AI Stock Warning: Buy these AI Stocks Instead.

Just about everyone is wrong about AI. Not in the sense that we have a bubble. I don't think we have a bubble, but a lot of people are dead wrong about AI right now and what stocks are going to ultimately benefit from AI from here. So, there's a couple of things we're going to talk about in today's video. Number one, the spenders and the receivers. What you're seeing right now should concern you for the receivers, the companies receiving the spending, the semis, the memories, things like that. Number two, we will talk about this compute air pocket that is coming and what mass enterprise adoption of AI is actually going to look like and how long that's going to take. It's not something that happens overnight. And number three, I will share with you guys where I actually think the value is within AI stocks right now. I think this is the weirdest market that you're probably ever going to experience because the clear and obvious AI winners over the next two to three years are trading at massive discounts. And I will share some of those stocks with you in this video. And I do think this is a very important video as we even head into this upcoming week. You're going to have ASML earnings, TSMC earnings, and this is going to be a big deal. And we know that earnings expectations for this quarter are the highest they've been at since Q3 of 2021. Wall Street's expecting 22% year-over-year EPS growth for the markets this quarter. And while I think we're going to exceed that, it's it's going to be harder to impress. Okay. So, first and foremost, I want to talk about the relationship, the actual stock relationship between the spenders and the receivers. The spenders are Mag 7 and the hyperscalers. The receivers are the semis, the memory, the GE Veronovas of the world, right? Some even your industrials are in there, your your Caterpillars. What you'll notice recently is pretty uniformally speaking the spending has stopped being rewarded. So Microsoft down over 30% from highs, but even some of your others like Amazon down, you know, I I messed up my finger really bad last night, but down 12%. At one point it was down about 19% recently, but you've seen them start to come back and we'll talk about that in just a moment because that's very important to point out. Google, you know, look at Google, right? Down what, like 14% from highs, something like that. Let me get the actual number. Down 13% from highs. Oracle, that's just smoke. That's down like 50%. But even Meta, you know, in the last couple of days, they've kind of started to try to compete with OpenAI and Enthropic for some of the coding agents. Um, but that stock's still down 16%. And the hard truth is if we over the next couple of weeks hear from these companies that they're going to be spending more money on AI, their capex budgets go up, their stocks are going to come down. That's the attitude. That's the relationship in this market. If Amazon comes out two weeks from now and says, "Yep, we're going to spend $30 billion more than we thought this year on AI capex, nothing else is going to matter. Amazon's going to come down." Well, this is very important to understand. If we look at the dot bubble, so a lot of the So, during the com bubble, there were a couple of companies, telecoms and a, you know, handful of companies that were buying Cisco's products. Well, like WorldCom for an example peaked on June 21st, 1999 at around $64.50. It was already in a decline well before the broader market top. Global Crossing GX peaked around May 14th, 1999. Significant decline started in mid to late 1999. So investors started to sell the companies and the stocks that were actively spending the money on Cisco's products. So Cisco actually peaked a year later after some of these companies, but the the signs that it was really about to slow down were not in the Cisco, they were in the spenders. The long story short of this is if Wall Street continues to punish hyperscalers for spending more, eventually they will be pressured to cut spending, which will ultimately hurt the areas that everyone loves right now, the semiconductors, the memory, so on and so forth. And that's why I do think over the next couple of weeks, as we get the hyperscaler earnings, it's going to be very important to see does Microsoft raise capex again. their stock's down 30 plus percent. If they do and the stock comes down or you know Amazon, Google or Meta for that matter, that's not going to be a good sign for the AI trade. So that's number one. You are seeing some parallels right now between the spenders and the receivers in the same kind of way that happened in 1999 before the market peaked. Now, I don't think we're going to have some kind of drastic reduction of capex, but the problem is people are expecting capex to go from 700 billion this year to 1.3 trillion next year. Really, if you want upside in semiconductors, memory, AI stocks, the receivers of this money, you need spending to come in at like 1.6 1.7 trillion. And if Wall Street's going to punish the Mag Seven for spending that much, you know, potentially doubling CAPEX next year, that's not going to allow them to spend that much. They will ultimately cut back and AI stocks, semiconductors, memory, they're going to be really bad investments. Now again, I don't know if that's going to happen, but the process of, you know, beating up the spenders has already been underway, which is kind of the first sign of this. When everyone's money drunk and, you know, hyped up about something, companies, yeah, if if you're going to be rewarded for spending more, which these Mag Sevens were for a long time, and that flip-flops, you're ultimately going to have to slow down. Point number two that I want to make here is enterprise adoption of AI. So if we think bigger picture here, all of these hyperscalers, they're building out compute. They're building data centers. If you