now, remember, this is the service PMI. I just want to highlight that the service PMI matters a lot from prices paid. And aside from this period in 2022, up at 71, you're talking about the highest level since 2010. Now, this CPI print, we're now up at 3.2. So, we're up here. We have two very easy comps coming over uh, uh, the next two months. It is highly likely, based on what's going on, that we will be well above 4%, maybe even let's assume 5%. At 5% CPI, if we get there, we're talking about the highest number since just before the great financial crisis. If somehow or another we get above 5.8 and up to six, you're talking about the highest number since the 1990s. So, again, I think it's very hard to look through what's happening on CPI.
Now, for core, even though the core came in at 0.2 today, the PCE core, which is the Fed's chosen part, we just had our third month in a row of greater than 0.3. The only time over the course of the last 25 years that we had three months in a row of greater than 0.3 was during the inflation boom here. It's the only time, guys. We haven't even had back-to-back months, except for this one little point at the end of '24. So, we've had three months in a row of core PCE above 0.3. A lot of this, again, has to do with technology. Core PCE was firm in February because of goods. Goods inflation was 0.84. Okay. Again, if you exclude '20, it's the highest run rate since 1991. You cannot exorb ignore this. You just can't. Uh, so, here we were, 0.2, 0.9, 3.3 on the headline, 2.6 there. Market initially liked it, and then I think we finally have reached a point. The wedge between core goods in the CPI at 0.2 and PCE very high is due to the computer software and accessories category. This is a problem, guys. And look here, the software category was 6.6 five in February, has 40 times more weight in the overall PCE price. As I go through this, I'm just going to tell you that compute prices are only going to get worse going forward. Uh, cell phones, computers, we've got shortages in Mac minis and Mac studios that are extending further, and we're going to have a server problem, CPU problem. All of these things are coming.
Uh, the Fed expects little change to underlying inflation. It's great. We've got a 0.9 versus a 0.2. We put it post a couple more months of this, and the Fed's going to leave that alone, which means rates are going to stay below inflation for a period of time.
Look at the ISM PMI report, and just go to the little section in the report. Commodities reported up or down in price and short supply. So, if you think there's a bull market in commodities, look, up and down in price and in short supply. Aluminum, every single commodity here. This is not just oil. This is everything in here. Nothing is down in price. Commodities in short supply, electrical components, bearing components, electrical components, memory, rare earth. All of this is AI related. This is only going to get worse, guys. This is going to be a problem.
This is one of the positive things for Bitcoin. For all of you uh, who follow me with pump and your your Bitcoin lovers, uh, and all you institutional people that don't embrace Bitcoin, start to embrace it. Uh, if you've seen some of my calls, and if you've gone through this, and you think I've had a decent uh, a decent look at AI, especially with some of the names, I'm telling you this is the thing Bitcoin is made for. Um, you're going to have negative real yields very soon on headline inflation, and I'll show you why that matters. But, you're also going to have problem in a lot of countries based on what's happening with the Middle East in terms of EM inflation and PPI prints saying we're going back to double digits globally. You're going to start seeing some devaluations in some of these currencies over in emerging market land. Um, and we'll go through the ones that I still like, like Brazil.
Uh, again, abundance versus scarcity. That's what that whole thing is for. I can't say it any louder. We've got inflation over here. We have deflation over here in terms of software problems. Financials are fitting into it. The problem is you can't print your way out of this one. So, for the scarcity side, it's not getting any better because where you going to find the the supply? It either comes through demand destruction, uh, which I don't think is going to happen cuz I think they're going to deal with the energy side. But, just remember, this whole physical upgrade trade, first inning.
Jeff Currie on the structural trade, and I mean, he's getting in it. I I could not agree more. The bargain that trades US security for dollar dominance is breaking down. So, think about what happened with NATO in terms of the arguments. Think about what Trump posted on Saturday and how it just seemed to be this flailing thing uh, about what was going to happen. So, we've backed off. That's all well and good, but the question is, does the world trust the US to be there when it needs it? The nations didn't come in. This There's a much bigger story here that impacts the odds. Regardless of what your political beliefs are, you have to view this as where does the dollar sit in this whole thing? So, again, the dollar rallied during a crisis mode. It fell right back off during a period here. I agree with Jeff on this thing. And one of the things he adds, if I if Currie could put all his money into one technology, it's batteries. The military case, the grids case, the security case, all converge on storage.
