Dave:
Are we in the midst of an AI bubble? The technology, it’s clearly incredible. It has already started to reshape our economy, our labor force, and it has been the primary, some would even say the only catalyst for economic growth and the stock market for some time now. There is generally speaking, just a lot of excitement about AI and for good reason. But is the hype about AI getting ahead of its actual usefulness? And as such, could we be in an AI fueled bubble? Or on the other hand, are we just at the beginning of a massive economic boom? Today and On the Market, we’re going to dive into the AI economy, what it means for our country as a whole, and what it means for real estate investors in particular in the year to come.
Hey everyone. Welcome to On The Market. I’m Dave Meyer, the chief investing officer at BiggerPockets. And today we’re going into a topic that I have been wanting to discuss for a while now. We’re going to be talking about AI and whether or not we’re in a bubble. And it’s taken me a while to research this. I needed a lot of time to dig into this because there’s so much to it, but I have done it. I spent weeks looking into this topic. I’ve learned a lot. And today I’m going to share with you what I’ve learned because as you have probably heard and seen, the financial news is full of pretty contradictory opinions about AI and its role in the broader economy. Many people seem to be all in. They are betting big that AI can power the stock market and the entire US economy to even greater highs than it’s at right now.
Others though feel that AI spending and the valuations in the stock market is creating a bubble. And should that pop, it would have dire consequences for, of course, the stock market, but also potentially for the entire economy, real estate included. So which is it? Is AI a growth engine that’s just starting to get going or is there some irrational exuberance going on in the broader market? Today, we’re digging into the AI economy. We’re talking about what’s going on, where all this money is flowing, what might happen in the coming year, and how it could impact investors. Let’s do it. So if you watch any financial news or really just follow any media, you know that AI is really all the rage these days. It’s not actually that new of a thing. People have been experimenting and talking about AI for a long time now. But what’s really changed and what I’m going to focus on in today’s episode is basically what’s happened since ChatGPT launched in November 2022.
Because that was, at least for the average person, the birth of AI or generative AI, which is the type of artificial intelligence that we’re all using when we use Gemini or Claude or Grok or whatever you’re using, those are called generative AI models. It’s a type of artificial intelligence that can do exactly what it sounds like. It generates stuff, ideas, research, sentences, images, videos, that kind of thing. And this mainstream introduction of generative AI when ChatGPT came out is really what has changed behavior in the stock market, has changed behavior with businesses, in the labor market. It was really the first, or at least the most tangible tool the world, at least the average person got to see the power of AI. And since that happened, things have really started to go crazy from there and with good reason. I probably don’t need to tell any of you this, but of course the potential impacts for AI in general and generative AI are enormous.
I do not think it is an exaggeration to say it could be one of the most significant technological advancements in human history. Everything from research, medicine, law, art, all of it is ripe for augmentation by AI. So naturally, businesses and investors are pretty excited about what’s going on here. To many investor types, this is basically like the start of a new internet era. If you remember back then, if you were alive for that, you might recall that investors and entrepreneurs, when the internet came around, saw huge potential and they actually realized a lot of that potential. Many people built enormous fortunes riding the wave of the new technology that was the internet. Workforces, systems, processes, all of them were completely recreated. And in reality, although there were some dips and bumps along the way, the internet lived up to the potential that it had when it first came out.
And I’ll say personally, I believe that the potential for AI to do the same thing is there. AI has enormous potential. So when I’m talking about whether or not AI is a bubble, what I’m talking about, at least in this episode, is not about long-term potential. I’m not really debating even if AI is going to be useful or not. What I’m talking about is whether investors and companies are getting too excited too quickly. Are they getting ahead of themselves? Are the decisions that businesses and governments are making about AI logical or are they irrational? That’s really the question that I want to dig into today, like what’s going on with the AI economy today and in the next few years, not decades from now, because as investors, that’s what’s going to matter to us for our portfolios in real estate and in the stock market.
It’s going to impact how you make financial decisions because as you’ll see throughout the course of this episode, the AI economy has gotten so big that whether it succeeds or fails, we’ll touch almost every other part of the economy. And that’s why we’re digging into the short-term impacts of the AI economy here and now in this episode. So to get into that, we got to first just kind of take a look at what is actually going on in the AI world here and now today. There are, of course, tons of different companies getting into the AI space, but unsurprisingly really, the spending is really dominated by just a couple of massive players that are known in the industry as the quote unquote hyperscalers. These are some of the biggest companies in the country or in the world. We’re talking about Microsoft, Alphabet, Google’s parent company, Meta, Amazon, OpenAI, Oracle.
