Listen to the full episode, here.
John Norris (00:30):
Well, hello again everybody. This is John Norris of Trading Perspectives. As always, we have our good friend, Sam Clement. Sam, say hello.
Sam Clement (00:35):
John, how are you doing?
John Norris (00:36):
Sam, I’m doing fantastically and I hope you are as well.
Sam Clement (00:39):
I cannot complain.
John Norris (00:40):
I’m glad. And the thing is people really can’t complain. At least people who are owners of AI-related stocks because Sam, as you know, it seems like over the last couple years, that has been the story driving the U.S. markets is just what these AI-related companies have been able to do. You’ve seen their stock prices just go through the roof and you just got to wonder how much longer it can last.
Sam Clement (01:02):
I mean, it really has been almost the only topic it seems like for some time and just the massive size of so many of these companies and the flow through or downstream effects, whatever word you want to use, has just engulfed the conversation around markets.
John Norris (01:19):
And so I mean the rally in some of these names, I’m not going to name them, you already know what the names are. The rally and some of these names are so great that of course people are going to start trying to compare it to the tech bubble that was in the late 1990s that came crashing down in 2000 to 2002. And while there are some similarities, Sam, I think there are also some differences too. The similarities I see is just really a new technology, if you will. AI, of course, probably has been around for a while, but now just spending so much money on the infrastructure in order to make it happen. Same thing with internet. The internet didn’t just come about in 1999, 2000, but back then, and I’m old enough to remember it, a little bit older than you are by a couple years, but back then really in the late 1990s, all you really needed to have was a business model or business plan that said something about the internet, you have a line of credit from the bank and you’re almost an instant millionaire.
People didn’t understand really truthfully what the internet was or how it was going to change our lives and how we conduct business, but we knew that it was going to. And so people just started throwing money at it. And so that is what happened while back then. And now it seems like people are throwing a ton of money at AI. I would say the difference between what was happening back in the 1990s and what’s happening now is the companies which have really seemed to benefit the ones that make all the headlines, the ones that you’ve heard about in the large cap tech sector, these are real companies with real earnings and they’re real market leaders. Whereas a lot of people that got burned back in the late 1990s were those dot-com bubble stocks, stocks that had no earnings, no revenue and were valued in the billions of dollars and people got burned.
So right now I’d say one of the major differences is a lot of these companies that have been flying so high have really been generating revenue and earnings.
Sam Clement (03:23):
Not only are they generating earnings, but I think what’s so unique about it or unique about it when we’re comparing the two rather is just the overall market leaders being… I mean, it’s not only those market leaders, but there’s such a trend amongst almost all the market leading in terms of market capitalization all taking part in this as well.
John Norris (03:43):
Yeah. And so I mean, taking a look at it, some of the names in that large cap name sector and some of the biggest companies out there in the world, Nvidia and some of the other ones, Sam, I would say I think they’re still going to be around another three to five years. I don’t think they’re going to vanish. Whereas some of those tech names back in the late 1990s, I know that they vanished. And then also something that we don’t talk about when we talk about the tech bubble back then was there was almost a known expiration date because from 1996, 97, all the way through the end of 1999, everyone was gearing up for something called Y2K and Y2K was going to be, I mean, my Lord, we were worried about it all because some computer programmers back in the late 1970s or early 1970s tried to save some space on a binary code and at the end of 1999, everything was going to go back to 1900 as opposed to go to 2000 and it’s going to throw the entire world into a dither all that stuff.
And so businesses were spending a ton of money on technology in late 1990s to get ready for this, what turned out to be non-event, but the Y2K and then not surprisingly in 2000, 2001, businesses didn’t spend as much money on technology because they had spent so much leading up to Y2K. Now, Sam, there’s not really an expiration date here. I mean, AI is out there. It’s not like, okay, we got to get all of our AI investment in by the end of September of 2026, or we have to get it in by the end of 2027. So I think the lead time, if you will, the runway, whatever you want to call it, is just more infinite, if you will, with AI than it was with spending and even internet spending heading up to end of 1999.
Sam Clement (05:32):
Which I wonder if it almost complicates it a little more that there isn’t end day, right? Because the race is almost against your peers and so-
John Norris (05:41):
There’s a race with no finish line.
