SPEAKER_00 0:00

In today's episode of the AI Investor Podcast, an eye on micron, micron technology. We got micron earnings.

SPEAKER_01 0:06

Micron earnings.

SPEAKER_00 0:06

Micron. Micron. This talks up more than 25% in the past five weeks. It's Micron's world. We just live in it, and I'm I'm just happy to be here. Eric, we are back with the AI Investor Podcast. You and I are both traveling this week, so I'm sorry if our audio quality and backgrounds are not quite up to our usual standards here. But how are you doing? How's vacation with the family? You're near a beach. You guys were doing a little golf carting, a little beaching.

SPEAKER_01 0:38

Yeah, we're down in Charleston. We go here every single year. It's a lovely place for anyone looking for some uh new vacation destinations. But Austin, you know, I'm on vacation and I'm looking at the news, and all I can say is a memory, it's a memory world, and we're all just living in it. Micron's up 18%, even though it's a trillion-dollar company. It it feels like, you know, last yesterday I was following Micron's earnings, and back in 2024, whenever NVIDIA reported, CNBC would have countdowns to earn earnings.

SPEAKER_00 1:07

It was the Super Bowl moment.

SPEAKER_01 1:08

Exactly, the Super Bowl moment, like the world would end if they missed earnings. And that's a bit what feels like with Micron right now. So I think most of this episode, sometimes we say, hey, it's gonna be a little shorter episode. As you know, we're we're both on vacation. We'll probably keep this one a little tighter, talk about the events, but we really just wanted to talk about kind of what's happening in memory today, and then we'll kick it over to a few listener questions and we'll close it up for this episode.

SPEAKER_00 1:34

That sounds great. Yeah, this memory moment is singular, and I I know we're gonna get into this, but I got to get two things out of the way up front here. Micron had more in net income this quarter than they had revenue last quarter, which is something I'm gonna have to repeat multiple times before I fully grok it. But I repeat, that is their net income this quarter exceeded their revenue last quarter, and that's largely an explosion. I mean, driven by a lot of things, but an explosion of gross margins from 39% to 85%, which I know we're gonna get to. One quick request to our listeners, though. We don't do ads on the podcast. We do this for the love of the game. Um, and Eric, your picks have uh performed very well for our listeners here. So I ask any investor or any listener who's got a couple moments here to go leave us a comment or review on Spotify, YouTube, Apple, wherever you listen to this podcast. We really love hearing from the community. We are going to do a quick QA at the end of this episode, and those are questions that we curate from those comment sections. So if you're interested in, you know, if you have any stocks you're interested in, you just want to provide any feedback, please YouTube, Spotify. It really makes all the difference. We appreciate that. Okay, Eric, into the news. Uh market volatility has now become the norm. And there have been multiple days where we've seen major $100 billion, trillion dollar stocks moving six to seven percent. And the Van X semi-index is really the hotspot of this. And I don't know what what the specific volatility of that index is, but it was down 6% in a day. This was the Korean market puking roughly 10% on memory stocks. It looked like maybe the memory stock trade had peaked. And then the next day, right, we got a four to five percent rebound. We're seeing micron, I think it traded from over $1,000 a share down to low 900. Then now we're well up, you know, $1,100 as of earnings right now. So, and this is all in the span of like six, seven trading days. So, what is going on with this market volatility? Why is it so exaggerated right now?

SPEAKER_01 3:30

Yeah, it it definitely feels like over the course of this podcast, we've had like 48 moments of the AI trade is dead, mostly quickly resolved. And this week started when I when, like I said, I'm on vacation, but while I'm checking my Twitter, it was like a full-blown debate around memory being essentially dead. And that was led by price action of the Korean stock market, causing a large and fairly painful uh drawdown, especially in semiconductor stocks. As you know, the Van Eck, the SMH, which is widely tracked, or the SOX, they're both down more than 6%. Um, I believe that was on Monday and then Tuesday followed with another 2% loss. So, Austin, what was going on there? Number one, it's it's some level of rotation. A lot of investors might say, Oh, well, you're talking about six percent losses, my portfolio wasn't down that much. Well, people were rotating into software stocks, hyperscalers, um, away from these popular trades. So if you have a lot of hyperscalers, NVIDIA, Amazon, Meta, Google, et cetera, in your portfolio, you might have actually been up on those days. But as you noted, what we're seeing this morning now, semiconductors, the market's open. I will quickly type in the SMH, we'll look, and it is up 4% at the open. And Micron, as you noted, let's look at that. It is up 19% at the open to uh $12.50 per share. And Austin, some of the background factors that I think are leading to this level of volatility that we're beginning to see. We we mentioned several of the sell-offs have been caused by the uh Korean stock market, seeing massive declines. Korea, of course, houses two of the three biggest memory companies, an SK Heinnix, and Samsung. Well, Austin, their entire index is up more than 100% year to date. And it's it's it's a it's frankly, it's kind of a gambling market. Um, when you look at leverage, there's reports of the most popular stocks trading in Korea, and right now it's just dominated by leveraged ETFs. The thing about leveraged ETFs, Austin, you we both know prediction markets, Colchi, uh, poly market, their whole pitch is we are built to manage risk. Um they're not managing risk, no one's using it. You know, there'll be a story every once in a while of a bar offering free drinks if if three goals are scored in a game.

SPEAKER_00 5:54

Oh, that was a good one. That was a JP Morgan bar, I think.

