저자: Kevin · 외국인
요약 보기
SK하이닉스 IPO(ADR, 약 29억달러 규모)를 앞두고 메모리 반도체 과잉공급 우려를 제기. 메모리 업체들이 2026~2030년 대규모 증설을 진행 중이며 공급 부족이 해소될 것으로 봄. SK하이닉스는 PER 8배 수준이나 월스트리트 컨센서스상 성장률이 사실상 정체(연 1.6%)라 밸류 부담이 크다고 평가. 상장 직후 밈성 급등 가능성은 있으나 장기적으로는 메모리 공급 과잉으로 10년간 비관적. 단기 매매 기회는 있을 수 있으나 장기 보유는 비추천. 마이크론은 PEG 0.9로 싸보이나 성장 지속성 의문.
다듬은 글
Oh, oh, time to talk about another IPO that could potentially double on opening, and after that, um, we need to talk about this . But let's make sure we know. Well, now that we're talking about the big IPO coming up on Friday, we need to review some of our recent decisions. One of the latest IPO decisions was Cerebrus.
In fact, I made a video called “Cerebrus IPO Warning, ” and one of the things I said in it was this. The income will disappear over time. It quietly looks like a little pump and dump to me. Oh, I was a little pessimistic about the stock, let's put it this way. And not only was I a little pessimistic about the stock, but I think we were rightly pessimistic.
" Puppies" opened with a high of $386. They are currently at $198. You can see the full analysis we did last time. This is about half as low. And today we will analyze the next juicy IPO. But that's of course only after mentioning what we said about SpaceX. SpaceX, we said, “Hey, if you can buy this for $135, you get shares, okay.
Buy. If you had it before, fine." The venture capital firm I run also has shares in SpaceX before this, so we invested in that, we're still locked into that, okay? We made it clear, "Hey, if you've had this before , you're probably okay." But otherwise, the free float of the stock is so low, it is expected to burst and then leak out.
And, oh my God, look at that jerk and it will leak. Anyway, the bottom line is that IPOs are really interesting because they are a big money grab. And generally, IPOs are intentionally designed with very low free float to pump them up, and that's what brings us to the next uptick, I mean the next big player in this industry that's going IPO this week, and that's SK Hynix .
SK Hynix is now public on the Korean stock market, but it is not available for investment in the United States. So many people are very excited that SK Hynix is going public. More full-size AI memory , baby. Yes, exactly what we need. More memory. Because, after all, Micron is simply amazing. Uh , SanDisk, which was a spin- off from Western Digital, is simply amazing.
But the last few weeks have started to make people scratch their heads. And the timing, perhaps unsurprisingly , is not long after SpaceX's vacuum cleaner started sucking up a lot of money. It started to see some peaks in the memory sector, which may be temporary, but Micron is down 25% from its peak.
And SanDisk is also down about 24% from its peak. Probably because they realized we were about to have an IPO of SK Hynix, which would represent about 1/3 the size of SpaceX. So you might think, "Oh, it's just, you know, another small IPO." No, it's quite big. 1/3 the size of SpaceX, which currently holds the trophy for the largest IPO in history, that's a huge amount.
They're expected to raise over $29 billion, and a lot of people want to know, "Kevin, is this the memory stock that's finally worth getting into?" The rating is actually quite low." Well, let's find out. First, Micron also shows us that its valuation is actually not that high when compared to other companies.
You see, we usually like to evaluate companies by their earnings at the end of that year. I don't care about previous earnings or last year's earnings. What do we think the profits will be at the end of this year? And then, from this point on, what do we think the growth of the company will be? And then we can figure out how reasonable it is compared to other companies?
For example, Palantir, I think, is about twice as expensive as SpaceX . Which means that to me, stocks of companies like Palantir or some of these software companies are twice as juicy as stocks of companies like SpaceX. And of course, there are other companies that are even juicier. But the point of this video is to focus on memory.
And the Micron actually looks relatively cheap compared to this whole package. Micron potentially looks, I actually think so , looks cheaper than Palantir. Why is that? This is a memory game. Why is it trading cheaper than shares of companies like Palantir? Why are potentially stocks of companies like SanDisk trading cheaper if that's the case?
We'll look at the numbers for SK Hynix, sorry, in a moment. Micron shares are trading lower because markets are starting to factor in that growth is likely not going to be sustainable over the next 4 years. Assuming that current analyst expectations for Micron are correct, the next 4 years of growth for Micron will be fantastic , and you really don't need to worry about valuation because it's cheap.
