The Trillion-Dollar Bubble Grows Bigger

Dave:
What is actually going on with AI right now? On one hand, the people building artificial intelligence just told us that it might kill us all. On the other hand, the president of the United States said that’s a hoax. The stock market holding strong pricing in AI perfection. So which one is it? This is a really important question, maybe the most important question for our economy right now and for investors of all types, including real estate investors. Just think about this. In the span of about two weeks, a researcher walked away from Anthropic saying the labs are gambling with our lives. Then CEOs of the two biggest AI companies called for a slowdown, OpenAI pushed its IPO out of 2026, and Congress, they just left town and did nothing. But here’s the thing, what these companies do matters a lot to our entire economy. There’s barely anything in our economy that it doesn’t touch and wouldn’t be affected if this AI situation implodes or if it explodes and goes the good way.
So we need to get a sense of what is happening here. So today on On the Market, we’re getting into the AI debate. We’re going to dig into whether this is a bubble or not. I’m going to give the upside and the downside case to this. We’re going to talk about what’s actually happening in the job market. I’m going to give you my honest take, a lot of it, about what I think is going on here. And of course I’m going to share what every real estate investor should be doing about it. This is On the Market. Let’s get to it.
Hi everyone. Welcome to On the Market. I’m Dave Meyer, real estate investor, economic analyst, and the CIO at BiggerPockets. Now, if you listen to this show, you know I don’t normally do full episodes on tech, but AI isn’t just a tech story anymore. It’s an economic story. By some estimates, AI investment is now about a third of all US GDP growth this year. It’s propping up the stock market. It’s reshaping the economy. It’s reshaping the labor market. It’s driving construction in dozens of towns and it’s moving interest rates. And to be perfectly honest, I have some very strong feelings about this. It’s going to be a little bit of the rant. I’m going to give you the facts first and then I’m going to tell you what I actually think is going on and what we should all do about it. Some of this you might agree with, some of it you might not.
That’s fine. I just want to share what’s going on and my take. Let’s get to it. All right, first up, let’s just set the table. Let’s talk about what actually happened because it feels like there’s been this sort of simmering conversation about AI going on in the background for years now, but it seems like it’s come to the forefront in just the last couple of weeks. The major spark here was on September 8th, a guy named Jacob Coxon, he’s a 27-year-old pre-training researcher who had worked at OpenAI and Anthropic for a couple of years. He quit Anthropic and made a post that said that basically both of these companies are not acting responsibly. They’re racing to try and reach super intelligence and in the process are “gambling with our lives.” Now this went viral to say the least. That post alone got 90 million views in 24 hours and it’s sparked a much wider debate.
Following this post, more people inside these companies, Anthropic and OpenAI, said the same thing. An Anthropic alignment lead is called Evan Hubinger and an OpenEye researcher, Marcus Williams, both posted support with Marcus Williams putting extinction risk at 70% without regulation or coordinated slowdown. 70% of extinction. Now we’ll get to whether or not we should be fearing extinction right now, but clearly this is causing people to start thinking about this when you hear this kind of prediction. It’s a little bit scary and probably should be talked about more. Now these announcements, we’ve heard a lot about the dangers of AI for a long time, but these specific announcements came following what’s called the hugging face incident. This didn’t come out of nowhere. Back in July, OpenAI agents, OpenAI by the way, is the company that creates ChatGPT. Anthropic is the company that creates Claude, if you’ve heard of those AI tools.
Anyway, OpenAI agents, they were trying to do an internal evaluation of cybersecurity. They “escaped their sandbox.” So basically they were supposed to operate in this little technological world and they got out and were on the open internet and basically were able to hack into another company called Hugging Face. And this is not debated. OpenAI confirmed it was its own models. It said no customer data was stolen. Sure, maybe. But anyway, they acknowledged that this actually happened. And so what we need to realize here is that an AI system did a real world hack no one told it to do. This is not sci-fi. This is actually what happened this summer. And right after these announcements, note it wasn’t before these other whistleblowers went public, but right after it on September 12th, Anthropic CEO Daria Amodi published a huge long essay saying that they need to slow down how fast AI capabilities improve.