I've been talking to a lot of you about batteries. It's a hard thing to play. Lithium is on the critical minerals list. I have a bunch of uh, battery stocks. A lot of these have fallen, like EOS have fallen sharply with the software unwind and the growth unwind. I still think the battery trade, and Tesla's fallen off, one of my favorites, too. Uh, all the way back down to where it was in the summertime. The battery and energy trade for these names is going to happen again. Uh, just got to wait for a little momentum to start going, but we saw some change this week. I think this is going to be a bigger theme coming out of this.
Computer equipment investment just hit 1% of GDP. Just look at these numbers, guys, and just look back in history. Business computer equipment investment as a percentage of GDP. This is why being long consumer or being long software or being long any of this stuff about the way GDP was in the past is completely irrelevant. We are now in a commodity hardware bull market, software, all of that.
Last week, I put something out on transports. Just like I did when I put a Marvell piece out, and I got pushback, and I got a pushback on a whole bunch of things. I love pushback right now. Transports, there was pushback. You guys are missing the boat on this. For people who are permabears on AI, suffer the consequences. For people that are bearish on the economy right now, stop looking at housing. Stop looking at autos. Stop looking at all of the things related to the consumer. The consumer is fine. The wealthy people have more money to spend than you can believe. Transfer payment people are getting 5-plus trillion a year. We have ex- expenditures almost up to 20% of healthcare, which is non-cyclical. This is not 2006 before the baby boomers retired. What we have right now is a boom in this white line, which is capital goods X air and defense. This is a straight line up, and this is overlaid with the transport sector. So, everyone who sent me something when I showed flatbed rates last week and said the economy's not going, you're missing it, blah blah blah. I think you guys have to do some homework right now and start looking at more than some people that you like to read who are bearish.
Uh, Goldman Brack Goldman Sachs embraces the picks and shovels of AI with more capex ahead. AI trade in full full full full-blown bubble right now. Uh, all right, I'm going to write a piece about Nvidia this week. Uh, it's it's in at least one, I think two of the baskets, but definitely one of them. So, it's on my thematic list. As we've had multiple compression as a big story this year, we've now taken Nvidia's 2027 PE down to the lowest level in the last decade. Stock is having this nice consolidation here, in my opinion. Uh, in my opinion, it's going higher. Earnings just continue to grow while the PE has sunk down to 20. We are at an agentic side, and if you don't believe me, go listen to this podcast.
Uh, part of the reason I'm writing this paper on Nvidia is Jensen Huang has been a gift that's kept on giving. I run almost all of my AI models on things that he says during the week, and the names that I get, like a Marvell, like Coherent, any of these ones that he's made investments in, he tells you way ahead of time to put money in. Not literally, but if you use AI and you create a knowledge base, go look up Andrej Karpathy and figure out what a knowledge base is, you can go out and you can start getting data, but you have to go through this to figure out where we are in AI right now, because everything he said, the market got inference wrong. And when he says the market, it's pretty much everyone, guys. Inference demand is exploding because models are no longer just answering prompts. They're increasingly reasoning, planning, searching. This was not supposed to happen this quickly. So, it's not just that people were wrong, they were wrong this soon.
So, Open Claw, Cohere, Perplexity, compute, every single one of them is now agents. Agents are running everywhere. We don't have enough compute to deal with all these agents, and now enterprises are adopting at a pace far faster. If you've seen Anthropic's number, go back to what people were saying. There was no way to monetize AI. That was 6 months ago, guys. And now we were at a run rate of 30 billion, up from 14 billion, up from 9 billion. The numbers keep growing now in terms of this because they're taking in tons of enterprises. So, when you go through this, we need more CPU, we need more memory, we need more switching, we need more networking, we need more power, we need more cooling, we need more storage, we need more software. All of this stuff is what you can invest in outside of software. He only says that because a lot of his customers are in there in software, but we do need more software. It's just going to come from smaller businesses. Everything is about performance per watt and heat, everything. Computers are turning into AI factories. It's becoming the center of the system and the center of the business model.