build it, they will come, right? You've heard that statement before. And that is largely true. But if we think about where compute demand is going to come from, where is compute demand going to come from? Why are hyperscalers building out all of these data centers? enterprise AI. That's where the compute demand is. Yeah, sure. You typing in or me typing in queries to chat GPT and Gemini and you know Grock and all these models claude that's some compute demand but it doesn't justify trillions of dollars of spending. You need massive enterprise adoption of AI. That's number one, which will happen. But number two, robotics, realworld AI applications. That's going to take a lot of compute and a lot of, you know, memory and and and all of it. That's ultimately why I don't think this is like a 2000 style bubble. But I do think almost everyone is wrong in their expectation of of compute demand. People on Wall Street right now, let me just put this into a little bit of a like a visualization for you. Okay, people on Wall Street right now, they're seeing compute demand that literally just went vertical. So, they're assuming it's going to continue to go just vertical. That's not the case. Okay, I believe we are about to see an air pocket of compute demand. I've been talking about this on X a lot. I've mentioned it on the channel before, but if you look at companies today when Claude, you know, Enthropic kind of changed the game. Everyone went out and started experimenting. What can AI do for my business? How can we make things better, have a better ROI, improve margins? that yeah, you've seen an exponential rise in compute demand, but there's a big difference between renting the car for a week and buying the car. Right now, you're in the renting the car phase, or you have been for the last couple of months. Everyone went out to rent the car. Eventually, they're going to buy the car, but there's a lot of things that need to happen before we go from renting the car to buying the car or before we go from experimenting with AI to massive realworld deployments of AI that are actually taking a lot of compute demand. So, I've been talking about this on X. I think personally by late 2027, by late next year, you're going to see start to see the signs of massive enterprise AI adoption. By 2028 through 2030 is when the actual massive enterprise adoption of AI actually starts. But since you know my opinion is my opinion, I asked Google, when will enterprise AI AI adoption hit mass penetration? It says enterprise AI adoption is projected to hit true mass penetration defined as more than 50% of mainstream enterprises capturing measurable bottomline value in production between 2027 and 2031. So in this range it says while a staggering 78 to 80% of global 2000 companies have at least one AI workload or agent sitting in production the ecosystem is currently stuck in an adoption gap. Companies are spending heavily, but only 29 to 32% report sustained, measurable P&L impacts. The market has passed the peak of ungrounded hype and is actively working through foundational operational hurdles. So it says here 2026 through 2027 embedded UI and agents. 80% of tools add AI text basic task automation 2027 through 2028 is the slope of enlightenment. high ROI workflows go mainstream, data gets fixed. So what what Google says here right for next year is we're going to have massive data restructuring. It says here the reality mass penetration cannot occur with fractured corporate data. True business penetration will scale here as companies achieve AI ready data infrastructure driving crossindustry production success rates past the 55% mark. What this means here if you look at companies 90% of their data is unsiloed or unstructured in data silos. You can't do anything with that. It's a lot of random data here and there. This is why companies like Snowflake and Data Dog and Rubric's Anna Pererna offering they just started offering are going to be so successful in the next year or two is because these companies say a Walmart they're taking 90% of their unstructured raw data and they're structuring that data categorizing that data to actually be able to use it with AI before you have agents out there just doing things for your business in massive numbers. You need the data. You need the agents to be able to basically understand the data. So there's there's a data hurdle here that is going to take some time to work through. Says here from 2029 through 2031, AI enters the plateau of productivity. The enterprise AI software market is projected to skyrocket to over 270 billion by 2031. The reality deep mass penetration at this stage complex multi- aent frameworks autonomously execute entire back office processes procurement supply chain routing endtoend financial auditing with zero human prompts required. So it says here data fragmentation AI models break because legacy enterprise database architectures cannot feed accurate real real world information safely. It says current business impact up to 60% of enterprise AI projects are stalled or abandoned due to poor data readiness. the trust and safety wall generative AI is inherently probabilistic the best like best guess right whereas enterprise finance and operations require deterministic 100% exact accuracy and this is important um when we talk about some of the stocks that I like like a UiPath right UiPath their RPA technology is deterministic AI agents are probabilistic that's a that's a big deal keep that in scale remains restricted to lower liability sandboxes like internal knowledge bases and copywriting tools. So you can't really use AI right now for anything important basically right because data is unsiloed or I mix the two words together data is unstructured sitting in silos and probabilistic models just mean even if you have a hundred agents doing one task a piece at least one of them is going to make a mistake you know even if they're 99% accurate and you have one mistake That's a problem. And that's where a company like Rubric is uh very valuable in actually being able to undo mistakes that agents make. And we'll talk about that in just a moment. But it says, "Vanishing ROI metrics. Teams are building impressive software pilots, but they struggle to track hard cost reductions or real topline revenue additions. 