We’re also talking about companies that support them and build infrastructure, companies like Nvidia or AMD, those kind of companies. In just the next year alone, these US tech hyperscalers are expected to spend a massive $527 billion on AI capital expenditures alone. That’s just one year in 2026. That is triple the level that was seen of investments before ChatGPT came out. So this is massive scale in just the last three years. Now, I know that billions of dollars, if you read about the financial news or the government, hundreds of billions of dollars gets thrown out a lot these days. So I just kind of want to put this in perspective, just how much money this is. If you just took the money that these couple of companies are spending on CapEx for AI, just isolated that spending alone, it would be the 31st biggest economy in the world.
It would be about the size of Norway, which is a very wealthy country I should mention, Norway’s entire economy. But the Unitited States economy is just frankly so much bigger than every other country. When you measure it, this is only one and a half percent of the Unitited States GDP, which is not that crazy. Now, total overall spending alone is not really enough to understand these issues. We kind of have to dig a little bit deeper and talk about where this money is actually going. As real estate investors, maybe you picked up on this, but a minute ago, I said that the hyperscalers are spending over a half a trillion dollars on capital expenditures. Does that term sound familiar? It should, because we have this in real estate investing. In real estate investing when you’re talking about a rental property, CapEx or capital expenditures, usually talking about things like renovations or putting in a new roof or an HVAC system.
And in the AI world, it’s kind of similar. We’re talking mostly about infrastructure, which for AI at this stage basically means massive data centers. The majority of this money is going into data centers. So much of it, it is kind of hard to fathom. Data centers in general are basically needed to create the computing power that AI and LLMs large language models need to run off of. And frankly, right now, these hyperscalers are all competing with each other to try and build the best models, the fastest models, iterate on these ideas so quickly, they need as much computing power as they can get. And they’re just building these data centers like crazy. You probably hear about it. I don’t even know where you live, but there’s probably a data center getting built somewhere close to you because they’re getting built everywhere. And we’ll talk more about this in a couple minutes if this actually makes sense.
But just for now, note that companies aren’t really pouring money into hiring or into software. They’re putting money into hardware. And this hardware is massively expensive stuff. Some companies like Microsoft, one of the most profitable companies in history, tons of money, same with Google, right? They can self-fund these things, but there are also other companies like OpenAI, a relative newcomer onto the scene, they’re taking out massive debt to just build, build, build. Before we keep going, I just want to caveat this to say to all my friends here who are real estate investors, don’t get too excited about data centers. People ask me this question all the time like, “Oh, there’s a data center going into this neighborhood. Should I buy property there?” No, not just because there is a data center going in there. That might be a sign of other good things happening in that area, but data centers alone are not really long-term job creators, right?
At least as of now, they help construction in the short term, but these are largely automated facilities that might need a couple dozen people to manage them, but this isn’t like a major corporation moving in and moving their headquarters to an area. So I personally wouldn’t buy real estate near a data center solely for that. It’s just not a big enough long-term driver of housing demand. So I just want to caveat that because I know a lot of people are listening to this episode thinking about, what does this mean for my portfolio? I just want to say right now, I don’t think it means the data center building should really impact your strategy all that much. Anyway, back to the question of whether or not we’re in a bubble. I just talked about sort of the spending side, but we also have to figure out what is happening with revenue.
These companies are spending a ton, but are they actually making money from their investments? And the answer is not really, or at least not that much. There has been actually quite a lot of high quality analysis on this topic because these companies are mostly publicly traded companies. There are a lot of professional analysts digging into all the publicly available information about these companies. And from what I’ve gathered from these analysis, in order to justify the 500-ish billion dollars in annual spending that these hyperscalers alone are committing right now, these companies need to generate roughly $2 trillion in annual AI revenue by the end of the decade. That’s roughly how the math works out when you’re talking about stock valuations and whether these companies are going to live up to the expectations that stock investors have into them. So just in a couple of years, they need to get to $2 trillion, but as of right now, as of late 2025, early 2026, the data that we have shows that actual end user revenue, people buying products like ChatGPT or partnering with Amazon or one of these other companies, that revenue is estimated to be under 100 billion and still a lot of revenue, $100 billion, but we’re saying that the industry needs to 20X its AI revenue in four years, possible?