SPEAKER_01 5:57

Yeah, yeah, but you know, that that's zero percent of the use case, 99.9% of the use case is is is just gambling. And leverage ETFs are similar. So what happens is you have a lot of money piled into these, they're often short-dated, they're actually target dates for only specific days. They don't actually track over long time periods, and they just cause capital to move relatively quickly. So, right now, and we we can get more into microns earnings. Um, what happened was we had an exaggerated panic about the memory trade unwinding at the beginning of the week, which was following price action at the back half of the week, we get to reality, which is micron's earnings, and we see what we've talked about with this memory trade right now, Austin, which is it's probably gonna be 48% of hyperscale, almost half of hyperscale spending next year. That that's very much intact. And we've gone through waves since we started the show. First, it was NVIDIA, we saw photonics, we saw bottleneck mania. It's memory mania right now, Austin. And the reason it's memory mania is because it's such a percent of spending. And um, you know what what we saw in Micron's earnings is this isn't something that's going to suddenly hit a wall. In fact, it's it's going to be more enduring than the the market's given it credit for at nearly any point in time.

SPEAKER_00 7:15

Yeah, I I know we're gonna spend a let's dive into Micron deeper um right after this. And one of the big things that people have been watching with Micron is the gross margins, right? And you've talked about this before. It's like, look, this this bottleneck is such that it's not a question of whether or not Micron's going to make money, it's how much money are they gonna make at which at what margin because they're fulfilling on the memory bottleneck. So it's like if they can only fill 50% capacity, maybe it's at an 86% gross margin or 80% gross margin. Maybe they can, you know, as they add supply, they can fill 70% of capacity and that's at a 65% gross margin, right? Like it's it's it's so these two things trade off. And I know we're gonna talk about that a little more in a second, but I wanted to get to your point on levered ETFs because we had talked about how one of the biggest risks in the market right now is interest rates, and I agree with that. I wouldn't necessarily call levered ETFs a risk in the market, but if you're an owner of any of these underlying equities, it's just adding volatility into the system. Maybe this is top of mind right now. I actually re-watched the big short last night. So maybe this is uh, you know, credit default swaps and MBS uh bundling all over again, uh, and it's top of mind. But I mean, the day SpaceX IPO'd, right? The biggest IPO ever, we're already seeing multiple levered ETFs launch. And I think as of right now, I don't know how many there are, but I know I've seen Pro Shares, I've seen Direccion, I've seen Granite shares all have SpaceX levered ETFs. For our listeners, we do not recommend levered ETFs. They're they're a fundamental short-term trading instrument. Maybe the promises for managing risk, the reality is that they introduce volatility and leverage, which is not a bad thing if you're a long-term holder. It just makes the ride a little bit, you know, you know, more exciting on the ups, scarier on the downs. But there's a natural erosion because of how these levered ETFs um manage their underlying positions. So just as a market commentary, not as an AI commentary, the explosion in popularity of levered ETFs uh is something we have not talked about on this show, but it definitely feels like it's sort of this like broad-based like volatility tide that's sort of you know raising some of the pressure in markets a little bit or the the excitement at a minimum. But let's uh let's talk about this micron moment because this feels like NVIDIA in 2024, and that's exactly what you saw, the the entire market is hanging on these earnings. And I had actually seen something from Fact Set earlier, I believe, that really underscores this, which is if just if you remove just NVIDIA and micron from the entire sec uh tech sector's Q1 earnings, earnings growth goes from 51% with them in to 28% with them out. So it's really, you know, these two companies really are carrying this entire sector right now, which is probably why Micron now feels like NVIDIA in 2024. What else are you seeing here? And and what is what is driving this uh 86% gross margin, you know, moment? It's it's incredible. To to sell a product at 86% gross margin to make 28 billion in net income in a single quarter feels it feels impossible.

SPEAKER_01 10:14

Yes. I mean it's it's higher gross margins than NVIDIA ever saw now, which is a remarkable thing. You look at Nvidia with their servers and the kinds of competitive advantages that you go into in the software ecosystem around CUDA, and I believe they peaked at 79% gross margins. So the idea that Micron has higher gross margins than what you'd see as kind of the most durable franchise right now in AI is remarkable. But Austin, last quarter the numbers were a blowout. Uh that was to be expected, but they were more of a blowout than I think even the whisper numbers you hear about on Wall Street. Revenue is up uh to 41.5 billion. I I don't know if you saw I I put in my notes I had sent you. Maybe you took it from me, but the fact revenue is higher than net income. If you saw that and uh it saw uh said I I I have to get that in the show is because it is so striking.

SPEAKER_00 11:10

I actually didn't get it from your notes, but I'm sure we saw the same breathless tweets about it, and like our jaws collectively hit the floor. It's just like it it it is hard to describe to people who are not long-term investors how singular that fact is. I mean, I'm I'm I'm I I'll I won't I won't try and belabor the point here because it's hard to express how impressive that net income growth is.

