Why Memory Stocks Are Falling

Hey guys, welcome back. As I said last night in the video, today was going to test our conviction. But in actuality, the last couple weeks, it probably tested our conviction because we have seen since June 25th, a staggering $1.5 trillion dollar in market value get completely erased from the semiconductor sector. Micron alone has lost billions of dollars in market cap. It vanished in a flash. And we see the stock right now trading back in the low 900s today, slamming down from where we saw it at the late at the high at the late end of June. With Micron Sandis and the Roundhill Memory ETF all falling more than 20% from their recent peaks, the mainstream media right now is screaming that the red high redhot AI memory trade has officially cracked and plunged into a technical bare market. Amateur retail traders right now are staring at their screens and they are uttering panic. They're asking, "Is the AI trade over? Was this the absolute top?" But on this channel, we definitely bypass the surface level news. We look for the signals in the market. We look at the raw fundamental data. The correction right now, if you do look at the raw fundamental data, it makes absolutely no sense. Just about 24 hours ago, Samsung dropped preliminary Q2 earnings onto the street and their operating profit was so staggering that it didn't even look real. An 1,800% increase year-over-year, beating Wall Street's consensus and outearing Nvidia, out earning Nvidia's most recent record-breaking quarterly profit of 53.5 billion. Samsung made more money in the last three months than it did in the previous two years combined. The fundamentals did not collapse this week. So why did the stocks why did Samsung slump 7% Asian trading and triggered an emergency circuit breakers in the KOSPI index? One institutional portfolio manager summed it up perfectly today when I heard it. He said, "Great companies can become poor investments when expectations become impossible to satisfy." That is exactly what I think has happened this week. If you want to navigate the tape, if you want to navigate this market without losing your shirt overall, you have to realize one thing. The market isn't repricing the business today. It's repricing the expectations. the structural shortages or the bottlenecks, they haven't changed, guys. But the market's patience apparently has changed. And that's what we're going to dive into today's video. I'm going to look at a few factors that came out today that I think also will help us understand why we're seeing these stocks collapse the way they did today. But there is optimism at the end of the video. So, make sure you make it all the way through. I'm going to give you my checklist, what I'm watching over the next two weeks, and we're going to walk through some stories that hit the tape today that also, I think, as I mentioned, can explain what we are seeing right now across the memory sector. But if you haven't done it already, make sure you hit like. Make sure you hit subscribe to the channel, guys. Let's get started. Let's dive into the data. First, let's start off by taking a look on an article that I think will prove around the expectations are getting impossible to satisfy. You can see this article in Barons that Samsung's huge earnings beat wasn't enough for the stock and what it may mean for Taiwan Semi. So, let me scroll down real quickly. I want to call call out a couple of pieces here. You can see the report what it mentioned that Samsung's preliminary earnings that it operating profits came to 89.4 trillion which is about just 58.5 billion for the three months ending June. That would be a 56% jump over the previous uh re quarter's record profit and a 19fold increase from the same period last year. Quarterly revenue is forecast to have more than double from the prior year to a record 171 trillion. They go on to say that these figures were not disappointing. In fact, Samsung's profit outstripped Wall Street's expectations, but a 382% runup in the past 12 months have left Korean investors looking for excuses to take profits. That's part of the story. But part of this article I wanted to call out, it says right here, if you look at Missou's Jordan Klein, what he says, by the look of global semis and the momentum factor, you would think that Samsung missed their quarterly and guided down. It is what he calls momentum in reversed and we're going to see it across the well-owned semih hardware sector of tech. Klein added. He goes on to say that Samsung stock sharp fall and global response also present a difficult setup for Taiwan semiconductor manufacturing and ASML holding which will both report earnings next week. The bar is set high for those two companies just like we saw with Samsung as well. So there is a risk for sure that beats and raises get sold in those stocks. So just a warning to all of you holders for Taiwan and ASML out there as well. But I want to jump over and talk about something that I think is important that we're seeing in today's market that help explains it a little bit and is this concept around multiple compressions. Because to understand how an asset can plunge 20 20% or more while its underlying business gets historically stronger, you have to understand the fundamental market mechanic called multiple compression. So let me do a quick kind of master class we'll call it on how institutional desks actually price these names overall. So a very simple graphic on your screen. You can see that a stock price is essentially determined by two variables. how much money the company makes and the multiple or the