The Dividend Cafe - Keeping an AI Bubble from Hurting You

Episode Date: November 14, 2025

Today's Post - https://bahnsen.co/3LGaAg6 AI Investing Bubble: Unpacking the Risks and Realities In this week's episode of Dividend Cafe, host David Bahnsen discusses the potential investment bubble surrounding artificial intelligence (AI). Bahnsen explores the phenomenon that started with the rise of ChatGPT in late 2022 and analyzes its implications for investors. He cautions against the euphoria-driven market, emphasizing the risks of high volatility and speculative investments in AI companies like Nvidia and Oracle. Bahnsen underscores the importance of adhering to sound investment principles and avoiding the pitfalls of chasing short-term gains in the overheated AI sector. 00:00 Introduction to This Week's Topic 01:06 The AI Bubble: A Growing Concern 06:37 Mathematical Fundamentals and Market Volatility 13:34 Historical Context and Future Implications 28:02 Conclusion and Investment Principles Links mentioned in this episode: DividendCafe.com TheBahnsenGroup.com

Welcome to the Dividend Cafe, weekly market commentary focused on dividends in your portfolio and dividends in your understanding of economic life. Hello and welcome to this week's Dividend Cafe. I am your host, David Bonson. I am very excited to be talking to you this week about a story that no one has been talking about all year, and that is AI, this thing that no one has mentioned in the investing world for quite some time. but it's sort of a big secret, and I'm going to crack the code on it here today. Of course, I say this in jest because it has been one of the biggest topics discussed,

not only across the culture, but particularly in investment markets this year and last year, and really since late 2022, when this little thing called ChatGPT sort of hit the airwaves, if you will, and we have been living in a time that I think, is profoundly important for investors. Now, if I were doing a Dividing Cafe today where I just basically wanted to unpack some of the noise, I've first of all done that before. It also is, I think, a little bit less exciting of a topic. There is a lot of noise around all this, but that's not my point today. I want to make the case to you all that this AI story is a bubble, that there is a bubble going on that is going to do great damage to certain

investors and that at the same time, I do not mean by that what I think a lot of people might mean. I'm not suggesting a systemic market collapse. I am not suggesting a particular time or season to exit. I am not suggesting that AI will prove to be a dud of a technological innovation. I believe it is entirely consistent and, in fact, more cogent to understand the legitimacy of artificial intelligence, the opportunity that exists in AI as it unfolds in a more monetizable way to real-life businesses in the future than is available now. And at the same time, understand that when you get into a moment where euphoria is driving investor return, and expectations that there becomes a moment in which it is very difficult to make money in that

as an investor. Now, there are always fads. There are always noisy things. And I have talked in the past about shiny objects where people get a little theme or narrative and they go pile into it and eventually get their face ripped on. And if we were just talking on another one of those things, then I probably wouldn't devote as much dividend cafe time and space as I have this year and we'll continue to. We did a dividend cafe about a month ago now where I tried to unpack why I found it utterly insane. I mean, intellectually, but also morally insane. For an investor, let alone an investment professional like myself and my colleagues at the Bonson Group, to change our belief system based on a particular moment in time to say, well, we actually want to put a pause on what our principles

are so that we can try to extract a certain opportunity for a season and then we'll jump out before things get a little hairy. I did not decide to write on this subject this week as NVIDIA has now gone down over 10% from its high just a couple weeks ago and other AI darlings like Palantir down over 15%. I wanted to write about this more. extensively a month ago. When I first wrote in defense of our principles, and then the next week wrote a sort of contrarian position about what AI will end up doing to jobs in the society. I had more of a cultural bend to it. But right now, I want to write about math. I want to write for investors in a very practical way. Because if this were just noise and

some people are going to get a little carried away, it doesn't matter very much. much to a lot of people and it wouldn't matter very much at all to our clients who are of course the primary audience in Dividing Cafe. But the reason why I believe this is more than noise is I think that we're now talking about a legitimate bubble where there is significant dollars involved that if almost every bearish thing, every negative thing, every concerning thing does not happen, it may still be very, very difficult for investors to make money because of the math of the gyrations, the high volatility of how a lot of these things are going. I'm going to make that case for you here today in the Dividing Cafe. So let me just basically restate the conclusion

and those you want to tune out now can tune out as long as you rate us five stars and give us a thumbs up and for a positive review or whatever it is that you might want to do. But, no, hopefully you'll stay to the end here. But I will not bury the lead. I've stated, I believe AI is in a bubble, but the risk is not to me in what AI does. It's not when it does it. But rather that there is a moment in our investing culture taking place that I've seen several times in my adult professional life, that a sensibility is replaced with euphoria and the decision making that results.