SPEAKER_01 11:33

Yeah, and looking ahead, guidance always matters more for these companies, guidance up to 50 billion next word. They're gonna beat that. It's it's gonna be higher. And margin guidance to 86%, which is the number we're talking about, which is absolutely insane. Most of the conversation this morning, Austin, is um they're signing these uh customer agreements basically guaranteeing a level of memory prices. And we've talked about this word a lot durability of earnings, durability of earnings. How long will you know this current wave of AI not only last but continue to grow? And this is what Wall Street loves because the question is when is Micron going to see a peak and and how sharp could the correction on the the other side be? And what they've said is that they've got agreements now running through 2030. It's currently about 20% their their memory volume for DRAM. They expect it to reach more than 50% of revenue, so they're gonna have more than half their revenue tied to these agreements. We got to work out some specifics about what exactly they mean. But for Wall Street, this is what they're focusing on this morning. Austin price targets. I think it was maybe like four podcasts ago. We let off talking about how Mike Ron had gotten an upgrade up to 1625, and it was the whoa, everyone was could this possibly be happening? And this morning we've got 2200 from Melius on reframing the memory trade, JP Morgan to 1540, Sesquehna to 2000, DA Davidson to 2000, and that was just me looking about 7 a.m. this morning. I'm sure we have a lot more, so we have price targets zooming above um what a few weeks ago was uh in the estimations of most market watchers, an insane price target. And now it's it's probably lower than what the median price target will be shortly. Also, another one that may warm your heart. I saw this, I said, I need to mention this for Austin. Uh, another uh name check to robotics, which Micron has done on several recent earnings call, but they um were a little bit more explicit this time in terms of memory needs, and they said the memory need of a robot's 10 times greater than the memory need for self-driving cars. And they said to expect a multi-decade memory cycle to begin from robotics later this decade. So, you know, this this does raise another point that as you may see some slowing in data centers themselves, will these other markets pick up, which is the point you know, NVIDIA itself has been making.

SPEAKER_00 14:06

Um be still my be still my beating robotics hard. I'm so glad you brought up that point. I saw that, I saw that stat as well. And actually, I I I guess I just had it, it did not occur to me or I had not seen good data on memory usage for self-driving compared to robotics. But that was a striking statistic, even if it's only half true. Let's say through you know algorithmic efficiencies and gains that are to come. If a robot only needs, uh, you know, fully autonomous robot needs 5x more memory capacity than self-driving cars, that's still an enormous amount of memory. And it reminds me of a underlying dynamic that we sort of touched on a couple of times over the last two years here, which is as capabilities increase with this technology, with artificial intelligence, new markets become viable, right? They become possible, they become uh affordable. And as these prices collapse, then more and more, you know, in these various technologies, more and more markets can then take advantage of AI and you know, the convergence of powerful AI, powerful batteries, um memory all converging at a point where these things are now possible. So this is not the micron story, it may not, as you're saying, literally just be limited to this moment in time with a data center build out. And people are saying, well, how long can data centers continue? It's like, well, look, once we get 24 months in the future, there's now other technologies that are creating this appetite for memory as well. So uh, I mean, we we it's just it's micron's world, we just live in it, and I'm I'm just happy to be here.

SPEAKER_01 15:29

Well, and it it's also a point, you know, you gotta sharpen your pencil, and this requires a lot of work on modeling it. But one of the biggest threats to a lot of the data center companies would be well, what happens if a lot of the intelligence moves to models on the edge? Um, you know, your things like smartphones or devices that, you know, you don't need to be going to the cloud for all this uh capacity, and you've got smaller models on your phone, for example. Well, that's a significant driver to memory as well. Would it make up for all this demand? It it's hard to say, but it's kind of a heads I win, you know, tails you lose kind of situation for memory in many ways. But you know, Austin, I think the bottom line for these micron earnings is uh these agreements they basically lock in free cash flow probably to 2026 levels, um, which you know a lot of modeling had them going below 2026 levels as soon as 2028. Um it's it's a it's a massive um narrative reset. Now the the bigger question, of course, though, and this is something we've talked about, is memory being half of spending, is that good for the total AI trade, right? Because what this effectively becomes is it's a tax across everything. Um, you know, in in many ways, AI as a total trade would be more healthy if memory wasn't so expensive. This morning, I I don't know where they opened, but I was looking pre-market. Most of the hyperscalers were down. And the reason is because, well, what what's good for memory and and has in many ways read-through for other semiconductor stocks is bad for the company spending.

SPEAKER_00 17:10

I was I was just I was just gonna say, is that what it is? You know, Micron's gravity to you know command a free cash flow grab of all these hyperscalers is that much greater now, right? They're they're even more important than they were three months ago and six months ago, and therefore all the free cash flow that the the Mag 7 and related companies have been known for and lauded for that's made them wonderful investments, they have to give it up now, right? I mean, they if they want to play in this game, they have to give it up to Micron. Correct, or memory or other memory companies.

SPEAKER_01 17:38

Yeah, and it puts a pretty sharp point on the fact that all these companies went from throwing off unprecedented levels of cash to zero cash production, and their cash is increasingly funneled into one former commodity that they don't seem to have any recourse against, right? You know, they're they're locking themselves into supply agreements, which is kind of an admission of, well, they're essentially capitulating, right, at this point.

SPEAKER_00 18:03

Um, Austin though, how many companies in the last 15 years could get uh Apple, right? The most powerful cash cash-rich company on earth, arguably, you know, top three cash generating machine of all time. How many companies on earth could ever get Apple to capitulate? Be like, yeah, we just got to pay more. And yet they are. I mean, Tim Cook literally said prices are probably gonna go up on smartphones because of memory. Um I can't.

SPEAKER_01 18:26

Yeah, they had they had another announcement this morning about it as well. So oh, really?

SPEAKER_00 18:30

I must I must have missed that, but to the point, I mean, it's yeah, too. How important and vital do you have to be to get Apple to publicly be like, yeah, we get we have no other option here?

