Key Takeaways
- The AI rally is far from over, with investment and demand still exceptionally strong, says Carmignac fund manager Kristofer Barrett.
- Hardware stocks remain the clearest AI opportunity because of higher barriers to entry and lower competition, he says.
- While “there is more to EM than AI,” not owning South Korean stocks has made life much harder for emerging market managers.
Karen Gilchrist: The AI boom has transformed global equity markets, driving extraordinary gains for select stocks. However, as concentration risk grows, portfolio construction is key. I’m joined by Kristofer Barrett, head of global equities and fund manager at Carmignac, to discuss the outlook for global equities, where he is seeing AI investment opportunities, and what he sees as the direction going forward. Kristofer, thank you so much for joining us.
Now, in your view, where do you think we stand in the AI investment cycle currently? Are we still in the infrastructure stage or are you now seeing tangible opportunities more in deployment?
Kristofer Barrett: So I’m not even sure we could call it a cycle. I mean, there’s been a continuous growth path since Nvidia’s March 2023 guidance upgrade, where they were starting to sell these GPUs for AI use. The growth has been explosive and it’s been constant since then, but with a lot of worry in the market and sentiment shifting. Right now, we’ve seen the hyperscalers continue to invest, unabated by price hikes and hardware shifts. Demand is very high still, and compute is sold out. This is what we see, and I think these last earnings reports showed that.
Nvidia, Broadcom, and TSMC Seen As AI Winners
Karen Gilchrist: You’ve mentioned that technicals appear stretched, with crowding in momentum stocks, and the potential for some of these abrupt rotations that we have been seeing. But you do remain confident, in the AI theme long-term, and I know that you view Nvidia NVDA, Broadcom AVGO as some of the safer players. So how have you been adjusting your positioning?
Kristofer Barrett: It’s been a very volatile market. I think a few months ago there was a view that AI is going to continuously grow, and the multiple you put on these stocks does not matter. I think we’ve had a healthy pullback and a bit of a reminder that stocks don’t only go in one direction. There are a few reasons for this, it seems, that are technical and leverage-based. The reality is the underlying investment has continued, as it has throughout other cycles. But the magnitude is very, very, very large, and there is a debate on the ROI [return on investment], especially for the hyperscalers.
The opportunity set now is in the hardware space, where there are shortages, where there are differentiated companies. The quality AI names—such as Broadcom, Nvidia and TSMC—they are larger, their growth rates will not be as big as these smaller names, but they’re more likely to be these ultimate winners. The other debate is on the software side, where we need to see new applications grow, and that’s usually what takes the slowest. It’ll be a multiyear path—both the public software names, but also private names coming up—growing that meaningfully.
Karen Gilchrist: Do you have any sense of conviction that software stocks could endure this kind of AI transition? I know that we saw that big selloff, but then there has been something of a rotation back into them.
Kristofer Barrett: So you don’t want to conflate all software stocks as being one thing. There are some that have done very well and bounced back. Atlassian TEAM is a name that we hold, which has proven that it can accelerate growth, even though it was seen as being one of the most potentially hardest hit names, and it was down more than 50% year-to-date at some point. But there are other names that have not been able to bounce, which we don’t own so we won’t talk about them. But then the big growth vector and competitive vector is on the private side, where we have startups and the OpenAI and Anthropic models themselves potentially competing. So that’s definitely a space that will play out over a longer period and there is more uncertainty.
Karen Gilchrist: You mentioned TSMC 2330, and that’s the largest holding in your global fund and the second largest in your tech fund. Clearly hugely strategically important, but it also does sit at the center of geopolitical tensions as well. How do you weigh those kind of stocks that do sit between those two sides?
Kristofer Barrett: So it’s an ongoing debate. Nothing has really changed. Taiwan supplies to the globe. Without TSMC, there is no Apple. There is no Nvidia. There is a balance: I think Xi and Trump are meeting in a few months’ time or a month’s time. This will continue, this debate, but this is much too important for the global economy. And I don’t think anything is going to happen because it would be existential for both sides.
Currently, China has their own AI ecosystem. They’re investing heavily. They believe they can compete, whether it’s on hardware without the leading edge but in different technology vectors there, but also in open-weight models, which is one of the debates that has also been worrying for the large Frontier Labs. An interesting space. We think they will all sort of succeed; there’s going to be a balance. There’s going to be frontier tokens usage. There’s going to be a bifurcation of thousands of different use cases and a smart use of tokens, eventually.
Under-the-Radar AI Stocks Carmignac Owns
Karen Gilchrist: There are, of course, these concerns over concentration risk in some of these big mega-cap tech stocks. So where are you seeing genuinely differentiated opportunities? I know you have holdings in the likes of Nittobo 3110, this Japanese fabrics and textiles firm, and also Taiwan’s Asia Vital Components 3017, which is another sort of separate AI play. Where are you seeing some of these opportunities?