Yeah, it’s possible, but you see why the debate exists, right? Because a 20X revenue increase, while it could happen, it is far from a sure thing that is extreme growth, even for an incredible technology like AI. And so this is the basic argument. Is the investment worth it or is this going to go to waste? Are these companies just wasting potentially hundreds of billions of dollars? In the last year, valuations from AI companies have gone sky high and basically people want to know, is this sustainable? Can the technology actually deliver the revenue and the profits and live up to the investments and the valuations they are getting right now? Or is this another dot com bubble where investors bet too much too fast and a reckoning came, right? You can see why people are comparing this to the 1990s because everyone knew in the 1990s that the internet was a big thing, right?
No one was debating that, but so many people just put tons of money into the stock market because they thought, “Oh my God, the internet is going to raise all ships.” And although the internet eventually did deliver on its potential, it wasn’t without paying. There was a bubble that burst before the true tech expansion and wealth really hit the market. And so people are wondering, rightfully, is this going to happen again? Everyone agrees AI is great, but are people putting too much money before we know if there’s a winner or is this time different and are these companies in great positions actually to be able to win the AI race and capture all the revenue and wealth that will very likely come from an AI boom? Now, we got to take a break right now, but when we come back, we’re going to discuss this in more detail.
We’re going to talk about the bull case, the people who are optimistic about this, and we’ll talk about the bear case, people who are fearful of a bubble. We’ll talk about why some investors think AI has tons of room to run and valuations are going to go crazy while others are fearful of a bursting bubble in the coming years. We’re going to get to that and I’ll also share my thoughts on what this means for real estate investors after this quick break.
Welcome back to On The Market. I’m Dave Meyer talking about AI and whether or not we are in a bubble right now. We talked about just what was going on with spending, but now let’s get into the bull and bear cases. And we’ll start with the bull case. Basically just means the optimistic case, people who are really favorable, optimistic about AI right now and think that the stock market and the economy has tons of upside beyond where it is today. I’m going to talk through a couple of the arguments that these folks make. And the first is pretty simple. Many people believe that AI will just grow the overall economy, that it is so efficient, it will add productivity to the economy, which is the way that you grow an economy without adding more people. It needs to become more productive. And they think that AI will make the economy just overall more productive.
Vanguard, the financial giant, has actually done a study on this and they predict there’s an 80% chance, very good chance, 80% chance that global growth will outperform consensus estimates due to AI. Specifically, they believe that there’s a 60% chance that the US economy hits 3% real GDP growth, which is great. Generally, we average about 2% real GDP growth, and Vanguard is saying that they think the US, this is going to grow. It’s going to go up to 3% real GDP growth, which may not sound like a lot, but going to 2% to 3% is actually pretty good. Furthermore, there are other economists, Mohammed El Arian, he is a very well respected economist who works at Penn Wharton. He believes AI will quote raise the speed limit for the economy, which is sort of this interesting metaphor that I kind of like. He’s basically arguing that the ceiling for GDP growth can basically just go higher.
In an economy without AI, there was just kind of limits to how much the economy could grow in a given year. And he believes that limit is kind of getting taken off and GDP growth can go even higher in an AI era, which would obviously be good for the entire economy. It could justify the spending and the super high valuations from AI companies. So that’s number one, right? It’s just going to grow GDP. The second argument is that the dotcom bubble is very different from all the investment that is going into AI because the hyperscalers that are putting all of this money into the stock market are profitable companies, right? These are like some of the most profitable companies out there, Meta, Alphabet, Amazon, Nvidia. These companies have hundreds of billions of dollars to spare. It sounds like a lot of money, but most of these companies are valued in the trillions and they can, quite frankly, this is crazy to say, but I think that if Amazon missed on a hundred billion dollar bet, they’d probably be fine.
And it’s the same thing with Alphabet or Meta, right? These companies are so big. They are so profitable that they are not as vulnerable as the companies during the dotcom era. Everyone always makes fun of pets.com. It’s kind of like the thing that it’s the stereotypical thing people point out, but there were tons of companies that were getting huge valuations before they were even profitable. Tech kind of got a bad name for highly valuing these kinds of companies, but that’s not really what’s going on here. The companies that are driving the S&P 500, the stock market forward are profitable companies. And so that is very different from what was going on in the late 1990s and early 2000s. So even though stock valuations are very high by historical standards for these companies, many Wall Street experts argue that it is justified. The third thing is that there’s general consensus that AI is going to be very disruptive and whoever wins the quote unquote AI race is going to make a ton of money, right?
