AI Investing Got Weird

So, right now, Open AI and Anthropic are adding revenue faster than I think any company in the history of American business. But something is changing because over the past 3 weeks, I've sat through more than two dozen earnings calls. And every executive is starting to sound exactly the same. And here's the crazy thing. A year ago, all these executives were trying to get closer to OpenAI and Anthropic. they were trying to namerop them in a conference call or they were trying to get out a press release with the company's names in them that they're working together. But today, the biggest companies in the world are quietly and kind of quickly trying to build as many ways around these AI frontier labs. The reversal is really only taken about eight weeks. And look, by the end of today's video, you're going to know exactly how to spot what I believe is going to be the next winners in AI. And I'm calling this the AI layer trade. So, here's what I'm going to walk you through on today's video. Number one, I've pulled up several key moments from conference calls that I've listened to over the past few weeks that is going to tie this AI layer framework that I have together. Number two, there's a big giant paradox that makes some of this that I'm going to talk about honestly kind of confusing because like I said at the top, OpenAI and Anthropic are growing faster than ever before while their own biggest customers are literally trying to actively replace them. I'll show you why both of those things can actually be true at the exact same time. Number three, I have a six layer framework. This is really important that I'm using to decide which AI companies that I am going to buy and I'm going to recommend to subscribers. This framework I think is it's not only going to help you pick stocks to invest in. It also helps you hold on and have conviction around stocks because as we've seen over the past few weeks, some of these AI stocks can go down a lot very quickly. And so if you have high conviction, you can buy more and you can continue to hold for the next leg higher. So, one quick note though before I kind of get into these things because I I think this is going to help you understand my perspective here. So, I've spent the past 25 years building stuff on the web. This is front end. This is backend. These are websites, database applications, even a handful of websites that I scaled up and I sold them on Flippa when that was like kind of an easy thing to do. Now, it's probably too easy and there's not enough money there. So, I long story short, I'm not just an like a user of this technology. I'm not just an investor. I I literally rely on it every single day. It's paid my family's bills for a long time. And I think when you actually build with this stuff, when you build with AI, I I think you have a pretty good read on which technology is worth paying for and which technology you can eventually stop using. So, let's jump into the earnings calls because I I think this is what's really important. So when you sit through like over 20 of them like I have over the past two and a half, three weeks, you start to notice patterns showing up company after company, different industries, different businesses, they're all essentially starting to say that the same thing. It's it's crazy. When Microsoft reported their earnings a few weeks ago, they told investors that this major shift was happening with clients that they were working with. In fact, I think the example they gave was like Levi Strauss or something. So instead of Microsoft Copilot relying on a single model like it did when it launched, customers started to use multiple AI models depending on the task and how much that model was costing them. Here's the CEO of Microsoft and this is what he said on the conference call. Since the start of the year, we have seen 5x increase in the number of customers building with models from multiple providers. >> So I think as investors and maybe even consumers in technology, we're somewhat used to one dominant operating system like Microsoft or one search engine like Google. But Microsoft's customers have stopped committing to a single AI company and they're actually starting to use several at the same time. This is what they want. And Microsoft actually took it a step further than just making this swap possible. On the same earnings call, Microsoft announced that it built more than a dozen additional AI models, ones for image generation, voice transcriptions, coding, cyber security, all these different types of things. So Microsoft is no longer selling its customers just Open AI's technology. Microsoft is selling essentially an AI store that carries a bunch of different AI models. So, and Microsoft wasn't the only one. This version of companies wanting to use multiple AI models, it literally played out conference call after conference call. The company that probably said it the most bluntly was Palunteer. Palunteer builds, as probably you know, some of the most widely used AI software for corporations, government agencies. I would consider Palunteer one of the largest builders of real production AI custom systems. And here's what Palunteer's chief technology officer told investors on its conference call. >> The assumption that the Frontier is actually the best performing is just not borne out in practice. Within 24 hours of bringing Neotron Ultra into our stacks, we found five production tasks where a standard Neotron Ultra model without post-training beat Frontier models. So, there's a lot he said there and let me