저자: JimCramer · 외국인
요약 보기
메모리 반도체(샌디스크·씨게이트·웨스턴디지털·마이크론)는 데이터센터 수요와 장기공급계약으로 마진이 구조적으로 개선됐고, 자사주매입 확대와 낮은 PER로 아직 상승 여력이 있다고 주장. 마이크론은 자선신탁을 통해 매수했으며 추가 상승 가능성을 언급. 엔비디아는 밸류에이션 우려가 있지만 펀더멘털 훼손은 아니라며, 미중 AI 경쟁과 중국 화웨이 생태계 확산 우려를 언급.
다듬은 글
My job is not only to entertain, but also to explain. So call me at 1-800-743-CNBC. Tweet me: Jim Cramer. This market surprises me almost every day. This weekend, in my regular column for CNBC Investing Club members, I wondered why the same memory stocks kept rising. Many have stayed away because they believe they have already missed these movements.
So I have to ask: what if not? So, today the Dow index fell by 273 points, or 0.52%, and the Nasdaq fell by 0.32%. This is actually a pretty unpleasant session, mostly due to rising oil prices and interest rates. I want to explain how these great movements happen and why they may not be over yet. The focus is on companies that produce memory and storage for data centers.
SanDisk has grown by 653% since the beginning of the year. Seagate jumped 261%. Micron grew by 254 %, and Western Digital by 211%. These figures are incredible and certainly crazy. These names in the field of memory have become risky. For decades, they have been risky cyclical stocks . There were periods of incredible growth, followed by terrible declines that wiped out all the gains and even more.
I have always found the management teams in this business to be financially suicidal, because they built capacity in good times while sowing the seeds of their own downfall. I've always hesitated to recommend them to you because of these crazy up and down cycles. But data centers appeared, and the world said that everything had changed.
The demand for memory chips now seems endless. All of these companies have learned their collective lesson. This time they are not rebuilding capacity. They want the deficit to continue. Memory is so scarce that even Elon Musk said it has become the main bottleneck for the development of data centers.
Meanwhile, memory manufacturers have shifted to new business models with long-term agreements that guarantee huge gross margins for years to come. They actually only produce to order. They raised prices so much that Apple complained: it makes your new phone more expensive . Can we trust these companies and their stocks?
Now three of them are buying up a huge amount of their own shares on the market. SanDisk is conducting a $15.5 billion buyback. Seagate announced a $ 5 billion buyback, and Western Digital approved a $ 4 billion buyback program this year . They take this money and give it to you, the shareholders, instead of investing in new capacity.
Only Micron does not conduct share buybacks. Capricious. Previously, these were smaller- cap stocks, but in the two years since data center construction began in earnest, everything has changed. I want you to hear about these impressive gains. When SanDisk spun off from Western Digital 18 months ago, it was a $7 billion company.
Now it is a company with a capitalization of 262 billion dollars. Seagate's market capitalization grew from $21 billion to $225 billion over the same period . House of pleasure. Western Digital was worth $17 billion, now it's worth $185 billion . Only Micron was big— $106 billion, but now it's a trillionaire.
It was easy. And you wonder why I think you should own individual stocks alongside your index funds. Now you might think that all these memory company stocks have gone up too much. This is the consensus I hear. But their margins, which are huge, are largely locked in for years to come thanks to these deals.
Seagate's gross margin was 52% in this latest quarter, a record. Last year it was 37%. Western Digital's gross margin was 54%, compared to 41% a year ago. This is incredible stuff, people. SanDisk's gross margin increased from 26% to 85% in one year. Micron reached 85% compared to 39%. These are monumental numbers, staggering, unheard of.
I remember if Intel could ever reach a gross margin of 64%, it would be amazing. Now listen to what these guys are doing. It is difficult for people to comprehend this endless demand for memory. They don't believe that customer agreements will be fulfilled. Their biggest fear is that they think Samsung will break ranks and build factories to flood the market with supply.