The winner is going to be very, very successful. And what a lot of stock analysts and economists believe is that there is sort of a moat because of how expensive this is. I could call that how expensive it is to build data centers, which you can argue both ways. Some people will say that’s making these companies spend way too much. Other people, the people who are more bullish about AI say, “This is actually not a bad thing because these companies, these seven, eight companies are the only ones who can afford to build this stuff. So they are more likely to win the AI race.” Whereas in the past, it was pretty easy for a company to start a new website and compete with pets.com or all these apps were very easy to build. What they’re saying is that AI is so hardware and CapEx dependent that the winner of the AI race is more likely to be an established, big, profitable company than it is to be a small disruptive company like it happened often during the dotcom bubble.
And argument number four is kind of interesting. It’s just about infrastructure spending. I said earlier that massive spending over $500 billion is huge and that that comes with some risk, which it certainly does. But I want to call out that actually there is historical precedent for this kind of investment. This is not some sort of spending that we have never seen before. And you’d actually find some people out there who say that we are underspending. We’re under investing in AI, which is kind of crazy when you think about that number. But when you look at the data, historically, the amount our country between public and private sectors have invested into truly transformative technologies. I’m talking about things like railroads back in the 1800s or electrifying the country around the turn of the century. Infrastructure spending on those huge projects peaked at about two to 5% of GDP.
And AI investment right now is about 1.5% of GDP, suggesting that the boom could keep going. Now, I’m not sure, and we’ll talk about this later, that that’s justified, but I’m just saying that if you have a transformative technology like railroads or electricity, spending one and a half percent of GDP on building out the infrastructure for it is not unheard of. Now, I went on a whole rabbit hole about how the US actually overinvested in railroads and there’s kind of a crash there. So keep that in mind. But still, there’s precedent for this kind of spending. So to summarize, the bulk case is AI is massive. It’s going to grow the economy overall, and it’s likely that one or several of these hyperscalers that are spending all this money, they’re going to win and they’re going to get the revenue and the valuations that the people who are investing in it are expecting.
And they also argue that mostly these companies who are investing tons and tons of money, they have the money to do it. So that is a credible case, right? But what about the bear case? What about the people who are more pessimistic about it? Their arguments go as such. Number one, is spending in the right place? Data centers are massively expensive, but we don’t really know that much about their utility, right? We’re building these huge things. Are they going to become obsolete in two years, in three years, in 10 years? We honestly don’t know because the technology is shifting so rapidly, it’s not really that hard to imagine like, “Oh, we build this entire data center with all these NVIDIA chips and we’re spending billions and billions and billions of dollars.” And then two years from now they’re like, “Oh, actually we need a totally different kind of data center.” You can kind of imagine that happening, right?
And so bears are saying, “Yeah, we’re spending, maybe the total amount of spending is right, but we don’t even know if we’re spending it on the right thing.” And so some would argue that a lot of this money could potentially be wasted. That’s argument number one. Argument number two is really the revenue thing. I brought that up a little bit earlier, but basically these companies are spending so much money, but without really the revenue to justify it right now. And this is really, in my opinion, the most credible bear case right now because the spending, like I said, it could be justified. There’s historical precedent for it if it was generating revenue, but it’s really just not. If you look at OpenAI, the creator of ChatGPT, they maybe, we don’t know, they’re a private company, but the estimates are that they might have about $20 billion in revenue in 2025.
That’s a big number, right? Any company would probably be pretty happy to have $20 billion in revenue. However, maybe you wouldn’t feel so good about 20 billion in revenue if you had $1.4 trillion committed to infrastructure spending in the next eight years. It is just 1.4 trillion is so crazy. That is so much money, it is kind of mind blowing. Similarly, Meta, who’s spending tens or hundreds of billions of dollars has admitted they are not seeing any direct revenue impacts due to their investments in AI right now, but those are just two of the companies. If you look at Microsoft and Amazon, they are reporting positive returns on their investments. So it’s kind of a mixed bag right now, but by no means our company is saying, “Hey, we’re investing in AI and we are getting an immediate ROI out of this. This is so great.