SPEAKER_01 18:38

Well, it shows how aggressively this happened. Apple almost always has a plan B, right? As as China production started encountering problems, they they already had a plan B for India and Vietnam. They've they've always had plan B's for the supply chain. They're famous for definitely playing supply chain names off each other. And this is the one area where they've been caught the most flat-footed. Almost, you know, we were lucky, right? At the beginning of 2025, we had named memory our biggest trend for the year. We had recommended micron, we had recommended SK Heinex, we had recommended um supplier chains as well. At the beginning of 2026, we once again pounded the table on it. We expanded our recommendations to things like ACM research. We've we've made a lot of recommendations across the memory space, but I mean, I never expected Micron would be up more than a thousand percent from where where we recommended it. It has been shocking. Um, you know, the the takeoff on the specific trade. And it it is micron, I think it's up more than 300% year to date. So if you had just purchased at the beginning of the year, it's still up more than 300%. But you know, Austin, it I just wanted to quick if for anyone out there who's listening to this, and I know it's a number of astounding numbers. Like I said, memory is kind of sucking the air out of the room from everything else right now. It is kind of the trade in AI. And if someone's watching that and and wondering what to do next, I I've pounded the table several times saying one of my top ideas in recent podcasts is something like Anto Innovation that's gonna benefit from the memory supply chain continuing to rise. Onto is up more than 30% across the past month. Um, we'll talk a little bit more about later, but ACM research is one I'd recommended in March, near that bottom. You look at CXMT, which is China's main memory supplier, it's adding three times as much capacity as Micron this year. Prices may be staying elevated in part due to some reticence from the big three, SK Heinrich, Samsung, Micron, to build more. There's only incentives to build more from the Chinese suppliers. Um, so you know, the not the only number, Austin, that Micron missed on last quarter. They missed on one number across all their financials, and that was their capital expenditures were low because they're not they're not building capacity as fast as they want. So if you are an investor and you're looking at it, this comes back to even if you're reserved about memory, uh, the fact that they're going to there's gonna be a lot of incentive to begin building more, and and you can track through supply chain estimates, it it is gonna increase a lot. And that leads to these supplier plays, which we have a number of in the portfolio. So even if someone doesn't want to buy a micron today at these prices, um, I I think there's gonna be a lot of derivative plays, of which, you know, I just ran through, we have made and and they have been very profitable so far.

SPEAKER_00 21:41

And you know, going from strength to looking like continued strength, you know, now that Micron has their strategic customer agreements locked up through 2028, they've got guaranteed CAC to control their own destiny, which is a luxury not always afforded to the memory space in the last 50 years, right? Where it's just been this complete boom and bust um cyclic. Industry. So the fact that they can very easily and more reliably forecast their cash position and revenue two years out, three years out from today, um gives them sort of an unprecedented and unusual level of planning autonomy and control here. Whether or not they can use that to command the same gross margins they have, or whether or not they use that to increase capacity, you know, we'll we'll have to see. But um a phenomenally strong position. Congratulations to Micron and everybody who celebrates Earnings Day. Um we you you we had talked about some of the hyperscalers being under pressure. I want to look at Alphabet slash Google specifically here. So they've uh Alphabet was kind of left behind early on in this AI race when it was just considered you know OpenAI's world and GPT was eating everything. Gemini made some amazing advancements with you know nano banana and audio and video in particular, but also just broad model capability. They they uh for a long time, maybe still today, were considered to have the most efficient model on a per token basis. Uh they've got their easy integration with the Google Suite Enterprise, which you know sort of takes a bite out of some of the you know, maybe Microsoft anthropic integrations here. So they were left for dead early on, completely rebounded. I don't have the numbers on the top of my head, but I think they doubled in you know some short order. But now they're down again. So the company's down roughly 15% since their mid-May high. And I'm wondering what is happening here. Is this Micron hoovering up other cash flow? Is there changes with model rankings? And now Gemini is seen to be less impressive than it was. What's going on? Not that anybody's buying Google for their exposure to Anthropic, but they do have this interesting dynamic to where they own a very large part of Anthropic, ad says Amazon as well. So, in one way, it's like if the story here is the strength of Anthropic's model, you also in a weird way get exposure to that through Google, although it's a small part of their portfolio. But why this change in tone? Because Google was sort of the can't catch up kid for a while, then it was the can't miss company. Now they're off pretty big, double digits. Why?

SPEAKER_01 24:00

Yeah. By the way, I am just we're recording this right as the market opens on Thursday morning. It did quite the reversal, Austin. Uh Nasdaq was up more than 2% last night. It's already down 1%. It's basically a straight line down in the morning. Uh there was inflation data this morning, but it it does show how quickly things can change. Microns give up half of its gains, and it's those damn levered ETFs.

SPEAKER_00 24:23

It's those damn levered ETFs.

SPEAKER_01 24:26

Yeah, and we'll get away with it next time.

unknown 24:29

Next time.

SPEAKER_01 24:31

Next time. No, you know, it is uh I will be it is such a question of right now there's a lot treated as what's what's good for micron is good for the mark. I I will be curious if that narrative holds. But with with Alphabet itself, it's down 2% today. It's it's down 15% um since since mid-May. So I think there's three items worth noting here. Um, some high profile uh exits. The biggest was Noam Shazir. Um, he's gotten a lot of credit for Gemini's Rebound. It it's always hard to tell as an outsider what a single person means, but he he's moving to open AI. Second Austin, everything's about coding right now. We we talked last week about uh in our SpaceX episode about them making an acquisition for cursor to get coding data. OpenAI has pivoted to make codecs uh at the forefront of their business. Anthropic obviously built kind of their current empire on the strength of Claude and coding. And this is just an area that Google's relatively behind. So I think there's as far as AI itself, there's some um kind of narrative that Google is continuing to fall behind right now. They've they've got a big model release that's been pushed back into July. We'll we'll see if they kind of leap back um kind of to the forefront with that. But third, Austin, it's just hyperscalers. I know we talked about this. They remain under so much pressure. When you look the past month, and this is before today, with the market um apparently seeing another big sell-off as we're recording this, Microsoft's down 12% across the past month, NVIDIA's down 7.5%, meta nine percent, Amazon 12%, Alphabet 11%, the NASDAQ as a whole is 4%. So I know you were just referencing leverage ETS, but you know, it is a question at what percent of this is people systematically maybe moving away from them as um, you know, a number of factors like needing needing to move towards equity and debt for the next phase of AI's build-out, um, fears around basically them having to pay this incredible uh memory tax, or third, how much of it's just money flowing away from these stocks in favor of new names, you know, as as interest built towards SpaceX, as interest built towards Micron, as interest moves towards um indexes like the SMH, that does need to come from somewhere. And there's been a lot of talk of these hyperscalers essentially being not only funding shorts for hedge funds, but also retail moving away. So I think I think there's some alphabet-specific issues here, but it is worth checking in on. The stocks that have powered the market for the last decade are seeing some of their worst relative performance in 2026. And that's a new, emerging, and very powerful story as well, because these are the stocks that if you index are the a huge portion of your portfolio, and for most people are stalwarts of their portfolio. So, you know, if if you look at underperforming versus some of the individual stocks, it's it's probably because most people are just heavily weighted into large mega caps that are suddenly stuck in mud in a way we haven't seen across the past decade.