Sam Clement (05:43):
Right. That picture lining up for a race with your competitors, these people that are, I mean, you’ve been competing against forever and we say, “Hey, you’re all running a race. We’re not telling you what the end is. ” And that’s going to create some probable confusion along the way. And so I wonder if that’s part of it where the pressure is internal. I guess it’s external as well, but it’s not this, “Hey, we need to get it done by this date.” So I think it can create some more confusion rather than clarity.
John Norris (06:16):
Yeah. Well, I would tell you there is definitely still some confusion out there. So what the end game is or what the end date is and when we finally get there and go past the end date or the end game, we’ll look back with 2020 vision and go, “Aha, there it was. ” Just like we did with Y2K going, “Aha, there it was. Of course we were going to spend less money on technology, but we didn’t really understand it at the time. We thought we were going to continue to spend money just forever.” And now you have U.S. businesses spending increased amount of money on AI and all the technology software, even hardware that’s associated with it, building out all these data centers and at some point all these businesses are going to need to see some form of return on their investment.
All this money’s been spent, they’re going to have to see some new products, new services, or at least some increased efficiencies in order to justify how much money the US economy is spending on AI. And Sam, when do you think this will be? I mean, when are corporate analysts or investment analysts, financial analysts going to say, “Okay guys, you spent the 10 billion, whatever it was last year, we’re going to have to start seeing an increase in EPS.” When does that happen?
Sam Clement (07:34):
Again, I think the difficulty is that there’s not this defined end date in time, right? And so the efficiencies, to me, that’s the biggest one that is the flow through effect that makes the earnings spigot turn on is what allows you to be more efficient and hire less people going forward and to allow margins to maintain or increase and especially given the environment, the macro environment over the last few years, really going back to COVID and the boom in hiring and then the don’t hire, don’t fire environment where you’re still probably a lot of companies are overstaffed, but not wanting to fire. We really need to start seeing those AI efficiencies come into play, especially given the conversation that’s intensified over the last few weeks about cost related to AI for the end users and companies, large companies, publicly traded companies getting through their AI budget for 2026 by the month of April.
So that’s where you have to start seeing that if the costs are going to be as outrageous as they
John Norris (08:37):
Are. Are you making an argument that maybe corporate America at some point moving forward are going to use layoffs and job cuts they’re going to say that those are AI-related even if they’re not in order to make it look like they’re getting something.
Sam Clement (08:50):
Well, it’s funny that I think you’re already having some layoffs due to AI, but it’s not that the AI is making it more efficient. It’s that the AI is so expensive that they need to lay people off. You think that’s what it is? And I’m 100% confident that that’s when you’re seeing AI layoffs right now, again, there’s- We’ve
John Norris (09:07):
Spent so much money
Sam Clement (09:08):
On AI. There’s so much money. We’ve got to cut costs. And again, we’ve talked about this ad nauseum with different topics, but the easiest way to cut costs is, you hate to say it,
John Norris (09:17):
Tech count. So going back to what you had touched on was we perhaps overhired at some point in time that AI is indeed creating efficiencies by getting rid of deadwood and it’s because not necessarily the technology itself has created efficiencies, it’s just that the expenditure around money is such that the CFO’s office is that we
Sam Clement (09:41):
Got to
John Norris (09:41):
Cut some costs.
Sam Clement (09:42):
And I say that to say it feels slightly anecdotally like cost is starting to be a little bit important to the end consumer of it. And for at least a year now, we’ve seen these headlines of companies and CEOs and what have you, pushing use of AI. They want AI throughout, they want these downstream efficiencies and then now they’re getting the bill for it and saying, “Okay, maybe this cost is not where it’s worth it yet.” And so that to me is that first step into really needing to see those return on investments is when people start caring about cost, right? If you don’t care about the cost, you’re probably not cared about the return on it. And so you have to start pushing back on costs a little bit before you start really demanding that return on investment.
John Norris (10:29):
So are you maybe implying that corporate America is spending too much money on AI maybe too rapidly and just how much longer can this continue to go on?
Sam Clement (10:40):
Yeah. I mean, again, whether AI is a bubble or not, I mean the first step of this whole bubble cycle is this fervor and zeal around something and not caring about the cost of it yet.
John Norris (10:55):
Tulips?
Sam Clement (10:56):
Whatever you want to say, railroads. I mean, what have you, UK railroads, I mean there’s a point in the cycle early on where costs don’t matter. Yes,
John Norris (11:06):
That’s right.