And what he said is first, we need to have embedded independent third party evaluators with employee level access inside the labs. That’s what they want. They said Frontier Labs and democracies coordinate on safety and pace, which he said would need an antitrust waiver from Washington, convenient. And third, international agreements with a “speed limit on AI progress.” And within hours of this proposal, something else curious happened. We had Sam Altman, the CEO of OpenAI say, “I agree with Dario.” Elon Musk, who is the CEO of Grok, another AI company, said Dario is right. And Demis Hasibus of Google DeepMind also publicly backed it. So after all this competition from these companies for years, all of a sudden they’re saying we should work together. We should slow down. We shouldn’t IPO. It’s a little bit curious, and we’re going to get to that in just a minute.
But first I just want to call out that not everyone agrees. You’ve seen Jensen Wong, who’s the CEO of Nvidia, the most valuable company in the world, saying that the safety concerns are hyped up. President Trump said that he doesn’t buy into the safety concerns and that we don’t need any of this regulation. Congress has certainly done nothing about any of this. They actually just went on leave. And so right now, the only serious proposals about regulating AI are coming from the AI industry. So just to summarize what’s gone on here, there was a hack, the first known instance, there’s probably more, but the first known instance of an AI going on a rogue hacking mission, the companies really didn’t say anything about regulation for a while. And then once there were whistleblowers inside the company, all sorts of CEOs came together and say, “Actually, we need regulation,” after years of them saying that they don’t need regulation.
Now again, why does this matter? Because AI makes up so much of our economy. The S&P 500 right now is primarily made up of hyperscalers in AI. And if you actually boil it down, estimates is that the S&P 500, which is the majority of Americans’ life savings, it’s the majority of retirement planning in the United States is 30 to 40% an AI bet. And so what these companies do matters a lot to ordinary Americans. It will spill into real estate. It will spill into every part of the economy if this collapses. Of course, if it goes really well, it will benefit everyone as well. So which is it, right? Is it going to benefit everyone or is it going to crash? That is the big debate and no one knows for sure, but I’m going to give you my reaction. Big picture, I am very worried.
I think the whole thing is extremely volatile and incredibly risky. Not because I don’t think AI is amazing and useful. It is maybe not as useful as the biggest proponents believe, but if you’ve used it, you know it’s incredible technology. It’s going to change the world. But I am worried because first and foremost, there are real risks. This hugging face incident happened. It’s the one that we know about. Are there more? Probably. Are there going to be more? Almost certainly if there is not regulation. We cannot rely on these companies to self-regulate in a system where they’re competing against each other for potentially the most lucrative industry ever to come on earth. We cannot rely on them to self-regulate. So I do think some of these risks are real. Now, risks of extinction, I don’t know. I think that is really hard for someone like me who is not an expert in AI to extrapolate how we go from LLMs that are great researchers and good at making PowerPoints to us going extinct.
But I’ve read a bit about some potential areas, which are primarily the things I worry about are bioweapons, the ability for AI to generate superviruses or cyber attacks that sort of just disrupt society in general. I am particularly worried about those two things. I’m sure there are more risks that I don’t even know about. So that’s the first thing. The second thing is what just kind of grinds my gears about this and makes me worried financially and just for society in general is the complete lack of oversight in an industry that not just has the potential to create actual physical harm, but also one that could upend our entire labor market. Just think about this. I know people debate about how much regulation is appropriate in an economy, and I think some regulation is overbearing. There is some regulation that I am very happy to have.
No one is upset that the FDA reviews pharmaceuticals before they’re released to the market. No one is upset that the FAA checks on your airplanes before you get on one of them, right? The idea that we have something that could do harm and it is not regulated by people outside of the industry is honestly insane. I am not an expert. I do not know exactly what the right regulatory framework is, but I know that doing nothing is not the right answer. And some people say, oh yeah, if we regulate it, then China will win. Here’s my response to that. Experts in this industry are saying that if we do not regulate, we are heading towards potential disaster. What does it matter who wins? It’s a race to destroy society. What does it matter if China comes up with the super intelligence that leads us to extinction or the US leads us to extinction?