premium investors are willing to pay for those earnings. Two weeks ago, the sentiment overall in the tech sector and the memory sector was at a fever pitch. Investors expected absolute perfection from this upcoming earnings earnings cycle. And because of that euphoria, they were willing to pay a massive forward multiple for every single dollar of future memory profit, assuming zero friction and infinite scaling. Fast forward to today, the structural data hasn't changed underneath. The fundamentals remain rock solid. The earning power is still there. But because global macro jitters have crept into the tape, investors decided that they are only willing to pay a smaller multiple for those same exact earnings. Nothing fundamentally broke inside the Silicon Fab. The multiple just got compressed. And as we analyze the rest of today's tape, keep this framework in mind. The market doesn't need bad fundamentals to sell a stock. It only needs expectations that become impossible to satisfy. And that's kind of the setup we see right now in the market. It's what we saw with Samsung. You have to remember that Micron's earnings just blew it out of the water. It kind of shocked everyone. We set this impossible bar for Samsung. Even though they delivered f like outstanding phenomenal earnings, those instant those whisper numbers kept creeping higher and higher and it got set to an impossible number. The fear is we may see that with TSMC and ASML next week as well. So let's jump over and take a look at the first article that I wanted to cover on today's video. You can see this in Morning Star. It says Micron stock falls as investors wonder if the memory market is near the top. And this is there's two pieces of evidence in here that I want to go go through through this article because as we established yesterday, Samsung's operating profit was historically strong yet the market punished the stock. There are two structural reasons behind this paradox. Overall, the first one is from Jeff Equity, Jeff Equity uh trading analyst Jeffrey Favuza. As you can see here, there's a couple pieces I want to call out. What he said was that most investors feedback continues to point to a skittish tape, elevated expectations bar, and underwhelming results, at least relative to relative to the major appreciation in Samsung stock. But he also pointed out right below that that the preliminary results did not offer investors insight into performance by business division and that those details will come in the full report later this month. So essentially Samsung's preliminary report as you guys know offered zero insight into the HBM to the semiconductor to the memory chip side of the business. So Wall Street was starved for granular data on the high bandwidth memory yields the shipping volumes. Institutional desks absolutely hate when they have an information vacuum. They have to wait until the full report drops later this month. I think on the 30th overall. I mentioned the 23rd yesterday, but I'm seeing results today that may be the 30th. So, they have to wait to the end of this month to get those answers. The second piece is all around it comes down here. What we also see from Richard uh Richard Windsor from Free Radio Free Mobile, the founder Richard Windsor, he called out everything around the whisper numbers. essentially he said that while Samsung's operating profits were massive you can see they were 19 times compared to the previous year they were only just ahead of the expectations and when you compare that to what Micron where Micron blew expectations away this set the stage for you like just a bit of a kind of a a disappointment I would say from what we thought we would see with Samsung let's be honest though Samsung's operating profits were massive but they were only a hair ahead of the highest whisper numbers circulating around the institutional trading desks. So, remember our framework that we talked about, Micron blew away, they completely blew their expectations out of the water last month. Samsung merely cleared a bar that was already set skyhigh to a near impossible level. And because revenue came in just a fraction below the most aggressive whisper numbers, algorithmic trading programs, they triggered immediate selling. institutional players use this massive historical fundamental milestone as a legal liquidity window to lock in generational profits. So there's definitely some profit taking that's happening across this sector right now. No question about it. But they didn't sell because the thesis failed. They sold because the good news was already 100% priced in. There's another piece that Richard mentions down here. It says in his view the answers, you know, questions of have we reached the top essentially. He said the answer will be determined by continued demand for tokens or the pieces of data processed and generated by the AI models as prices for them continue to rise in today's market. As long as demand continues to greatly outstrip supply, then we do not have a problem and memory companies will continue to rise and compute providers will continue to print cash according to uh according to Richard Windsor. overall. Now, right there, let's lay out the absolute strongest, most legitimate bare argument against this entire cycle. We need to look at this objectively on the channel. The bears are asking a valid question in my opinion. The question is, when do the hyperscalers run out of money? When do they run out of margin? Because Microsoft, Google, Meta, and Amazon, as you know, are on track to spend an unprecedented billions on capex this year. But hyperscalers