SPEAKER_00 27:47

Yeah, and you know, I did these are not official recommendations, but I'll just share with our listeners. I I've personally been adding to NVIDIA and meta on this pullback. You know, I like all the names on this list, but uh NVIDIA, the ultimate toll boost story still holds, and you talked about that in our last episode. So if somebody wants a good thesis there, they should go back. Not the most recent one, which shout out on your SpaceX episode. Fantastic, by the way, if anybody wants a deep dive on that. Um, meta being down roughly 10%. I don't, I do not know how it's gonna go, but and I know morale is very low at Meta right now, but this feels a little bit to me, it's got this has echoes of like Meta's H100 metaverse moment, where Mark Zuckerberg was betting billions of dollars on the metaverse and bought a bunch of H100s, which would seem like a cautionary tale since they ultimately wound down the metaverse. But remember, they were able to pivot those H100s to AI, get a re-rating. They eventually stopped the cash furnace that was the metaverse project. They can do it again. That is not that far going ancient history. You know, if Metaf realizes their, you know, their open source model is not the path, they can flip all of that capacity they've been using in the same way SpaceX is and maybe rent it if they want it and turn a cash furnace into a cash machine. And the valuation is so low right now, you're sort of paid to wait. I mean, this is one of the most impressive cash generating companies on earth. So you're not you're not paying a premium price for that optionality either, which is one reason I like that one a lot at today's prices. Um, Eric, any other final thoughts on these Mag 7 companies being down close to 10 over 10% in some cases, or do you want to move on to listener questions, of which we have many and we love to get to?

SPEAKER_01 29:19

No, let's move on to listener questions. I I appreciate the uh the check on the SpaceX one. I think you enjoyed not many other people did. That one had like one-third the listeners of the prior episode, which I get it. Lon Musk is a fairly divisive person. I'm sure a lot of people just don't want to listen to it for those reasons. But you know, I it does come back to the point you just made about Meta was something I kind of dug into there. I I think it's really underappreciated how big of a pivot SpaceX is trying to undergo, not not in space-based data centers, but what's happening on land terrestrially. And if you look at their playbook, um it does really lead you to what Meta could do um if if they come under pressure. If, you know, if if they aren't using their own capacity, that they're gonna have a significant revenue opportunity to rent it out as as the market stands today.

SPEAKER_00 30:12

Yeah, I mean, I as you had said in that episode, SpaceX is now an AI company and it's been an AI company for like a quarter at this point. Uh I'm glad, you know, I think you probably still don't own it. I don't either. They're now below their IPO price. That's not us taking a victory lap. We both deeply admire that company, their technology, everything they've done. There are just certain dynamics about the company's valuation and market right now that kept me on the sidelines. I think you as well. You talked about those in the last episode. It's a very low float. You have to underwrite it to a business that basically was non-existent six months ago, which is the AI business. And if we want to hold a company for years or decades, we want a little bit more time to see how that shakes out. But you have you have to admire the speed and magnitude of the direction shift here. I mean, no one else but Elon can pull that off. And you know, credit, credit to him for snatching victory from the jaws of defeat. Um, okay, on to some questions here. Uh, we have gold here from Denver saying, hi gents. I love how you explain AI and semi-supply chain in clear investing terms without too much engineering mumbo jumbo. Thank you very much. Two questions. For ACMR, that's ACM research, how do you assess the geopolitical risk factor today and what would make you change your mind? Says, I have heard you allude uh to the fact that you are not high on Bitcoin miners like DGXX pivoting to AI data center. Can you unpack that and what would make you change your mind? Thanks and keep up the strong work. Two great questions here. I'll just give a quick um tee up. For ACMR, one of the conversations we had had a year and a half ago that has some echoes here is ASML, right? This is the original bottleneck company. You know, uh two or three years ago, everyone was saying lithography machines and um ASML is the bottleneck company of all bottleneck companies. We did not invest in it, or you did not invest in the portfolio because they did not control their own destiny, right? They were this pond stuck between US and China. Now it's ended up being a great investment um since then, as some of those new bottlenecks have formed. Maybe people aren't as concerned about that dynamic any anymore. So I'm having a little bit of an uh you know a memory here of sitting out ASML for this similar dynamic. So I'm curious to get your thoughts there. And then this comment on DGXX pivoting to AI data centers. We just talked about meta pivoting to AI data centers, SpaceX pivoting to AI data centers. So is that a trend that we just like and any company that can do that is a buy? Or is it the case that like Bitcoin miners are a lesser version of that and for whatever reason, you know, a meta and a SpaceX are a better, have more potential to pivot?