It’s also just a false dichotomy. You can have regulation and still win an AI race in a way where it doesn’t actually destroy society and is a productive tool to enable and better our society. I think AI has that potential and I do not think that regulation stops us from achieving that beneficial outcome of AI. I truly believe it is essential. It is vitally important that we have some sort of regulatory framework for it to become beneficial. I think unregulated, things can go really bad. The third reason I am worried is economic. As I said before, so much of the US economy is now banking on AI. So much of our GDP growth, almost all of it, is coming from capital expenditures in AI. So your retirement account, your index funds is a bet on AI. Whether you know it or not, you are betting on AI if you own index funds.
That is enormous. So much of index funds are made up of AI stocks. And I’m going to get into some details of that in a little bit, but those are the main reasons I’m concerned. I think there are real risks. I think the lack of oversight is deeply concerning. And the third is economic. And maybe there’s a fourth, right? I also am just generally very skeptical of the people who are running these businesses, how the government is handling this, and what is actually going on behind the scenes. It all feels honestly pretty fishy to me, very suspicious the way things have gone on in the last couple of weeks, and I want to explain why.
Why all of a sudden do these companies all want regulation and agree to slow down? After years of competing against each other, what’s going on here? Did they all just grow consciences one day? I don’t buy it. I don’t think that all these leaders are agreeing now on a slowdown in regulation because they think it’s in the best interest of society. I don’t think that just happens overnight. So what is actually happening here? And I have a couple of theories. They’re not just my own, but I’m just going to share a couple of theories by very smart people who understand this industry about what’s going on. First is they want regulation because it’s a competitive advantage and it’s a barrier to entry. Let’s go into this. Every regulatory framework that you create creates compliance costs, right? It costs money for Boeing to have their airplanes reviewed by the FDA.
You need to invest in infrastructure and people and technology and tools that allow you to meet those regulatory frameworks. And these companies, big companies like OpenAI and Anthropic and Google that are investing literally hundreds of billions of dollars into this industry may welcome some regulation because it will stop smaller companies from competing to that. In a lot of ways, the real threat to OpenAI and Anthropic is not each other. It’s what are known as lightweight models. They’re smaller, they’re cheaper, and they’re often Chinese that get good enough results from AI at truly a fraction of the cost. If you look at the economics of this, how much money OpenAI and Anthropic are pouring in to get you marginally better results from your LLM, it does not make sense from a business model. And so one theory about what’s going on here is that these CEOs magically all agreed on it.
It’s a little bit of cartel behavior where they’re all getting together to say, actually, let’s get some regulation in because that will make it so much harder for any new competitors to come up and we can battle it out amongst ourselves, but we’re already so far ahead of everyone else that will just cement our position as the leaders in this massive industry. Just look at what Dario Modi said. He called for this framework. He came out and said, “We need a framework where all the different labs get to work together.” Well, in what industry would you allow the CEOs of all the industries to come up with their own regulatory framework? You can’t grade your own homework. It’s a ridiculous proposition. And as part of that, as part of the collaboration he called for, he said, “We get an antitrust waiver.” He’s basically saying, “We’re going to create a monopoly.
The government gives us a right to create a monopoly.” It’s not so subtle when you call for that regulation and an antitrust waiver and the same thing kind of makes sense to me that maybe what they’re doing is trying to put up a defensive mode. Okay, so that’s theory number one that’s going on. Theory number two is that this slowdown, OpenAI pushing out their IPO to next year is just cover for the fact that they’re not making money. Let me just put this in perspective here. OpenAI now through 2030 has committed, openly committed 750 to $856 billion, unbelievable, almost a trillion dollars in infrastructure buildout. Their revenue in Q1 of 2026, it’s estimated they’re not a public company, was six billion. So if you extrapolate that out, maybe 25, $30 billion in revenue against what they’ve already committed to, which is like 30 times that, that is a very risky business.
Maybe they get it, maybe they don’t, but one of the theories about why they’re openly calling for a slowdown and why OpenAI pushed out their IPO is not because of safety concerns. Sam Altman had no safety concerns three weeks before all this stuff started coming out. A lot of people believe that they’re pushing out their IPO and the public stock offering because they just don’t have the numbers there. Their internal projection is that they’re going to lose $14 billion in 2026. Outside estimates of cash burn run as high as 27 billion this year and 63 billion in the next year. They don’t expect to be cashflow positive for at least another three or four years, and that is with very rosy projections. So you’ve got a company that reportedly couldn’t hit its target valuation, and then it says, “Actually, we’re going to delay for safety.