I'm about to start thematic ideas on the edge device side, because we are now at the point, especially with Mythos, as I'll get through that. If you haven't read the paper I wrote on Mythos last week, uh, guys, you really need to subscribe and be a part of this just because I don't see the uh, street as being on top of this in terms of what how to go through it. This is really a macro theme that goes across technology. And when you go through the edge device side, there are a lot of winners that are not part of the current trades. This is why I spent time on the rack. The rack is part of it, but there's other places to go through cuz remember, as much as right now we've got AI disrupting software, it's going to get into the physical world through networks, vehicles, robots, satellites. That is going to disrupt other parts. So, it's not just a long side, you have to be on top of this for the short side as well, because a lot of the alpha this year has been traded by using hyperscalers on the short side and using software on the short side versus semis, but other longs.
I want to bring this up, too, because he also the only non-tech company he mentions is Eli Lilly. It's a gift for you non-subscribers. Eli Lilly's in my basket. The reason it's in my basket is because Nvidia and Jensen Huang talks about him more than anyone on this. Uh uh The chart looks like it just had a reset. They have earnings growing because of uh Ozempic or whatever their version is. Uh, but you also have AI discrub drug discovery, and I wrote this paper in November. This is still a theme. It's still real.
Here's the Anthropic chart. So, again, everyone that was fading AI adoption could not be more wrong. It's gone parabolic. This is the Open AI chart. Don't be surprised if Open AI starts to pick up, but regardless, uh, this number has gone parabolic. I will say, remember the last time Anthropic said what their earnings were, they had a problem because the margins were still coming down. They're going to have to raise prices, but we have a bigger issue. And the issue is, as Warren Pies shows in probably the best thing of measuring supply and demand, there is no compute available, guys. Uh, literally, there is no compute available. So, look at how far down this stuff has gone this year as the agentic side has kicked in. We can't keep up supply versus demand. That's why, as Axios put out uh, this week, AI compute wars. Now, again, Anthropic's server capacity isn't keeping pace with demand, leaving paying customers stuck on usage limits and outages. I can confirm that that has happened to me for the very first time. Yes, I had issues with Anthropic a long time ago, but once I upgraded to the highest package, I never had a problem. I did this week.
Dylan Patel, I highlighted this video that he did in uh, in terms of uh, interview with Dwarkesh Patel, his roommate, highlighted that Anthropic may be pushed towards lower quality computers. Open AI locks up premium supply. Commute compute costs are plummeting as efficiency in chips and uh, uh, increase. However, usage is skyrocketing faster, so total spending keeps climbing. It's Jevons paradox. Even at rack record capex levels, the industry isn't buying enough compute to meet full demand. On the 30 billion, this is what they had to do this week. Anthropic expands partnership with Google and Broadcom for multiple gigawatts. They're basically going to start using TPUs. This is what Dylan had talked about. They don't have a choice. This is what they put in there 2 years ago, a dozen customers paid Anthropic over 1 million uh, on an annualized basis. Today, that number exceeds 500. So, 2 years ago, they've gone from a dozen customers to 500. These are uh, Palantir-type numbers. Eight of the Fortune 10 are now Claude customers. Eight of the Fortune 10 are Claude customers. That is not good for Microsoft. That is not good for the software companies. Business subscriptions to Claude code have quadrupled. Uh, you guys can read this. They're getting into financial and data analysis, sales, cybersecurity, scientific discovery and beyond. In January alone, we launched more than 30 products and features. That's how quickly this stuff can come out. Think about Copilot. Think about how useless it is. 3 years plus since ChatGPT, and that's what we have. They put out 30 new products, including PowerPoint, including Excel. Uh, software companies cannot move at the speed. It's just not possible. Anthropic admits Claude code's users used limit way faster than expected.
Um, again, nobody was ready for this. So, this was the $200 one. They just unveiled manage agents. So, this is their competition for Open Claw. It's getting easier and easier to use agents, guys. Uh, if you doubted the token use from them, here's the token use going up through the roof. And remember, this is basically mainly models that are open source. So, the majority of this is open source. So, the frontier most expensive models have gone from 22% to just 4% today.
Alex Finn, um, a must must watch, but also a must listen to in terms of his ex-post and going through it. Something colossal is going to happen in the next 6 months. He's writing this all related to AI agents, which again goes back to the point. All of my 100 names in the thematic baskets, one way or the other, are directly connected to AI agents or at least AI adoption. They are all on the physical side, but they all fit in with the fact that we don't have enough compute. People keep fading this, but I'm just telling you that when you go through, he's also highlighting something I've said to parents, professors, kids at school, my son, as I've talked about, it has never been more critical that you were up to date on all the latest AI tools. The future is entrepreneurship. Jobs are about to be impacted. So, whatever you thought about the job situation, be prepared now cuz we're getting into a very different situation with Mythos, which again, I'll cover again. But Mythos is something you need to understand.