We want to just keep investing in it. ” That is not what people are saying. And in fact, not just with these hyperscalers, when you look at the people who are buying the products from these hyperscalers, so just regular businesses that are using ChatGPT or Amazon Cloud services or whatever they’re investing in, adoption is not so great. MIT just did a study and they said that 95% of AI projects get no ROI. Another report from IBM says that only 25% of AI projects are getting their expected ROI. So there are reasonable questions about what revenue these companies can generate in the short term. So generally speaking, there are projections that GDP will grow. There are projections that these companies will earn their valuations by increasing their revenue. There is a ton of talk and excitement, but the revenue just isn’t there yet and it has a long way to go to justify current stock valuations.
Some might call this speculation. Speaking of those valuations, I think that’s sort of what we need to get to because we’re talking about are we in a bubble? And the bubble could burst because people feel like the stock prices of these huge companies that carry so much of the S&P. These companies make up so much of our stock market. We have to understand how they are valued. Similar to real estate, people can value things on a cap rate or a cash on cash return or whatever. There are so many, dozens of different ways that you can value stocks or the stock market. But one is called the Schiller PE. It’s called the Cape Ratio, if you’ve ever heard of this. It basically measures stock prices against 10 years of inflation adjusted earnings. So the CAPE ratio right now is roughly at 39 or 40X. So it’s basically saying that’s 39 or 40 times those inflation adjusted earnings.
That probably on its own makes no sense. So let me just tell you that historically, the long-term average is 17X. So we’re at more than double. We are now at 40X earnings in the CAPE ratio. Normally it’s at 17X. And the only other time in history that the CAPE ratio exceeded 40 was right before the dotcom burst in 2000. So this is why bears are saying maybe we’re in a bubble because there are just some technical ways of measuring the stock market that are throwing off red flags. Every time the CAPE ratio has crossed 30 for an extended period, the market has eventually seen a decline of 20% or more. So that is another argument that bearers are making. The last one I want to talk about, which is a whole other big topic and it is complicated. So I’m not going to get super into it, but there is this big thing going on where all of these companies, it’s kind of incestuous.
They’re all like funding each other and investing in each other. It’s this giant web of companies spending in and investing in one another. And it’s kind of weird. You can look this up. You should Google it because I can’t explain it briefly in this episode, but it’s worth looking into if you’re interested in this topic. Basically, you see companies, let’s just use Nvidia as an example. They’re investing in AI companies and giving them money. They’re saying, “Hey, we’re going to invest in you, but with the money that we’re investing in you, you have to turn around and buy Nvidia chips.” So they’re saying, “Hey, here’s some money to buy my product with. ” It’s called vendor funding. There have been some pretty bad examples of this in the history where this has not worked out pretty well. Of course, it could be different this time.
It always could be different, but it does make the system to me at least feel a little bit fragile, right? The whole thing where they’re all funding each other, it just makes it seem like a little bit of a house of cards. Now, I’m not saying that’s necessarily what it is, but I would feel a little bit better about this if these companies were making their money and getting their money to invest from revenue, not from one another and sort of trading and propping up the whole industry as a whole. I should say that people who are really bullish about the market think that this is a positive. Some people say it’s a strength because profits are getting reinvested back into the AI ecosystem. But bears, famously, Michael Burry of the big short fame, who famously called the 2008 housing crash correctly, has pointed at this as the reason that he is shorting Nvidia and that he’s getting out of the market because he thinks that this whole thing is going to collapse.
This is not my expertise. I recommend you look into it, but this is a big thing that a lot of experts in this field are pointing to when they’re making their bearish case. So just to summarize the bear case, you would say there’s really no revenue. We don’t know if the money’s going into the right place. Valuations are already near all time highs and can that be sustained? And there’s all this vendor funding. Basically, stocks are priced right now for perfection. When I was reading some of these analyses and reports, that’s the thing that kept coming up is that the way stocks are priced right now, it’s like these companies have to just execute perfectly for the next couple of years to justify them. And bears think that that’s unlikely and that’s why they think that we’re in a bubble. Whatever side that you’re on, I think you can see, I feel at least there are logical arguments on both sides.
And of course, no one knows for certain, but I will share with you my thoughts about all this and which side I’m falling on and how I’m planning my own financial decisions right after this break.
Hey, everyone. Welcome back to On The Market. I’m Dave Meyer. Today we’re talking about the AI bubble. I’ve shared with you a little bit about what’s going on with spending, the bull case and the bear case. And now I’ll just share with you sort of how I’m feeling about after spending several weeks digging into this topic. And I approached it as unbiased as I can. Everyone always is biased, but I I genuinely just didn’t have a real opinion on whether we were in a bubble or not and just started digging into this. And as I’ve done this research, overall, I lean pessimistic about the AI bubble. I’m not saying that it’s not possible that things keep going. As I do with the housing market, I’m going to do the same thing here and say that I don’t like saying X is going to happen or Y is going to happen.
