ETF Edge - What’s next from Roundhill’s DRAM 5/11/26

Episode Date: May 11, 2026

Micron’s stock and Roundhill’s DRAM ETF have become the poster-children for the AI memory rally. But here’s what could be the next “phase” of the AI buildout.        Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

The ETF Edge podcast is sponsored by InvescoQQQ. Let's rethink possibility. Invesco Distributors, Inc. Welcome to the ETF Edge, the podcast. If you're looking to learn the latest insights on all things, exchange traded funds, look, you found the right spot. Every week, we bring you compelling interviews, thoughtful market analysis, and we break down what this means for investors.

I'm Contessa Brewer and for Dom Chu today. So what do semiconductors and prediction markets have in common? We're about to find out. Here's my conversation with Dave Maza, the CEO of Roundhill Investments, and Drew Pettett, director at City Research. That does it for ETF Edge, the podcast. Thank you for listening. And you can join us again next week,

or head to etfedge.c.c.com. Let's get started with Semizzo. I mean, Dave, we're seeing here, one of the poster children for this rally, Myron shares up 11 of the last 15 sessions. Since the end of March, the shares have more than doubled. And then we saw Friday major equity indices mostly flat while your DRAM ETF,

this was the Round Hill Memory ETF, up about 13%. What are you looking for? Yes, of course, semis have been on fire for the better part of this year. But what I think what's most interesting is now investors are waking up to the fact that the biggest bottleneck in the AI buildout is actually memory chips. And there's actually an incredible amount of supply and demand in balance with memory, which is one of the reasons why the stocks have been performing so well. And there really is actually a small number of companies that are involved in making

high bandwidth memory chips or DRAM chips, of which we find and purposely built into the DRAM ETF, which has now turned out to be one of the most successful ETF launches of all time. And you don't think that the bottleneck is going away anytime soon. I mean, is there any discretionary part of these memory chips? Well, it's interesting. So, you know, this is an area where memory has historically been incredibly cyclical. We've seen boom and bus cycles. And one of the reasons it was so cyclical is, you know, memory is actually found everywhere,

in your smart TV, to your phone, in your car, but it's actually been much more driven by consumer trends. What's changed is actually data centers and the growth and build out of AI. So if we use Micron as an example, you know, a few years ago, only about 15% of their business was going to actually data centers. Now it's 65. Along with that, it's changing from having very short-term contracts to longer-term contracts, which is making that business and others like Samsung and S.F. Hynix, incredibly less cyclical. So because of the demand for AI, the need for those chips, we're seeing actually XP, actually, for this supply demand imbalance, not just to extend into 2026, but perhaps to 2027, even to

28, because it takes three to five years to build a new fabrication plant to actually build these chips. So right now, as the AI demand continues to grow and the data center hyper-scaler build-out goes, all that CAP-X is actually going to memory. Drew, I know that you're sticking with these momentum names as well. What's making you confident that you're going to see the momentum continue on this trajectory. It's because the price momentum has earnings momentum backing. So this is the place where we have seen the best earnings revisions this year in the United States and globally. If you look at KOSPI for South Korea, again, driven by a couple of the big DRAM names,

just up and to the right when we look at analyst's expectations. In the U.S., a name like Micron, that's added five extra dollars to earnings for the S&P 500 this year alone. So to us, we don't like price momentum unless it has earnings momentum and improvement in the DRAM and semi-stocks have that. And when you're putting these semi-stocks through your screeners, even with the massive moves higher, up triple-digit percentages since March in three months, are you still finding that they come back from your screeners with a green as go? Yeah, the key is price performance is up that much, but we're talking about earnings that. let's call it a dollar of through cycle earnings before might be more like six or eight dollars this time. So yeah, if we're up 300%, but your earnings expectations are up six to eightfold for the next few years, it still comes back reasonably priced to us.

How does this thesis sort of play into how you're thinking about emerging markets? I mean, you mentioned the Kaspi in South Korea. Is that a favorite? It is. I got to give our global team a lot of credit here. they've really focused on the narrowing outside of the U.S. Again, where are the good fundamental stories, it's AI, and where do you find that, EM, but specifically in Korea and Taiwan, again, Korea and the memory story. Because they have a lot of exposure. Yes.

Yeah. Dave, is there a bigger story the market's missing here? Well, I think part of the story, which is debatable, is this re-rating story. So historically, again, because of the cyclical nature of DRAM companies, they traded at very low, price to earnings multiples compared to other semiconductor stocks. But what's changing now is because that the demand has become sustained and it's expected to continue to be built out, we're actually starting to see a re-rating of these names. So what's interesting is to Drew's point, these names have incredible price momentum, but that's coupled with earnings momentum, yet at the same time, they still

screen incredibly inexpensive relative to their peers, which makes this a very interesting setup. And one of the reasons why if we think about just what's needed for all these incredible new Navidia chips that are coming on market, we expect that the momentum for these names will continue because the demand is going to hold up. I mean, I feel like Nvidia has been grabbing the spotlight and the microphone. And, you know, Nvidia is the poster child for incredible growth. Are there other opportunities, Drew, that you're seeing in smaller or mid-cap companies? Yeah, I think so. Again, it comes back to the AI value chain.