95% of early deployments report negligible or completely unmeasured bottomline P&L impacts. So what does this really mean at this point?" Well, in the past six months, we've had a lot of playing around with AI. We'll call it companies seeing what AI can do for their business. Over the next one to two years, it's really about restructuring the data. Companies realize AI does have a place in their business and will be a big impact. But for the next year or two, it's really about restructurizing, categorizing unstructured data. And that's step that's step one really to actually using AI in your company. And then step number three is obviously actually deploying AI in a real way in your company. That's when the compute demand is going to skyrocket. So, it's going to look like an S-curve. If you want to look at this as the actual Scurve, yeah, by 2030, it says 2.5 billion AI agents will be in use worldwide, spending over $106 billion worth of consumption cost per year. Today, we're at 2026, right? There's a little bit of this happening in in low consequence, you know, areas. 2027, it ramps a bit, but it's really 2028 through 2030 that you see this mass adoption curve that begins. And this is why I believe there could be a little bit of a air pocket for compute demand as companies go, okay, we figured out what AI can do. Now we have things we have to internally figure out, restructure our data, so on and so forth. And before they can take that next step towards mass production, this will create a compute air pocket from likely now until probably the end of 2027. in this stretch, I don't think AI stocks, semiconductors, you know, this this loved AI trade right now is going to be the best area to be invested in. You know, and I think you're already seeing the signs of this. I mean, if you want to look at Meta that's went up quite a bit in the last couple of days or some of the other hyperscalers that have come back a little bit, people are saying, "Wait, maybe things slow down a little bit." You know, that's why Google's up uh quite a bit from the lows as well. Google's up 6 and a half%. Amazon um you know Amazon's up quite a bit as well from the lows. Let me give you an actual percentage here. Amazon's up 9%. But it's not even about the hyperscalers as much. Look at the Russell 2000. So if you if you look at IWM, small caps have outperformed the markets dramatically. Why why are small caps outperforming so much? Well, because look, if we're going to go through this air pocket, Wall Street's going to reposition into other assets before you actually even hit the air pocket. You know, hence why in 2000, the WorldComs and the the companies that were actually the buyers of Cisco products, their stock sold off first, right? Wall Street, they just tend to know when the tide is turning. And again, this is why you have XLV, healthc care near all-time highs. You have regional banks near all-time highs. You have transports near all-time highs. That's why you've started to see this broadening in the last month or so. It's because Wall Street, they know this air pocket is coming. Now, I also posted this on X. If you guys are not following me over there, go ahead and uh follow the real TCI. But I shared this last night as well, or this morning at 5:00 in the morning. Uh 2026 the shift to tasks task agents companies are moving away from simple chat bots to autonomous digital agents. These tools handle distinct workflows like marketing campaigns or invoice data entry. 2027 is infrastructure stabilization. IT departments will spend this year standardizing data infrastructure and compliance frameworks. This removes current bottlenecks around security and siloed corporate data. 2028 is true true critical mass. Multi-agent AI networks will become standard across enterprise software suites. Autonomous systems will manage complex customer service operations and integrated business analytics at scale. Now, the companies that will actually benefit from this are the companies that have been beaten up the most in the last six months, right? Because again, you know, Wall Street, they're not very forwardlooking. Most like 90% of people are really not investors. They're traders, right? They're hopping on the market trends. I I like to to find and invest in those beaten down stocks, but there's a caveat to that, right? Like Whirlpool, for an example. Look at that stock. It's it's been dog It's not the whole category that's dog It's Whirlpool. So you want to be careful of those stocks that are going through company specific issues. Like you don't want to buy every stock that falls 80% in a year, right? You're going to probably lose a lot of money if you do that. But when you have whole sector groups being thrown to the, you know, curb on trash day, that's when you want to say, wait, is Wall Street really right about a whole sector of stocks? And that's where I think over the next you know couple of years here as we you know progress next year to data categorization and you know continued advancements with AI but ultimately work towards mass enterprise AI adoption from 2028 through 2030. There are a lot of companies that stand to benefit dramatically here. First and foremost Zeta Holdings. I've talked about this company a lot on this channel. They are kind of like a palenteer now. They're they're you know Palanteer is the internal data side. They use companies data to tell companies what they should do, right? Um Zeta is doing that for the consumer side. Zeta has a seven-year partnership with Palenteer. Stocks trading like 17 times forward