Uh, Michael Dell spoke. He talked about how much memory demand is surging, and again, he did this at a Bank of America event this week. Uh, I just wanted to highlight how big it is and just that it's all happening at the same time. The demand is just picking up. He noted that investment pullbacks by memory companies amid the 2023 downturn have resulted in supply shortages. Again, this is a very similar story to what's happened commodities. There's just been underinvestment in the space because no one wanted to get hurt again hurt again in terms of overinvesting. Everyone, investors, uh, semiconductor companies, people underestimated how quickly inference would get here and go to the demand. The reason is because of this. Not just because of Mythos, but once Opus 4.5 came out, the world changed. I'm going to show you more charts on that. Make sure that from an investment standpoint, you understand the reason why it's important.
The Mythos signal. I wrote this for 22V last week. Again, for the subscribers who saw it. Mythos matters not just for what it may be, but for how it surfaced that's evidence that frontier may be moving faster in private than the public fully Recursive self-improvement basically means that the models are able to make themselves better, develop things on their own. Here are the podcasts you should go listen to to learn more about Mythos. Uh, I listen to this one regularly, the AI daily brief. I haven't really mentioned it too many times on here. Um, no real reason. It's a short one, but it's every day. But they went through should we be scared of Anthropic's Mythos? They also put one out the calm before the AGI storm. It's here, guys. We are at that moment. That is the reason why software dumped this week is because now people are starting to come to terms with Mythos, which again came out last week or at least was released leaked last week.
So, speaker says Anthropic has officially confirmed Mythos, its most powerful model yet, but it is not releasing it publicly. It is flaming It is framing it as powerful enough to require a controlled rollout because of the cyber risk. The reported benchmark jump is unusually large. The biggest gains are in coding and agentic tasks. I'm not going to read all this other part. It is capable of highly autonomous behavior, which gets back into the recursive self-improvement side. This is the calm before the AGI storm, guys. Um It's faced backla- Anthropic faced backlash over user limits. Uh Gemini 4 is presented as a major open source model. The next generation of models may represent a real step change with speculation around new Open A multi-modal systems and Anthropic's Mythos level advances. The biggest conclusion is that the industry may be approaching a phase where incremental updates are no longer enough and a much bigger transition towards superintelligence is starting to come into view. Something that people didn't believe in last year, but didn't think was anywhere near where there.
Why Anthropic's most powerful AI model Mythos preview was too dangerous for public release? Anthropic model scare sparks urgent Bassant Powell warning to bank CEOs. He called for an emergency meeting. When we have emergency meetings, and this was so sudden that Jamie Dimon couldn't attend. So, this is when they saw the model and the Treasury Department wanted to have a meeting on it. If you guys are not alarmed at this or not paying attention to how important this is, then I'll keep going.
Anthropic teams up with its rivals to keep it from hacking everything. Project Glasswing was created. We're announcing Project Glasswing, a new initiative that brings together Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JP Morgan Chase, blah blah blah. It's all in an effort to secure the world's most critical software. For people that doubt how strong the model is, you don't do these types of things. They showed it to the government. The Remember the government's at war with Anthropic. And Bassant still calls a meeting. AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities. I'm not going to give you the details on what it was able to do, but let's just say it was able to break into things that were thought safe for 27 years since the time they went up.
OpenAI is working on a cybersecurity model to rival Anthropic's Mythos. Google interview with Sundai Sundar Pichai on Cheeky Pint. I recommend this one. I'll just read this one. These models are definitely really going to break pretty much all software out there. You guys can go listen to the rest of it. This all this led to software stock sell-off. Palantir, Microsoft drop on Anthropic's Mythos. I've been pitching Palantir relative to Microsoft. It was up 25% over a month during the IGV unwind. It gave it all back this week on the fears that even Palantir would be disrupted. I think when we get the earnings reports, uh, they will still have massive growth on the enterprise adoption side, so I don't think this is going to hold, but I am not one to sit there and bet that software isn't going to continue to be under pressure because this is about model compression. And the one thing Palantir has anyway is high multiples. And when Nvidia is sitting at the lowest multiple of the last 10 years and they're in the hardware side, there's just better places to put your money right now than hope that Palantir is going to bounce back. Uh, relative to Microsoft is still the way I'd play it, but it's giving back everything and here it is right there.
