On today's episode of the AI Investor Podcast, we're diving deep into a presentation I recently created for InvestiCon where we break down the biggest trends across not only 2026, but what we're looking forward to in 2027 as well. All that and more on today's episode. Hey everybody, just Eric this week, as we had promised, this week is going to be a little bit different. We are going to go through a presentation I had assembled for the Investicon uh conference I was at. A couple quick notes. Number one, I do think this episode would be better in video form. We're gonna put it on our YouTube. I think we're gonna try getting the video into Spotify as well because it supports videos. Unfortunately, I don't believe Apple Podcasts, which is kind of the third major platform uh the AI Investor Podcast goes out to, supports videos yet. They they also don't support comments. Apple kind of kind of letting podcasts languish a little bit in terms of their various offerings, but that's all right. So if if you're only able to listen to audio, I know a lot of people listen to this podcast while um, you know, going for a walk or commuting or, you know, however people enjoy podcasts, that's okay. I I tried making this so that the audio version is going to stand on its own as well. But I do think if you have a chance to watch the presentation, that it will be kind of the optimal way to watch this specific episode. Again, you can find that on our YouTube channel. We'll put that in the show notes if you're not currently subscribed. And it's really three parts across this presentation because I was talking to an audience that I know if you're a listener to the show, you're probably very well versed in the rate of change we've seen in AI this year. But for someone who is a generalist and hasn't been following something like uh a podcast specializing in AI, it's hard to grasp how much change has happened this year. When we go back a year ago, the expectations for spend on data centers next year were 700 billion. Now they're 1.4 trillion. The boom in AI, this growth has doubled versus expectations, which again, for a lot of people, AI has now been a massive consumer product since the end of 2022 with the release of ChatGPT. And it feels like this kind of uniform growth. But the reality is the amount of spend we'll see in 2026 is higher than we saw in 2022 through 2025 combined. And then we'll see another wave of massive growth in the year ahead. Um, that I once again I believe few people can grasp the ambitions of what these major companies are going for. So in the first part of this podcast, I really structured it, what's happening in 2026? What exactly is the agentic wave? What are some statistics about the growth of agentic AI and what does that mean? And then I followed through with some of the key trends. Why, why is agenc AI having such a massive impact on memory? Why is it having such a massive impact on CPUs? What is kind of behind the breakthrough in coding and how is that fostering a lot of these um agenc applications like Claude Code and uh um open AI's uh codecs? So we cover that. And then in a second section, we look at what the big trends for 2027 are. Now you've you've likely heard about these on the podcast before, but we go into more detail. What's next in AI networking? What are the growth rates for AI networking? What's next for behind the meter? Why are suddenly, why suddenly are 50% of projects expecting to go behind the meter when just a few months ago that was closer to 30%? What is the reason for that? And what kind of acceleration could we see? And what are some stocks that could benefit from this? We've got a lot of stocks across this episode, I should note, and and some deep dives into a couple names you you probably haven't heard about before. I haven't at least dived deep into before in the podcast, names like FTAI and and Vicor as I talk about some power delivery trends. And then we conclude this presentation with a third section, which is just looking ahead beyond 2027. I've talked about this idea that the reason AI stocks right now, they've they've essentially flatlined since the beginning of June. And the key reason behind that is just the fact the market is working through whether or not 2027 represents a peak for this build out, or whether or not there will be further growth ahead in the years that follow? You know, the market is sorting out. Was all this growth just condensed into a very narrow window and we're gonna see not as robust of growth rates ahead, or could there actually be significantly larger numbers in the markets pricing in right now? Um, could someone like NVIDIA CEO Jensen Wong be right that we could see three to four trillion dollars in spend by 2030, which would be significant growth over even the optimistic um projections for 2027 today? So I'll go over what that future looks like, what the big questions are, what the risks are, but also what the opportunities are. What could happen in AI, what developments are happening that could lead to a continuing surge in demand for um for agentic AI and kind of the next era of what the next wave of AI developments looks like. So again, this this presentation, it's about a hundred minutes long. So this is this is a little bit longer. Sometimes Austin and I will say, Hey, we actually made this a short episode, and we get a lot of comments saying, Yeah, I wish you guys would do a longer podcast. Well, if you're one of the people who's been commenting that this this episode is for you because it will probably be the longest episode we've ever released. And um, like I said, I hope people enjoy this. And and if this presentation isn't your cup of tea, we're gonna be back to more regularly scheduled programming next week. We've been working on getting an interview with someone who's uh one of the most well-known names and biggest experts um in the energy space, where we're gonna be able to dive into some things like uh the impact on natural gas from AI. So that should be that should be really exciting. And we'll also do some kind of broader uh commentary looking at what's going on in the market and AI stocks in general. So we'll we'll return to kind of a normal format next week. This is just a one-off, but I hope you really enjoyed this presentation. I I spent a lot of time putting it together and I think uh I think it's one of, if not the best, resources for someone who wants to look at a presentation and get caught up on what's happening in AI. Um I just don't think there's necessarily a better resource. So if you've got someone in your life too that you're you're wanting to show what's happening in AI and and it feels impossible to catch up on such a big space. Uh, you know, again, this isn't concise at 100 minutes, but for something as big as AI, this is, I believe, one of the best summaries of what's going on in the space. So with that, I'll kick it over to the presentation. Just how big could AI get in 2027? And what are the key trends in the space? We're going to break it all down today. I'm Eric Bleaker. I am the host of the AI Investor Podcast. And what you're watching today is a presentation I built for an investing conference that I believe is one of the best resources on where AI has been in 2026, the kind of rapid takeoff we're seeing, what we're looking for in 2027, and how trends can change. A lot of people have been chasing trends across 2026 that have seen phenomenal returns in areas like memory. But what is next? What does the next phase of the AI build out look like? And what are the stocks that are going to be writing that we'll talk about the day? We're going to talk about a lot of different stocks. I've got a focus on three stocks. I've got over 20 stocks in the AI space across this presentation. And we're going to end it looking at what are the key themes as well across the future in 2028 and beyond. I think that this is the best AI presentation for people of all levels. This is probably something like a graduate level look at AI. So whether or not you are currently invested in AI, you are considering investing, or you're just a broad investor who knows how important AI is and wants to get a hold of it. We're going to cover all the major topics across AI that you need to know. We're going to cover where it's going in the future, and we're going to cover that long-term perspective today. So I think that this is going to be a great look for investors. And as I said, you know, we also have plenty of stock picks here. And it's going to follow a presentation that I built for one of the world's best investing conferences, and we're putting online for the first time today. So let's let's kick this off. This presentation is the next year of AI. It's all about what's happened in 2026. And then we'll move on to the key trends in 2027, and then we'll finish it looking at what is expected in AI in 2028 and beyond. So here's a little look at some of the areas we'll be covering. I hope this isn't too small to read, but we're going to talk about some of the booming areas across 2026. We're going to talk about some of these stocks to watch in 2027 and also the key themes. And then we're going to end it looking at the big picture, whether or not AI is going to be able to continue growing off today's levels. So let's begin with the market recap. What happened in 2026? Because it's truly been a sea change. But around 2022 is when AI first had its massive consumer moment. And what we're seeing in 2026 is AI having its first enterprise moment, where it is broadly applied across businesses, across the world, thanks to the growth of agents. I'm going to have some statistics here that I think will blow your mind a little bit. Some of the growth that we've seen across this past year is something that you don't always see just looking at stock charts for individual companies. And once you put this all together, you're going to really understand why 2026 has been so transformational for the AI trade itself. So let's move on to the next slide. And this really is going to hit home. How much expectations for AI have shifted across the past year? If I was giving this presentation in late 2025, I would have told you that we are expecting 577 billion in spend in 2026, the year we're currently in. In 2027, that number would be rising to 687 billion. And as we got to 2028, we'd have about 763 billion in capital expenditures, building out these data centers that are fueling the AI revolution. Well, when we move forward to today, expectations have absolutely soared. In 2026, the expectation is now 845 billion in spend. That's 46% higher than what was expected a year