Howard Chan, Kurv Investment Management

And when we actually came on this channel two years ago, exactly two years, because we're celebrating the second anniversary of KQQ, that was our very first basket idea in a particular sector. And so now I'm happy to report, I think when we were first on, we were talking about a theory, financial, like how we would work. And now we have two years of history that performance that I've actually proven out those theories that we have been managing, applying the same techniques we have with these large institutional clients to an ETF.

That we're in, more and more people are needing to depend on investment income, whether approaching retirement or during retirement to offset either negative cash flow or just simply, needing more from beyond what they have in paycheck. And so, yeah. Oh, please go ahead. No, please. And there are lots of assets that you want to own, but don't generate income, right? So, for example, growth stocks, which has been larger and larger part of indices. We've talked about that in people's portfolios, but it's kind of an exposure you can't escape, right? That's where, if you look at the S&P 500, it's really the tech companies who have positive earnings versus more traditional consumer staples. They're actually relatively flat as a performance or actually slightly negative, right? But the problem is, is that growth stocks don't distribute dividends.

They, because of their return of investment is so high, they rather, either reinvest it back into the company or they do stock buyback. Or there are other asset classes, which we have moved now, like, where, so precious metals like gold and silver, right? Bonds, as we know, is less and less good of a hedge in a portfolio because that negative correlation between equities and bonds is, less negative. So other assets that are highly, more negatively correlated with equities are providing better hedges in a portfolio, but there also precious metals do not provide income. So we're really trying to figure out ways in where these are really anchors in people's portfolios that we think you should have going forward. And how do you build an income portfolio out of

Tend to look a little bit forward and see what are the next bottleneck? And I don't know if you recall, there were a couple of years ago, there was like this article on the New York Times where a reporter started talking to chat GPT and he kept, he talked to, to the model for like an hour or two. And it went to very weird places where it, I think at the end, confessed love to him, something like that. And that was a common effect called a hallucination. And the reason is because when you ask the model a question, it doesn't remember the previous question that you asked. So there has no context. It's asked, it's answering your question as it was like a new question every time. And

Fortunately, we, I think most of the models have left this hallucinated stage, but the reason why they're able to fix this problem is because they have actually put in more memory into the, uh, into the models so that the models are able to retain past histories of your conversation. So just like you're talking to a friend, you have history, so it can better answer your question over time. And that is made possible because, um, of memory chips. And in fact, it's not just the models. It's even NVIDIA's motherboard requires more, uh, memories to be accessed in order to actually help the compute as well, uh, as well as, um, all of these, um, data centers are being built. Uh, they require a service requires a lot of memory. So this is what we feel was, a really big

Bottleneck in the AI built out. And so that's why we launched, uh, KMEM as a dedicated focus, uh, uh, uh, an investment case for that particular segment.

Commoditized that somebody who comes in with a lower price can undercut the volume that they were selling. Uh, but with the AI built out, actually, um, they're actually creating a moat around their business. They're creating these specialized chips called HBMs, which is specifically their, their memory is essentially stacked on top of each other's that gives more capacity to the AI build out. So they have actually going back to the mine analogy, they have sold out most of their inventories up to 2028. Uh, so, people's like, well, if their demand is so big, why don't you just make more? Right. Uh, but the, the problem is, is that this is the physical constraint I was mentioning about. You hit a physical limit in order to make more chips. You need more fabs,

Fabrications, uh, uh, factories, uh, they take, 30 to 50 billion of fabricate, a facility to build out. Right. And they take time, physical labor to build out, the factories, et cetera. And then there's additional bottlenecks. Okay. Now you have the factories that's in place, but you need the equipment to make these chips, right? So a lot of these use very, uh, high end technology, like the lithography, which another bottleneck is only one company in the world, uh, that makes that as an ASML, a Dutch company, and they already have years of, uh, back, uh, wait list to get their equipment. So all of these really distinct bottlenecks makes it impossible to increase, uh, supply, at least in the near term. We actually think this