kind of break it down for you. Neotron Ultra is an AI model that Nvidia built and gave away for free. We'll put free in air quotes because anyone can download it and run it on their own expensive Nvidia hardware, but other than that, it doesn't actually cost anything. Palunteer took that free Nvidia model, didn't change anything about it, and it pointed it at stuff its customers were already paying the frontier labs like OpenAI and Anthropic to do. And inside of a single day, Palunteer found five real legitimate production jobs that the free model that was made by Nvidia actually produced better results than the models built by the most valuable AI companies in the world. Then Palunteer went on to explain why it would rather do business going forward in this way. >> We built a partnership with Nvidia. We're expanding our application layer. we are going to uh enter the market and already entering it in the classified space as Sham alluded to of of fine-tuning models. So the models actually fine-tuned by us in our enterprise on Nvidia stack outperform Frontier models and you own the weights, you own the alpha, you own everything. >> So owning the weights is essentially owning the AI model itself. It's kind of the same way we used to own music, like we used to own the DVDs and the CDs and those types of things. Rather than streaming them and renting them, Palunteer is telling its customers they can own the model outright. And that that honestly is great for governments and their types of clients. So instead of renting access to somebody's else's model that can be changed, repriced or even shut off without warning. And all three of those things, models changing, models changing price and them being cut off and shut off without warning. All three of those things have happened over the past couple of months. And Palunteer's customers can't deal with that. So, the largest technology companies in the world are adapting to a world where AI models can easily be replaced, which should mean these freaking AI labs like Anthropic and Open AI are in big deep trouble, right? Except that's not what's happening at all. Open AI and Anthropic are growing faster than ever right now. And it's really any point in their company's history. It's actually been pretty good and well documented at this point that Enthropic ended December of last year with about $9 billion in annualized revenue. So less than a billion dollar per month in revenue. Now outside researchers and anthropic probably leaking the information out into the public. They're at over70 billion dollar of annualized revenue in July. I I can't think of a company in the history of American business that has added that much revenue at that speed. I mean, it is remarkable. Openai is also seeing a reaceleration of its business, particularly on the consumer side. At the end of July, OpenAI said it passed 1 billion active users. So, the two companies whose products are supposedly becoming replaceable, and this is what the executives are saying, they're actually growing faster than ever before. And over those same weeks, their largest customers are actively trying to spend less with them. Even firms outside of big tech are kind of seeing this trend and trying to participate in it. Here's the chief executive officer of Uber describing this on the company's earnings call. >> A few years ago, many expected AI to converge around a single foundation model. Instead, multiple frontier models have emerged alongside a growing open source ecosystem. >> Now, look, I I I realize we live in a world where it's always good versus evil. It's Democrats versus Republicans. It's Lakers versus the Celtics. It's thumbs up or thumbs down. And by the way, if you're still watching this video, give me a thumbs up or a thumbs down on the video. Either one, however you feel. But here's the crazy thing about AI right now. The demand for artificial intelligence is real. I think we can all agree on that. And I think we can all say it's off the freaking charts right now. The companies that make the AI models are capturing an extraordinary amount of revenue and growth right now. and their own biggest customers at the exact same time are building ways around them. All three of those things are true all at the exact same time. It It's literally the most extraordinary thing I think I've seen in technology. And also I think this makes investing in artificial intelligence a very hard thing because the horse race of what model is ahead and what companies are using what and it's difficult to track all of this stuff in real time. So a while back I actually stopped trying to figure out what model was the best and which model people are using which one's in the lead. Instead, I'm starting to track where the money settles in real time. I'm doing this. So, the way I'm doing this is I'm starting to treat the whole industry as a stack. Every piece of it sits upon another piece. The way kind of like a house is built. So, you have the foundation below the building, then you have the floors, you have the walls, and eventually you've got the roof. the kind of the original version of this framework is is not mine. I'll admit that. It actually comes from Jensen Wong at Nvidia. He laid out what he called the five layers of this stack in an interview at Davos earlier this year. And I thought it was actually one of the cleanest explanation that really he or anybody has given about what's happening right now. He described AI as a five layer cake. Now, since then, I've actually added a sixth layer of my own. And this sixth layer is