As an analyst, I am trained to think in probabilities. That’s what we do. And I just try to think about what is the most likely thing to happen? Not saying that the alternatives can’t happen, but I think the most likely thing to happen in the next couple of years, I’m not saying in 2026, but in the next couple of years, is that there is going to be a correction in the stock market, a fairly significant one. I don’t know the timing of that. We’ll talk about that in a minute. But ultimately, here’s why I’ve come out this way. Number one, it’s kind of a simple argument, but just we don’t know if this is going to work. We just don’t know. People are so excited about it, which I get. I’m excited about a lot of AI things too, but we don’t know if these companies are going to be able to pull off what they are saying they’re going to.
There’s not really that much evidence of it. Yeah, ChatGPT and Gemini are super cool, but businesses aren’t really adopting them. They’re not making tons of revenue. They’re not saying, “Oh my God, ChatGPT has totally changed my whole business.” Sure, there are individual instances of that, but that is not happening at the scale that they need to justify the stock valuations that they are. And I am not saying, do not get me wrong, I am not saying that I don’t think AI will work in some way, shape, or form. I definitely think it will. What I’m saying is that the AI tools that we have right now is what I would consider a V1, a version one of generative AI. And if you think back and look at history, how many V1 technologies have failed? So many of them, right? How many electric cars failed before Tesla finally got it right?
How many social media sites failed before Facebook took off? Remember when people were investing in Blu-ray or LaserDisc or whatever, only for streaming to take over? We just don’t know what the final form or at least this growth form of AI is going to be. And yes, there are several reasons, there are good reasons to bet on these US-based hyperscalers, but this is what people always say. They always think that the incumbents are going to be there forever. If you asked people 30 years ago, will GE still be one of the biggest companies in the world? Will their investments pay off? Probably everyone would say yes. Look at GE now. Everyone always thinks GE or Sears or whatever are going to be there forever because they can’t envision something that hasn’t happened yet. And then something new comes along and shocks the entire world. I’m not saying it would be wise to bet against these US hyperscalers, but to assume that they’re going to win and win with the current technology framework and infrastructure and investments that they’re making today, that is a really big if.
Because in my opinion, even if Amazon wins, they might have to rebuild every data center they have. They might need to go from LLMs to something called a world model, which is a total different way of building AI. We just don’t know. And I get that they may nail it. They may. But if you’re saying that we are valuing stock so highly because we’re so confident that they’re going to win, it lacks evidence. And to me, as an analyst, that’s why I trend pessimistic, because until you show me evidence that these companies are going to nail the revenue side of it and earn these valuations, I’m going to lean pessimistic. And I just kind of want to build on something that I said before, because again, not saying AI won’t work in some form, but these large language models, these things like ChatGPT where you’re typing in and communicating within AI, what they call them agents, right?
This is just one kind of AI. It is only one structure, architecture for building AI models. There are totally different ways that you can do it. There’s something called a world model that I was starting to look into, and a lot of AI researchers think that’s actually the better way to get to agentic AI, the ultimate holy grail of AI that all of these companies are trying to get to. Some of them, a lot of the leader and researchers thinks that LLMs and all these investments that they’re making is not the right way to go, that there’s a better, different infrastructure for building AI that’s smarter and more efficient. Now, I am not smart enough to know which one is right. I’m just saying that LLMs are not the be all end all. I don’t think anyone agrees or thinks that LLMs in their current state or the end state of AI, that this is the best it’s going to get.
We’re always going to be typing to ChatGPT and writing prompts and getting them back. No one thinks that. So we just don’t know how we’re going to get to agentic AI. Some people think LLMs can get there. Other people think that they can’t. And so I just want to show you that there’s a lot of doubt about the right way forward with AI. And that means that some of these companies could be wasting hundreds of billions of dollars building infrastructure that they don’t need. The thing I kept thinking about when I was doing this research is like, what if OpenAI? Super exciting company. I use ChatGPT all the time. I use Gemini all the time. I use these things. I’m not saying anything bad about these, but I kept thinking, what if OpenAI is basically like the Blackberry of the. Com era? Does anyone remember the Blackberry?