is never ever good. The stakes are higher because of the dollars. And I think that what a lot of people believe is, well, if I avoid owning the pets.coms of this moment, you know, that was a situation people bought terrible companies. They went down 100%. And there were people who lost all that money and they moved on. And I don't think there are very many people that put 100% of their portfolio in pets.com. Maybe some did. They probably didn't have a very big portfolio if they did. But they touch the stove and those people are not going to be touching the stove again, Lord willing. However, I don't think that's what we're talking about. The low quality nature of some of the companies at the heart of the dot com implosion and bubble

make this story worse. It is not better because they are higher quality businesses involved right now. It is giving a full sense of security that I think is covering up for some of the real mathematical fundamentals at play. Let me just start with what I mean by some of the mathematical fundamentals, and it doesn't require you to have an outlook on AI. It doesn't require you to have a particular opinion or point of view on a given company. It just requires you to understand how math works in terms of percentages. NVIDIA was $212 on October 19th, and as of today, it was sitting somewhere in the 175 range. It's lost $500 billion of market cap, but as a percentage, that's only about 10%.

And it can add $500 billion in a minute, it can add a trillion, it can take away $2 trillion, It releases earnings results next week, and so you can expect there could be big volatility up or down around whatever happens. But that's the thing about the absurdity of when companies are 50 to 60 times earnings, and we're talking about trillions of dollars as if it's normal. The fact of the matter is that a company losing $500 billion of market cap in 15 days is not normal, nor is a company adding $2 trillion. in market cap in 10 months,

which is what NVIDIA had done before the last 10 days. That came after it added $1.5 trillion of market cap in the year before. These are absolutely unprecedented moves of valuation in absolutely unprecedented periods of time. Now, here's the thing I find so interesting. One year ago, NVIDIA was at $150, $150 a share.

It ended up getting down to about a hundred between the deep seek fears with China and Q1 and then the Liberation Day fears in the Trump trade war back in early April. And then, of course, it has since rallied way, way higher. And if you bought right at that bottom where you sit now, you're sitting very pretty. If you bought a year ago, you have a return. You're up. It's about in line with S&P. The fact of the matter is that there are points at which one can buy it,

points of which one can sell it that make all the difference in the world. If somebody bought two weeks ago, they're going to need a 25% return. Let's say if Nvidia gets to 170. They're going to need a 25% return just to get back to even. Now, maybe that'll happen easily. Maybe it won't happen for a long, long time. I have no idea. But that's just the math of it.

And that's with a pretty easy case. The problem is that when one has a high, basis. They are not sitting with house money. Someone who bought it years and years ago at 70 is not really likely to care about a 220 to 170 move. Someone who bought it at 215 does care. And I think that you have a very, very different group of investors, different types of hands holding it that is going to create a lot more volatility around the name and ultimately undermine the thesis that the popularity is the backstop. It has to constantly go up because there's always another buyer right there.

Now, let's use Oracle as an example, which has hardly been an AI darling, but then this year became one. And they announced a big deal with Open AI giving them an order, an Oracle giving Nvidia an order, and Nvidia giving Open AI equity capital. And Oracle went to $345 two months ago. And today it was sitting at $215, down $305. dollars down 38% and after being down 38% still trading a 50 times earnings so now it has to go up 59% just to get back to even for people who bought at that number now maybe it'll do that but the big announcement came and now here we are two months later and you need 59% to get back

to even then when investors face that kind of math that often historically causes them to throw in the towel. It makes people give up. It causes them to abandon their own investment strategy. So psychology and sentiment become extremely important and much more important than common sense and fundamental logic. Look, I really do hope that the story works out well for Oracle. I have nothing against Open AI.

There's a circularity in all of them. it that I expressed a month ago was concerning and is concerning throughout a lot of the AI CapEx story. Well, I'm just right now sharing not what I want to happen or don't want to happen and not what I'm predicting what happened. I'm just explaining the math that someone right now owns a big, big, big successful stock and they need it to go up 59% before they're back to even. And I find that to be a daunting situation.