ago. In 2027, the expectation is now nearly 1.4 trillion in spend. That is up almost a clean double, 99%. And 2028 is up 89% versus expectations a year ago. The bottom line here is there's been a transformational rise in expectations of what AI spend is going to look like in the near term. And this is a little bit of what we're going to unravel today. Can the spending continue into the future? Are we at a peak? Or um yeah, is is basically we have a lot of stocks priced as though where we are going to reach in 2027, 2028 is the absolute max. Um, and the question is, can can AI continue growing beyond that rate? So we will continue moving on the presentation. You can see the 2028 figure here now. And here's the companies that are driving this. You know, you see growth between 7x, this is between 2023 and 2027 for Amazon and how much they're spending billing data centers, all the way up to 30x for a company like SpaceX. And this group is collectively called the Hyperscalers. And they are what is leading to these massive projects. I have a picture here on screen of Stargate Abilene. That's a 1.2 gigawatt data center I visited. It's an absolutely stunning site. I think it took more than 30 minutes just to drive around the perimeter of it, covers more than a thousand acres. It's employing more than 10,000 workers. And pretty soon, when you look at that, that's a 1.2 gigawatt site. There's going to be dozens of these projects springing up across the world, largely driven by hyperscalers, which it shows you the ambition of this build out. I highlight on this slide right here, Alphabet, there's a lot of attention. Last quarter, they they reached their first ever quarter of negative free cash flow. And a lot of people are probably thinking with the headlines that were spun up about this, Alphabet must be at the peak of how much they're going to spend on the AI build-out. But that couldn't be further from the truth. Because as you see from the numbers here, which um are largely derived from research from uh JP Morgan, Alphabet, 205 billion in expected capex in 2026. Their internal projections are closer to 375 billion next year. So that's a significant step up. And we see this significant step up happening across all of the companies on this chart here. The rate limiting factor is not that these companies want to spend more, it's whether or not they will be able to spend, whether or not the supplies will exist in key areas like power and actually getting the chips inside the data center. So I think it's something, as I mentioned earlier, we've been sitting through this AI build-out, and it feels relatively uniform. You know, this began with ChatGPT first being released into the world in late 2022 and throughout 2023 and 2024, felt like AI started increasing in volume, and and it it feels like it's been this uniform rise, but the truth is actually opposite. It is in its most exponential phase right now, which I don't believe most people fully grasp. And you know, we have the numbers here to show what kind of an exponential rise we are headed into into 2027. And here's some charts that show what's happening in this takeoff. Uh because again, uh we had AI first in our new era with the release of Chat GPT. But the thing about Chat GPT, it was largely a consumer story. And it improved in the following years as we got more and more capable models. But the the future path of AI, if it's going to be as powerful as some predictions have forecast, isn't going to be that rate limited by college kids going in and typing in answers for their tests. The the idea behind what the final path for AI is, is that it's going to be embedded across the economy. You know, if you work at a company, AI would be embedded across your workflows. Um that is the final future in terms of AI actually boosting productivity across the economy and being kind of the technology that's that's been promised. And we finally saw the first signs of this thanks to agents in 2026. So if you look at this chart right here, this is agentic usage at OpenAI. And you can see here that when ChatGPT or GPT-5, I should say, was launched last August, uh, only about 1% of OpenAI's uh token usage was from agents. Across the past year, it's exploded to north of 64%. If we go to the next uh chart, what we'll see here is that token usage is absolutely exploding. Um tokens used by agents last year was essentially nothing. And now it's five times more usage from people, people going in and searching things. So what we have is uh agents, which again is uh being able to create autonomous usage for AI. It's it's it started with things like clawed code, um, and it's it's moved on to so many services. You have codecs now from open AI, and this is driving the usage of tokens. And tokens is again the output from these large language models, and it's going to be massive growth in tokens that's going to lead to a lot of the demand for AI across the future. I I put it in here, but it is a question of whether or not the growth of tokens themselves is the Moore's law for the current era we're in. You think about the past, Moore's law defined the past 50 years of technology. And it essentially said, and I'm gonna paraphrase here, but the the capability um of uh CPUs doubled approximately every 18 months. And it was this continuing growth of um the capability of processors and semiconductors, and that was enough to drive this era of technology across the next, or uh, I should say, the past 50 years. But when we look at what's happening right now, data from OpenRouter, which was purchased by Stripe and is a company that has a lot of access across um, you know, companies that are using a collection of different models, they're showing that token volume is doubling every 11 weeks. So if we continue to have this massive growth rate of consumption, well, that's that's where your supply-demand equation is. And this is why these companies, why Google is looking to go from 205 billion in Billing out data centers last year up to 375 billion in 2027 if they can get the supplies for it, if they can get the chips, if they can get the power. And what we see at the bottom chart here is this agenc usage is is actually spread across knowledge work in in a way that might surprise people. When you look at the growth of weekly codex users, which again is open AI's agentic product, legal usage has soared from February 1st to uh this this chart is as of the middle of August, I believe, 108-fold, sales 41-fold, people recruiting 41 fold, marketing 26-fold, healthcare 24 fold, finance 20-fold, project management 18fold, and engineering only fivefold, and maybe that's because you know they had a higher starting rate because it was the field that was using agentic AI earlier. But you know, the the bottom line here people will ask what's different if you're sitting at home and you are just plugging in queries into open AI, which is how most consumers experience this technology. It certainly doesn't feel different. But what this data that I'm presenting on these slides is showing is that there is a sea change that's happening, and it is due to agentic AI and building these autonomous workflows where you're able to give AI tasks and have it operate towards them. And that is leading to this massive token consumption. And this is where this supply demand imbalance has begun to grow, and this is where this insatiable demand from companies like Amazon, Google, and many of their customers that they're selling compute to is being created. And on this slide right here, I have the Economics of serving AI. And I've talked about this before if you're a listener of the AI Investor podcast, but as of right now, being a company with data centers is incredibly attractive. SpaceX recently rented out capacity at $31 billion per gigawatt to Anthropic. They rent out capacity to Google at $48 billion per gigawatt. The economics are such that building a gigawatt of data centers right now costs something around $50 billion. And that cost may be rising as some of the components like memory, which we'll look at later, are really increasing in price. But the point here is many of the current deals at these spot prices mean that if you can have compute capacity, you can get payoff to rent out in as little as one year to maybe two years, which until recently the expectation was more that would take closer to four to five years. So it's it's creating a bonanza. And the reason that this capacity is so lucrative is companies like Anthropic and OpenAI, who have these leading models, at the current rates for getting tokens from their APIs, they're able to earn something like $100 billion annually per gigawatt from the newest chips. So they're willing to pay premium prices because they're able to charge such premium prices to their own customers, right? Because companies that are beginning to build workflows through AI, it's increasingly spreading across many of these early adopters. So there's every incentive in the world for companies to continue building out compute as things stand right now. This is why SpaceX, I teased this earlier, but Elon Musk is looking to spend 300 to 500 billion in 2027 if he can. I don't know if he'll be able, I don't know if he'll be able to spend this much, but he is looking to spend 300 to 500 billion. I've got it down here. If you're looking at this slide right here, he he would like to build out to 10 gigawatts. He has 1.4 gigawatts available today. And he's racing to do it because he believes SpaceX can build capacity faster than its rivals so he can take advantage of these really lucrative economics. Google has seen their cloud business inflect from 32% growth in the middle of 2025 to 82% today. And this business is so lucrative they don't have capacity to really make their own models, which they have to train with their capacity competitive. So they're racing as much as they can to bring capacity online. Microsoft, they would like to build as many data centers as possible because they have a deal with OpenAI that allows them to serve OpenAI's models, allowing them to collect the same economics as OpenAI, a company that could very well IPO soon for multiple trillion dollars. And a company like Amazon, well, they're trying to protect their lead at the forefront of AI. They continue saying that they are getting outstanding ROI from AI. And their custom chip business, building chips for their servers, is now at an annualized run rate of more than $50 billion, meaning that Amazon is, it's it started as an e-commerce play. It tied on cloud computing. Um, it built an advertising business on top. It ties all together through Prime. And