SPEAKER_01 32:45

Yeah, so I as far as the second part of the question, I I bucketed a neo cloud question right after this. So we'll we'll we'll get to the neo clouds in just a minute. I think first let's let's unpack the ACMR question. Uh as far as geopolitical risk, the short answer is you can't you can't eliminate. Um some exposure is good. It's it's just it needs to be managed as part of an allocation. That's why you won't want to be 100% allocated into an Iwan semiconductor because you have a potential existential risk. Um, you you can have it as 5% position, 10% position, but you need to recognize that there is a level of existential risk that exists with companies and China. Um, some background, just quick ACM research. I mentioned twice now on this podcast. Uh, we recommended, I think, on the March 13th podcast. It's headquartered in the US, but it has a longtime operating in China. Uh, that's ACM Shanghai. You can actually buy that stock in Shanghai. I assume most of our listeners would rather just buy the US version. But what's nice about this is Austin, you and I were doing a lot of uh investing work back in 2010-2011 when a lot of the Chinese companies were coming to the US and they go under a structure called VIE, which is a variable interest entity, where you're not actually buying shares of the company itself. It's kind of a wonky workaround. ACM research is not that. You are actually buying the company. And and what's to like about this is I allude to it earlier, China has very different memory dynamics, and ACM Research, it's it's an equipment supplier, specifically with a lot of focus on the memory space. So, with China, first of all, the companies that are memory makers in China they have low market share to begin with. So being able to increase capacity is good for them because they're starting at like 3% market share. They only want to move up. They're not as worried about these top-level trends that uh Micron, SK Heinrich, Samsung are worried about. They're in share gain mode. Second, they're still not even meaning domestic supply before they can expand. And and Austin, one of the big headlines in the past week was companies like Google looking at whether or not they could use Chinese memory because it it it falls under ban list and other areas, but they're actively looking at this as a workaround. Third, China itself has made chips a strategic lever. They want to have domestic champions in memory. They've created several um stimulus bills around this, they've got more coming. So it's a space they're gonna continue pouring support into. And fourth, the fact that China has no access to EUV, or talked about ASML at the beginning of this, which is the maker of EUV, means they're gonna have to use workarounds, they're gonna have to get clever. They can't have these three nanometer chips. Um, so they're gonna have to use advanced packaging, they're gonna have to use things like hybrid bonding. That's a part of ACM's business as well. That's that's also a part of the business for companies like BE Semi, which is a past recommendation, would which has a lot of um exposure to China. So, Austin, when it comes to a company like ACM, how do I eliminate geopolitical risk? I can't. It's it's a part of the stock, it's it's baked into what it is, and where you need to manage your risk is at a portfolio level. So for this one, I think it is a good opportunity. Um, I I recommend it because you know I like the trends in memory and I wanted some exposure to China, and there's relatively limited options, and the stock's gone absolutely gangbusters. But the good news is as well, it's gone gangbusters. So have so has Micron and the other memory plays. And this one might have some trends that are um even stronger into 2027, 2028, and beyond. So when I recommend, I said this one's not gonna be a household name. Well, it's becoming a very popular name faster than I could have ever imagined, as a lot of investors have discovered it.

SPEAKER_00 36:59

And you know, just a word on these companies. As you said, you know, hey, it's good to go in eyes wide open. You can never eliminate geopolitical risk, it's just part of it. So size your position accordingly. It's not a perfect analogy, but anytime you're buying one of these, you know, ADR international companies, I know you and I are both fans of Mercado Libre, and you know, many years ago we invested in Baidu and companies like that. There's just a layer of risks that you and volatility, the extra risk and volatility that you have to accept geopolitically, but also currency risks. I mean, we we are not currency investors, we're not macroeconomic investors. A company can be executing phenomenally well in an underlying way, and just the dynamics of that country or that currency can mean that your position just doesn't perform that well. I think Mercado Libre is roughly flat over the last five years, despite fantastic execution. I don't say that as a reason to not invest in this stock, but just as an example of how many additional variables enter the mix. So position your size accordingly and you know, expect volatility in both directions, high and low, right? So, so so buckle up, um, but they can be really rewarding if you're able to hang on for the long run. On to our next question here. Uh, somebody said, uh, goaded podcast. Eric and Austin are legends, excellent content and depth. Something has been picking at me though. Everyone, including Eric, anthropic CEO, Gavin Baker. So he's the AI luminaries. Look at you on the Mount Rushmore of AI names, Eric. Um, uh, so all these people are talking about compute constraints. In this pod, we heard about Oracle, Core Weave, and how they're down. Why don't we ever hear about Nebbius? Land-based compute backed by Microsoft, Meta with prepayment, rising near all-time highs, cloud compute vertical stack built from the ground up for AI. We can't talk about SpaceX and Cursor without then talking about Nebbius. Eric, your thoughts would be great, and it's overdue. So, good transition from the last topic. And we had uh, so let's talk about neo clouds, let's talk about this opportunity from former crypto miners being able to now pivot to renting compute, SpaceX pivoting to renting commute compute, Meta's potential pivot to renting compute. We haven't not talked as much about neo clouds and Nebbius in particular. So, what are your thoughts here?