Sam Clement (11:07):
And so maybe we’re starting to get to a point anecdotally seeing companies start to cut back on costs, removing the companies had AI leaderboards to show who’s using it the most and they’re removing that because people were artificially using it. And so maybe we’re starting to get out of that first step of, hey, all AI is good AI, any use is good and now starting to reevaluate what’s the cost of this and what’s the return on this?
John Norris (11:33):
Well, how do you think businesses are going to be able to recoup their costs on AI investments when it seems as though there’s certain companies out there that are giving away their AI assistance, whatever you want to call them for free? Or is it just completely, am I just comparing apples to oranges? Because right now I’m sitting there taking a look at it going, “Hey, you’re going to spend a hundred billion trillion dollars and think you’re going to get it back up somehow.” How if Joe Kuh Public or what have you thinks of this as really a free technology,
Sam Clement (12:03):
Like a Google search replacement?
John Norris (12:05):
Yeah.
Sam Clement (12:06):
Well, the problem is, and this isn’t unique to AI, if anything is free, then you’re probably the product, right?
John Norris (12:14):
You make me feel like a sap with that.
Sam Clement (12:17):
But it’s almost missing the mark these people. We’ve talked about it when they view AI as a search replacement, right? That is not the intent of what any of these companies are building. And so I think that’s where we’ve had this issue when you talk about AI with different groups of people is you’re almost talking about two different things, two entirely different things.
John Norris (12:39):
So the only thing that, I mean, I guess AI can get is just information about me, like what social media does. It’s very pernicious or very tricky what social media does and keeping names out of this podcast, you sign up for something, “Hey, what do you like? Hey man, why don’t you tell us what you like, what music do you like? What restaurants do you like? Do you have any recommendations?” All that type of thing and it’s just you give away this information for free stuff you might not even tell some of your friends about your likes, but you’re just going to give your social media, “Hey, this, that needs so they can market to you. ” So it’s the information. You are the product. I was the product. And I guess with AI, if I’m using enough, whichever one you want to use, Claude, Copilot, ChatGPT or whatever, whatever the free versions of, you’re inputting questions, you’re inputting some data, whether you realize it about yourself or not, you’re putting stuff in there so they know a little bit about you.
And Sam, do you think they’re going to hold onto this information and just keep it private or do you think they’re going to sell it and try to make some money from
Sam Clement (13:41):
It? Well, the goal is to make these models more efficient, right? And they all have different levels of the models and based on the intricacies of what you’re trying to get done, use different ones and they use less or more power essentially and they’re trying to make all these models more efficient and have to use less compute essentially. And so the more that goes into it, the more efficient they get and the less compute they have to use theoretically.
John Norris (14:05):
And so if we’re getting to that point in the cycle where businesses are starting to think about costs and wanting to start seeing some improvement in efficiencies, all that stuff, how much longer can we really go on with it just throwing money at it? And some of the maybe it seems like circular financing and sweetheart deals and it seems to be maybe some of the companies are just swapping money back and forth. How much longer can this go on before the gigs up?
Sam Clement (14:36):
I think the short answer is quite a while. I mean, we’ve talked about it and again, this applies to a lot of things or not just AI, the irrational exuberance comment that is so famous from the dot-com. Shoot,
John Norris (14:48):
That was like 30 years ago.
Sam Clement (14:50):
But that happened years before the market turned over.
John Norris (14:53):
Yeah, it was like 96
Sam Clement (14:54):
When that was 30 years ago. And the biggest run up was after that. And so this can continue and again, we’ve talked about the differences that these companies are extremely profitable, not all of them across the board, but the largest companies are the leaders of it and are extremely profitable. And so if that can happen in a 1996, 97, 98, 99, you compare that to a time where companies are profitable and this is accretive to earnings every year, anything can continue for a while.
John Norris (15:28):
Well, do you envision maybe at some point, I guess to be, okay, well maybe there’s not a collapse in technology expenditures like there were in 2000 after Y2K, but corporate America just starts to spend increasingly less on a percentage basis. I’ve used this example previously in investment committee and presentations and talking with clients and prospects. Imagine Company A spends $10 billion on AI this year. It’s a lot of money, right? Next year they spend 20 billion. That’s an increase of 10 billion, right? Year after that, they spend 30 billion. That’s more money. It’s an increase of $10 billion, right? Year after that, they spend 40 billion, year after that, they spend 50 billion. You get where I’m going. Each year they’re spending an increased amount on AI technology, additional $10 billion a year. However, that growth rate by definition starts to go down from 100% to 50% to 33%, 25%, I think 20% and so on and so forth.