But, you know what makes assumptions. The assumption makes... Okay, you know the joke. We don't need to go into this in detail. But Micron's valuation is currently at about a 0.9 PEG ratio, which is very cheap. This suggests that something like 2 1/2 X is built into its estimate. But that's assuming the growth rate doesn't turn negative.
And SK Hynix is about to show us the problem. Before I show you the SK Hynix problem, you need to look at this diagram. Bloomberg Intelligence recently published this chart that indicates that the memory shortage may be a surplus. Where we exceed the percentage sufficiency ratio to support production at 100% capacity.
Right? The whole point of memory...Wait. Let me back up for a moment , because those were a lot of big words. Companies that make chips employ a lot of really skilled workers. You know, when you see pictures of chip factories, they're all wearing their PPE. You know, not to be confused with PPP loans.
I'm talking about those white suits, masks and all that. Okay, these people have skills. They're not just, you know, pushers at the factory. They are in the factory, but they are skilled button makers. You want these people to stay employed because you don't want to spend money on retraining them. This applies to all chip companies, manufacturing companies.
Unlike Nvidia, which simply develops chips, which is something you can't even think about. But anyway, the problem is that if you do n't have enough orders to keep them busy, you send them home, and then they go get another job. So you try to maintain 100% plant utilization. A zero adequacy ratio in this case would actually be equivalent to 100% load.
As you can see right now, we are below the sufficiency level, which means we are very busy on the factory lines and we are not supplying the market with enough for its needs, right ? Ideally, you want supply to meet demand. Is n't that right? So you have 100% capacity, so 100% supply, 100% demand, perfect.
Everyone is happy , you have balance. If you have too much demand, therefore too little supply, too little supply, prices skyrocket. And that is exactly what is happening now. Prices have risen rapidly. And when I say " grew rapidly," I mean insanely. We're doing about three to four times as much with some of these revenues or these increases in numbers.
In fact, if you look at Micron's recent earnings, you would say that this company has the craziest pricing power in the world. Look at this. Sales of DRAM products increased by 67%, mainly because prices increased by about 60%. In other words, we shipped only a few percent more DRAM products compared to the previous quarter .
Low to mid single digits for DRAM and NAND, but we increased DRAM prices by 67% and we increased prices by about 60%. And we increased the prices of NAND products by about 85 percent. It's quarter by quarter. So, in one quarter, from the first to the second quarter, they raised prices by about 60-80%.
Compared to last year, they raised prices three to four times. This is a crazy price increase, and it reflects the lack of supply right now. The problem is that Bloomberg is already warning us that this probably won't last long. But we already knew that. Why did we already know this? We already knew this because these companies are coming out with a lot of additional construction.
And that's a bit problematic. In fact, yesterday we did a pretty big video about some of the build-outs that are happening with Micron. And it was related to the inflation versus deflation argument that we were making, Katie Wood versus TS Lombard. And inside of it, we actually mentioned that Micron will probably increase its offerings by about 50-50%, or 100%.
So, somewhere between half and doubling. Um, that's because they're essentially building all these new plants, which are expected to be completed in mid-27 in Idaho, in '27 in Singapore, late '28 in Idaho, in the second half of '28 in Singapore, in 2028 in Taiwan, in 2030 for the first of the four in New York.
And this is Micron. Regarding SK Hynix, you might think, “Well, maybe they’re different. Maybe they're not going to create that much supply. Maybe they will limit supply and stay busy.” Um, no. SK Hynix is no different. They also expect to double capacity within the next 5 years. In fact, management in this case is literally telling us that they expect to double management or double their production capacity within the next 5 years.
They are doing this with a plant in Korea, which is expected to be completed in 2033. So this one still has some time. They just started. They start this next year. They have n't started this one yet. Um, they expect one of their expansions to be successfully completed, just finished. The M 15X expansion has just been completed in the first quarter of 2026.
Advanced production or, rather, packaging at the end of 2027. The NAND flash memory factory will be built in the first half of 2029. The advanced packaging plant in Indiana will be built in the first half of 2028. And then another complex in the southwest region will soon be announced later. So, again, capacity is expected to double over the next 5 years.
Why is this important? We have to talk about growing inventories and growing supplies in the production of memory chips. What are we doing, Gavin? What the hell? Understand this. Whenever you look at the risk factors of a memory company, you'll almost certainly find that this is the number one risk factor they list in the summary, and I'll see if I can find it.