now this is potentially another company-changing business line in terms of building these custom chips to serve AI. And that's what their CEO, Andy Jassy, believes. And he doesn't want to fall behind. So let's look at some of the businesses and um kind of trends that have led to some of the biggest booms in stocks across 2026. And I would start with memory. I've got a quote here from one of the early investors, I believe in SpaceX. And he says, memory, not compute, is the rate limiter of the agentic era. And Alon Mus says, few realize this. Memory has been on absolute tear recently. In the AI investor portfolio, we recommended Micron at sub $100. We recommend SK Heinex. We recommend several other derivative plays. It's been outstanding. We're early to this, we've seen a thousand percent plus returns in a little more than a year and a half across these positions. And here's the numbers behind this. The memory market in 2025, as you can see on the chart below, was about 72.8 billion. In 2026, it's expected to hit 356 billion. In 2027, it's expected to hit 899 billion, which is a remarkable figure. And there are some signs it's it's going to remain more persistent than than people imagined headed into 2028 to 2030. Now, let's break down, because we've talked about memory a lot in the past. Why agentic AI requires so much memory? There's two primary factors. Number one, model weights. The weights of models must be stored in memories, and they're getting massive. Mythos from Anthropic is reported to have about 10 trillion parameters. So this would require a single NVIDIA system, about 52 GPUs just to hold those weights in its memory. So that's that's almost an entire system right there. The second part is KV cash. Basically, what happens is agents use agentic loops where they plan, call tools, read results, and re-reason. And with each loop, the model creates notes on what's read so far. This is KV cache. It's a little bit of a hack to the fact that right now models themselves don't really have memory of what you've done. They they create these hacks to understand prior context, and these hacks are incredibly intensive, specifically on memory. So this is why memory is so far the rate limiter on AI. When you produce every single token, every single word across an LLM spitting out its outputs, all of the weights for a model must be reread, and all of those notes must be stored. So that's again, it's not every time you ask a question, it's it's every word of generation. And if memory is the bottleneck to a system, it means your other expensive components like GPUs aren't working. So again, it's important to understand this. This is why memory is the rate limiter of AI in its current incarnation. Another big industry that's been booming alongside this birth of agentic AI in 2026 is CPUs. So what's going on is that models themselves, they just produce text. They're not actually running anything. So when you have agents, what you want is them to actually perform tasks. And when they're performing tasks, they're going to essentially be using a regular computer with CPU and RAM to do so. So some examples of these tasks I've got right here: running code, searching the web, querying a database, calling others' software, posting a message to Slack, clicking around websites. And, you know, at the end of the day, GPUs were the story of the beginning of AI. But CPUs now are why AI can actually do things in this agentic world. And that's leading to a higher amount of CPUs relative to GPUs. So we have this chart here. If we go back to 2020, NVIDIA's DGX A1000 system had eight GPUs per CPU. If we go 2024, their HGX H100 system had four R CPU or GPUs per CPU. Today, some of the systems like GB200 and VL72s are a two to one rate GPUs to CPU. And it's believed then this next generation, we will hit a one-to-one ratio. So why is Intel such a story stock of 2026? Why is AMD? Why is ARM holdings? I've got the returns here. Um, you know, this was as of late August, but Intel up 122%, AMD 104%, arm holdings up 108%. Well, it's because of understanding exactly this growth of agentic AI and what that means for both memory and also CPUs. Another thing we need to talk about when we look at industries that have been booming in 2026 is um coding. I I don't know if I've talked much on the AI Investor podcast about harnesses, but but it's really important to understand them. Um again, models are very good guessers of what words come next. The differentiator increasingly in this AI age isn't necessarily just the model themselves, it's what's wrapped around the models that allows it to get work done. So, you know, you look at a harness, a harness decides what a model sees, which files, which search results, what they're allowed to do, what tools can they use, what applications, what rules do they follow, the action. A model decides to search for X, the harness itself goes and searches, and the stopping point, and this is important, harnesses decide when a task is done. So when we look at what happened in this prior year, we had the release of Claude Code, which we talked so much about, and it was a very big deal. But Claude Code is a harness. And what OpenAI was, or sorry, I should say, what Anthropic did was they were very good at basically, they were very good at coding, which allowed them to build harnesses across many different areas. When we look at Claude Code, that's a harness. When we look at Claude Design to build out presentations like this one, that's a harness. Claude co-work. So because they had gotten so good at their models building coding, they were able to build out this infrastructure to make their models far more usable and also make their models autonomous. Again, the story of 2026 is AI that goes from being able to be a chat window that provides answers into something that can do things. And when AI can do things, it becomes remarkably more powerful. And the reason it can do things is because of these advancements in coding from prior model breakthroughs and then being able to build these vastly more usable harnesses that we've seen for things like Claude Code. Here's a chart right here, which shows Codecs. Codex went from 1 million users, and this is from OpenAI, in February, February 5th, up to 20 million users by late August. I think they're up to 25 million. They've had an update since then. It just shows the incredible growth and how much the AI market right now is being driven by these agentic tools. SpaceX paid $60 billion for Cursor, which is a company with proprietary coding data. And now they're starting to catch up with open AI anthropic. As I've said many times, if SpaceX didn't make that cursor acquisition, they're probably toast. But now they're once again competitive. And we can see where the acceleration begins as we go to this next chart. Looking at open AI, they went from 19 billion in revenue in November up to 33 billion in June, which is a good growth rate, but it's probably not going to get them the valuation they need. Well, as of late August, the run rate appears to have accelerated to 48 billion. So from 38 billion to 48 billion growth of, I think that's around 50%. And it's due to the rapid acceleration in codex in the chart that I showed on the prior slide. We look at anthropic Clawed Code hit mainstream adoption at the beginning of 2026. At the time their annualized revenue run rate was 9 billion. At the end of July, they're up to 65 billion. So again, having these harnesses, thanks to advanced coding capabilities, is what is allowing these companies like OpenAI and also Anthropic to see massive revenue acceleration. It's what allowing, it's what's allowing companies themselves to build AI throughout all of their systems right now. So this is at the base. Like I said, if you are someone who just goes to Chat GPT and asks questions, AI hasn't fundamentally changed. It's gotten smarter, hopefully, and more capable, but it hasn't seen the step change that's being reflected in this revenue growth rates. And to understand why all these companies are suddenly willing to spend so much and bet their entire future on AI, you have to understand this agency step change that has been the story of 2026. Okay. So I hope I've covered 2026 well. What we're going to look at now is what's going on into the future. What's going on in 2027 and beyond? And there's going to be some really key industries that I think are going to find the next phase of AI. And we've got a couple of them on the screen right now. So first off, everything I just talked about in 2026, what has been the you know, end outcome from that? Well, you see, Sandisk. Leading stock in the market. Moderna, you could say it's a little more AI adjacent, but it's definitely a company that uses personalized medicine thanks to AI, is the second best stock. When we look at the top 10 stocks in 2026 so far, nine out of ten are due to AI. This is this is how powerful AI has been this year. And a lot of it in that initial period in kind of March, the end of March through June, where we saw a lot of these gains. As again, the market was ingesting the step change that was happening thanks to agentic AI. So the question is what what step changes are still misunderstood by the market? What is still going to happen in 2027? I think a few areas we're going to cover today. Sovereign AI, basically, how enterprises are going to be able to create how they want to use all of these models and how they protect their most valuable data. AI networking. Memory's been the story. Memory and storage has been the story of 2026. Networking is increasingly going to be a way to reduce the memory bottleneck. And the growth rates of AI networking remain outstanding. So we're going to talk about some of the key themes across networking, and we're going to talk about a stock Semtech that remains one of my favorite ways to play this trend. We're going to talk about behind the meter power. We all know about the power bottleneck, that building out all these data centers is going to create enormous power demands. And if you're going through the grid, number one, grid power can't be built fast enough. And number two, um, even if everything was going through the grid, you're going to have these downstream impacts on consumer pricing that is going to be really untenable, especially as a you know midterm political year in the United States. The solution to be able to get data centers online is going to be able to generate power on-site without necessarily relying on the grid as a primary source. So we're going to talk about some companies that could benefit from this trend. We're also going to talk