Now, that is actually one of the better and more benign stories between you talking about Nvidia and Oracle, consider a company like Corweave, which is a darling recently gone public, a darling of AI cloud and platform infrastructure. The stock was $187 a share five months ago. It is currently $77. And right now at 77, it is trading at negative 38 times earnings. So I will let you do the math on what the P.E. ratio was back in June. Now the stock has to go up 142% to be back to even.

Do you see the point I'm making? This is not a bad company. This is a good company. Now it's cash flow negative and earnings negative. There's a path to something being different there. The things are getting better and better for the company, and yet the negative returns have gotten worse and worse. If the stock goes up 20% per year for five years, which would be double the S&P 500's historical return, then the investor will be back to even from where they were in June of 2025.

So this has nothing to do with the company is being mentioned. I'm not expressing any bullish or bearish opinion about Corweave Oracle. and video, let alone a private company like Open AI. What I'm trying to get at is that these things can, for some investors, become economically existential, and they are always mathematically daunting. And yet this is, I think, where the story lands because of the high volatility and the need for entry prices and to have an exit price that all pan out. What is it that I'm worried about? Well, first of all, let's just understand the premise behind the whole A story. Number one, is that the hyper-scalers are spending so much money on their AI capability build-out

that regardless of how it works out for them, there is just massive profits coming to the people they're spending the money with. Fair enough. The capital expenditures of some companies are the income for another. However, the problem, of course, with that is it presupposes that the money will continue going, even if things go poorly for the spender, to the company receiving the money. They most certainly will not.

And then the companies that slow down in their spending or cease their spending have the leverage over the company who's high valuation, which is predicated on ongoing massive revenue growth, falls apart. So you do not end up with there's a winner and a loser. You end up with two losers in that situation. But the second premise as to why people are in the AI trade, besides their willingness to just buy people building out AI and not worry about the people spending the money on the build-out.

The second premise is that the people investing in AI are going to make so much money once they convert this newfound computing power they're investing into in real-life application. And they don't know how they're going to do it, and they don't know when, but it's worth going along for the ride because there's these huge new profits coming from AI application utility.

Now, if you find me sharing these two valuation paths, as something alluring to you or tempting, then you are not hearing me correctly. There is unbelievable risk, an unbelievable speculation involved in both of these stories. The build-out part of AI is where everyone's focus is right now. We've talked about it before. Many analysts call this the picks and shovels phase.

AI is at ground zero of AI infrastructure. They make the chips that make it go and so forth and so on. But at some point, that does have to lead to usage and application. And I think most people understand that, unless you are one of those few that are timing an exit. Now, for this to not need an exit, but to go into a transformative moment, we have to look at history. Tell me something that has ever gone from a massive craze of buildout to a utility and application phase without a huge, fatal purging of dead weight from that excess. along the way. You can look historically at railroads, at telecom, at my own beloved fracking and shale story of 10 to 15 years ago, radio, fiber optics, media. These things become legitimate

stories that have transformative impact on society, that along the way, the buildout phase goes to excess phase, and a lot of people get their faces ripped off. And it happens over and over again until there is big liquidation. Now, many of these things I just described are what my friend Louis Gogh refers to as productive bubbles. There is a productive asset that contributes to something productive in society that went into a bubble valuation. And when all said and done, the owners of it change. Someone loses money. Someone buys it cheaper. There's a re-rating or repricing. And then there's an ownership allocation shift, but there's a productive asset still standing. It's just that it used to be at X and now it's worth half of X or 20% of X, and it used to be

held by Smith and now it's held by Jones. Fine. However, that's different than what Louis refers to as a scarcity bubble, something that doesn't have a lot of productive use in society, that won't have a lot of residual benefit, that its bubble was based on the idea that it had such a scarcity value that it would keep going higher and higher. And then when all said and done, it bubbles, its bubble burst. And lo and behold, it doesn't have the scarcity value that people believed. And it ends up having to fade away in value for a significant period of time. Louis uses Japanese real estate, 1970s gold, even goes back a few hundred years at Tulip Mania. As his examples, I'm perfectly willing to throw out there the possibility.