It was kind of the first smartphone, but not really. It didn’t have apps. It wasn’t touchscreen. It had the little track ball and the whole keyboard. I had one and you would BBM everyone. And everyone thought like, “Oh my God, this is amazing. I can text, I can go on the internet, on my phone.” And everyone thought Blackberry was here to stay, right? Then all of a sudden the iPhone came out and everyone was like, “Oh, wait, this is a way better technology. That whole Blackberry thing sucks. I’m never going to buy another Blackberry. All I’m going to do is buy an iPhone or an Android.” And now Blackberry isn’t even a thing anymore. And I’m not saying for sure that that’s going to happen to AI, but this happens all the time, even with really exciting technology. And I think there is the chance that it happens again because history is frankly filled with exciting new technologies where the first mover, the person who introduced the thing doesn’t actually win and the winner actually comes out of nowhere.
And in some ways, this reality is why companies have to spend so much. They are 100% in a race to figure this out first and to try and crush any competition and beat everyone else to the end state. But to me, that means even though one of these companies very well could win, even one US-based hyperscaler could win, a lot of the other companies are not going to win. They’re going to be massively inefficient and they might spend hundreds of billions of dollars in total flops, which of course would negatively impact their stock prices and could pull down the entire economy. That’s my number one thing. It’s just like, we don’t know if this is even going to work, if this is the right infrastructure. And the second reason I kind of lean pessimistic is just the revenue thing. Maybe they’ll get to $2 trillion.
Maybe revenue will start to explode. But when I look at these adoption rates and what CFOs and companies are saying about their implementation of AI, they’re all saying they’re going to implement more. They’re not saying they’re getting great ROIs. And I don’t see companies spending way more because there isn’t a new tool. ChatGPT, yeah, it’s gotten better since 2022, but has it really changed all that much? Are people going to start opening their pocketbooks? I mean, maybe they’ll come out with new product. I don’t know, but we haven’t seen something that’s really going to start driving their revenue in massive ways. Amazon’s been more successful, but we’ll see how that comes out. But for right now, I’m skeptical because there’s just not revenue to justify these valuations. So overall, the way I’m thinking is that there is short-term risk. I’m just not sure we’re there yet.
AI is super exciting, but we’re betting on valuations. The stock market is basically saying, “We know that these companies are going to win. That’s what their valuations are telling us, and I don’t see it. I think they have a good chance to win, but I am skeptical about buying in at these valuations thinking that they’re going to go even higher.” So much of the stock market right now. So I think it’s about a third of the S&P 500’s growth is predicated on one of these companies winning and doing it perfectly. They have to nail it because it’s already priced as if they’re kind of going to win. So if there are any errors, the market could tumble. And I’m not saying that means AI failed, not at all. I just think this is kind of similar to the dot com bubble. People were rightfully excited back then, but they made irrational investments.
Ultimately, the technology, the internet, hugely impactful. And this could be happening again. People know AI is impactful and something is going to happen from it, but I think there is a risk that there’s some irrational investing going on right now. Now, of course, that doesn’t necessarily mean it’s going to be a catastrophe. Of course, stock market crashes are fairly common. We don’t like them, but they happen. And the market rises again. And I personally believe that even if there is a crash, the market will recover. But there is some really interesting data that suggests a crash now could be pretty bad. There was a recent article actually in The Economist by an economist named Gita Gupinath, and she basically says since more ordinary people are investing in the stock market than ended before and more foreign individuals, a stock market hit could be more widespread and have a bigger impact on consumption in the US, which I should mention drives about 70% of our economy, that it could have a really big hit on that.
She actually calculated that if the stock market takes a proportional hit as the dotcom bubble. So basically relative to its size, the same kind of decline, it would destroy $20 trillion in household wealth in the US alone. This could impact consumption, of course, 70% of GPP, like I said, it could impact retirement plans for the massive boomer generation whose majority of their wealth is in 401ks and in the stock market. And this is where the AI potential bubble spills into real estate investing for me because I think if we see the stock market crash, we could see demand for housing and consumer confidence decline. If this happens, yes, some things would be beneficial to the housing market. You would probably see mortgage rates drop, which would provide a floor. I’m not saying that there would be a crash in housing because of this, but I do think it could keep transaction volume low and the sort of very normal and expected human reaction of fear would start to take over because if people see their net worths decline dramatically, they might tighten up on home buying or buying cars or moving into a new apartment.