Uh, Meta finally coming with a new AI model. Again, this is all for the deflationary component of moving the prices lower, which is screwing the margin side. It makes the hyperscalers in a fight to get prices out there and deliver this stuff. They've got compute shortages. They're in competitive thing. Google released their Bloomberg competitive version. Uh, Perplexity has Perplexity Finance. Now you have Google Finance. This is just a snapshot of the page. It's good.
Uh, and again, as a reminder, this is private equity and IGV versus uh, SPX. We had more news in private credit this week. Wow, Carlyle, the new one with requests totaling 15.7%. They will limit shareholder redemptions. Uh, this won't stop and remember, these are not a minor deal, guys. This is a big deal. Uh, They have to Next time they come up, I mean, they have to come up with more money. So, the problem is fundraising for private credit is done and fundraising for private equity falls to the slowest pace in a decade. Remember, if you've got redemptions leaving, think about this like a Ponzi scheme. Private credit's not a Ponzi scheme, but they have a certain amount of money. They leverage it by a certain amount to get the returns that are up near equity levels. Let's assume they leverage it by about 1 and 1/2 to 1 and 3/4, which seems about industry standard. And they have about 25 to 30% of the money that they have, not the leveraged money, of the money they have from investors in software. Then you multiply that number times 1.7 and you end up with about 50% of the cash is in something related to software for a lot of these funds. And if money keeps leaving without money coming in, they have to sell those bonds if they can. They have to sell the loans. This is where the issue starts to become a bigger issue. So far, they seem to have been trying to plug the hole. Now we're going to start getting into this every quarter. This doesn't go away. I will keep saying it. Pro- credit situations and credit cycles do not stop on their own. This isn't a private credit issue that will continue. It's starting to come into part of the muni market. Software pain is about to get a lot worse as the public software companies go down and the private equity markdowns happen, obviously the debt has to get hit, too. And who wants to buy anything related to software right now? So, that's your issue.
Uh, I mentioned this before. I will keep showing it until it's no longer true. When financials are below the 200-day moving average and they are the worst-performing sector, which is where we are right now, bad things happen. Even with the bounce we've seen, financials are still there. And as I showed you, this happened this week. So, software went down, but we didn't get a bounce in private equity. What we did get a bounce in is Bitcoin.
So, for everyone there who loves Bitcoin and watches me and reads my stuff on Bitcoin, uh, this will be your favorite five or six slides. For the institutional side, you may not like this. Uh, by the end of this year, guys, you're going to be focused on it. Um, it was a big week for Bitcoin relative to software. It was a big week for the miners as well. It was a big week for Ethereum. We are starting to get to a point where we're getting a break in correlation. So, last Friday was the break in correlation between oil and stocks. We have a break in correlation right now between Bitcoin and software. It is incredibly noticeable. Straight line down, a move higher. This is two years of overlay between software and Bitcoin. Here's the relative trade. Almost back up to the all-time highs. I showed this chart many, many times and said this thing had done well even when it was down here relative to software. Once we get to the point Bitcoin has no disruption to AI. Every bank has a disruption as we're seeing to Mythos. Get ready for hackings. Your money inside the bank, as much [snorts] as you think it is safe, is now at risk of hack. Cryptography is far safer. Quantum is down the road. So, for all of the ironic nature of this, one hacking in a bank, I think a lot of money's going to move into Bitcoin. Get involved with the theory of getting things on chain and getting them safer. This is Bitcoin relative to software. Here is the weekly move. Whoop. Sorry, guys. Here's the weekly move. Largest in the last two years by far, 22% outperformance this week. Here's Bitcoin with MACD crossover, guys. The weekly MACD crossover. I would draw a trend line right down here. We are right on that right now. Uh, I think once we get above this high right there with the MACD's now crossed over in there and a break in correlation and the fact that I think we're about to get some really good news, which I'm going to show you now, I'd be ready.