about power delivery bottlenecks inside data centers. Nvidia is making a massive transition into how power is delivered to each one of the servers inside data centers. We'll talk about a stock idea in Vicor that might be a huge winner here. And finally, we're going to talk about the companies making the machines to actually make the machines, the machines to make chips. It's called semiconductor equipment. And it's been one of the most lucrative spaces in investing across the past 15 years. But a lot of people avoid it because it is really complicated. We're going to break down some areas in semiconductor equipment that could lead to really phenomenal returns in 2027 as a lot more supply moves in to try and catch up with this demand for basically AI chips of all kinds. So let's get started with this. First off, I want to talk about sovereign AI. Um, what you can see here is while the race for AI is the smartest possible model, the usage of the smartest and most capable models isn't always high. We see Fable 5 right here on the bottom. Um it has very low usage. This is the most capable model at the current time for Manthropic, but it's its usage remains low because it's relatively expensive. And what's going on here is companies are going to have to create their own mix of they're increasingly using AI across all their workflows. Do they want to rely on a single company? Do they want to have open source? How do they want to handle their own data needs? And this is leading to a massive surge in open source, uh, open weight, I should say, models. Open weight models as of July 1st were just 27% market share of tokens. And this is data from a company named Vercel. And recently, in August 25th, that number hit 55%. So a double in usage. And this is a really important trend because the companies that have really advantaged positions in this space are really starting to see their share prices soar. So Snowflake is a recommendation that we've had in the AI investor podcast. Um, it it it they recently report earnings, I believe that they were up almost 25% the next day. And the reason is because they're going to be an analytics platform that companies that want to protect their own data and be able to mix in models across all these companies are going to look to. The best of breed companies, it's starting to dawn on the market that these companies aren't necessarily going to be disrupted, but they're actually going to see their opportunities uh really expand from the usage of AI. So again, the leading companies in this space have seen excellent returns recently. And this allows people to diversify from if you're looking at AI, just hardware companies, to actually software companies that are growing as well. Snowflake's a recommendation we've had that's done exceptional. Cloudflare is another company we recommend that continues to see basically an increasing future of how they can be a toll booth for the agentic era. Both those companies are pretty expensive. One recommendation we recently made is ServiceNow. And this is a company that sold off dramatically, I believe at their peak there, down more than 60%. But the company is going to handle the orchestration layer for this agencera. And their AI revenue today is less than 10% of company revenue. But they expect that to grow to more than 30% by the end of the decade. This is just a company that is extremely well positioned, is relatively cheap relative to some of the other plays in the space. And it shows you how in the past, you know, we look at the trends for AI in 2024, 2025. It was really all these names, NVIDIA, AMD, all the all these Micron, all these companies making the hardware. There is a burgeoning trade that's opening up amongst software names as well. And a lot of investors who saw these companies just get burned for years on end don't necessarily understand the opportunity that's opening up. And that's one that I believe is going to continue as companies increasingly put AI at the center of their operations. Um, they're going to need the software orchestration layer. And these are some names that you should definitely consider. Now, let's move across to AI networking. This is another key trend, I believe, for 2027. The growth rates here are astounding. I mean, we can look AI networking in 2023, which was the year that AI really kicked off, was about 10 a $10 billion market. That's expected to grow to $245 billion by the end of the decade. Between 2022, when again the Chat GPT year to 2020 to 2030, this is a market growing at a nearly 50% compounded annual growth rate. Um, and as I've mentioned earlier, it's really a place that allows um solving for this comp uh, I should say, this memory bottleneck in a way. So you you've got three things with AI networking. Number one, scale up, and this is how many GPUs can be put in a server rack to effectively act as a single domain, a single chip while sharing resources like memory. For NVIDIA, they're trying to push from 72 GPUs in a rack to 576. This this might not sound like it has massive implications, but it does because again, the more GPUs you can get to act effectively as a single chip, the more you can share memory. And what you're going to do is you're going to try spending less on memory and more on AI networking to get all these chips to act as one coherent unit. Then you're going to have scale out, and that's where you're going to have 100,000, even more chips that are connected within a single data center. And then you're going to have scale across, and that's where you're going to be actually linking data centers across a region to basically continue pushing this the limits of what you can do with chips where you'll you can maybe only have so much power as a single data site. But maybe if you've got 10 data centers within a 50-mile region, you can actually link them all together. And again, this all just continues putting more demands on AI networking. So the stock I want to feature for this is Semtech, which if anyone is a longtime listener of the AI Investor Podcast, they'll know Semtech because I've now recommended it twice in 2027. But but here's the key pitch here SemTech has a couple major trends. Number one, the complexity of AI networking is increasing substantially. This is really akin to what happened with the release of the iPhone itself and smartphones, the last dominant technology trend. What you had was you had the first iPhone come out in 2007 at the time. I don't even know if 3G was necessarily an established network. I don't even think the first iPhone had 3G, but what we did for successive generations was we kept pushing the networking technology for smartphones from 3G to 4G to 5G. And what that created was the most successful company in the smartphone era wasn't Apple, which might be shocking. Apple was the most dominant and they are now a $4 trillion company. But the number one stock you should have bought if you want to make the most money was a company named Avago. And what Avago does, did and still does today, they are now uh Broadcom actually, was they built the filters that as networking technology, 3G to 4G to 5G, kept getting more complex. It required more and more of this radio frequency chip content that Avago was building. Well, SemTech has a similar opportunity here because I'm going to put a chart on screen that you're going to see we're about to move through networking generations much faster than we have in the past. We're going to start seeing speeds of 800G rising in 2026 and 2027. But in 2028, that's going to give way for 1.6 terabit. So we're seeing the speeds for networking technology in the data center continuing to increase and continuing to shift at a much higher rate, to where by 2030, we're going to see an explosion in the growth rates for the next networking speed technology, which is 3.2 terabits. And why is this important? Well, Semtech believes that their opportunity per each of these optical transceivers, these networking connections, is going to grow from tenfold from where it is today up to when we get to this 3.2 terabit generation. Effectively, they serve as somewhat of an analog to what happened to Avago back in the iPhone generation with mobile networks for today's AI networking opportunity. So as we continue seeing more optics in the data center at higher speeds, it's going to require more revenue from Semtech itself. Second, they have a technology called Active Copper Cables. As I put here, this whole market's expected to grow at 98% compounded annual growth rate between 2026 and 2030, which is phenomenal. You almost never see a market growing at this growth rate. But they have a specifically attractive relationship with Google supplying these active copper cables. And this is going to really matter as Google scales up its TPU program, its custom chips, especially with a partner named MediaTech across the next year. So what you're going to see for SemTech is across 2026, the tail end of it, and 2027, massive growth rates from this active copper cable opportunity. And then after that, in 2028, 2029, 2030, you're going to see a massive boom in opportunities from these networking speeds increasing. So this is a company, most people don't know what Semtech is. This is one of the most boring backwater technologies that have existed. But you know what also used to be a boring backwater technology? Memory. You know what every single investor is chasing today? The returns from memory and the unique technology position it employed that allowed it to receive these returns. And that's why I think SemTech is positioned across in the next five to seven years. Even if we see growth rates for AI beginning to moderate, the growth rates for Semtech itself and the products it provides is going to be significantly above the broader AI market itself. So that's that's the first of three different stock recommendations we have here. Another trend, this is our third trend for 2027, is behind the meter. Here's the whopper right here. In Texas, they have 466.5 gigawatts of grid connection requests. 