Hello everybody, welcome back to the Wolf Financial Show. My name is Gav Blaxberg and I am your host for today. And joining me for another amazing episode is Dan Niles, one of our favorite money managers to have on the show, always bringing the honesty and truth to light. We're talking today about the memory field, specifically because we have a new largest listed company in mainland China and that is CXMT. They are a memory maker and they are coming onto the scene. We lightly talked about this during the SK Hynix IPO and now it's a great time to follow up since they have actually gone live. Huge IPO, 400 plus percent off the bat. So let's talk about it. Dan, what do you make of this company coming to light and how this is going to affect the rest of the memory world?
Well, I think what it really shows is when you have a stock that's up 466% on its IPO, everybody pays attention, right? And it's, you know, half a trillion in market cap in mainland China. And so there's sort of this narrative that people have had that, oh, China doesn't really matter. You know, they're trailing edge. And so it's not going to really disrupt the memory trade, which has been arguably the hottest trade on the planet since, you know, at least over the last year, but you could say since, you know, you had the introduction of ChatGPT at the end of 2022. And you've had a lot of people saying, oh, you know, this industry is no longer cyclical. And I think with all that stuff I disagree with completely, but what this IPO does is it forces people to examine that, and then SK Hynix before that. Because don't forget this meltdown's been happening really since June 22nd, and I put out a note on June 20th saying I thought we were in for a speed bump because of token minimization becoming a big deal and companies trying to control their AI bills. So you can trace this all the way back over the last month. SK Hynix got some increased focus on it, on, hmm, wow, there's these companies in Korea, if I'm a US investor, that are actually getting pretty aggressive and having an ADR listed here because Samsung doesn't have one listed here. You know, that gets people focused, and then this one, just because it's up 466%, it gets you focused on China.
Yeah, it does. So let's dive deeper into that. You know, you pointed out the cyclicality potential within these pieces and how it makes people focus on this. One thing which we'll talk about is potentially a flood of cheap memory coming from China. I want to dive into both those topics. Maybe you can, you know, speak to each of them. Why this, you know, why you believe this is still a cyclical industry and then also how China could affect things.
Yeah, so a little bit of a history lesson here, and I know we talked about this and you had a chart that I'd given you when I talked about it on the SK Hynix. But if you go back, the US back in the mid-1970s had 94, 95% DRAM market share. Intel had 75% of that. And then Japan entered the market, and it was sponsored by the state. Obviously Japan does a lot of, you know, electronics, and so it made natural sense. If you want to start in semiconductors, the reason companies start in semis is because it's a commodity. And so you can really get good at making chips when they're commodities. And so Japan then entered the market and it was sort of people said the same thing. Oh, well, it's trailing edge and, you know, it's not the latest and greatest. It won't really matter. And then what happened is by the time you got to 1985, Japan's market share went from 5% in 1975 to 80%. And Intel basically was going to file for bankruptcy other than they decided to shut the business down themselves and focus on the sideline called microprocessors. Now, if you fast forward to the mid-90s, you had Korea actually in the, you know, late '80s, but Korea, South Korea entered the DRAM market, and they really started to ramp that up during the late '90s. And then by the early 2000s, they were forcing the Japanese companies basically out of business. And Japan today has a market share of 5% versus what they had in 1985 of 80%. Now, if you look at it in 2025, you had China around 5% share or so. That's going to continue to grow, and my belief is that I think CXMT, right now the estimate is about 8% DRAM market share. I think by the time you get to 2030, it'll be close to 30%.
And you can think about it from a big picture basis and say, well, hey, China actually consumes about 30% or so of global memory anyway through the fact that, you know, they have cell phone sales, you know, 25, 30% of global cell phone sales go in there, similarly with PCs, lots of other items. And so it would make sense that, hey, if you're just supplying domestically even, you could get to roughly that kind of share. And so I think that's kind of the history of the market, and every time, whether it was Japan entering or South Korea entering or now China, everybody always makes the argument, oh, well, you know, it's trailing edge, and it won't really matter because it's not the latest and greatest. It's a commodity. It always matters. And I think right now the market's getting a really good reality check on some of this, and, you know, we'll see how this plays out, but I firmly believe the industry is as cyclical as it ever was. You're just going through a very strong demand phase, but ultimately you're going to end up with a crash on the other side of this. I don't think this is the crash. I just think this is a correction like you typically see, and you saw several of those in the 1990s, and we can get into that if you'd like.