that maybe crypto will belong in that space, but plenty of crypto lovers are going to disagree, so that that's all fine. My point being that when a scarcity bubble pops, you're not left with the productive asset. It didn't just merely switch hands. The value story itself went away. I think that when we talk about AI as a bubble, that I am talking about a productivity bubble. I concur that there's a technology there will be useful. And I think if it does trade much lower in value later, and there's totally different names who own it, different people, because some investors lose their shirt and get washed out, that's still very different than having an unproductive asset, be in a bubble. The other kind of catch to this, though, is the difference

between how bubbles, whether they're productivity-driven or scarcity-driven, are funded. When something's funded by equity, people can lose a lot of money. But the people who lose a lot of money are the people who invested a lot of money, what we refer to as risk takers. And it can be very bad for them, but not bad for all of us. The analogy I use is it something you can go wrong in the pool, but it hurts the people in the pool,

and that's different than a systemic risk where it hurts the people who are outside of the pool and even people that didn't go to the pool party at all. And I think to get to a place where something bad happening in the pool hurts everybody that was even just out in their car, driving around is generally when you have a debt-fuelled bubble. And we have seen those before, those that were totally unrelated, disconnected in their own practice to the financial crisis

of 2008 were still substantially impacted by the impact to the economy, to housing, to stocks, to bonds, to credit markets from the contagious nature of the 2008 financial crisis. That's a classic case of a scarcity bubble condos in Arizona that were attached to debt and leverage as a funding mechanism. There's a lot about this AI story. I want to be and believe to be different. First of all, the productivity versus scarcity nature, but second of all, the equity versus debt funding. However, this is one of the things I need to say as I get ready to wrap up. I don't know that that's going to hold.

I don't know that this will stay in equity-funded situation. oracle's certainly borrowing a lot more money than some of the other players that we've talked about. You know, you do have these hyperscalers that print money in their core business. They are free cash flow generating machines like the world's never seen when you look at meta, Google, Microsoft, Apple. These are just insane cash flow generating companies, have the ability to lose a lot of money and equity without it going into something debt. They have unlevered balance sheets.

But there's a lot of questions about what the capital structure. of AI things are going to be. There's a lot of questions about counterparties. There's a lot of questions about triparty situations. I've talked about various economic transactions taking on a circular nature. But there's also a question of how any of it's going to happen at all without a buildout of power that will be debt funded, that will be debt fueled. And so I don't think anyone knows enough to know where some of those other adjacent risks are going to come. What I I would say to leave you a couple quick concerns as I finish my case for this being a bubble and then want to conclude finally with the practical reason that we're handling it the way we're

handling it. Number one, I believe that there is an awful lot of this whole story tied around a private company that no one knows anything about that is at the hub of a lot of issues as a buyer, as a seller, as an acquirer, as an acquiree, as a funder. This open AI company that makes chat GPT, I have never in my life seen a private company with no real balance sheet per se, be able to track absolutely infinite amounts to speculative venture capital, equity capital, and be a part of a counterparty two transactions of this magnitude of trillions of dollars. It's absolutely unbelievable. Maybe they just end up owning the universe. But you know, a lot of people seem to be missing the fact that the amount

of CAP-X going into AI right now, you cannot have everybody win. The amount of money spent would require everyone to win, but they can't all win. One of these players has to beat out another player. You're going to have failures. You have to. And I think that that represents a huge unknown, unforeseen, undiscounted risk that there will, in a maybe best case scenario, and an optimistic scenario, be some companies in the AI story that win, but that comes with others losing. And unless an investor is picking the winners and avoiding the losers, which represents a level of risk, I very much doubt they are aware of, I think there's going to be significantly

more carnage than people realize. I would like to add, I do not believe that we understand the risk of China in this. Now, perhaps the U.S., which is certainly winning in the AI story, so far, stays in the lead forever and ever on Met. I do not think China has better tools and applications that we have in the U.S. I do believe they're trying. And I do believe that certain recent history examples of people laughing at the idea of U.S. losing supremacy to a global competitor might want to go look at Intel's success in chips and semiconductors versus Taiwan's semiconductor and decide who's laughing now. Is it totally out of the question?

that there is some degree of risk from global competitors. NVIDIA does 40% of their business in recent quarters with countries like Malaysia, Taiwan, in Singapore. Is it really out of the question that, I mean, they're obviously selling their power to global competitors. I simply don't accept that there's no risk that American presumed dominance is vulnerable. I do not say that as the base case, but I do say it as a risk that needs to be factored in

when you start talking about these valuations and this level of euphoria, and I don't think it's being factored in. I also would add that I think an awful lot of people, both professionals and novices, involved in this story right now that are most bullish and euphoric on AI. They were not alive as professional investors or grown-up investors in 1997, 1998, 1999, to the 2000 bust, that there isn't the same degree of memory, of familiarity, of understanding, of risk appreciation, of humility that might be appropriate given these certain circumstances. Now, I am not suggesting that this entire situation is totally analogous to Pets.com, because as I said before, I do not think it is.