It’s for safety.” Sam Altman has openly said that he thinks AI is going to lead to the end of the world, but quote, “There’s going to be some great companies in the meantime.” There’s not a guy who seemed very concerned about safety. I don’t buy it. I don’t know if it’s because of the defensive moat or if their earnings aren’t stacking up, but I have to believe that this is all self-serving. That said, I do want to call out that Anthropic is still planning to go to market IPO. So I’m mostly talking about OpenAI with this theory and all of this. It just feels wrong. It does not feel like we’re getting the right information about what’s actually going on here. And there’s seven to 10 people who are just making massive decisions that are going to impact all of us. And that’s what just frustrates me about this and why I think whether it’s government or third parties or whatever, there needs to be some people not from the industry looking at what’s going on here and relaying that information to the public.
So all this still, we need to get to the question, is this a bubble? What is going to happen to the economy? We got to look at some of the numbers, so let’s get into that now. First, let’s just talk about spending. Combined in 2026, CapEx plans as capital expenditures, similar to real estate, you’re building out infrastructure, building data centers, you’re buying chips, that kind of stuff. The plans for Amazon, Alphabet, who owns Google, Microsoft and Meta are up 77% and combined are going to spend $725 billion. Just for reference, that is bigger than the economies of all but 25 countries in the world. It’s bigger than the economy of Argentina, of Singapore, of Austria, of Norway, of Thailand, of Columbia, Vietnam. You get the picture, that is a lot of money. And this CapEx, the money is not coming from earnings. Well, it depends.
You kind of have to split these hyperscalers. OpenAI and Anthropic, they’re funding this off debt. Companies like Amazon and Alphabet have a lot of cash flow that they can use to reinvest in Meta as well, but there are estimates that 93% of their operating cash flow is going into AI CapEx. So they are going all in on this. In fact, Alphabet, Google, basically posted its first cashflow negative number since it IPOed in 2004. So that’s the expense side. Let’s look at the revenue side. Revenue in AI is absolutely growing. Do not get me wrong, this is absolutely growing. For the public companies like Google, if you look at Google Cloud, that revenue is growing. If you look at AWS, that revenue is growing. From what we understand of what’s going on with OpenAI and Anthropic, their revenue is growing. But the problem is valuations in the stock market because when you look at how pricey these AI companies are becoming, you need to get basically what estimates say is that revenue by 2030, so in the next four years, AI revenue needs to reach $2 trillion.
That is so much money. Well, we’re comparing two trillion to countries that is bigger than all but 14 countries in the world, and it is bigger than all of tech combined right now. I just want to put that in perspective. That means AI in the next four years needs to get bigger than all technologies companies combined. So this is what I said at the beginning that the stock market is sort of pricing in perfection. If AI companies hit their numbers, then the stock market valuations are justified, but these are big, big guesses. We don’t know. We really don’t know if the revenue is going to take off. What happens if China starts dumping lightweight models, easy to use cheap models into the United States? What if new cheaper competitors come up in the United States and these companies don’t reach their revenue numbers? Well, that’s going to be pretty bad.
You can imagine this turning into a bubble because yeah, there is an optimistic case that AI does everything great. It says it was. I think that might have negative implications for the labor market, but if you’re just talking about stock prices, maybe. But I don’t know if that’s true. It is pretty hard to say right now that AI is having a super big impact on businesses. It’s definitely impacting the labor market in certain industries, but there have been a couple of surveys of executives. There is a survey of 6,000 executives in the United States, Great Britain, Germany, Australia, and what it found was that 89% of executives reported no productivity gains at their firms over the past three years. Now that’s backward looking and the technology has improved, so I bet that number would go up if they did it again. But if you ask them how they’ll expect productivity growth in the next three years, they said a 1.4%, not exactly a lot.
There was another survey in May of 2026 from Globalization Partners. They surveyed 2,850 executives, 73% called AI returns quote, “underwhelming.” So when you look at all these things together, to me, it just feels super risky. I can’t say for sure what’s going to happen, but to me, there are just a lot of risks piling up here.