So this is part of the story. So, yeah, semis, you need semis, you need memory, you're layering that into data centers. But on top of that, there's other enablers that play to the back end here. It's the industrial names. You got to lay concrete. You need plumbing. You need electrical work. When you're going to build out a new data center, a lot of those industrial.

names also have really good pricing power, and they're seeing better efficiency, again, because they can price higher when they do those jobs. It's really remarkable how many industries have exposure to the AI explosion. I mean, I cover insurance, and we hear insurers all the time talk about the opportunities in data centers. So, you know, when you're talking about what is it going to take to fuel this to miss those enablers, I think is the word that you called them on live television, I think is important. If memory is this next phase of the AI cycle, Dave, what do you think comes after this? What's next? Well, it's interesting. So, you know, are actively managed generative AI, TF, the ticker is chat, CHAT, we've got exposure to some

of these bottlenecks. So bottleneck trades. And so I think most folks are now aware that semis over software has really become not software eating the world, but semi's beginning to eat. the world. But there's actually other companies that play a role with that. So we think about sort of optical names. So these are companies like Credo, Coherent, Luminite. Those are all needed to actually make these semiconductor chips speak with each other. So it's one thing to go out and buy as many into video chips that someone can afford, but they're actually only as good as how fast those chips can speak to each other. And again, the amount of memory chips that are needed for Vera Rubin is significantly greater than even Blackwell needs.

So these bottlenecks, whether it's to Drew's point, some of these industrial names, which are playing a large role in building out these data centers, or just the equipment that's needed to make data centers work as effectively and as powerfully as they can be are all starting to play a role here. And the market is beginning to kind of look at what those next legs are. I think, of course, right now, front and center is memory. What do you think the next phase is? I think people are rotating back to some of the mega caps, like the Meg 7, for example. Like we saw a lot of money go back into, I would say, like mega cap growth, like the NASDAQ type of exposure last month.

And looking at our AI team's work, you're finally seeing some of those cash returns on investments for the hypers start to flatten out again. So you're starting to see better behavior from mega cap growth. And that becomes kind of your defensive trade when you're worried about higher for longer oil. I'm curious, well, an oil is another one we could get into here. But Dave, I'm curious, when you're talking about this bottleneck, it sounds like you're saying, even though the semi-space is crowded, it's not overcrowded. It's just about this log jam. How do you see that log jam easing, that the entry point or the portal becomes wider,

so there's room to spread out? Yeah, it's interesting. Well, of course, at some point in any industry, you end up, having supply meet the demand that's there. What's interesting that we're seeing with memory and then the phonetics names that I was speaking about a moment ago is that you can't just turn this on kind of overnight with a flip of a switch. These take many years to build the factories and equipment that's needed. And in fact, the equipment that's needed to make these names, that in and of itself is sold out. So I think it's going to be some time before we see some normalization

of this supplying demand in balance. But over time, I think data centers are going to become a part of the economy. If we look at the most recent GDP results that we saw from the U.S., a big part of why we saw growth was the AI buildout and the data center buildout. And so as that continues, and AI becomes to be integrating in our economy, I think we're actually going to see sustained demand here, which is a little bit difficult to forecast exactly what that looks like. But I think investors who are sort of sticking their nose out and saying, you know what, because these stocks have moved so much in such a short period of time, maybe missing out on exactly what's happening beneath the surface.

You mentioned oil there. Let's talk a little bit about geopolitical risk and the way that plays into, I mean, if memory chips are the thing, then supply chain is super important. How are you factoring in the Iran War risk, you know, the lingering. risk from Ukraine and Russia and the bigger platform about trade and international cooperation. Yeah, it's funny from a really high level perspective. It's kind of buy oil as the commodity, not the equities, as your tail risk hedge. I look at our commodity work. We don't think the tail risk, the extended Iran conflict scenario is priced in there. So that's a good hedge to kind I have on the side. When you think about equities, it's about buying where there's secular demand. So in EM, again, Taiwan, South Korea, easy place to do it. And then it's the U.S.

While those areas may be hurt by increased power costs and energy costs in the background, their demand source is still okay. In the U.S. especially, this is like our favorite geopolitical hedge right now, large cap growth in the United States. You're relatively, insulated versus the rest of the world. You're still impacted, but not the same way as Europe. Dave, what are your thoughts? Well, I think, again, we find ourselves in an interesting environment where there is a significant conflict.