I think the tech companies of today are higher quality businesses than many of those tech companies of the year 2000. But that doesn't change the fact that you need revenue growth rates, not just revenues, but the growth rates of the revenues to continue accelerating to rationalize this whole deal. And I simply do not believe they will. I believe that the revenues will continue to grow, but the growth rates will decelerate at a time that CAPEX as a percentage of revenue is flying higher. That CAP-X is a percentage of revenue will be going higher

while revenue growth itself is declining. And I do not believe that that risk in fundamental economic clarity is appreciated by enough investors. So there is a, not everyone can win dynamic. There is China global competitiveness dynamic. There is the investor base behind the story

represents, I think, a vulnerability behind it, and there is just the basic reality and inevitability of revenue growth deceleration, all attached to investment thesis that are themselves vulnerable, which is that the picks and shovels are enough, or that the utility is all going to come out fine, even if I don't really know how. You've got vulnerabilities sitting on top of vulnerabilities, trading at 50 times, 60 times earnings. You can go a lot higher, my friends. It can go lower. Maybe that level two weeks ago at Palantir and NVIDIA at all-time highs, maybe that will prove to be a generational top. And the weakness we've seen in this story the last week or two is just the

beginning. Maybe it's just a little dip and it's about to soar higher yet again. I don't know. But I do know that when the rest of the world buys into euphoria, the notion of being covetous of their short-term results generally is a very bad idea. that what you ought to be covetous of is the very, very thoughtful, disciplined, cogent portfolio that you have. That is definitely subject to volatility to up and down movements, to successes and failures, but not to something existential where you need a 142% return to get back to even. I don't want to cherry pick these examples.

There's plenty of better stories in there. And you could also choose to isolate stories on the bullish side of when someone got in at one of those dips and got out at one of the booms higher. Those things happened too. But I don't think they happen repeatedly. I don't think they happen intentionally. I don't think they happen as part of a solid philosophical investment plan. What I want to do at the Bonson Group is continue to advocate for investment principles that have stood the test of time that includes the time of dot com, the time of financial crisis, the time of great. world instability, the time of great peace and economic growth and expansion, times that have covered

decades of all kinds of diverse events, and yes, the time we're living in now, where more and more people seem to be losing their mind about one particular narrative that is not entirely defensible, that we'll end up having a lot of utility in the future, but along the way is exposing certain people to risks and existential pain that I am not interested in incurring on behalf of my clients. So to that end, we work. Thank you for listening. Thank you for watching. And thank you for reading the Dividing Cafe. Look forward to being with you again. The Bonson Group is a group of investment professionals registered with Hightower Securities LLC, member FINRA and SIPC, and with Hightower Advisors, LLC, a registered investment advisor with

the SEC. Securities are offered through Hightower Securities LLC. Advisory services are offered through Hightower Advisors LLC. This is not an offer to buy or sell securities. No investor process is free risk, there is no guarantee that the investment process or investment opportunities referenced Tyrion will be profitable. Past performance is not indicative of current or future performance and is not a guarantee. The investment opportunities referenced Tyrion may not be suitable for all investors. All data and information referenced herein are from sources believed to be reliable. Any opinions, news, research, analyses, prices, or other information contained in this research is provided as general market commentary and does not

constitute investment advice. The Bonson Group in Hightower shall not in any way be liable for claims and make no expressed or implied representations or warranties as to the accuracy or completeness of the data and other information, or for statements or errors contained in or omissions from the obtained data and information referenced here in. The data and information are provided as of the date reference. Such data and information are subject to change without notice. This document was created for informational purposes only that opinions expressed are solely those of the Bonson Group and do not represent. those of High Tower Advisors LLC or any of its affiliates.

High Tower advisors do not provide tax or legal advice. This material was not intended or written to be used or presented to any entity as tax advice or tax information. Tax laws vary based on the client's individual circumstances and can change at any time without notice. Clients are urged to consult their tax or legal advisor for any related questions.