Of course, it could happen, but not anytime soon. It takes a lot of time to build factories. And look, that's why Micron is selling for just seven of its projected fiscal 2027 earnings. SanDisk is trading at eight times its earnings. Now, if we look at fiscal year 2028, Western Digital has a ratio of 16, and Seagate has a ratio of 17.
Overall, these stocks are cheaper than the market because no one believes them. No one. Well, maybe some. People make fortunes. Last week, we purchased a position in Micron for a charitable trust because I truly believe that this industry has changed for the better. That said, I admit that I'm uncomfortable buying stocks that have risen so much.
Fortunately, we purchased them at a significant discount, as they were trading at the prices of Korean Samsung and SK Hynix, not Seagate, Western Digital, or Sandisk. The timing was exceptional. Club members should be satisfied. Although at the time I thought I was risking everything, and of course I was worried that the price might go down further.
You see, I don't know anyone who would be completely comfortable buying stocks at such a high price. I consider Micron more of a growth stock than others. So I can convince myself that maybe it wasn't just the production restrictions that caused the price to rise. But we all know that this memory deficit has to end someday, right?
I don't deny that. However, I have doubts about the timing of any downturn. I think Micron could double again before the boom ends. Some of you are talking about data center slowdowns. I know it's a risky statement, but these long-term deals are impressive for Micron's margins. Hey, and by the way, listen here.
We're going to visit Micron later this week to interview CEO Sanjay Mehrotra at their research and development facility in Idaho. "Fab" is short for " fabulous". Not really, but I like it. I think it will be informative, don't miss it. I can't wait. It has to be...It's going to be incredible. Of course, we know I need steel-toed boots.
I think I have them . market and AI analysis. If you want more in-depth stock analysis, you can also join our channel membership and get premium analysis created specifically for members. Now let's move on to the main topic. I recently came across a headline saying that Nvidia's valuation was giving off alarm bells.
And honestly, this headline immediately catches your attention because Nvidia is no longer just another stock. . But when you actually look beyond that headline, two main factors that cause concern. The first is the relationship between Nvidia's valuation and earnings expectations. Nvidia's business is growing extremely quickly, but analysts don't always raise their earnings forecasts at the same rate.
So, the stock may appear expensive on certain valuation charts even if the earnings forecast continues to rise. This discussion matters. The valuation schedule alone does not necessarily indicate that a business is overvalued. The second factor is the ownership structure. Nvidia shares are already held by a huge number of institutional investors.
This means that some large institutions may simply have less opportunity to increase their stake in Nvidia because they already own a significant portion of the company. So the argument isn't necessarily that Nvidia's business suddenly has a serious problem. It is specifically about how stocks are valued and how realistic additional institutional demand is in the market.
And this is not at all the same as claiming that Nvidia's fundamentals are sounding the alarm. In fact, recent comments from Nvidia indicate the opposite. Jensen Huang continues to talk about the huge demand for AI infrastructure with the potential for very significant growth. So the real debate isn't just about whether Nvidia's business is slowing down.
The bigger question is whether the market is correctly translating this growth into a value that investors are willing to pay. There is another factor that is sometimes overlooked. Nvidia has invested in a number of private companies across the AI ecosystem. So when people only look at innovative traditional assessments and metrics, they may not consider the full value of these investments.
This doesn't automatically mean the stock is cheap, but it does mean that the headlines and valuations that are sounding the alarm bells aren't telling the whole truth. Now let's move on to another, much bigger question. The race in artificial intelligence between the United States and China. The argument is that American advanced models remain ahead , while some Chinese models are accused of using techniques such as distillation to learn from stronger models.
At the same time, the United States currently has a significant advantage in computing infrastructure for AI. The chips used to train and operate these models are a critical part of this equation. But at the heart of all this is a strategic issue. So, American AI technologies are exported all over the world, including China.
The argument from some industry leaders is that cutting off access completely could create unintended consequences. Rather than slowing down China's AI ecosystem, this could encourage Chinese developers and companies to build everything on domestic alternatives. And Huawei is an important part of that conversation.
The concern is that if American companies voluntarily cede a significant portion of the global AI market, Chinese technology could eventually gain greater influence outside of China.
Comments 0