90% of those are for data science. They've approved 8.9 gigawatts, and um they've only uh have 3.9 operating today. Here's the bottom line, which shows just how crazy the drive is to be able to connect to the grid in Texas. The request to connect to ERCOT in Texas at 466.5 gigawatts. That is, I'm doing the numbers in my head here. It is 36 times more than the power that's consumed by the city of New York. It's eight times the Tokyo region. And here's what's amazing. You think about San Francisco, the city where all these AI companies are headquartered, the requests into Texas for new grid connections are about 500 times the annual power usage from San Francisco itself. I I've said in the past on this podcast, um, on the AI investor podcast, I should say, Eden, which is a global supplier of data center equipment, they've said that the US backlog's 307 gigawatts. At 2025 build rates, that's a backlog of 15 years. So here's the problem. We're gonna we're gonna put it, I guess this chart right here, where it's kind of broken up, but you see at the bottom of the screen, US power generation is roughly flat since the year 2000. And now it needs to grow significantly to be able to handle these these power needs. So what is the play here? What is the play? If if you are Elon Musk and you want to maybe build up to 10 gigawatts next year, and so far Texas has only connected 3.9 gigawatts to the grid from from data centers, how are you gonna do this? Well, Elon Musk has a plan, and his plan is to generate power capacity at the sites themselves using things like jet turbines and natural gas generation, natural gas engines. And and what's happened is as it becomes clear to everyone who is building all these data centers just how much demand there is, everyone's realizing at the same time they're going to have to shift from connecting to the grid to getting all the supplies possible to generating power on site. I've got the chart right here on screen, but six months ago, about 70% of projects were planning to connect to the grid and 30% behind the meter. Today, in a shift of six months, that splits closer to 50-50. And that data comes from people inside the supply chain itself. So, what we're going to see is a rapid shift towards companies looking at these behind-the-meter solutions. So, what is a company that could benefit from this? Well, FTAI Aviation is definitely one name I think investors need to consider because it's such a weird thing that if I told you how big AI was going to be back in 2023, you would have never guessed. You would have never guessed, oh, I need to go buy jet engines. But this is the way that these interconnected areas of the economy work, that the reality is often stranger than fiction when we play out these trends. And like I said earlier, one of the key ways to be able to generate power is essentially taking jet engines to generate power and using them to generate power for these large data centers. And this is what FTAI aviation is going to do. SpaceX, as I noted, they have 1.4 gigawatts of compute today. They're expected to go up to, they're trying to build up to 10 gigawatts in 2027. As again, you can see on the screen. How are they going to do this? Well, they're going to look for behind-the-meter solutions because what behind the meter does, it allows you to once again skip the grid. But number two, it allows you to skip a lot of other bottlenecks. If you've got these jet engines that are essentially generating power on site, what it allows you to often do is also skip a lot of the other backlog components like transformers of certain voltages, electrical substations, all these other things with years of backlogs, because you're not having to once again go through this grid pipeline. You are simply generating power that you're going to be delivering on site. So again, you're not needing to go into those areas. The problem, the backlog for these jet engine companies that are converting towards uh power needs is substantial. The number one company in this space is GE Vernova. They're currently shipping about three gigawatts of turbines per quarter against a backlog of 116 gigawatts. So that means that their backlog now stretches into 2031, even with some growth in what they're able to produce. Here's where FTAI Aviation comes in. This is a company that their legacy is essentially doing uh, they're a platform for leasing and uh repairing uh jet engines. And and what they realized is that they could make a pivot towards entering this market and they would be able to create capacity that currently does exist because the backlogs go to 2031. And what we're seeing here, as you can see in the chart below, is their revenue is expected to grow from about 3.9 billion this year. And as they scale up, it's expected to hit more than $10 billion by 2029. So here's what's interesting this company is currently at about 10 times their expected 2029 earnings versus GE Vernova, which is twice that multiple. So they're in a very attractive industry. There's a lot of capacity to absorb with the plans. You know, I showed you the plans of all these hyperscalers earlier. Well, again, they're going to need to find some solutions. And FTAI recently signed a massive contract, which is believed to go to SpaceX itself. And again, uh a potential threat to FTAI is that Elon Musk sees this industry as so essential. He's going to begin trying to build his own jet turbines to be able to power his data centers. So, on one hand, that is a threat. On the other hand, he sees this as so essential. Um, it says something about the market itself. And another beautiful thing is that if the sales actually decline for FTAI in the future, well, one thing is you get a large install base. Well, you're gonna have maintenance contracts on these as well. So you do get some recurring uh profits behind. So this is a company I think is really exciting. I've recommended some other companies in this space, though. I I've never recommend Bloom Energy, but I see it as one of the most disruptive companies in the behind-the-meter space. You'll you'll see uh their ticker and logo below. Solaris Energy Infrastructure. This is another company that um SpaceX is relying upon for their extremely ambitious build rates. Um, yeah, they rent mobile gas turbines to data centers. Um, again, you know, their complete package, it's fans generation, distribution, emission control, storage, operations, and maintenance. Um, it's it's a company that pivoted. So many of these companies are pivoting. They are seeing, you know, you could say, well, these companies are pivoting to something that is about to enter its peak. But if you understand the insatiable demand happening right now for AI, and you understand how limited the grid is going to be in terms of solving power needs across the coming years, you see how lucrative this can be for these companies for years. You know, it is something more akin to memory, these stocks in memory at the beginning of 2024, than where they are today, where this situation is largely played out. You know, we are looking to get into trends like behind the meter right now, when the market, in my opinion, poorly understands the kinds of growth rates these companies could see as we look at the build-out across the coming years. And lastly, I'll just say if you weren't even looking at companies like FTAI for its um, you know, its pivot uh towards jet engines as um, you know, power generation, bloom energy, if you weren't interested in Solaris Energy, which is SEAI, SEI as a ticker, you could even look towards natural gas companies. As a lot of these behind the meter um you know, data centers begin operating. It's going to be incremental pressure in very tight uh natural gas supplies, right at a time that natural gas is going to see more global exporting things to a lot of liquefied natural gas that's been created. So you could look at companies like WBI, execute, once again, all on the bottom of the screen right now. So this is this is, you know, I do think this is behind the meter. It's it's been something that if you are really in the weeds of AI, it's been talked about for two or three years. But where it really hits is acceleration and the need for it becomes really acute truly begins in 2027. So I'm I'm really excited about some of these opportunities. Another really interesting trend that is beginning to take shape in 2027 is Nvidia's pushing for one of the largest architecture shifts ever in how power is delivered. We talk a lot about delivering power to data centers. Well, that's one side of it. Another side of this is how power is actually delivered inside the data center. And that's an area far too few people have been thinking about because it is an area with extreme amounts of growth. So, what we see here is NVIDIA is pushing that basically they want to currently you're going to take power at about 54 volts to servers. Um, they want to take this up to 800 volts and they want to eliminate a lot of conversion steps between alternating current and direct current. They want to take direct current directly to the servers. Now, what are the downstream implications of this? Well, number one, it's going to improve efficiency, it's going to make data centers consume less power. That's that's good because, again, we just saw in the last section how much of a bottleneck power is. But another thing, it's going to lead to significantly less copper usage. Uh, NVIDIA has an estimate that for a gigawatt data center, um, this transition will lead to 200,000 fewer pounds of copper needed. But the other implications on stocks is going to be you're going to essentially move from a lot of revenue going to companies that make these um specialized kind of power distribution units into chips themselves, power management chips. And the kinds of power management chips that exist are going to use entirely new compounds. Instead of silicon, they're going to use things like silicon carbide, uh, gallium nitride. And there's companies with specific expertises in these spaces. So, you know, what are some of the companies that we're looking at here? Well, a lot of them have seen runs throughout kind of April, May, and June, but a lot of them have seen significant fall-offs in recent months as many momentum stocks hold off across the summer, which makes their valuations far more attractive. These are often companies, a lot of a lot of these kind of um power management chips for higher voltage were only previously applicable to EV uh markets. So these companies are often priced and thought of as EV companies. But as this 800V shift happens, um, again, you we're only gonna see, I've got in the prior slide, 2% ass tires at 800V uh estimated in 2026, 37% in 2027, 59% in 2028, and then 74% in 2029. So we're effectively going to see a widespread shift happening across the coming years. And that's going to lead to booming revenue for these power management chips for companies like Infineon, ST Microelectronics, which is Orex. Recommendation. And when you look at a more speculative idea, something like Navitas, which again, this is a speculative company, you'd want to position to only a small position, but a company that maybe could see significant benefits from