actually one of the biggest reasons why I'm making this video. And I think it's really important you understand this. So, let's go through the cake right now. At the bottom layer of the AI cake or the AI framework that I have is the power industry, the utilities, the nuclear operators, the solar companies, anybody building electrical or electrical equipment or anything that moves power into buildings. Companies that are popular here are like G Vernova and obviously the power providers. These are pretty popular stocks and I don't think you need me to tell you which ones to invest in. But I will say over the past year I've actually recommended Eaton and Emerson Electric to our subscribers and both of these stocks have comfortably outperformed the S&P 500 this year. So above the power layer is the semiconductor chips. This is obviously Nvidia, Broadcom, AMD. You also have the companies that manufacture the equipment. This is Taiwan Semiconductor. This is ASML Applied Materials. You even have the memory makers. Now, this is Micron and this is Sandex. This layer is extremely difficult to replicate. And the competition in this layer doesn't just come out of nowhere. These firms have very very durable moes around their business. Now, above the semiconductor layer is the data centers. So this is Amazon Web Services, Microsoft Azure, Google Cloud, Oracle. There's also the smaller NeoCloud operators like Coreweave and Nebulus. So above the data center layer is the layer that I've added myself. I call it the infrastructure layer. It's easily the best performing stocks in what the market often refers to as quote software stocks. I'll come back to this in a moment because these aren't software stocks. This is way better than software. Now above the infrastructure layer are the AI models. Open AI, Anthropic, Google's Gemini, Meta is doing some stuff. Elon Musk is doing something with Grock. You also have all the open models coming out of China, Nvidia, other companies out there. And finally above that is what I will call the application software layer. These are the applications that you click on that you visit and that you download. This is Salesforce. This is Service Now. This is Adobe among many others. Now, here's the key thing to this entire cake or this framework, however you want to call it. It's the test that I run on every layer before I decide if the company is worth investing on. And you can run this test yourself in like 10 seconds if you understand how all this works. If the company disappeared tomorrow, this is how you're going to run this test. If the company disappears tomorrow morning, what happens to artificial intelligence? Let's start with ASML. This is the one company that builds the world's lithography machines that print the most advanced chips. If ASML vanished tomorrow, the entire leading edge semiconductor industry and the AI industry essentially would stop in its tracks. Let's take Nvidia. every frontier model on Earth, including the open- source stuff being built in China, it's largely trained and built on top of Nvidia hardware and software. Honestly, remove Nvidia and you set the whole AI industry back years probably. Now, let's take Enthropic. Enthropic could cease to exist tomorrow morning. In fact, some of its models have actually had that happen to them. And the AI industry didn't even blink. The work would route to Google. it would go over to OpenAI or they would download a free model probably by the end of the week, maybe even by the end of the day, remove Adobe and people would just find new ways to edit videos and you know edit photos. It wouldn't be that hard. That is the whole framework in one single test. And it honestly if you do that framework across all of technology, it produces a pattern. the money, all the profits, all the profits, all the cash flow is accumulating at the bottom of the stack and it starts to thin out the higher you start to go. The reason is because of substitutes. There is one company on earth that makes the machines that makes all the chips. There's only three or four companies in the world that can actually design and manufacture leading edge chips for AI. there are like four or five companies that can actually afford to build data centers at scale. And so by the time you reach anthropic and open AAI at the model layer, there's dozens of options and several of them now are free open-source. And so the very top is the application software layer. There are literally thousands of companies at that layer and you can now vibe code and build replacements for them internally and that's what companies are doing. The higher you climb in this AI stack, the easier you become to replace. Easy to replace means you don't get to set your own prices. It means another company can come out of the woodwork and start competing with you. And I honestly don't think it's where you want the bulk of your money as a tech investor. Now, this now brings me to the layer that I've added personally. Between the data center and the AI model layer is a layer that almost nobody pays attention to or they do the mistake and they put it at the top layer. They put it up with software. It stores and organizes your company's data. It watches the entire AI system. It watches and makes sure AI agents are behaving and have permissions and doing all those types of things. It secures all of this as well. And it connects everything together in a way so all the pieces can talk to each other. This is databases. This is data warehousing. It's monitoring. It's cyber security. It's integration software. This