Well, I mean, we could look back at that because I think it goes to this idea of cyclicality. One thing that people point to today as to why it's potentially different is in the past you didn't have years and years typically of memory being purchased in advance, and so people are saying, hey, you...
Was it all people could get? The same nowadays. It's like, hey, we'll buy everything you can make. We'll buy it years out. We'll buy it at premiums.
No, that's not even true because you can go back to COVID, right? And you may have forgotten this, but during COVID, you had long-term supply agreements, not as much for memory, but for other things, you know, lots of other semiconductors. Because if you go back and you look at COVID, right? What happened? Well, everybody was forced to go online. We couldn't do anything in person. So you had this humongous ramp in businesses having to get onto the internet, do all their business over the internet, etc. You had this huge surge in demand for semiconductors, and the semiconductor companies weren't ready for that. And so you had a lot of long-term supply agreements that were actually struck in 2022 by companies, and Microchip is probably a great example of this because they were probably the most aggressive with long-term supply agreements. And they basically forced their customers to take the chips because they had signed these long-term supply agreements, whether they liked it or not. Well, what happened when the customers came off of that? They stopped buying. Revenues went down over 50% year-over-year when you got into 2023 in some of those quarters. So you've seen what happens just a few years ago when companies thought that they were completely sold out and demand was going to be good forever and they were signing these long-term supply agreements. And the way long-term supply agreements work in reality is the following, right? You've got these massive customers, whether it's Apple or others. They're signing these agreements. When the prices go up, these massive customers expect the suppliers to honor those agreements. Then what happens is when prices collapse in the spot market, what they'll do is they'll go back and say, hey, we want to renegotiate. And then the semiconductor company goes, well, you signed this agreement. And they'll say, okay, fine. If you make us take it, then when this agreement runs out, we'll do what happened to Microchip. Because there's always another customer that's willing to supply. And so we're just going to cut you off, punish you, and go to your next, you know, whoever your competitor is. And I guarantee—well, I guess you can never guarantee anything—but I would be highly surprised if that didn't happen on the other side of this. Because as I've been saying for a while, if you believe AI is the most transformative technology in history, which is why people are saying, oh, well, you know, it's different this time in memory.
Well, if it is the most transformative in history, then that means that you're going to have more people that can possibly make it out on the other side over-investing because they know if they're the last man standing, they're going to make a fortune. This happened whether it was the railroads or canals in the late 1800s and 1900s, whether it was electricity, TV, radio, fracking, or most recently the internet. And I believe this is one of the most transformative technologies we've ever seen. Therefore, I believe it's an over-investment, and therefore I believe just like every other great transformative technology, you're going to have a bust on the other side of it, and memory is going to be at the epicenter of that.
So, how are you positioning right now because we've had a decent pullback in some names. Some not so much. Micron.
Some names are down 50, 60% in the last month.
That's not a big deal because here's the thing, you can go back to 1995. Okay? So, when did Netscape Navigator come out? End of 1994. That was the first mass internet web browser. It was sort of your ChatGPT moment for the internet. Then if you go to late 1995, the semiconductor index had over a 50% drawdown that started in late '95 that extended for almost a year, roughly a year or so into 1996. That's a drawdown. That's of the entire index.
Semi index is only down, I think, from peak around 20, what is it? 25%? That's nothing. Now, you can go to 1997. Similar thing. Semiconductor index in late 1997 had over a 50% drawdown that lasted roughly a year then as well. Now, the index from the end of '94 to its ultimate peak in March of 2000 was up 850%. So, this is a pullback. This is actually not that severe at an index level. Yes, for some companies, are they down a lot? Yes. But are they still up a ton year-to-date or over the last, you know...
Absolutely. So, for me, when I hear people complaining, oh my god, I go, well, that means you're not managing your risk well. Because you bought this hook, line, and sinker, you've forgotten your history, and you're not, you know, I don't know if you're permabull or what the situation is, but I lived through this in the '90s, and I go, you're going to get this type of stuff happen. And what's scary, what I like today is that people are starting, there's starting to be a lot more questioning because of the SK Hynix, you know, ADR, because of the CXMT IPO in China, of, oh, okay, well, maybe we're getting a little ahead of ourselves.