this transition. But one stock idea that I haven't discussed before on the AI Investor Podcast from Power Delivery is going to be Vicor. And this is Ticker VICR. It's it's a company that relative to some of the others that we would be looking at is maybe a little bit more expensive. But what it what it has is it might solve kind of the next bottleneck caused by distance. And that's really important because we think about high bandwidth memory. Again, this high bandwidth memory, this memory that's used in AI servers, high bandwidth memory went from $23.9 billion in revenue in 2025 to an estimated $173 billion in 2027. Just phenomenal growth rates. And the leading stocks in high bandwidth memory are companies like Micron, which across the past year is up about 700%. SK Heinex, which is up 540%. Again, these are recommendations we've made. And what high bandwidth memory does is basically it puts memory as close as possible to the compute die. It puts memory about five millimeters away, where conventional memory sits 30 to 55 millimeters away. So this putting memory closer to the computing die leads to more bandwidth per watt. And in an AI world, this is just absolutely essential. This is an absolutely massive idea. And when we go down and look at the next trend that we've talked about, co-package optics. Well, in AI networking, there's a similar dynamic that we need to basically get the distance down to millimeters. And what we see with co-package optics is this market is expected to be $100 million in 2026. Essentially nothing, but expand to 15 billion. So what's that growth rate? It's a lot. It's a lot. It's 150x or so by 2030. And once again, the distance is millimeters because the current dominant technology is these pluggable transceivers that plug into the backup servers, and then you have to convert from a light signal to an electrical signal and run about 15 to 30 centimeters. Well, what co-package optics does is again, it takes it down to essentially a distance of millimeters, um, where you're going to have basically just this electrical transition be as small as possible. And that's going to have incredible efficiency gains. And when we look at the companies with dominant positions in co-package optics that we have right here, Lumentum up 617% the past year. Coherent up 218%. You know, people are looking for the companies that can solve this technology bottleneck, right? We, you know, so many investors are looking for bottlenecks of where things are sold out. What you're actually looking for is what's the next technology bottleneck? What's the next architectural shift and what companies are best positioned for that? And we might have something here happening with power delivery itself. So what Vicor has is their vertical power delivery modules. I'm going to try and move this here. Actually, let's exit here, and I'm going to move this slide so you can see it better. But what Vicor is going to enable is that they're going to basically move again power delivery closer to the die itself. And again, this is going to take something that is moving from uh you do these advanced packaging, the space on chips is getting more and more premium. And you're going to wind up at a situation where you're going to need a power delivery unit that's as close as you can to the die. And this is a technology that Vicor is dominant in. And as we see higher voltage delivered directly to chips, this is going to become more and more essential. So Vicor today, it's a little expensive, especially relative to some of the other stocks that we've talked about. But if it's technology, where we're seeing the shift happens, I believe is going to follow what we saw in high bandwidth memory, what we saw in optics, where it's who has the best technology for being able to essentially get as close to the compute die as possible. And in power delivery, I think that company is Vicor. And if this does play out, I think this is a company with significant upside, similar to what we've seen in areas like memory and areas like optics. So is this a little bit more asymmetric where it either plays out or doesn't? I think definitely that is that is the case. But if it does play out, I think the upside is relatively substantial for this trend. So let's move on. We've got one more trend for 2027. And I just wanted to talk a little bit about semiconductor equipment. If you've been listening to this podcast, the AI Investor Podcast in the past, you know that we believe that the growth rate of semiconductor equipment is going to outpace expectations in the coming years. And what's what's so interesting about this is this is a space that's extremely well positioned for the coming years. And it's a space that's been kind of the hidden winner across the past 15 years of investing. When we look at the top performing stocks, which I have on screen right now, you see names like NVIDIA, Comfort Systems, Tesla, Broadcom, Micron. Many of those names are things that investors would likely know. But you continue down the list and you've got Lamb Research, KLA Corp. Applied Materials. The key idea here is supplying the equipment to make chips has been kind of sneakily one of the best industries on earth. You know, you don't see an Apple here again. You don't see an Amazon, you don't see all the names that most investors know. These companies kind of toil in the background. But as technology becomes an increasing part of the economy, the companies that make the systems to make the chips, well, it's incredibly complex. It's it's an area that it's it's hard to get into. And when you have an expertise in a specific vertical area of building systems to make chips, you generally are able to hold on to the advantages for long periods of time. So, you know, I think when we're looking at what the key opportunity for the coming years is, it's something I just got into in the last section, talking about power delivery itself. That advanced packaging, you know, in the past, how capable chips were would be you'd be going down from uh seven nanometers to a five nanometers to three nanometers, or just increasingly shrinking the size of the transistors themselves. But where we're going to be looking in the future is we are small enough that this shrinking capability is going to be less of the progression of chip capabilities. And how you actually package chips, again, how you how close you put things to compute die, how you're able to stack memory, all of these areas is going to increasingly become the driver. So the companies with expertise in how to do this packaging are going to be the biggest winners. And also, it allows some new entrance into this space. So you're going to be able to see companies that have their new kind of niche areas that they dominate in establish themselves. And as I showed, saying that three of the top 15 stocks across the past 15 years, um, each with 5,000% returns or above, are semiconductor or equipment companies. We could see the birth of some of the new winners here. So some of the stocks I really like here that are in our portfolio, Anto Innovation, Camtech, uh, a larger one that maybe less upside, but less risk would be a LAM research. And again, as as we come into 2027, 2028, and it becomes clear how much more of the semiconductor equipment is needed to be able to build out the ambition around how big these data centers are. I think the growth rates for these companies is going to be significantly in excess of what Wall Street expects. So, you know, this one, you know, I know these are all kind of in the weeds trends in the sense that, you know, behind the meter and and uh power delivery and uh semiconductor equipment, they're they're not as well known today as memory. But again, you just need to go back. Today, the most searched for thing in investing is areas like DRAM ETF in memory. Two or three years ago, that would have seemed deeply boring, right? Deeply boring. Um and today it's an area of intense interest among investors. And I think we have some next steps there. I covered five of them across the last section. I also gave you those three kind of featured stock ideas in addition to a lot more. Um, so you know, I think I think this is just an area that is going to see a lot of increased attention. So let's look beyond. So I just covered some of the big topics in 2027. Let's let's look beyond. What happens in 2028 and beyond, right? What is the big picture? What are some trends that are beginning today that maybe they're not going to see the catalyst and the acceleration in 2027 because they still have a little bit of more work happening, but there are trends that can change the world. Um, we'll we'll cover some of those here. So I think question the first question I want to cover is what I'm calling the $345 billion question. And this is, you know, the number one question you get asked when you're covering AI is is it a bubble? That's what everyone wants to know. Is it a bubble? And where the bubble begins, question begins is are companies like Google going to be spending more than they're capable of? And if the demand for AI slows down, they're going to be caught in a situation where they were spending essentially beyond their means and they're going to have to pull back extremely rapidly, and we're going to see the market basically implode on itself, like the dot-com bubble, right? That is the question. And so this $345 billion question I have it outlined here, that is the gap between what I believe most of these large AI companies are going to try to spend next year on CapEx in 2027 and what they're going to earn in operating cash flow. Google next year, Wall Street has them at $266.5 billion in operating cash flow. And their internal ambitions are to spend closer to $375 billion. So how are they going to do this? Well, number one, I think there's a few ways that we can look at this. Earn more revenue. What's happening right now is Wall Street's likely underestimating how much revenue these companies will earn. You know, these these companies, all, especially an Amazon, a Microsoft, a Google, they have cloud computing units, and Google's cloud computing units now growing at 82%. If demand for AI continues, they're going to earn more money. So maybe, maybe they'll earn instead of $266.5 billion next year in operating cash flow, maybe it's closer to $330 billion, $350 billion. And suddenly that gap isn't so large. Second, I think we're going to see a lot more financing vehicles. Nvidia recently announced a $500 billion special purpose vehicle, essentially turning computing these DAS hours into an asset class. Broadcom is in talks for $100 billion