Well that at some point in the future, I mean, don’t know when it’s going to be, but I think it’s out there that investors going to go, “Okay, the growth rates, hey man, we’re still spending a ton of money on this stuff, but the growth rates have slowed down. I’m going to go on and I’m going to take my marbles and go to try to find the next great growth story.” Although we’re spending money on this, it’s going to be here, it’s here to stay, it’s going to fundamentally change our lives on how we conduct business, but the growth rates are now in the mid teens to 10%, I’m moving on to something else. At some point that almost has to happen, just do… I’m going to not use this correctly. Large numbers out there that it’ll have to happen. It doesn’t look like 2026 as if, but is it 27?
Is it 28?
Sam Clement (17:14):
Also the difficulty with it is it begs the question, what’s priced in, right? What’s the real expectation? Because if it does go from 10 to 20 to 30, the example you used, well, if that’s better than what the market’s expecting, things are pretty good for a while and can continue even with that slowing growth rate. You can read pro formas, you can read whatever you want, commentary from CEOs, but the market’s expectation is going to be different and it’s almost impossible to know. So that’s the first question is how does it compare to the market’s real expectations, right? Because we’ve seen it with some of these large cap companies and cloud services, which has been a trend for years before AI and those margins have come down, but it’s a matter of how are they coming down versus the expectation?
John Norris (18:01):
Well, I mean, lost in all this and I think lost in maybe some of the irrational exuberance or exuberance, whatever you want to call it, is a lot of this assumes there’s an infinite amount of money or an infinite amount of capital out there and we’re seeing a lot of IPOs getting ready to come out. Is there a finite amount of capital for AI?
Sam Clement (18:25):
Well, it begs the question of where when there’s an IPO, where does the money come from, right? Money always has to come from somewhere. At the end of the day, supply and demand matters for all of these IPOs. And you’re right, we’re talking about several, not trillions of dollars of actual securities coming online, but trillions of dollars of market cap coming online this year and billions of dollars of new issuance coming online. And so that begs the question, where is that money coming from? And is it coming from other stocks? Is it other sectors? Is it bonds? I mean, people talk about cash on the sideline. I don’t really buy that argument because when you buy a stock, you’re giving someone else cash. So where is that money coming from? But at the end of the day, supply and demand has to matter at the most simplistic level.
John Norris (19:13):
Yeah. So for me, with all of this that’s going on, and you take a look now, what is today? June the 8th, 2026 and take a look at technology and communications combined being what, 40%, I’m just going to spitball and say 40% of the SMP 500, I think combined it’s even greater than that if I’m not mistaken. I mean, if this continues on and people keep on investing like they have over the last couple years, essentially what they’re making is the bet that these two sectors will be 42% of the S&P five, 46% of the SP, maybe 50% of the SP, what you get catch my drift. If these
Sam Clement (19:50):
Sectors are- Just by the, if you outperform,
John Norris (19:52):
You’re right. If you outperform, you’re growing your market share. At what point do you think individual investors will go, “Hey, listen, I understand tech is a long-term story. I understand all the long-term story, all that stuff, but there’s just too much value. There’s too many good companies every year generating good earnings that are way discounted compared to the market as a whole.” At what point do investors don’t want to jump off the bandwagon and go on back to just the old tried and trip?
Sam Clement (20:23):
It’s a tough question to answer because it’s a trend that’s continued for quite some time, right? You would
John Norris (20:30):
Think it’s out there, but
Sam Clement (20:31):
It doesn’t seem like. You had people probably balking at tech being 30% of this a few years ago and saying, “This is insane.” Well, it’s 40% now, pretty quickly over a few years. And so what’s to stop that trend from… And when again, we’re talking with such large numbers to go from theoretically, I’m not forecasting, but if it went from 40 to 50%, well, that’s trillions of dollars flowing into this. That’s a huge jump in how companies like that would perform. So we’re talking about a lot of money flowing into different areas and if something’s going to outperform, that’s a lot of money flowing into it.
John Norris (21:08):
So it seems like we’re almost talking about two separate things. Investors buying stocks in the secondary market and then also companies spending money on AI trying to squeeze a litle bit more efficiency, a little bit more revenue out of it. What is the eventual end game here and what does it look like when it comes to
Sam Clement (21:26):
Investing in IO? Well, I will say they are two separate topics, but I think they’re very intertwined with stock performance and where companies spend money. We saw this post- COVID with companies that were… They were rewarded for huge amounts of CapEx and every time if they announced they were spending more, well, the stock went up. It was this self-fulfilling cycle that if it kept going up, they kept spending more and kept going up until all of a sudden in 22, a lot of these companies started to be punished for spending too much money. And so as much as they are two separate topics, I think they’re vastly intertwined with the outcome of both the economy and the direction for the markets.