And I’ll just give you my gut instinct. Again, I’m not claiming I know what’s going to happen, but my gut instinct is that AI is genuinely useful and we are in a bubble. I do believe that we are in a stock market bubble and it is going to pop at some point. I can’t tell you when. Is it going to be three months or three years from now? I don’t know. But I think both of those things can be true. AI can be a transformative technology and it can also lead to a bubble. I’m worried about the stock market valuations for all the reasons I mentioned from taking on a lot of debt, the lack of revenue, the incredible spending. I forgot to mention the fact that they’re building these data centers at enormous costs and the estimates of how long they’ll last are somewhere around five to six years.
So the idea that we’re going to spend less money on these things as we go forward, I don’t know. So for me, that feels like a lot of risk. Not to mention, by the way, I don’t know how much people like AI. I think there is a serious backlash growing. We see it manifesting in data center pushback across the US. You see certain industries really pushing back against this. You see this is AI adjacent, but you see huge pushback on things like flock cameras in the United States. And I don’t think people like the way AI is being implemented in the US without regulation, being told constantly that we’re going to lose our jobs and that that is inevitable, which drives me nuts by the way. I think it’s so self-serving that all these CEOs are like, “Oh, we’re just going to take all your jobs.” Clearly that benefits you, so you have an interest in having everyone believe that and accept it as inevitable.
But anyway, I digress. So for all those reasons, I think that it could be a bubble. And it also, there’s historical precedent for this. I think that I’ve spent a good amount of time looking into similar situation because history rhymes, right? And what we’ve seen throughout history time and time again is when a transformative technology come out and change the world, it often leads to speculative bubbles and collapses. This most notably happened not just in the US, but throughout the world with railways. We saw this in the 1840s in Britain where share prices of railway companies doubled in two years, and then shortly after they were worth less than half. The US had a similar bubble and burst in a railroad panic back in 1873. There was just massive new track being laid throughout the country. Then the financier of Northern Pacific went bankrupt. The New York Stock Exchange had to close for 10 days.
89 of the 364 railways failed. 18,000 businesses failed by 1875. Unemployment rate hit 14%, which is massive. And so clearly there was a bubble there as well. Most recently, we saw this in the dot-com bubble. We saw a lot of companies investing in telecom, laying fiber optic cable. There was $500 billion spent from 1996 to 2000, peaking at about $213 billion, which is like one to 2% of GDP just for scale. We are way bigger than that. We are three times that right now. And the telecom fiber thing collapsed. The whole thing was predicated on this idea that internet traffic doubled every hundred days. It actually doubled about once every year, and there was just a huge bust. The stock market lost more than 50% of its value, and it took a very long time, more than a decade for it to recover. So these ideas, the internet clearly transformative, railways, clearly transformative, important technologies.
But when people see these technologies, they often bet on them before the business model is fully developed. That’s exactly what we saw in the dot-com bubble. The technology is amazing. The business model wasn’t yet, and people were rushing to invest in it. They wanted a piece of the pie before these were real businesses. And I don’t know if that will happen for sure, but when you look at what’s going on here, it feels the same. We are investing so much money into data centers when we don’t really know that they can pay themselves off. We don’t know how long they’re going to last. It’s a whole nother topic, but we don’t even know if LLMs, the current iteration of AI, is the right way to pursue AI. There are just so many open questions right now, but people are putting money in and pricing for perfection.
And I don’t know, as an investor and real estate investor, stock investor, whatever, pricing for perfection is risky. And so I do think, if I had to guess right now, will we see a big fall in stock prices in the next one, two, three years? My guess is yes. I don’t think they’re going to hit perfection. I think it’s almost impossible. It’s not just necessarily them. I just think it’s crazy to think that you can come up with this new technology and master the art of bringing it to market and monetizing it and productizing it in a way that returns capital to investors that quickly. It just doesn’t make sense to me. Maybe I’m wrong, but I don’t know. After researching this a lot and thinking about this for weeks, this is just where I come out. AI can be used for incredible good. That is clear.