is Snowflake. It's Palunteer. It's MongoDB. It's data bricks. It's Data Dogs. It's Crowd Strikes. It's PaloAlto Networks. On the large cap side, you have Microsoft, Google, and Amazon. They all have these integrated tools inside of their software and their data centers as well. This layer sits a little higher on the stack. And by that logic, you'd probably think h maybe it's not worth that much, but it is worth a lot. And here's why. Everything above this layer is generic. and in some cases open source or easily swappable. Everything inside of the infrastructure layer is yours and hard to switch away from it. Now, I've been putting money behind this framework for a while now. And and I'll tell you how it went, including the the part that was somewhat painful for a couple of months. So, I recommended Snowflake to Equity Empire subscribers last year, and it was entirely based on this framework logic that I've presented for you here. Not because it was an AI company or it was software. It was because it sat in the layer that I just described that I knew after 25 years of building web applications, it was very hard to replace a company like that once you started working with them. And look, I was early. the stock went down and it when I say snowflake went down, it went down a lot and it stayed down and subscribers started to question the call and started questioning if I was still committed to it. And I said publicly on this channel that I completely overestimated Wall Street's ability to understand what this company and any company in the infrastructure layer actually does. It's not software. And so I averaged down the entire way. Fast forward to today and Snowflake is up I think it's over 80% in six months. It's I think literally the best performing quote software stock in really almost the entire stock market. Now I'm not telling you this to brag or take a victory lap. I'm telling you because the framework is what helped me make the decision to invest in this company. Here's the most important part. The framework allowed me to hold and buy more when it got really ugly. The framework gave me the conviction to buy, hold, and keep buying until investors saw the critical piece of infrastructure software that Snowflake and many others are. So that's the system. It's got six layers to it. It's got a single test. You can run it on any stock that you're looking for. Now, what I do for Equity Empire subscribers, if you're curious, is that's the part that comes kind of after this framework. Every company I work with is kind of mapped, at least in my mind, to what layer it sits in. I also give recommendations on what prices to buy, a buy range, give you the conviction when it starts going down like snowflake to like, no, we've got to keep adding here. The price also, I would sell this. There's risk management, there's stop losses. We don't just hold and hope. But in the case of Snowflake, we actually set the trade and the investment up for some volatility. It was actually pretty well executed other than being a little early. Now also what I do when earnings come out or new technology emerges. I go through the conference call. I make videos. I record videos. We look at the technical charts. We look at everything and I let the subscribers know where I think the company is. Have they moved up? Have they moved down? Has new competition caught up? All those types of things. The framework though that I gave you on today's video is free. It's yours to keep. You can do whatever you want with it. The implementation on what to buy and the deeper analysis, the videos, the specific earnings videos, that's what I provide to subscribers. There's always a link to that in the description below. Whether you click on that or not, it's completely up to you. But here's what I want you to take away from this video. Take any AI stock that you own. put it in one of the layers. Figure out where it goes. Again, it's a little bit easier if you've spent the last 25 years of your life in these layers and digging in them and figuring out where everything goes. But I think most investors like yourself can do that. Then ask yourself, what happens to the AI industry if this company disappears tomorrow morning? If the answer is everything stops in its tracks, then it's probably a great place to put your money. If the answer is somebody else would do the job by Friday, then you're near the top of these layers. And I I wouldn't honestly allocate as much as my investment into companies at the top. Most people are going to spend though the next two years doing just that. They're going to buy the top of the stack because that's where all the headlines are. Those are the companies making the new software, the new thing that you click on, the new thing that consumers are using. That's where all the headlines are going to be. But all the money is going to trickle down to the layers I described at the beginning. So, I hope you have a better understanding of this. It's a crazy time. AI is on fire and at the same time, companies are trying to figure out how to not use the AI labs. It's unbelievable. Now, we've just gotten past the major earning season. I've recorded a ton of videos for paying subscribers. That means it frees up a lot more of my time. I'll be back later here on the channel to help us walk through it and try to make sense of it the best we can. Thanks for tuning in today's video. If you like my content, please subscribe. Please like the video. Please tell a friend. I appreciate it. I'll see you guys again soon. Good luck with your investments.