So, this is Fed funds versus CPI. Uh, right now we've got the Fed funds contract here. CPI actually bumped up today. It's up to 3.3. We're about to cross over. That's See this period here? This is when we had CPI above there. CPI above Fed funds rate doesn't happen too often. Here's three-month bills, real three-month bills. So, this is using year-over-year CPI and three-month bills, which are currently around 364 and you've got CPI at 3.2. So, that's where you get this point four number. Uh, the the red line here is zero. The reason this matters, here are the returns for Bitcoin using the matrix of are real rates positive and are the Fed funds rates rising. So, think of this as positive real yields. So, we've got the Fed funds rate above uh, there and we have Fed funds rate rising. You don't want to be in there when the Fed funds rate is rising. You don't want to be in there when you've got positive real yields. Now, if you've got negative real yields and the Fed fund is on hold or easing, here's your return. All of the return is basically in that period. We are about to enter this quadrant very, very soon. The next CPI print will put us there. That's the time you want to own Bitcoin. Uh, just showed you that one. No need to show you that.
All right. So, I finally got my thematic list out, 100 names. This is five themes. For those of you who are new subscribers, this is going up on the website. You also have all the thematic list that make up the 100 names. That is a bunch of them, which I'll go in and show you. Here is the uh, Bloomberg chart of that basket. So, the 100 names uh, made new all-time highs. You can see how it's gone straight up. It's up big for the year while the market is down to unchanged. Uh, there is a technical scoring thing thanks to John Roehke who did all 100 names as of uh, Thursday. So, you can go through, pick your names. This is just a small snapshot. These are all Japanese names, but there's a lot I mean, they're mainly US names. But you can go through a lot of these are chemical names, by the way, and and uh, memory names.
Here are the individual themes. You have the rack. You have chemicals. You have optical fibers. You have advanced packaging and you have power. New highs for four of them. The only one that didn't make a new high, ironically enough, is the energy power one. But I would expect that one, too. So, all of these are different themes and this is what I want to show you with the AI trade. What I'm trying to find you guys are themes that are durable, that are all going to benefit from AI agents. Every one of these was put together from November on. The first one being the optical fiber one when I pitched Corning and included all of the other names that are in there. All of these lists, which total about 95 of the themes, uh, are there. And then there's five names including Lilly, including EWZ, including Palantir, including uh, silver and then one other one that I included in there, which I can't think of right now. But you guys can see all 100 names. This is the way I think you guys should be invested at this point in terms of an AI trade.
And what I want to show you is what's happened since right around the time of Opus 4.5. This was the turning point in everything. So, the green line here is the S&P software index. So, again, Opus 4.5, the rumor start coming out, it gets released couple weeks later, this breaks down. Um, this here is my thematic side. This is when the transfer goes into the hardware side. The PMI's break out in December. All of these trades, all of these themes are what I consider the AI theme. Now, you also have this uh, orange line here, which is the hyperscalers. The hyperscalers to me, even though they're making numbers, are going to continually be under multiple compression, which means best-case scenario, they kind of go sideways. They are great short relative to this basket. The white line here, not to pick on Dan Ives cuz I like Dan Ives and we got to know each other and I think he does good work. The problem is the where Dan and I disagree, which is the important thing for this, is in software. Dan has a high component of this related to software. So, his software names have dragged him down. We have some similarities in names up here. They're mainly on the semi side. This is the difference, guys. Dan is a technology person. He has viewed AI as a software trade. This is not Web 2.0. I repeat, this is not Web 2.0. AI is not a technology. It is electricity. It needs power. It needs copper. It needs optical fibers. Those are not his strength. This is what's going on and this is why if you have money in these types of things, you want to start figuring out a different way. If you go to my list, you'll have a better way to go through this. You can run it on your own.
EWZ's included in there. Made new highs. We are getting close to five-year highs on EWZ. EWZ Brazil benefits significantly from what's going on. I wrote a paper on them uh, a month and a half ago. Subscribers can go see it. Brazil in the green compute arbitrage of the AI decade. Highly recommend people understand why Brazil is such a key component of AI, but it's only gotten more important since I wrote that paper because the data centers that are being built in the Middle East are at risk. In the early days of this conflict in the Gulf, Iran made a strategic decision. It struck two AWS data centers in the UAE. Middle East security and data centers, the Latin America factor. Everyone's starting to focus again on how safe Latin America is for data centers, but more importantly, how Brazil already has clean energy and is abundant in it. Mexico and Brazil lead data centers boom amid US policy shakeups. Brazil energy demands and data center growth.
Finally, silver. Wrote a piece on silver this week. My favorite commodity in this. You can go learn about why it is a critical component of the AI trade. That's it for me for May this week, guys. Uh, I appreciate all the shout-outs. Keep reaching out. Send this to your friends. Subscribe. Subscribe at the at the site. I'll see you guys next week from Brooklyn.