for its own platform. I think we'll see companies do a lot more financing vehicles, essentially turning compute into an asset class. And third, they can issue debt or equity, which we've talked a lot about on the AI Investor podcast. But, you know, recent debt issuances continue to be oversubscribed. There's an incredible amount of demand. A company like Berkshire Hathaway, uh, with Buffett still has a chairman, is leading the charge to buy Google's debt, which shows how much quality someone like a Buffett believes this company still has, even as they are issuing debt to continue funding these large build-out plans. And so that that continues to be a way that they can continue getting more operating cash flow. And then I think this leads into the next question: how how big will the AI market be in 2028 to 2030? Because what we look here is a chart that shows how much each of these hyperscalers are spending by year. You know, I've displayed this several times throughout this presentation, but this really shows how much of the takeoff. You know, if we think, if you think that 2023 to 2025 were crazy, you you only need to look at the period here on the chart between 2023 and 2025. It's absolutely nothing compared to where we're headed next, right? Um, and again, can we keep growing? Is 2027 a peak year for AI, or can we keep growing beyond it? You know, I think a few things. Number one, open AI and Anthropic. How much revenue are leading companies like this going to be able to generate? Anthropic, as part of their IPO processes, point towards exiting 2020 about $200 billion in what you call annualized revenue, ARR. Some of their VCs have said that's closer to $250 billion. And some third-party VCs, like a Gavin Baker, who I've talked about on this podcast, have claimed the number could be $400 to $500 billion. Where these companies hit on their lofty targets will say a lot about how big this market can go. Because again, if an open AI and an anthropic are hitting numbers beyond $250 billion in terms of their revenue by the end of next year, they have essentially become companies the size of Microsoft. Companies the size that are just slightly trailing behind a Google or an Apple. And that's going to significantly fuel how much revenue is coming into AI itself. When we look at the numbers, here's the numbers to know in the coming years. $845 billion. That's what's expected to be spent by these hyperscalers in 2026. $1.37 trillion is their expected, how much they're going to spend on data centers, their outlays for capital expenditures in 2027. And $1.446 trillion. That's how much they're currently expected to spend in 2028. Now, what does someone like NVIDIA CEO Jensen Wong believe they can spend? Well, he's put the number by 2030 at closer to 3 trillion to 4 trillion annually. If we did hit that number, most of the companies today, an NVIDIA, a Broadcom, these large AI companies, are going to look extremely cheap in retrospect because they're essentially priced, and we'll get more into this in just a minute, that growth rates are stalling from where they hit in 2027. How could we continue growing from here? Well, it really comes down to how good these models get. We have some potential breakthroughs, and there's been a lot of signals from the actual labs themselves that they're about to see a next wave of AI growth. One area would be continual learning. Today's models, they're they essentially train and they're frozen in capability, and they get ideas like memory from what we've talked about earlier, basically these hacks that AI models today, in a sense, are very inefficient. And they they need to use KV cache to essentially take notes from your prior conversation and access that KV cache when it's generating every single new additional word itself. Um, what continual learning would do would it would allow models to actually update from experience. So it would make them not only significantly more capable, but think about this, think about the switching costs that would exist here. If continual learning is a concept and you're using predominantly open AI or anthropic, suddenly the memory that the models have would become significantly more valuable. So you have to think about how much of today's rates amongst companies like NVIDIA suddenly getting far more involved, buying Hugging Face and buying, you know, companies in open source models. How much is then maybe seeing where this next generation of model development is going and knowing how sticky these switching costs could be? Um, you know, this is a company like NVIDIA on their earnings call, they have been nothing but bullish about where the future of AI companies is going and the revenue opportunity for companies like OpenAI and Anthropic. And I do believe a big part of that is where they see the next shift for AI models coming from. Another area is I've got right here, number two, long horizon agentic tasks. You know, this is again today, because these agentic loops are so memory intensive with the current design of LLMs. Well, if we are able to make them significantly more efficient, we could have tasks that could go for hours or days. It would be, you know, I've got here, but a task today, you say, how do I migrate a database and get suggestions? A long-horizing agencies would be you say to a group of agents, I want to migrate a database, and they would work for days and they would come back and say, we've migrated the database, we've tested it, we fixed all errors, and it's fully deployed. Um, this is this is where you see how much this impacts workflows and where AI becomes the center of workflows, especially for broad knowledge work, which is my third point, the broader knowledge economy. Um I took a lot of pains to talk about the growth of coding earlier in this presentation and how much that's driven the incredible growth of AI across 2026. Well, the reason that coding was kind of a first area that AI really revolutionized it and um, you know, has completely changed how coding is is deployed across companies is coding is what you call verifiable, right? You you write code and you see whether or not it works or not. Um math is verifiable. There's there's a lot of domains that are verifiable. A lot of broad knowledge work is not verifiable. It requires ideas like judgment, but there's a lot of signs that AI is becoming extremely more adept at these non-verifiable domains. And if that happens, the total addressable market for AI increases substantially because you know, advanced economies like the US and Japan and you know, these developed markets, their core foundation is the broader knowledge economy. It's the knowledge economy. And if AI is suddenly much more applicable to this broader knowledge area, again, where you can use agentic AI and Where it can have meaningful additions and massive productivity boosts, you know, at 10x from the levels today. So when Jensen Wong says, I believe AI can hit three to four trillion by 2030, what is he seeing today that's different, right? If he had said it would have sounded just as ridiculous if he had said a year ago, I believe AI could hit $1.4 trillion in 2027. The market was expecting $700 billion. It would have sounded completely ridiculous above all expectations. But what the market would have missed was how capable AI became. And there is something really interesting here. It does come down where a lot of people fear, you know, the initial deep seek moment for AI was largely that these models would get so efficient you would need to spend more on AI. This is Jevins' paradox. You've probably heard of it. Making AI more capable means it's going to be more broadly used across nearly every single use case. AI will be much more broadly used across companies and processes if it has things like continual learning, if efficiencies from recursive self-improvement, essentially these models being able to improve themselves, allow us to create these long horizon agentic tasks. All of a sudden, the application of where you can use agents for AI processes explodes by a hundredfold. So at the end of the day, whether or not a company like NVIDIA is worth three times as much when we get to 2030 will largely depend on whether or not we stall at current levels, whether or not we hit this 1.4 trillion and spend a year and we're not really moving, or whether or not we get to that scenario that Jensen Wong has outlined, where it's three to four trillion dollars annually. And what's going to drive that is whether or not we see these capabilities, right? Because what people missed, if you went back to mid-2025, the market broadly did not understand how much agenc capabilities was going to change the demand structure for AI in the year ahead. So the question you need to ask yourself, what is the market potentially missing today? What are the future breakthroughs that make these models significantly more capable? Because that is what ultimately will decide whether or not we are going to be on a path towards three to four trillion dollars in spend by 2030. And if we do that, like I said, most of the companies we discussed today are going to see likely incredible returns still to come. We will have truly only been in early innings. Well, 99% of the media narrative around AI is the idea that we are in a late stage bubble. So that is the question. Finally, I've got some questions that keep me up at night. Some bull versus bear, which you know I was just kind of pontificating in the last section. But I'll go through a few things that if you are an AI investor, you're interested in investing in AI, or you're just a broad investor who knows how important AI is. These are some key questions you need to consider. First of all, what are you paying for 2028 earnings? Nvidia just reported earnings on August 26th, and they said that their growth rate for 2027 is going to be 70%, and it would be higher if they weren't supply constrained. This shocked the market. Well, if you were looking at the statistics I had in this presentation, it shouldn't shock you because you would see the growth that's consistent with a broader growth rate for AI spending expectations from 2026 to 2027. We have it grown from about 845 billion to about 1.4 trillion. That's more than 60% growth in 2027. NVIDIA is forecasting 70% growth. The amazing thing about NVIDIA after these earnings is they're projected to do about $409 billion in revenue in 2026 to $695 billion in 2027. Well, what does that do to their multiples? Well, in 2027, NVIDIA is trading as I filmed this for about 13.1 times earnings versus an SP 500 average of 19.4. If we go to 2028, NVIDIA is trading for about 