All of that could weigh on the real estate industry, especially if this potential bubble combines with some other labor consequences of AI that we haven’t even gotten into in this episode. That’s a whole other topic. But I think everyone knows that many people, even the CEOs of these hyperscalers are saying that AI is going to have massive impacts, not good, on the labor force. And so if you combine a potential bubble and decline in $20 trillion of household wealth with a bad labor market, that could really subdue appreciation and rent growth in the housing market in the short term. So that is something to keep an eye on. And I’m not saying that that’s going to be a disaster for real estate investors. I actually think when these things happen, better buying opportunities exist. And so if you’re in it for the long run, that could potentially be good.
I mean, people look back on 2009 to 2012 and say, “Man, I wish I bought then.” That was during an era of fear when not a lot of people were buying and investors had an opportunity to buy good assets and good prices. So I’m not saying this is a disaster for real estate investors. I’m just saying that it’s something that could happen. And why, even though this is a real estate investing show, I am paying so much attention to the AI bubble because it is so big that it could really impact the rest of the market. One other thing I want to call out how this could relate to real estate investors is that if the stock market does decline by 10, 20, 30%, whatever, institutional investors could slow down because a lot of these quote unquote institutional investors are things like pension funds, they’re endowments and they have actually, they have kind of rules.
They have these allocation buckets. So like they say, “We’re going to invest 20% in real estate, 80% of the stock market.” That’s kind of our philosophy, our investment thesis. So if the stock market drops, that actually it’s kind of just this math thing, but it overweights their percentages. They’re over allocated into real estate. So that means they might slow down on buying real estate just because the value of the stock market drops. And that means they could stop buying new properties. They might even sell some assets to rebalance their portfolios. As we’ve talked about in the residential market, these companies own about two to 3%. So I’m not saying that would be crazy, but it is something that you should keep an eye on. So after all this research, hopefully this has been helpful to you, but where I’ve landed is I am still a little bit torn, but I lean a little bit pessimistic about the stock market and whether these valuations can be sustained.
I am skeptical. Anytime our economy or the stock market is so dependent on a few companies, it makes me a little bit worried. Anytime stock valuations are honestly speculative, like let’s just call it what it is. They’re speculating that these companies are going to earn revenue. This isn’t like, “Hey, they had great earnings this year and we are justifying our valuations based on that. ” Some of it, a company like Microsoft or Amazon, obviously a lot of their valuation is based on actual earnings, but the run up in their valuations over the last year or so has been largely speculative. And so that worries me. Even if these are amazing companies, some of these are incredible companies doing amazing things, but the margin of error to me is just small. And so that introduces a lot of risk. And as with everything on this show, I cannot say for certain what will happen.
My goal with episodes like this is just to explain the risk. Just explain that that risk is out there so that you know, because I believe personally, my philosophy on investing is that risk isn’t your enemy. You can invest with risk, but you have to know that it’s there. You need to make decisions and underwrite your deals, understanding all of the risk that is out there. And to me, this is a risk that is out there for the economy and it could spill over into the real estate market. And that’s why I’m trying to share with you the risks that I’m seeing so that you can plan accordingly. I should mention, I’m still heavily invested in the stock market, but I have made my portfolio a little bit more defensive because even though I do think a retraction is likely a correction, we just don’t know the timing of that.
And that’s what’s so hard about stock investing. I’m not going to give stock investing advice. That’s not my expertise or my purpose of this show. But I will just say this, that even though I think that there’s a correction coming, valuations might go up another 30% and then crash 20%. We just don’t know. That’s why I personally just take a dollar cost averaging approach to investing in the stock market and put money in at regular intervals, but I have shifted to a little bit more defensive. I want to be in the stock market in case I’m wrong and things keep going up because I’m 38 years old and even if the market crashes, I think it will come back by the time I want to retire and maybe live off some of my stock investing. I don’t want to get out of the market. I’m not panic selling or anything like that, but I am making it a little bit more defensive.
I am willing to forego some potential upside to protect the downside because after all this research, I do lean a little bit pessimistic, like I said. So that’s it. That’s my analysis of the AI bubble potential as of right now. There are good arguments on both sides, but I’m leaning a little bit pessimistic right now just because I think a lot has to go right almost perfectly for these valuations to be justified, and that just rarely happens. So that’s how I am thinking about this and I’m going to plan my own stock and real estate investing, but I’d love to know what you think. Are we in a bubble or not? What should we as a community here at On the Market be thinking about in terms of AI? Let me know in the comments below. Thank you all so much for listening. I’m Dave Meyer.
I’ll see you next time.