eventually will face strict return on invested capital limits. The infrastructure must eventually generate massive tangible software and services revenue to justify their spends. If cloud customers realize they aren't monetizing AI fast enough and decide to temporarily slow down their hardware deployments because memory and compute costs are just too high, memory demand could moderate significantly. It is a sophisticated argument that the Bears are laying out, but there is a counter case that the Bears are missing. And even Ted calls it out in the article. Let me pull that back up and pull show you the highlight that Ted calls out here. It's right here in this in this paragraph. Ted pointed out that to massive spending by the hyperscalers and said that the return on invested capital degrades to the point where they cannot pass on component component prices. the memory cycle could pause until supply reaches par with demand. That's not expected until 2028 or 2029, he added. This is the piece that the bears in my opinion are missing in their argument overall. Probably their strongest argument is the fact that this monstrous demand will be will demand will continue to outstrip supply into 28 into 2029. The physical architecture of the next generation large language models means that the demand for memory density is scaling faster than compute itself. Hyperscalers cannot afford to stop buying or to blink first because doing so means losing the structural infrastructure race. We aren't at peak capacity. We are simply navigating the shortterm friction or hyperinflated hardware stack right now in the market. And the third piece of evidence I want to call out from today's tape was something that came out. You probably saw it yourself on Barren, but you can see it's an evidence. It's a piece of evidence related to the Chinese manufacturer. So Reuters reported that Chinese AI champion Deepseek is quietly developing its own proprietary AI chip. If we scroll down, you can see this ear an early fall uh deepened after Reuters reported that Chinese artificial intelligence company Deepseek was developing its own AI ch AI chip which could leave it less dependent on Nvidia hardware citing people familiar with the matter. The chip is designed for inference running AI models rather than training according to Reuters. We know that Nvidia gets less and less of the revenue from China overall. So that is not an issue for the company. You can see it here. Nvidia won't necessarily lose significant amount of business from deepseek. It uses its own if it uses its own inference chip chips with China becoming a smaller and smaller portion of its revenue. But the news is just another example of the competitive pressures that Nvidia faces in the market today. And guys, what does this mean for the memory sector? Why did I call that this out overall? Because I think we when you see a Chinese startup building its own custom memory chip, you know, you start to factor it in with the other stories and I think this is where the bears are going with this because when you see open AI introduces its custom inference chip, the Jalapeno co-design with Broadcom, you see Anthropic, it it's exploring its own custom silicon and now DeepSeek is designing custom chips explicitly optimized for AI inference. The bears are trying to argue or in my opinion will try to argue that these massive AI software labs when they build their own custom silicon they will cut out traditional hardware gatekeepers and tank industry margins overall. But independent semiconductor analysts quickly point out a flaw in that logic. You know designing a custom inference chip it takes years heavy capital and immense physical infrastruct and an immense physical infrastructure loop. More importantly, custom silicon does not bypass the laws of physics. Ironically though, what we have seen and Apple showed it earlier with their relationship with Broadcom because we were worried they would try to use their own custom chips. What we're seeing ironically is the more software companies try to optimize their own proprietary custom silicon architectures, the more physical advanced packaging facilities and high performance memory footprints footprints are required to actually run them. Deepseek isn't destroying the memory fortress. They are knocking on the door asking for specialized highdensity memory allocation. Custom silicon chips still need memory and they need it in massive quantities overall. So that was another article I think that the bears would use or could push down this sector and just say we've reached its top. But in my opinion, I think it just makes the memory sector even more fundamentally strong. But I want to jump over and take a look at how the market actually did perform. So let's pull that up. Let's look at a couple different things. First, we'll start off with just kind of the major indexes. And if you look at the broader macro map that you see here, where the actually money went today, you'll see this wasn't a marketwide crash. The S&P 500 lost only about 0.5. 045%. It closed just over 7500. The Dow Jones Industrial Average barely budged. It lost I think uh 0.25. 