Summary of risk factors. Risk Factor Summary and I can't find it right now. But, uh, I'll find him. The number one risk factor they usually give you is, "Hey, memory." That's it, I understand. "it's really cyclical, and it's almost always the first one they give you." The memory semiconductor industry is prone to cyclical fluctuations.
In 2023, Micron lost billions of dollars because it sold memory chips at a price lower than the cost of producing them. Why did you do that? Because you don't want to lose those workers in white suits. So you're going to keep the factory lines running and producing memory chips, even if you lose money on every one you print.
Seems crazy, doesn't it? Do we think SK Hynix was different? This is an important question. Was SK Hynix different? No. They also switched to negative cost of goods sold in 2023. Here is their profit and loss statement, which shows that they spent 33 trillion won, generating 32 trillion dollars in revenue.
Well, that's exactly what you need. So, this shows the cyclical nature of the memory business. And perhaps SK Hynix's biggest problem is that not only are their margins worse than Micron's, but they also have about 60% gross margin, which is pretty good. You know, don't get me wrong. 60% earn for every $100 in gross profit.
That's pretty good. You are at the level of gross profit of Apple and Microsoft. Not quite on Nvidia's level, but they're good. Not net profit, but gross profit. Er, Micron has a good margin. Micron's latest margin was 84.5%, which means that this company has 24.5 percentage points more margin than SK Hynix.
So Micron actually has better performance than SK Hynix in terms of margins. Quite impressive. As for forecasts , and this is a big problem , in terms of forecasts, SK Hynix markets , unlike Micron, are already pointing to actually steady growth over the next 4 years. So, even though it's an ADR, which is an American Depositary Receipt , which is just a fancy way of saying it's a foreign listing.
And this is where things get a little complicated, as the projected earnings per share (EPS) for the year-end is technically $207, but the ADRs are worth only 1/10 of SK Hynix shares in Korea. So you need to divide earnings by 10. They are projected to buy back a share at year-end for $20.75 in earnings, based on current Wall Street estimates, which means if they IPO at around $166 a share, that means they are only trading at about eight times earnings.
Somewhere between six and eight times the profit, they'll probably hit the market. But the growth rates that Wall Street already assumes for them remain virtually unchanged. We expect growth of 46.1% next year , 7.7% the year after that, minus 30.3% the year after that, and minus 20.21.9% after that.
That leaves me with 1.6% growth over the next 4 years. So if I have 1.6% growth and a P/E ratio of 8, or, sorry, I have a P/E ratio of 8, if I take 8 and divide it by 1%, I actually get a P/E ratio of 8, which is extremely expensive. Does this mean that stocks will fall because SpaceX was also extremely expensive?
No. Because only about 2.5% of the outstanding shares will be put up for sale. Which is really not enough. SpaceX was worth about 4.1%. Uh, SK Hynix has a core corporate that will own about 25%, which goes to the company out of that 2.5%, and if I take 2.5 and divide it by 0.795, that effectively means that's 3.14% of the shares floating.
This is the floating point number pi. Er , a mathematical joke. Sorry, that's so stupid. In any case, about 3.4%, so they will have less free float than SpaceX. This means you remember what happened to SpaceX? Of course, we remember what happened with SpaceX. They invested because they only had 4.1% of the shares outstanding.
It's very small, and people bought shares. This is not SpaceX. This is a memory manufacturing company. So, to be fair , it will probably trade in unison with Micron and SanDisk. A lot of money is invested in Micron and SanDisk , and some of those people will want to allocate some of their money to SK Hynix.
So they will probably sell some of Micron and SK Hynix. And that's why stocks are falling, because momentum has slowed and traders are getting out. So, momentum slows down, traders exit as people try to diversify, and what happens is, you know, the momentum movement reverses. Or the markets finally realize, “Oh, [ __ ], we probably won’t grow these companies as much because we’re massively expanding capacity and demand is n’t going to go to infinity.
Everyone should know that after Elon Musk said, “ Usually when a company says demand is infinite, or an industry says demand is infinite, that’s a major signal. Elon Musk told us at the peak of Tesla pricing in 2021, Tesla's price point has peaked. He told us, “The demand for Tesla is practically endless.” Which turned out to be completely wrong.
And now we're hearing that memory companies, in particular, well, Sand... Excuse me, let me explain. SK Hynix doesn't see peak demand until 2033, they say, and Micron says they don't believe supply will ever catch up with demand. In other words, even if they, you know, massively increase capacity, they don't think there's going to be a tipping point where supply finally meets demand, which is kind of a way of saying, "Oh, there's more demand than we can ever produce." “It’s not as wild as saying that the demand for memory is limitless, but it’s probably the closest we can come, because I have to give these people credit.