10.5 times expected earnings versus a market average of 17.9%, meaning NVIDIA, despite growing at a 70% growth rate, is trading at, well, depending on the year, a 30 to 40% discount to the market. Companies like Broadcom are very similar. It trades at a similar rate. So the question is, what is the market pricing in for these companies right now? For NVIDIA, it would appear the market is pricing in that 2027 is going to provide a peak year, that this is the most we'll see for this kind of growth of AI spend. I just gave you the scenarios where that maybe doesn't come true. The question is, what's your upside versus your downside? If AI spending were to relatively stall and you've got a company that's trading for a 40% discount to the market, how much do you have to lose? But if the other side would be that we did hit that $3 trillion to $4 trillion in spend by 2030, NVIDIA were to re-rate to trade closer to market averages, maybe the upside is 2x to 3x. Maybe you have asymmetric return potential from these rates today. The bottom line is many leading AI companies, there's this kind of belief across the market that they trade for extremely expensive rates. The opposite is now generally true. And on the other side of the coin, if you are buying a company that's trading for 30 times earnings in 2028 or maybe even higher, do you want to take that risk on when you can find companies trading for much cheaper rates in the AI trade? If a company is trading for that much, you want to make sure it has opportunities where it can grow at significantly higher than broader AI spending. And we talked about what those opportunities might look like with a company like Semtech earlier that could see growth rate significantly in excess of broader AI spending in years like 2020 to 2030. We talked about that with Vicor, again, a riskier stock, but one that could see a takeoff as we see a technology shift across that timeframe, even while maybe growth rates for other companies decelerate much more. The next question I would ask is what moats will exist in five years? You know, if you're a longtime investor, one of the most famous investing ideas is the remote. You know, what I think it was popularized by Warren Buffett, and that's, you know, it's it's a castle with a moat surrounding it. And how protected are you? Um companies with no motes or narrow moats uh, you know, lack um features that would uh allow them to persist, even if they're, you know, they were to execute poorly. And companies with high moats have generally advantages that would allow them to win uh regardless of surrounding changes to their business. Uh you know, high moat companies in technology have traditionally been things like Google, um, Microsoft. But we are seeing a really changing world. Uh OpenAI recently announced a new chip named Jalapeno that they um essentially it looks like design with 99% AI usage. And and suddenly the question is what is the mode around chip design? Um how persistent will that be? And and what are the companies that, regardless of how the future changes, you need to imagine into the future. If you're looking at what company has the highest probability of being dominant 10 years from now, what company broadly benefits in the most ways from a future? Where intelligence is almost unlimited and it's incredibly cheap. And I think when you go through that exercise, it leads you to companies like Amazon, a company that supplies the actual computing, thanks to its cloud computing division, has advantages in areas like robotics that could take off that would transform its business and how it's able to extend its lead. And um, you know, you'd also look at a company like Taiwan Semiconductor, that if intelligence is rapidly exploding, they still make the chips themselves in an incredibly hard to disrupt company or something like a Quanta Services, which, you know, I mentioned this in a recent podcast episode, but Quanta Services is a company that if this AI future continues progressing, while companies will build behind the meter today, there would be no question we we do need to get back to improving the grid. And a company like Quanta Services is a leader in basically grid modernization that if if that becomes a priority in terms of reorienting our company uh country's economy to this kind of new digital world with AI, Quanta Services wins in as many scenarios as possible. And you know, the third area that I'll put on screen right now um would be what industries take off next. I I talked about some of the areas today that seemed the most attractive in 2027. I talked about behind the meter. I talked about the growth for some of these uh software companies in sovereign AI and and um large enterprises increasingly wanting to protect their data while also using a variety of different models from AI companies. I talked about AI networking, I talked about power delivery inside the data center, I talked about who makes the machines that build chips. I talked about all these things that are going to see, in my belief, some near-term acceleration in 2027. But but what are some topics that maybe won't see the near-term acceleration in 2027, but have massive opportunities in 2028 and beyond? I think one of the biggest areas is autonomous labs. Um, we recently saw Moderna announce a positive phase three trial using personalized treatments uh to combat one of the most deadly forms of skin cancer. And what autonomous labs are is the concept where AI designs experiments, uh robotics runs them, and analytics feeds back results so that you're able to shape the next run. And this could be truly transformative for a space like biotech with companies that could be some key beneficiaries. So thinking in your mind what the next progression of AI looks like and what impacts that could have in a space like healthcare and biotech. Also, robotics. I've talked about robotics forever. It is the holy grail of AI. At the end of the day, most of the world's economy is still a physical economy, and robotics is the most addressable space for that. Uh, the the the problem is there's a lot of insatiable demand for it. Unitry, which is a Chinese company, recently IPO'd, and their IPO is 8,000 times oversubscribed. A lot of people want this to happen at a much faster timeline than maybe it's delivering through as it relies on a series of breakthroughs to truly be something that um changes the world. So it's something that I am incredibly bullish on across the next 10 years, but have some worries about expectations built in across the next one to two years, but it is always a space I'll be looking at. And if you are looking at that kind of AI bull case where it is building in, you know, it's not something that just generates investor wealth, but it is building in a legitimately better world for us to live in. I think robotics needs to be a key part of that. So I think, you know, we have reached the end. It has largely been a marathon. I hope this has been interesting for everyone watching today. And I'd like to mention a few things. First of all, if you enjoyed this, subscribe to 24-7 Wall Street for more AI stock ideas. We're going to be putting out a lot more videos just like this. Um, I hope that today, if you are someone who's been watching AI and really trying to get a grasp of what's going on in it, I hope today was something that, yes, it took 90 minutes, um, which is a long time for me to be talking to a camera, but it it really gave you a look at why we've seen this inflection point in 2026 and what this inflection point pretends to in the future, what we've seen so far today, and how that will be built upon for some of the biggest trends across 2027. We put out weekly videos on our YouTube channel. We're going to have more YouTube exclusives diving into specific technology areas. We're going to cover just investing in general, 24-7 Wall Street. I'm an owner of it, and it's my goal to make this the number one space for investors to be able to come to and be able to get news on their companies and um, you know, continue um, you know, following what has become a 24-7 news cycle. That there's there's always something that it, if it's if it's not, you know, company earnings, it's that we have uh investors on CNBC and we have social media and a constant news flow that can become overbearing. So having a place that kind of takes that signal from all this noise and delivers it to you is more valuable than ever. Also, the AI Investor Podcast, we put that on our YouTube. We also have it on all the leading podcast services: Spotify, um, Apple, Amazon, you know, you name it. We are distributed there. We also manage our portfolio there. The average recommendation since we started in late 2025 is up more than 125% across more than 50 recommendations. So you're not only going to get the latest AI news, you're going to also be able to follow that portfolio. And finally, if you want to follow me on X, I'm at Bleaker Tech, as you can see on the screen. My website is 247wallstreet.com, which I would love for you to visit. Make a preferred source on Google News. And if you want to email me, I also have my email address on there, but I won't say it out loud because I don't want it going in the transcript for bots to see, which will be unavoidable regardless. But I hope you've enjoyed this presentation. I hope you continue following 24-7 Wall Street. And also, I will put out there, if you are watching this video, if you're new, put a question below. We'll do our best. We're going to do some um listener QA um sessions where if you subscribe to the channel, you put a question below, you're going to see us be able to give an answer to it because what we're trying to build here is not just a YouTube channel, but also a community for people who are really interested in this next wave of investing with AI that I believe will be kind of similar to the birth of the internet. You know, the internet wasn't just a technology. Every company became an internet company. It was a transformation that existed across the next decade. And if you weren't really plugged into what the internet was doing, you were just behind on investing in general. I believe the same thing will happen as we've shown today. Whether or not you're using AI a lot in your own life or you are using Chat GPT to do queries, there's something much deeper, much more fundamental, a transformational shift that's going to impact the entire economy. So with that, I thank you for tuning in. I hope you'll join us in the future on 247 Wall Street. The AI Investor Podcast is for educational purposes only and should not be considered investment advice.