Bottleneck is a two to four year thesis. That's why even, I think the elephant in the rooms, like we've had a good, big correction in semiconductors. Why is this still interesting? Is that because this constraint is going out two to four years, um, we see this actually continue to be an investment case. The reason is because you've actually seen that these, uh, memory companies who used to make commoditized products actually have pricing power. So SK Hynix just reported earnings. Their revenue went up 250% and their profit went up 500%. So they are price setters instead of price takers before into the market, right? Of course, there's other incumbencies that are coming in, uh, Chinese company, which what I think caused a little bit of the pullback, but

How they're setting the prices for their chip is at the same level as the Korean company. So it's not a cheaper alternative, uh, that people thought it was going to be. But I also want to mention, there's a secondary effect. So I'm sure everybody's heard of, Apple raising prices for their products, uh, MacBooks, iPhones, Xbox raised prices. Uh, all of these consumer products are raising prices because memory chips costs has gone up. There's a secondary effect, which is that because these three companies, Samsung, SK Hynix, Micron want to build a memory chip that has the moat, it's actually crowding out them making out commoditized memory chips. So there is another bottleneck in just consumer electronics, uh, in terms of all of these productions. So I think this

Actually has implications for many parts of the portfolio. Um, so we, we think this actually is a very interesting space for people to continue to look at despite, we actually think that the pullback is actually a good entry point into this particular space.

Of the backlog are they going to be able to resolve? Because many of these memory chips, again, as I've mentioned, has already two year, three year, uh, uh, backlog on, on orders. And then if you look at the hyperscalers, when they say they have another two, three years of compute backlog for people who want to use this, right? So there is this chain effect that, ultimately, um, uh, happened. So we think that if there is increased supply or slower demand from the hyperscaler, it's just going to compress the backlog, but they're still going to have pricing power. Uh, uh, the other thing that we are monitoring also is new entrance in the space.

There is a behavior in semiconductors where in the past, uh, we've seen this in the nineties and two thousands where, semiconductors chips become really hot. A lot of new entrance come in underpriced existing incumbent. And in fact, that's what a lot of the Korean companies did to Micron. Micron was like the incumbent in memory chips. And then, uh, most of memory chip production moved to Asia because they were actually able to produce it, uh, at a lower cost, uh, and undercut it, uh, because it was a commoditized product. Right. But I think in this cycle is a little bit different because we've seen a bifurcation in the market. There are large capable companies, the three companies who have the capital to focus on the specialized chips. And that's very much

Dependent on the hyperscalers and their spend. And then they're, they're moving out of the, the commoditized product, but there's still a bottleneck there. So really what we're looking at is a new entrance and also in the commoditized space, uh, because that feeds into everything we do. And I, another example is, I don't know if you people recall during COVID car prices went up, not because there wasn't enough metals to make, but because there were not enough memory chips in the cars to actually make what, cars are really just like computers on wheels nowadays. Right. So, um, so those things will filter out actually to more of the broader economy.

So that's another thing that we take a look at. And, and, and we think there's probably still going to be the cycle and semiconductors, but the, the peak and the trough are flattened a little bit more because just because there's more products and, and, the need is so much greater. There's, there's also upside risk. the next hottest thing is sort of robotics, right? How do we automate those also requirement? So there, there could be competing demands, offsetting those of the hyperscaler. So I would just say that there, there are risks definitely, but the space is so dynamic right now. I think we're kind of in this innovation push, uh, the 1900s industrial revolution was sort of this innovation push. I think we're kind of in that

Other cycle. So, uh, risk is, doesn't always have to be downside. It could be an upside case. So we're, we're kind of monitoring that space, uh, quite closely. So I was looking at the holdings as we've,

Actually the thing maybe also worth mentioning is that, um, there are many different types of memories. There's NAND memories, there's solid state. Um, there's also storage memory that I, Western digital and Seagate, I, there's a big thing in the 90, you go to Fry's and get like these two terabyte, um, uh, uh, uh, storage. All of these are being eaten up because all of these are different parts of the data center and, and AI infrastructure, um, built out. So what we felt was there was most pricing power in the top three, but as things become cheaper and as, market has pulled back, we we've actually seen some other cheaper names that we're actually also reallocating into. So we're very much an active management firm. So when there's opportunities,