SPEAKER_01 39:06

Yeah, and I don't want this episode to just be Eric Reacts to news that happened seven hours ago at Market Open while you're listening to it later. But Austin, I I am looking more at this, these Apple headlines, um, and they are raising prices by as much as 25%. The stock is down uh 5% today, which shows the following uh knock-on impacts of this continuing uh memory situation. Well, it makes a select group of stocks winners. Um, it is probably a net negative for most portfolios with what's doing to widely held shares. Um, but to the question itself, which now that I've gotten everyone to forget what it actually was, it was about Nebbius. Um at the end of the day, I I would say I had a little bit of a fundamental misread on neo clouds. I thought we would have enough exposure through hyperscalers. We had just recommended when when Core Weave IPO'd and a lot of the attention was first shining on them. We had just recommended Amazon. So I I thought we would have exposure there without taking on the risk of smaller companies. Austin, as we've talked on this podcast a lot, what matters more the big idea or getting granular down to companies, and and I think in this market, the big idea matters a little bit more that if if if you start out with the overall concept that compute demand is going to remain insatiable and above expectations, um, the companies that are have progress to stand it up are going to see outsize demand in a significant way, um, which is what we believe, you know. So I guess you know, taking taking a little bit more wax at some of these companies, even though we knew they were speculative, would have been smart. We did talk about them in um, I think it was in uh August episode last year about some YOLO stocks. Um, but you know, Austin, I I I think on the extreme speculative end, we had been asked about Bitcoin companies in the last one. If if the question is, you know, okay, compute is incredibly supply constrained. Any company that's gonna be able to stand it up as an extremely valuable asset, you either look at a company trying to beg a whale, like Fermi America. This is a company that's trying to stand up a 17 gigawatt data center and I believe the Texas panhandle. As a side note, that's three times the power consumption of New York City, which is a crazy stat. Or, you know, the thing about the Bitcoin miners is often that um they're going to have access to GPU infrastructure from previously being Bitcoin miners, and a project that's near hundreds of megawatts can be a catalyst. So, one one person that's been doing this, we've talked about him in the past, Leo Ashenbrenner. He he runs situational awareness, he's become extremely famous because of his returns and the amount of money he's managing. He recently bought 5.6% of Nebius, which is what the question's about. So let's just talk a little bit about Nebbius. I I've long thought that they were one of the more high quality names in the space. Um it's on the surface of ridiculous valuation, Austin. If you looked it up, they're expecting $3.4 billion in revenue this year, but they're worth $65 billion, which sounds insane. But their scale is expected to amplify extremely rapidly from 3.4 this year uh to 45 billion or so by 2030. And Wells Fargo sees them at 7.5 gigawatts by that 2030 time frame. So if they can stay on target with some of these builds, you you could see a runway to $100 billion or or even more in revenue. Um the bottom line for Nebias is is why they should probably be the most um high quality of the Neo Clouds, is because they're the closest thing to a hyperscaler. They're the closest thing to a hyperscaler because a lot of these companies chasing these big deals, they're doing essentially bare metal. They get the deal done, they they ship the bare metal out to hyperscalers. That leads to projects that often have more tenuous ROI. Nebbius, they spent a lot of their initial time doing what hyperscalers did, building a diversified customer base, designing their own rack systems, creating their own software layer. So that puts them in a very good position. And the stocks up 200% year to date. Well, it's because they're doing the thing the market was asking for. They were being more conservative while they were building up these long-term advantages, but they're now kind of going for higher levels of growth. They're signing some of the deals that were mentioned with Microsoft and and Meta. So, Austin, the reality is um, you know, do I wish I'd recommend Nevius? Yeah, it's up 200% year to date. This is one that there's a lot of details to go into the finance. For these companies. There's a lot of where they're building, what the advantages of their technology are. And every time I've looked at doing a deeper dive across neo clouds, I'm a little intimidated. And it's sometimes, you know, you need to be able to pull the trigger before you have all the details. Because often singing and waiting for all the details allows an opportunity to pass you by. I think we've done, I've done a little bit of that with neo clouds. And maybe, maybe in the future as well, it's it's an opportunity we'll dive more into. You know, there are some risks happening right now as well. You talked about SpaceX ranning out their capacity, meta potentially becoming a neo cloud itself. So there may be some new competitive vectors in this space. Um, and it's really just gonna rest on is demand for compute higher than the supply? If so, they remain an extremely advanced environment. If not, there's gonna be pain throughout that sector.

SPEAKER_00 44:59

Um, really good responses there, and I'll just pause a beat on there. I I've also completely missed the neo-cloud moment. So, you know, we we can't win them all. Uh, but as you said, it's really complicated, and there's some things that just make this space a little they they it's probably a space I need to add some exposure to. Uh, but there's some dynamics now that just make it seem very sort of moment-in-time convenient. You know, one thing on Nebius in particular, it just also has a bizarre backstory, right? It was spun out of Yandex, and then I think it's only been a public company for like three or so years. It's located in the Netherlands. So it might have just sort of flown under the radar for a while, just because it's sort of got an obscure story, and it's and and and now that people have caught up to it, it's you know, that sort of pent-up demand is mooning it. Um, but beyond that, you know, obviously auction brenner loading up on it is a validating sign here. But one of the things that I worry about is what happens as we get to the next chip cycles with like Ruben and Feynman. And I and I know a lot of these neo clouds have locked up supply deals with NVIDIA, but as you get to these next cycles of more and more chips, you we know that like AMD and NVIDIA want their chips to be used. And at some point, if they're going to be selling these chips and they're still themselves limited, they're going to want to sell to companies like Meta or Alphabet that have incredible cash piles and can finance at cheaper rates and more reliably buy and deploy those chips than NeoClouds. Now, that's that's this is not me, this is not a sour grape somewhere. I completely misread the neo cloud opportunity here, but that dynamic is still what's kept me on the sidelines a little bit, where it's like it feels like this industry that sort of emerged from nothing that is benefit from this current supply moment, and that maybe three or four years in the future, we we sort of get on the other side of that as capacity comes online. And are are they really going to have enough of the leading edge leading edge chips to continue to rent compute at these premium rates, or are those leading edge chips going to go to the companies that are larger cash flow generating, have additional businesses from which they can finance debt off, and then they become the neo clouds here, or using these chips to train their own models. That's that that's been in my mind, and I haven't been able to get past it. But hey, credit to anybody who bought Nebius and looked through those clouds because you you did a great job and hope that uh there's more gains ahead for you. Let's look at this uh next question here. It says, as usual, Eric and Team, great podcast. We are getting a global sell-off in AI tech names today. Would love a discussion on the next podcast on some names that ran up after the gem of an episode on March 13th. Boy, thank you to our comment commenters this week. Lots of lots of praise. This feels good. Thank you guys. Uh, the gem of an episode on March 13th, and where good entry points would be for stocks like SMTC, air test systems, and STM, and what could be a brief buying opportunity. Eric, you went on a shopping spree then. I think you bought 13 stocks on that dip. They I think they've all done very well for you. Um and then this commenter said, I mentioned those stocks because I missed those great entry points. So basically, uh we're seeing a bit of a global sell-off. Are there additional stocks that you'd like to pick up on the weakness we're seeing today as I distill that question down?