earnings. I mean it it's it's it's a very impressive company. They have data profiles on 92% of the US adult population trading at a third of other software price to sales multiples. Like the stock could triple and be considered fair value. It's pretty impressive. So Zeta is number one. Rubric, this company is incredible. And I think tomorrow I actually want to make a video on Rubric specifically, but they really have three different vectors. So number one, they have cyber resiliency and backup. So Rubric can identify the exact moment a bad actor or bad files enter a company's database. Go in, remove those files, back the company up to when that breach happened. So companies are no longer held for ransom. This is missionritical in the day of AI. So that's number one. That's a that's mission critical. Number two is basically doing the same thing for AI agents. AI is always going to be probabilistic. It's never deterministic like RPA like a like a you know this happens do this or that happens do this. It's probabilistic. Rubric also makes sure that companies like their you know AI agents that are ultimately going to go into other companies you know databases are good. They're verified. They're not bad actors. Right? So again AI agents they're going to be probabistic. They're going to make mistakes. there's going to be bad AI agents. Rubric can reverse the mistakes that a that good agents make and keep bad agents out of companies operations. Again, that's mission critical. Also, they have Annaperna, which is categorizing data. So, like a snowflake and data dog, they're kind of like the data lakes, right? But Anaperna is in a a different vector where they're actually going to help categorize data before it goes to the data silos. So companies right now, if they want to recategorize their data, they have to move all of their data out of a silo, recategorize it, and then put the data in categories in a silo. Basically, I'm really simplifying this here, but a rubric is going to do that all within place, right? So, it saves companies a lot of cost. And this, while it's not missionritical, it's going to have a lot of demand over the next couple of years. Now, something like a snowflake, you know, data dog, these actual data stories, these data lakes, data analyzers, right? that's going to be just in massive demand for the next couple of years and probably for the next decade or so. You could say the same for something like a MongoDB as well. You know, Palanteer just like Zeta, they are on the other side of the coin. business intelligence is only going to continue to grow whether it's internal business intelligence like what Palanteer does or external business intelligence like what Zeta does predicting consumer trends you know finding the best area to build a a restaurant or build a new factory right um this is all part of Zeta's business intelligence becoming an AI infrastructure company because they have so much data so Palenteer that's an obvious winner you know um something like like a UiPath. Again, back to the picture that I shared to you guys. UPath, they're not the company that benefits from experimentation with AI. UiPath is a company that benefits from actual AI agents being adopted long-term from companies. So a UIP path, yeah, it's going to see more traction next year, but really for something like a UiPath, it's 2028 through 2030 where things are really going to get exciting. So these are some of my favorite companies and kind of a little bit of a crash course on why I like them right now. There's I'm sure there's many others like Back Blaze, that's another really good one that competes with Amazon's S3. They they have like uh 63% gross margins. They build their own racks. Like phenomenal company, billion dollar market cap. I think that's got a long future ahead of it as well. But um yeah, this is why everyone is wrong about AI. They're not wrong that AI is revolutionary, that it's going to change the world. Like that's correct. People are just wrong right now in who's actually going to be the biggest winners over the next 5 years or so. It's not the Nvidas at this point. Until true mass enterprise adoption happens and robotics happen, you kind of want to stay away from the the hardware side of the AI trade in my opinion. But then again, I think some of these software stocks, that's the vector to actual enterprise AI adoption. You know, companies are not going out to vibe code their own solutions. I've talked about this before on the channel, but Walmart, most analysts think they spend about $10 to $30 million per year on their Service Now subscription. It would cost Walmart $4 to $60 million to build a Service Now competitor, right, to to use in their company operations. But it would cost a hund00 million or more per year to fix bug bugs and pay for tokens. If Walmart spends 1020 million on Service Now per year, why would they go out and spend 40 to 60 million to make their own Service Now and then spend over a hundred million a year to keep it up and running? What? So it costs five times more to do what Service Now was doing? It makes no sense, right? And that's the case for really all all of these software companies. Um, you you gotta you got to look at software here. If you're someone that's trying to take um a little bit more risk, potentially risk is in the eye of the beholder. I I wouldn't say it's more risky to invest in software over hardware, but some people would say that. And uh you know, if if you're really looking for those 5, 10, 20x opportunities from here, I think they're going to be in software. And I think over the next, you know, 6, 12, 24 months, you know, software stocks are really going to shock everyone. So, let me know your thoughts on this down below in the comment section. Hit the like button as well as subscribe to the channel if you guys have not done so already. Have a fantastic rest of your day and I will see you in the next