00:00 Speaker A
As investors digest Meta’s expanded partnership with chip companies, they are focusing on the increased capital spending that big tech companies are investing in their AI infrastructure. Wall Street is eagerly waiting to see the winners and losers of this deal. And so does our next guest, Thomas Carroll, Steffel Equity Strategist. He says the AI industry is becoming divided between hardware winners and service losers. Thomas, nice to meet you. So let’s start there.
00:26 Speaker A
Well, as you say, we’re currently divided into winners in AI hardware and losers in AI services. Please explain that to me, Thomas. Please explain it to us.
00:35 thomas carroll
yes. Thank you for having me on board. Well, the crux of this note was that if you think about the market over the past three years, AI trading as a whole has actually gone long. But what was interesting for us in the past was really last quarter, as you said. It was a breakup of AI services that were perceived to be losers. You and the previous guest were just talking.
01:03 thomas carroll
And the perceived winners, uh, AI hardware, the names have, you know, safety attached, and the market is really moving towards them. But the gist of the memo was that if you believe the market opinion about the overall losers of AI services trading, the overall security is really questionable.
01:25 thomas carroll
So, to talk about what we’re actually seeing in the AI services space, as I talked about with my last guest, there’s a lot of pressure on these business models as we move forward, and we can see that very easily across the credit curve. Whether it’s very safe stocks, high-yield sectors, or high-tech sectors, credit risk premiums tend to rise significantly.
01:49 thomas carroll
And we think that’s really what’s holding down the name of many AI services. Well, I’m concerned about the large amount of CAPEX they’re doing with the uncertain ROIC in the second half of this CAPEX. The question we have is whether the market can believe that these AI services companies will survive. If you look at the main revenue from AI hardware names, it’s capital investment from AI service names. So,
02:14 thomas carroll
The two are interconnected, and this split in trade has made us somewhat cautious. This is consistent with the overall view of the S&P 500 this year.
02:25 Speaker A
Thomas, you’re actually comparing this moment to the late 1990s, which is interesting because we saw a sharp decoupling of hardware and services, where we saw volatility explode. What do you think are the most important similarities between now and then?
02:41 thomas carroll
Yes, I think I got a little goosebumps on the back of my neck when I first observed that particular study. When we look back at the late ’90s, we just remember the tech bubble, the boom, the bust, but it was actually obviously a very long period of time. And while the Nasdaq actually peaked in early 1999, the market held out for about nine months.
03:02 thomas carroll
So what we observed then is very similar to what we observe now, except that back then GPT was the Internet. Obviously, you know, we’re talking over the Internet. It changed all of our lives. No one doubts the importance of AI, but it was very interesting at the time that hardware was spreading in a non-exponential pattern compared to internet services. And when that trade peaked, it was up 95% in nine months, and when it peaked, it was kind of an all-time high.
03:28 thomas carroll
For the market at the time. So we’re seeing a GPT similar to GPT, perhaps even bigger than the Internet. You know, we’re seeing a very similar split in services related to hardware.
03:41 thomas carroll
Oh, I kind of get goosebumps observing this similarity, especially considering how thin the S&P 500’s overall margins have been so far this year.
03:51 Speaker A
I thought it was interesting, Thomas, that you note here how AI services are trading at relative valuations that we haven’t seen since the global financial crisis. So I think as an investor, you’re trying to determine at that level, is it an opportunity, or is it not, is it a value trap? How do you make that call?
04:13 thomas carroll
Now, I think the question is why did this happen? Because even though the future earnings of these companies continue to rise, we are talking about a 10-year valuation of price-to-earnings ratios. So I brought up the overall credit risk that is rising within the system. I’m really attached to these names. And the question we have is: Is CAP X starting to exist for these companies’ business models?
04:36 thomas carroll
That’s pushing this valuation down, and if that’s correct, it could be a value trap. If you know, credit stress starts to grow. I think the bigger question we have is: Will it extend to the general economy? Because the whole AI space is not only very important to the U.S. economy through the K-type consumer wealth effect, but also very important to the industrial production side.
04:57 Speaker A
Thomas, I want to go back to what you mentioned earlier, the increased credit stress in some parts of the technology industry. Thomas, what the hell are you looking at? Well, what signs, what signals? How concerned do you think investors should be about that?
05:12 thomas carroll
Yeah, um, for us, it’s, if you look at the whole AI services space, it might include the biggest hyperscalers. They’ve also seen credit risk go up, you know, Oracle’s chart has been looked at by a lot of investors, and that’s really true for most hyperscalers in different frameworks, but it’s also true, you know, in the broader credit risk, high yield space. This has almost doubled since the second half of last year.
05:32 thomas carroll
Overall, it was what was called a high yield credit spread in the market at the time. Well, that was a key driver of reputation compression that we saw in the AI services space. Yeah, that’s something that we often see in the early stages when there’s stress building up in the system in terms of all areas of the economy, but especially here with technology.
06:03 Speaker A
Thomas, it’s great to be on the show today. thank you.
06:05 thomas carroll
Thank you for being here. appreciate.