25% after hitting a fresh all-time intraday record high earlier in the session. In fact, the advancing if you look at the advancing stocks or advancing uh uh stocks in the S&P, they actually outnumbered the falling ones by a 1.3 to one ratio today. So, this was not a marketwide crash. In fact, if we look around the NASDAQ 100, you can see there were there was a lot of red on the board, but we also see a lot of green. If you look at those green patches, you see Microsoft, Google, Meta, Amazon, they all held their ground and closed green. In fact, Zachary Hill today from Horizon Investments noted that the real story over the last few weeks is a healthy mechanical sector rotation. The AI hardware names have had a blistering historic run. He noted in the first half of this year, expectations got so high that they almost are impossible to beat in the near term. money is temporarily flowing out of the high-flying semiconductor names and rotating into defensive highquality value sectors that lag behind. If we look at the physical tape from Micron today, you can see it on the board. Let me close this down. And how it just gapped down. Sorry about that. How it gapped down from the opening today. It's what we expected. We said we were going to see a choppy start to the morning. the Samsung panic. It just triggered an automated stop losses kind of algorithmic selling causing this kind of cascading waterfall that you see in the first hour or so of the session. It actually pushed Micron down to a low intraday low here of 8.91 a share. Just over just over 8.9150 uh a share for the stock. But look what happened at the VA. Look what happened at that point when it hit the low. As soon as it got dipped down below 900, a massive wall of institutional buying stepped in to lift it nearly $47 off the low and it closed just about 93839, excuse me, 939 today. That late day reversal that you see, those last few candles right here, this late day reversal and that kind of market on close, that last five minute candle, it tells you everything you need to know. The weak overleveraged retail hands were completely sh they shake shooken out of the market this morning while the long-term institutional value hunters treated the pullback as an accumulation window overall. And let's just see what Micron's doing in the after hours. You can see on the screen what they did today. Closed just about 939 down $46 just over a little over four and a half% but if you look in the after hours right now we're seeing a bit of a positive response. We're up $9.50 50 cents fluctuating right around that 1% increase right now trading about 948 across the board. If we look at the other stocks we see SanDisk is up almost $50 in the after hours right now. Western Digital is up about just under $5 and Seagate down here up almost just over $12.50 one and a half%. So we are seeing some capital rotate back into the memory sector right here in the after hours. But we have to ask ourselves what is the final verdict? You know, are we crumbling? Is the memory sector? Is the memory fortress crumbling under the weight of a newly minted technical bare market? My honest answer is absolutely not. Today was a classic mechanical sell the news liquidity flush. Probably a little bit of profit taking as well. The structural shortage across the global supply chain is completely real and the capacity bottlenecks remain locked in. The structural shortage hasn't changed. Just the market patience, the market's patience right now, it looks like is what has changed. So, right now, every investor is asking the same exact question. What happens next? Instead of guessing or panicking, there are three tactical things I'm watching over the next two weeks just to know when the true bottom is in. And let me show you those three. I'll pull them up on the you can see the three that I want to walk through seeing your screen. First is the Samsung's full earnings. Honestly, we have to get a look at that divisional business. We need to see the exact divisional breakdown, the granular reality. We need to get a forward guidance from them as well just to validate what we see in the market. That'll help clear up the data vacuum. So, I think that's on July 30th. The second piece is the DRAM, the NAN contract pricing updates. We need to verify if trends force if trend force projected 13 to 18% Q3 DRAM price hikes are tracking to reality proving that physical demand that that will prove that physical demand continues to outpace supply. And then the third one is just key technical levels and institutional volume. I want to see if the major players like Micron can reclaim and firmly hold those psychological support levels. In this case, it would be the thousand level for Micron. Can we hold that on heavy institutional volume? So, this Friday, as you guys know, SKHix is listing in the US market. Their ADA, their ADR officially begins on Friday, and they'll begin trading on the NASDAQ. It arrives as a it'll arrive as the ultimate lititness test of institutional sentiment. The question no longer is whether these companies will print record profits. We've seen Micron, Samsung, all print record profits. Samsung just proved they dropped historic profits. They beat the highest quarterly earnings that we've seen it from Nvidia in the past with 53.5 billion. They've outpassed it historical quarter. Probably it is the largest for a private company across the board. So the question now though is just how fast expectations can normalize for the sector. Just remember, the smartest money in the world doesn't run from a liquidity flush. They use the transition into a technical bare market to accumulate assets inside of a structural for fortress. They accumulate these assets at discounted prices. So guys, keep your eyes on the data. Ignore the noise in the market. There's tons of it right now. Ignore it. Follow the true signals overall. Keep listening to this channel. I'm going to drop a video each and every day after the close of the market. Hope you enjoyed this one and I'll see you in the next one.