Artificial intelligence is reshaping the composition of the S&P 500 and challenging traditional ways of assessing large-cap U.S. stocks. A cluster-based approach may offer a clearer view of the changing index.
BNN Bloomberg spoke with Scott Chronert, U.S. equity strategist at Citi, about grouping companies into growth, cyclical and defensive clusters to evaluate performance and fundamentals.
Key Takeaways
- More than half of the S&P 500 can now be linked to artificial intelligence, extending the theme beyond a handful of mega-cap companies.
- Semiconductor and hardware companies have increasingly benefited from capital spending by hyperscalers, while software faces a more complicated impact.
- Growth stocks now represent more than 50 per cent of the index, compared with less than 20 per cent about 30 years ago.
- The shrinking weight of cyclical stocks has weakened the S&P 500’s direct connection to economic conditions and other macroeconomic developments.
- AI-related earnings growth could support higher valuations, but stock performance will increasingly depend on each company’s position in the buildout.
Read the full transcript below:
LINDSAY: Throughout the AI-fuelled bull market, investors have been talking about the Mag 7 to assess large-cap growth dynamics. However, our next guest says this construct is now a bit out of date, and he has a different framework to understand the AI infrastructure buildout. So, here to tell us more is Scott Chronert, U.S. equity strategist at Citi. Good morning. It’s great to have you join us.
SCOTT: Hey, Lindsay. Thanks for having me.
LINDSAY: So, let’s start with you thinking the Mag 7 construct is outdated now. Why is that?
SCOTT: Well, I think, you know, we’re over three years into this AI tailwind, if you will, and the Mag 7 very deservedly has been a mechanism for assessing how that is playing out from a fundamental perspective, as well as from a price-performance perspective. Going into this year, we began to argue that you should expect the Mag 7 to begin to trade more idiosyncratically. That is, not all of these stocks will be beating to the same AI drum that they had over the previous couple of years.
The first-half performance underscores this. The Mag 7, at the end of the, at the end of the first half, was actually flat, if not down, for the year, while the broader market had moved higher and the AI tailwind storyline had really shifted over to the semiconductor space.
So, our point here: The mega-cap component still is relevant, but we’re expanding the definition now to various other industry groups within the S&P, such that we would argue that over 50 per cent of the index can be attributed to AI. We need to monitor that whole set of companies to get a better sense of where the trend takes us from here.
LINDSAY: Okay, so why do you think segmenting the market into growth, cyclical and defensive is a better way to kind of understand the AI buildout?
SCOTT: Well, it’s really changed because what’s happened with the AI buildout. You know, we’ve been talking about AI enablers and users for some time. That’s been a storyline for a couple of years. But within the enablers, you’ve gotten this demarcation now between how one thinks about the hyperscalers, those spending specifically in the capex, and those benefiting from that spend, which is going to be a lot of the semi and hardware names. And then, somewhere off to the side is how the market perceives this to impact areas like software.
So, all of this is unfolding in a way where it’s confusing the bigger picture. If you look at the price action so far this year and the earnings growth for the S&P 500, it’s really been driven by a smaller cohort of players that are attached to semis, but in particular the commodity semis.
So, essentially, the way we’re identifying this is that times passed versus 30 years ago. Our growth component has more than doubled in size. Thirty years ago, it was less than 20 per cent of the index. Now it’s over 50 per cent. This has dramatic, dramatic implications for the trading behaviour for the broader S&P 500 index.
LINDSAY: So, how much exposure, then, do you think people should have to growth stocks versus cyclical and defensive? Like, how risky are growth stocks right now compared with the others?
SCOTT: So, what’s going to happen as this unfolds, right? So, what you’re going to get with a bigger percentage of the index having this growth connotation, you’re going to get more realized volatility. You’re going to get higher-beta swings in the market. So, you have to be prepared that, as the S&P 500 shifts from a construction perspective, the fallout is going to be what we’re seeing almost daily, is a lot of volatility.
So, what we’re suggesting in terms of our view is that, yeah, you want to be tethered and have core exposure to this growth cohort. The AI theme and trend isn’t going away, but how it’s going to play out at a stock-specific level is going to be more and more differentiated between where you fit in this AI buildout.
Now, at the same time, you’ll think about the rest of the index, cyclicals and defensives. What we would say is that, you know, we’re very all-in on the broadening thesis for the second half of the year. That set, that sets up for, we think, ongoing rotation into cyclicals and perhaps some areas of the defensive part of the market.
LINDSAY: Just going back to that scale-up of the growth cluster in the past couple of decades, like, what are the implications of that, or at least some of the implications?
SCOTT: Big implications. So, there’s the growth mentioned that with the volatility component, but the flip side, the flip side is that the cyclical component is less than 30 per cent of the S&P 500 right now. So, when you think about many of us get very attuned to assessing the macro implications of whether it’s economic conditions, Fed rate perspective and so forth on broader U.S. equities, well, be careful, because what this argument is telling you is that only a relatively small subset of the index is actually directly exposed to that, right?
So, we want to be wary that the macro connection is lessening versus history. As I mentioned, we’re probably going to see a higher level of ongoing volatility. But at, at the same time, a lot of people will look back at historic valuation connections and say, “Gee, isn’t the S&P 500 expensive versus history?” And I’m going to argue, “Well, versus history needs a big asterisk next to it because the nature of the index has, in fact, changed.”
The growth dynamics that come with this have changed. It supports higher valuations, but it also puts more and more burden on the future growth expectations of this AI playbook to really drive the S&P 500.
LINDSAY: Just lastly, do you have any concerns that the valuations of some of the companies within this growth cluster have gotten ahead of themselves?
SCOTT: No, I don’t have concerns on the valuations, Lindsay. What I’ve got concerns on is, as I mentioned earlier, you get to a point here where a lot of the fundamental action becomes a bit of a zero-sum. What’s good for the semiconductors in terms of aggressive hyperscaler capex spend, as an example, is bad for the hyperscalers because the market keeps asking the question: What about return on investment, and will the revenues be there to support the spend?
So, you get this dynamic where you get a shifting playbook within the growth cohort. I don’t have concern about the valuations as much as the longevity of the earnings growth profiles. That’s really what this all comes, what comes down to. And I’d say right now, in general, in general, this growth part of the market is discounting fundamentals for 2026, probably through 2027, maybe 2028, but not out through the end of the decade.
And the Citi house view, for example, is for ongoing AI capex spend to continue to ramp through the end of the decade. So, we still have a ways to go in this playbook, but how it’s going to show up stock by stock is going to be ongoing idiosyncratic.
LINDSAY: Okay, Scott Chronert, U.S. equity strategist at Citi. Appreciate your time. Thanks for joining us.
---
This BNN Bloomberg summary and transcript of the July 22, 2026 interview with Scott Chronert are published with the assistance of AI. Original research, interview questions and added context was created by BNN Bloomberg journalists. An editor also reviewed this material before it was published to ensure its accuracy and adherence with BNN Bloomberg editorial policies and standards.