We will, we'll actually go into those particular links. Interesting. So there are, as,

Difference between us and a few other ETFs is that we can actually get the Korean denominated securities. Uh, and so, uh, some just focus on, uh, U S securities. Uh, some focuses on some Asian securities. One of the, I think difficulties is that Korean equity markets is actually highly regulated. It's largely open to institutional investors. You have to be registered to get those kinds of exposures. So we actually, it took us a while to launch it to actually get those access to get those securities when we think there's value. Uh, uh, and then we have also access to other parts of Asia's where a lot of the memory chip, um, ecosystem is. So I think what we want to do is not just pick the right names, but from an execution perspective, get them at a price that makes sense

Have done it done. So is because essentially they're monopolies, they, they dominate their particular verticals, and then they move to the next adjacent vertical to induce that growth. Because as a company gets larger, you expect the growth rate to slow down because they're part of larger part of the market. They can't grow more anymore. Right. Amazon dominate retail. They go to the next adjacent data, AWS, et cetera, et cetera. So what we wanted to try to do is to get this essentially growth premium momentum that exists in these particular names. And, uh, most of these names do not distribute dividends. So we wanted to use options to generate income. However, we were also trying to solve a third problem. Uh, this is our kind of tech focus is that most option income

Strategies underperform their underlying because, uh, you're giving up some of your upside to write option positions to generate the income, but we don't want to do that. We want to provide price appreciation and income. So two years down the line, I will say that we have kept up with the underlying and in fact outperformed it. And then it generates about 15 to 18% of, uh, distribution, depending on, where implied volatility is in the market. Uh, and also, um, it's not just, um, uh, in exposures to the hyperscalers is technology company. We define technology companies as those companies that have used technologies to transform their particular industry. So there were moments where we had Oracle, when they had the first, um, pop from, from the data center, um, uh,

Built out, um, Netflix when they were growing subscribers, less of this year. Uh, and we actually were able to move from the mega caps to semiconductors this year. That's actually what's kept the portfolio outperforming, uh, relative to actually almost all peers, uh, in the category. Uh, so when we first came on the channel two years ago, we were talking about, we look at momentum waiting because that's a characteristic of tech. And then we, um, the perfect time to do cover calls or call spreads is when the underlying is moving sideways or downwards and we'd be more tactical about it. So we've done kind of what we've talked about and that's actually proven out to have good total return as well as good, uh, income distribution.

Company than a traditional company. So it's been falling, but this is actually the perfect time in which if you write, uh, call spreads or, you can clip the full coupon because the underlying never hits the strike, uh, but versus it would have been a very poor choice, uh, sorry, we, we do this with, with Microsoft, uh, this year, Google as well, because these, um, essentially have been falling this year, right? It would have been a very poor choice if we had written calls on semiconductor names like micron that's in the portfolio or AMD or et cetera, because you would have constantly capped your upside. So being able to work first and foremost, active managers. So we have to have an opinion on the underlying and, and then what, then it's the question of the

Implementation of those views, right? Uh, we can decrease the exposure for name by, uh, writing a call or, um, buying a put, et cetera. And so we want to, in the truest sense in active management, collect the premium where we can fully, uh, realize, uh, that implied volatility and then

Dividend equity portfolios invested. So now you can pair up dividend equity with K triple Q. You, the combined portfolio actually has no, gets closer to the sector weighting of S&P 500. And you get actually a 5% pickup, uh, from, from the combination of the two. So now it actually evens out your portfolio in terms of getting us large cap exposure. It's not just us large cap value. Now it's actually almost like us large cap value plus us large cap growth. You get the full sector, uh, in, in, in the exposure. So we think this could be a, uh, a core, a compliment. And then K mem, I will say is more of a specialized view. It's, it's thematic in nature and us large cap

Already has large part of technology, right? So this will be probably in part of your equity, a portfolio, but it would have to be coming out of your growth portfolio, uh, uh, allocation. Yeah.

You the exposures that you need in a portfolio, but in a very, very tax efficient way. So that those are actually all coming out, uh, uh, in fact in the last week and in the coming months as well.