And in some ways, this time is different because it involves the Strait of Hermuz, I ran oil. But to Drew's point a moment ago, I think the market, of course, sold off significantly when we saw the escalation of conflict. But when, as it's had a time to take a step back, actually absorb the incredible earnings that we are seeing, and in particular for the case of Magnificent 7, expected to continue to see,

we actually see that the stock market is insulated from some of these macroeconomic concerns. And now that doesn't make it any easier for what people are paying for prices at the pump. But mega-cap tech runs on chips. It doesn't run on oil. I want to switch gears a little bit to a subject that is near and dear to my heart,

and that is prediction markets. I'm following along as federal courts weigh in the validity of prediction market offerings, especially on sports, and you have states and tribes fighting back and saying they're not. And at the same time, Dave Roundhill has one of three applications before the SEC in this space. The SEC has just postponed it. I know you can't talk specifics, but big macro view. talk to me a little bit about putting a prediction offer, an event contract out as an ETF. Yeah, exactly. So Roundhill is one of three ETF issuers that have filed for this. In fact, we're actually the first ones to file for prediction market ETFs, which are based off of event contracts.

And the ones that we've focused on are focused more on the political side. So on the presidential election, so Republican or Democrat, the Senate and the House. And one of the reasons why, while a novel instrument event contracts, prediction markets have been around for decades. The University of Iowa has actually had a prediction market for some time. So they're not actually that novel in and of itself. What's different, though, I think, is the utility of the potential for a prediction market in TF. You know, for years, you know, strategists like Drew and his peers look for baskets of how investments can take advantage of whether a Republican is going to be an office or Democrat or hedgehog. But tools like this actually allow for a more direct way for investors to express a view or hedge their portfolio against potential risks that they may have depending upon the outcome of an election. A bipartisan bill has been introduced in the House that would prohibit prediction markets from offering event contracts on the outcome of elections and on government action, which would take out a lot of what has been popular to talk about would also prohibit sports.

for what it's worth. But when you look at that, is there still, if there were no elections that were available in prediction markets, would there still be a reason to offer, say, a hurricane hitting the West Coast of Florida ETF? Yeah, I think it's interesting. With prediction markets themselves, of course, there's an incredible amount of iterations for what actually is being predicted or what event people are looking to express a view on. But there is also at the same time, a significant number of contracts or events that can allow people to express use, the potential for a recession or not. So even outside of politics, you know, it's set aside sports for a moment because I think that's a different story, at least for the time being,

I do think the power of prediction markets where people can express a binary to you can be very powerful, which is one of the reasons why, you know, we've seen the ETF vehicle itself provide access to a multitude of exposures. At one point, having stocks in an ETF was incredibly novel and new, and then fixed income, commodities, cryptocurrencies. I think this also has a potential to be the next frontier. Yeah. Drew, you have not filed an application on this yet. But when I've talked to other companies who have,

they've said, look, the other thing is, much of America's assets are in their brokerage accounts, and it makes it easier to trade on a question like who's going to win the White House or weather or CPI. It makes it easier to trade on that if it's already in your brokerage account. What's your high level view on the future of prediction markets and how it fits in with ETF strategy? I'm going to step way, way back on this one and really talk about just portfolio construction. And over the years, people always look for uncorrelated assets to put in their portfolio. access to uncorrelated strategies,

it's really been a driver of that sleeve in portfolios. That used to virtually be zero in the ETF landscape. Now we're seeing, I would say, kind of two to maybe five percent as in differentiated uncorrelated assets. These types of products would fit there. I think people are just looking for different types of risk, different types of exposure. ETF is just a vehicle that they like to consume it in.

What do investors need to know if they're looking at an uncorrelated asset. What risk are you trying to manage? What characteristic are you trying to generate? Is it cyclical risk? Is it growth risk? Is it just something that markets typically don't have a really good way to trade? As Dave said, we try to make a bunch of baskets,

and it's probably why I have more gray hair. Over the last few decades, technology has transformed our world in amazing ways. Through it all, InvescoQQQEF has continued. connected investors to the forefront of innovation. Access the future today with Investco QQQ. Let's rethink possibility. There are risks when investing in ETFs, including possible loss of money. ETF risks are similar to those of stocks. Investments in the tech sector are subject to greater risk and more volatility than more diversified investments. The NASDAQ 100 index includes the 100 largest non-financial companies listed on the NASDAQ. An investment cannot be

made directly to an index. Before investing, consider the funds, investment objectives, risks, charges, and expenses. Visit investco.com for a prospectus containing this information. Read it carefully before investing. Investco Distributors, Inc.