These Koreans are a little more conservative than Elon Musk. Okay, anyway. Uh, they expect to raise $28 billion , so they probably have green shoes, they'll probably expand it a little bit as well . The company is obviously spending money to expand its factories. So they have free cash flow of 26 trillion won, which is about $17 billion.
That's pretty good. So their cash flow in 2025 was, you know, about 2/3 of what they were going to raise in the IPO. So they will make good money from this IPO. They will indeed receive a certain flow of liquidity. Especially considering that last year they actually borrowed funds to pay off their dividends.
Not much, but they borrow money. They paid off the debt in 2024, which was quite... » Okay. Although they had to , they effectively just paid off what they received in 2023. Because in 2023 their net cash flow was negative. They lost 4 trillion won. So they are still borrowing. It would be good to see them pay off a little more.
Although their balance is actually not that bad. They have about 34 short-term loans, 18 accounts receivable. Their, ahem, short-term debt is about 37.3 and 18 long-term loans. So this actually puts them on an even balance. So they are doing well. The balance is quite stable. They're actually going to strengthen it even more in this IPO.
Their profitability is normal. Um, stagnant growth is the biggest risk factor here. You have a jury waiver included in the ADR, which is an interesting thing that you're starting to see more and more often. Elon also tried to stuff it into, ahem, the SpaceX IPO. We'll see if this, um, holds up. Um, of course, the cyclical warning is number one.
We already know that. As for the company itself, they have a decent market share. In DRAM, which includes high-speed memory, they rank second in revenue with a market share of 29.1%, and in the high-speed memory market, they rank first in the world. Oh, this is about our, um, fancy Nvidia chips with a market share of 56.4% in the first quarter of 2026 .
And they are the second largest NAND flash memory supplier by revenue with a market share of 18.5%. So not bad. So, in summary, what do we think about this company? Well, this is another case where memory liquidity sucks into the pie. And, in my opinion, it all comes down to the market's ability to find new leadership now.
That's why the NASDAQ hasn't hit its highs right now, because we've been hit by a huge liquidity drain from SpaceX, which has taken so much money out of the market's ability to sustain momentum. We need new leadership. Maybe it will be memory, but memory on the daily chart SanDisk is still...a transition to a new level, because Micron is transitioning to a new level, and SK Hynix is more of a case of we need to distribute the wealth that is there and it's assuming an even bigger transition to a new level.
Not ideal. Obviously there's a bit of a push today because Nvidia had some bad news last weekend: " We're going to delay the release of one of our server racks potentially for a whole year." “This is a big plus for AMD, and that’s why AMD stock is up 9%, which is what we talked about ad nauseam. That's why, if you're not already a member of the "Meet Kevin" program, I think you're missing out on a lot of opportunities by not participating in these live broadcasts every morning.
Nevertheless, this is my opinion on SK Hynix. Bottom line for SK Hynix: uh , the valuation is very high because Wall Street has already factored in flat growth. Micron probably looks like it has a low or cheap valuation because Wall Street has n't yet factored in this growth transition. But, uh, that's the expectation of Wall Street analysts.
The market probably is. The market says, "Yes, but we don't want to pay a higher PEG ratio for Micron because it will probably go under, just like SK Hynix." "So the actual stock predictions, the stocks will obviously say, ' Wow, SK Hynix doubled on the first day or something.'" “They always do that, you know, price the stock at $79 and then IPO it for $150, $160 or whatever.
Okay, great. It will be next to SpaceX. What will happen after this? It has meme-making potential. Meme-making potential, if it comes out, it will drag Micron and SanDisk down with it. The problem is that once that initial momentum wears off, they will probably all fall . Will another part of the market take the lead?
This is the hope that the indices can continue to rise. We'll see, because even if SpaceX joined the Nasdaq 100 tomorrow, only less than 1% , you know, market cap would be a part of the whole of SpaceX, which would represent less than 1% of the Nasdaq 100. We have a lot of liquidity issues, including lock-up expirations.
And the same will be true for SK Hynix in the future. So buckle up. " I wouldn't buy this stock after the IPO unless I was trying to quickly turn it around and free the stock from meme trading." But I wouldn't keep it at those levels, especially given the growth projections. And I personally think that in the long run we will have a significant surplus of memory.
This makes me very pessimistic about these stocks over the next 10 years. A trading opportunity in the short term? Good. In the long term? Not with a 10- foot pole. This is my opinion. Thank you .
Comments 0