SPEAKER_01 48:07

Yeah, I I think let's let's focus mostly on Semtech of that category. This is a stock that's up more than 100% since we recommend it. Well, a lot of other optic stuff has gotten a little bit more stuck in the mud, which which is fine. You know, you go up 500% in a year. It's it's fine to have some digestion period. And obviously, I talked about in my personal portfolio, I was coming back some allocation to optics stocks, expecting we we might see a moment like that. But SemTech, one of the good things about them is they went across every channel. What's what's led to some pullback in some of the optics names is we had talked about um two episodes ago that there's some reports that the next kind of um technology shift in optics, which is co-package optics, might be delayed. Semtech wins, whether it's co-package optics, whether it's transceivers, whether it's copper. As long as networking needs are getting higher, speeds increasing, semtech's products are gonna be under more demand. But we've also talked, Austin, about winners of this TPU, which is Google's chips versus Nvidia, and which companies are gonna have massive tailwinds from TPUs. Semtech is kind of a sneaky winner. I don't know if we've ever talked about from this angle before, but there's two versions of TPUs, one from Broadcom, another from MediaTech. Broadcom has long been handling Google's chips. They don't like how much control Broadcom has, the kinds of margins Broadcom can have. So to hedge their bets, Google's been working with Media Tech as another company in the space to basically get something where they can call the shots a little bit more. This media tech version, Austin, um, I've name checked him before. There's a deeply technical blog, name irrational analysis. He does great work. And one area that he's put a lot of work into is how much Semtech could be a massive winner from this media tech um TPU. Just this week, we saw some reports, they're gonna get huge amounts of volume, especially into 2028, um, which that's that's gonna potentially pour a significant tailwind into uh Semtech itself. So if if you're looking for a name that you haven't added, I would maybe, and something that's had recent catalysts even after the gains, I would maybe look towards SemTech. That's ticker SMTC. Um it's it's a stock that um once again, it's it's just no matter no matter where the wind blows, as long as networking's in more demand, they're extremely well positioned for it. And it looks like their total addressable market might be growing by a material amount from uh these media tech CPUs. So maybe of the ones that you missed, I I would I would look to add that name first at these prices.

SPEAKER_00 50:59

Uh great recommendation there. Now, Eric, you you are the guy who can see around corners and your pattern recognition is better than anybody I know. And yet I have to call you out. Every single time we do a podcast episode and you say this one's gonna be a short one, it's never a short one. Here we so we we've we've still got a couple questions here at the close. Do you want to do you want to do two more?

SPEAKER_01 51:19

No, no, actually, I I had put those at the bottom of the notes. I think what we're gonna do in the future is we do get a lot of questions. We would love to answer more, but you know, questions sometimes I I don't know if a good podcast is always an hour of questions. So we're probably gonna move more bonus QA into our YouTube channel. So if you're not subscribed to our YouTube channel, if you if you want more content, more questions, um we'll we'll increasingly begin putting more QA. So we'll we'll next week, I think what we're gonna do is we'll film our episode and we'll take some added QA. I I know a lot of people will write into questions at 24-7 Wall Street. We we deeply appreciate those. One problem, we get spammed so hard in that email address. It's truly hard to find what are questions. I dug through it. I I I found the ones I thought were real listener questions, and and we're gonna go through some back catalog there. So if you sent an email into questions at 24-7 Wall Street, um, you know, subscribe to our YouTube and and we're gonna get to a longer list of those questions um in you know the coming weeks.

SPEAKER_00 52:22

Yeah, definitely less spam on the YouTube, Spotify, and podcast comment section. So if there's a question you would like us to answer, please, please ask it in one of those channels instead. We'll do our best to answer them. We can never get to all of them, but uh, as you've said for years and as we continue to see manifest itself, this is a community. Some of our listeners are now our writers, and we love having them. And we love every one of you, as our listeners, who takes a moment to leave us a rating or review or to ask us a question in one of those channels. So, Eric, let's call that a wrap for today. You're on vacation. Uh, I'm in the mountains of New England, so let's go enjoy this beautiful summer weather. Thank you for your thoughts on Micron. Congratulations to all the Micron shareholders out there. Volatility is in the arena, and uh, hope all of our investors are buckled up for a continued exciting ride uh as AI investors. Until then, we'll see you next time. The AI Investor Podcast is for educational purposes only and should not be considered investment advice.