Kristofer Barrett: Yeah, definitely. So the hyperscalers are in question now because of the capex, because the free cash flow is very, very low now. We think, ultimately, they will get a good return on investment. But I’ve been asking recently, are they investing because they have to or because they want to? And it’s probably both. And the increase in investment is going in one direction; it’s going to hardware and semiconductors. We’re trying to find the names that are either reasonably valued in this space, or that have absolutely unique technological advantages that we don’t think are easy to replicate.
Hardware is harder to replicate: it is chemistry, science. You need to have experience. You need to actually make these things, not just write code for them. So definitely a space where there’s a vector of less competition and higher barriers of entry. You mentioned Nitto. We also have JX Materials 5016 in Japan. Sumitomo Electric 5802 is a bigger conglomerate, but they’re specialized in cables and indium phosphide, which is very interesting for the optical buildout. That’s in Japan. And in Taiwan, you mentioned Asia Vital. We have Lotus 1795. We have Grand Process Technology 3131. Semi-cap [semiconductor capital] equipment is going to be very interesting, given, for example, the Terafab announcements, Intel INTC raising money recently, and TSMC starting to even more aggressively grow their logic foundries.
Karen Gilchrist: Fascinating. And, again, I notice in your tech fund and your global fund, you also have significant emerging market holdings. Now, of course, some of these are comprised of some of the Big Tech names, SK Hynix 000660, etc. But are there other EM opportunities that you’re also looking at?
Kristofer Barrett: I’m happy I’m not an EM manager anymore. I used to be; it’s really tough. I think over 50% is basically AI-related. The big trade there is the memory names. So the Koreans. If you didn’t get that right, given the volatility, that’s been super difficult. There is more to EM than AI, but it is kind of tough to actually find these names because of the negative correlation. When AI has been rallying, these other names have not been. We’re interested in the consumer space. We have a meaningful or relevant position in MercadoLibre MELI. We’ve built a position in SEA SE and Allegro ALGM. So basically online platforms for consumption, And I think consumer names are probably a good space in EM, given demographics and other issues like that.
Financial and Healthcare Stocks in Favor
Karen Gilchrist: Absolutely. And I know you’ve mentioned also a “barbell” approach in your strategy and looking at that element of diversification in your investments as well. I know you have significant holdings in some financial services firms. What are the opportunities you’re seeing there? What might other investors be overlooking in some of those?
Kristofer Barrett: So it’s been a truly unique market, where high beta tech has gone in one direction, and a lot of other names like healthcare and financials have actually gone in a different direction. I.e. the beta has gone to almost zero versus the broader market. This has been extremely challenging, because it means there’s very, very high volatility, but it’s also given opportunities in names that don’t usually trade at lower multiples.
Exchanges in financials is a space we like. Business services that have data and unique moats. Healthcare is also interesting, more as a balancing factor—especially distribution companies, where you’re actually active in the physical world. But also biotech, where we have a basket of mid-cap biotechs. I think it’s important in an overall portfolio approach to be balanced. We’re running a global equity fund, it’s not all in on any one theme. We don’t know the future. The key is to be valuation conscious, and most funds, especially if they’re growth-oriented, they lose sight of that. The other important thing is also just timing; buying the names when they’re down or selling them when they’re up. And that’s easier said than done.
Karen Gilchrist: Of course. And just taking a zoomed out lens, looking at the impact that we might see from AI, say five or 10 years down the line, just how transformative do you think it will be? Any particular sectors that you really think are going to be most impacted, for better or worse?
Kristofer Barrett: This is the multitrillion dollar question, and it’s felt like it’s the only question that’s mattered in the past years. And it might still be that way. There’s been a debate between HALO, so hard assets, not being able to be disrupted by AI and basically soft assets, what used to be the best companies with quality, no capital needs, the ones seen at risk. I think this is the job here and now, right, to figure out which companies can adapt, can use AI, can use—I think importantly—people and AI, which is where I think the big differentiator will be.
There’s a lot of excitement, I think, into R&D and investing and change, but it takes longer than people hope and realize. To me, AI is the second wave of the internet. The internet’s collected all the information. What AI is basically organizing it, so you can ask a question, and synthesizing all of the information that we humans have actually thought of and written up. And there will be some novel ideas there. But I think the real value is the novel ideas with the people. So I think we like companies that have big R&D and that are leaning into this.
Karen Gilchrist: Kristofer Barrett, thank you so much for your time. For Morningstar, I’m Karen Gilchrist.