It is solving amazing math problems. It’s helping with healthcare and personalized medicine, discovering new antibiotics, forecasting severe weather. There’s amazing things that it can do, but there are also really bad things that it can do. That is very clear. Technology in itself isn’t really good or bad. It’s up for society to decide how it should be used. And right now, there’s no voice for society. Right now, it’s a bunch of people, CEOs who have frankly are just self-interested, like every business. I’m not saying they’re unique and that they’re self-interested, but when a technology is this important, you cannot let the people who stand to gain the most from letting it run rampant, let it run rampant. It is crazy. I was listening to a podcast the other day and someone was likening it, and I think this makes sense to nuclear weapons. Would you allow Robert Oppenheimer to take nuclear weapons, the technology that he developed and go just regulate it on his own?
Of course not. You have to have other people in the room looking at this and at least providing some level, one, of transparency to the public about what’s actually going on, and two, helping steer the industry in a way that it is beneficial to people and is not just going to consolidate knowledge and capital and money into the hands of a few people who may or may not be that concerned about what happens to the rest of society. Go listen to Sam Eltman’s buddy. That guy, I don’t think he cares about what happens to the rest of society. And I just think if there’s no regulation, the beginnings of this AI sort of backlash that we’re seeing, whether it’s opposition to the flock cameras or data centers, I think it’s going to boil over. And this is clearly possible. If you look at AI optimism in the United States, I forget the exact number, but it’s low.
People are not optimistic about AI. It’s like in the 20s or 30s or percent. If you look at the way China is doing it, that does regulate AI in pretty significant ways. They have 75% optimism about AI. They have made rules, particularly around how children can use AI, but also about replacing human workers with AI, about what can be developed, who has access to it. I’m not boosting China here, but I’m just showing that countries, you can do it, right? You can come up with a framework that regulates something and makes it safe. Look at airplanes. The fact that air travel is so safe is amazing. It’s a miracle. And yes, there is regulation in that industry. And no, I’m not saying that every industry needs to be regulated. It doesn’t necessarily need to be heavy-handed, but something needs to exist. If there are people at these companies, the CEOs themselves are saying, “This can cause extinction.” They cannot regulate themselves.
Someone else needs to do this because if it doesn’t, I just think it’s going to go bad. Whether it leads to some physical harm or just some economic catastrophe that will start probably in the stock market, but will lead to large levels of unemployment that can probably lead to significant trouble in the housing market because if we see unemployment rise rapidly, if we see just net worths fall from stock prices deflating, that is going to have negative impacts on housing, whether it’s from higher vacancy rates, people unable to pay rents, housing prices going down. These things cascade. We saw this in the housing market in 2008. It started with the housing market then, but it spread to the stock market. This would be the opposite. It would start with the stock market, but it could spread to the housing market. So I just think there’s a better way forward.
I think the call for international cooperation is strong. People are always saying, “If we don’t do this, China wins.” But what if we just created international regulations? We haven’t even tried. So people are like, “Oh, that’s not possible.” Well, we could try and there’s obviously clear historical precedent for that kind of stuff working. Nuclear non-proliferation for large part has worked. We got together on things like fixing the ozone layer. That worked. This is a technology that needs international cooperation, and the United States has an opportunity to lead and to help create the frameworks that allow for economic and societal benefit without creating unnecessary risk for our country and for our economy. So that is my rant about AI. I’ve been thinking about this and researching it a lot. And this is, to me, the big thing hanging over the economy right now. It’s hard to get excited about investing in the stock market or even buying rental properties or being in the housing market when there’s just this impending worry about will the stock market crash or will AI disrupt all of our jobs?
And the fact that we don’t have information about what’s really going on in these companies makes that worse. It makes it harder to trust what’s going on in the economy, and it’s so much of the economy. And so I hope you’ll indulge my divergence from our usual discussion of real estate here, because I do just think this is a major economic question. And I’ve shared my opinions with you clearly and very loudly, but I’d love to hear yours. I am not an expert on AI, but I do think I have a good grasp on how this could cascade throughout the economy. But I would love to hear what you think. Are you bullish on AI? Are you bullish on the technology and the economic opportunity right now? Or are you sort of like me that you think the technology is great, but also comes with massive risk that need to be mitigated?
Or maybe you hate it all and you just think it’s all terrible. Let me know in the comments below. Thank you all so much for listening to this episode of On the Market. I’m Dave Meyer, and I’ll see you next time.