John Norris (22:07):
I don’t know how I can improve upon that answer, so I’m not even going to try it. So that takes us down to the last point, the last question that we have for today is, are we in a bubble or not, or is it really that easy?
Sam Clement (22:20):
I don’t think it’s that easy, but I lean towards no at this point.
John Norris (22:27):
Oh, you got to explain
Sam Clement (22:29):
Yourself. I would say no because I think there’s real money flowing to it. We’re seeing some impact from it already. I personally don’t believe multiples are at some lavishly outrageous level right now. If companies are growing at the rates they are, they probably deserve to be rewarded with a higher multiple than historical collaborations. Of course. And so you have to take all that into consideration when you look at these whatever ratio you want to look at for whether a market’s in a bubble and it’s really not that simple, but I tend to say no would be my
John Norris (23:01):
Answer. I would tend to say it’s not that easy and I would say that the household names with which people are familiar, the answer is probably not, probably not. Whereas there are a bunch of names down in the Russell 2000 and names that you’ve never heard of that don’t have any earnings, don’t have any revenue, start to talk about AI, like what we saw back during the tech bubble and there are some companies down there in small cap and micro cap land that just don’t get the headlines. Yeah, those types of names might be in a little bit of a bubble and some of them micro belly up and all that stuff, but it should not. Even if those things go by the wayside, it shouldn’t cause the same type of reaction that we saw from 2000 to 2002 because the companies which have been leading this, again, are real companies, real big members of S&P 500 with real revenue, with real earnings and some analysts out there might be going, “Ah, come on, it’s not real,” and all that stuff, hey, they reported it’s real revenue, it’s real earnings.
So I think that’s what the difference is. I think Maybe that on the microcap level, you might be in a little bit of a bubble there. However, as a whole, I think this might have a little bit more legs, a litle bit more room to run, but at some point that growth rate will slow and I don’t know exactly how investors are going to respond at that time.
Sam Clement (24:31):
Yep. It’s hard to know.
John Norris (24:32):
All right guys, thank you all so much for listening. We always love to hear from you all. So if you have any comments or questions, please by all means, let us know. You can always drop us a line at or you can leave us a review on the podcast out of your choice. Of course, if you’re interested in reading more or hearing more of what we got to say or how we think, you always go to oakworth.com, O-A-K-W-O-R-T-H.com. Take a look underneath thought leadership tab and find access to all kinds of exciting information, including links to previous episodes of trading perspectives, links to our newsletter/blog commonsense, as well as all the good stuff that Mac Frazier and the advisory services team puts out there as well. So with that, Sam, I’m going to ask you if you have anything else today to talk about on this exciting topic.
Sam Clement (25:15):
All I’ve got.
John Norris (25:15):
That’s all I’ve got today. Y’all take care.
This podcast is for informational and educational purposes only and should not be construed as personalized investment advice, or as an offer to buy or sell any securities or investment services. Listeners should consult their own financial, tax, and legal professionals regarding their specific circumstances. Any reference to specific securities is provided for informational and illustrative purposes only and should not be considered a recommendation to buy, sell, or hold any particular investment. It should not be assumed that any such securities discussed were or will be profitable. References to political or policy matters are provided solely for purposes of economic analysis and do not represent any political views or endorsements by Oakworth. Information from third-party sources is believed to be reliable; however, Oakworth does not warrant its accuracy or completeness. Past performance is not indicative of future results.
Oakworth Asset Management, LLC (“OAM”) is a registered investment adviser owned by Oakworth Capital Bank Inc. (“OCB,” or together with OAM, “Oakworth”), Member FDIC. Equal Housing Lender. This podcast may describe the services of both OAM and OCB, but please note that they are two separate entities that provide different services. Advisory services, investment management and financial planning, are provided by OAM. Because of the ownership relationship and possible involvement by OAM associates with OCB, there exists a conflict of interest to the extent that either party (OAM or OCB) recommends the services of the other. Investments offered by OAM are not FDIC insured, may lose value, have no bank guarantee, and are not insured by any federal or state government agency. For additional information about OAM, including its services and fees, visit adviserinfo.sec.gov.