저자: JimCramer · 외국인
요약 보기
AI 반도체 과열 조정은 펀더멘털 훼손이 아닌 포지셔닝 붕괴이며 AI 설비투자는 지속된다고 주장. 엔비디아·애플은 보유 대상, 마이크론은 HBM 수요로 메모리 사이클 논리가 바뀌었다고 평가. 시장은 실적 호조에도 과열 해소 과정 중이며 주도주 변동성을 기회로 봄.
다듬은 글
day. I am afraid that you will suffer losses and move to other forms of investment, such as cash, especially those who trade with leverage. I am, of course, not against owning shares in technology companies. Hey, my biggest positions in the investment trust are Apple and Nvidia, for God's sake. Sure, they have n't caught up with SanDisk or Western Digital lately, but they are unique, great companies that are making a fortune.
Apple shares are making progress . Nvidia shares currently. Part of that is because it doesn't sell enough in China. I think it's because Nvidia's customers tend to both praise it and work against it. And the media likes to report it every time a customer tries to develop their own chips to get away from Nvidia.
But it's really cheap, and Nvidia is probably, I think , the best-managed company in the world. As I always say, Apple and Nvidia stocks should be held, not traded. I think the Apple brand will allow it to pass on higher memory costs to telecom operators, while benefiting from Alphabet's spending on AI.
Nvidia , my position on it has become controversial. I wanted Nvidia to be the chip of choice for China, for the Chinese to use our technology. This was a controversial position. The Chinese military would n't use them anyway because they don't want to depend on an American chip manufacturer. But once Nvidia started hitting a wall in China, I no longer wanted American companies to give away trade secrets to the Chinese for cheap AI.
Hey, why don't you look at the case that Micron filed in 2017 against a Chinese company that stole so much of Micron's intellectual property. This case has become a criminal investigation by the Department of Justice . A case in which a Chinese company admitted its guilt. I'm not making this up. I have great respect for Nvidia and don't want to listen, but not when national security is at stake by giving China secret access to the data of many of our companies.
I don't think that's reasonable. So, here's my advice. If you want a memory chip manufacturer, I like Micron. If you want GPUs, for me it's AMD or Nvidia. If you want CPUs, it's Intel, racks called Dell, Optics. Make it trivial. We always forget that when you get a big jump in semiconductors, it's a good private bank.
That sounds pretty conservative to me . And I saw what Jamie was saying. You don't base your opinions on the chairman's feelings about the stock market, do you, Stephen? No. I think you have to remember that great banking management teams look at risk through a different lens than the risk management of long-term investors.
But you have to admit that there are certain bubbles in the market that you ignore at your own peril, right? Well, I think if you look at it, yes, there are some bubbles, but we are seeing a very healthy expansion of this market. Did it start, because it started, then stopped, and then started again.
And is this exactly what we want? And we see it . Yes. Aha. The conversation we had back in December was about Mag 7 and hyperscalers. Now look what happened this year. " Mag 7" are falling, while the market continues to rise. Well, it all depends on the technology sector. And then we saw a shakeup in the software industry .
And now we are seeing a significant pullback in semiconductors. And yet, despite all this, the market is close to historic highs because we are seeing expansion. Profits continue to grow. To a large extent it's technology, but if you look at industrials, utilities, finance—it's a broader story, and I think it's really great .
Have we seen enough of this quarter's results to know that there won't be a major disappointment across the board? Well, the first results look fantastic. Almost 90% of companies exceed expectations. And so I think the results will be strong. However, we must be aware that, due to what we have seen in the markets and due to higher expectations, companies that do not meet forecasts or are not ahead of them enough are likely to face a big drop.
You have to be prepared for that and, frankly, look for opportunities because volatility can be your friend when you get a little bit of that volatility. I don't know if anyone has ever thought about a market crash or if anyone has been able to determine the cause of it before it happened. Later, looking back, it always seems obvious what the cause will be.
What do you think the reasons will be this time? Will it be inflation that proves too stubborn? Will this be what will force the Fed to raise rates? Will this be somehow related to capital spending on AI? Well, you know that there is no money or that it has been cut. What could happen this time? I think there are two main risks that concern us right now.
The first is the capital expenditure on AI, because while we are seeing expansion, a lot of it is related to the creation of data centers, and if we were to see a significant reduction in that expenditure, I think it would have broader implications for the market. So, this is something we are watching for, but we don't see it happening.
Another point could be a repeat of 2022, when rampant inflation would force the Fed to tighten policy much more than markets expect. Note that at the beginning of the year, markets expected three rate cuts, and now expect two increases. We disagree with this . We believe the Fed will refrain from making changes.
And in these conditions, profits continue to determine the situation. We've seen, uh, many technological revolutions in the past. And, I think, there is a fear that with AI everything could come to naught, that the development could be excessive. At this point, I really don't think it's like, for example, a metaverse— the metaverse has come and gone, I think, pretty much.
This will not happen. It feels very different, but the important thing is that the benefits of AI cannot — for this to work — they cannot accumulate only in a narrow circle of hyperscalers, we need to see the positive impact of AI spreading not only to the technology sector , but also to other sectors.
Do they need to see returns before they complete construction, or will they just move at breakneck speed no matter what? AI stock trading has just endured one of its biggest tests, and interestingly, the May sell-off was not so much about deteriorating fundamentals as it was about something much simpler— positioning.
The semiconductor market has become extremely overheated. The popularity of credit ETFs has exploded. Activity in the options market reached extraordinary levels, and traders increasingly used short-term contracts to chase momentum in the same few AI stocks. When this positioning began to collapse, the sell-off became much more aggressive than the fundamental changes warranted , so stocks like Micron, Broadcom, Marvell, and other similar companies lost a huge chunk of value in a very short time.
But here's the most important thing. The AI spending story didn't suddenly disappear. Hyperscalers were still spending money. Demand for computing power continued to grow , while demand for memory remained stable . And companies continued to report strong earnings. So the market had to answer another question.
Was this the beginning of the end of the AI cycle, or simply the end of overheated trading? There is growing evidence that the latter is the case, and this distinction is extremely important, because if fundamentals remain unchanged and speculative positions disappear, the next move may be even healthier.
Instead of investors chasing every point of growth, one can see earnings catching up with valuations. This is especially interesting in the case of Micron. The traditional argument against Micron is quite simple. The memory market is cyclical. High margins attract competition. Eventually, new capacity appears, prices fall, and the cycle repeats.
But AI is forcing investors to reconsider this approach. High- bandwidth memory has become a critical infrastructure for modern AI systems. The problem isn't just that companies need more memory. The problem is that the industry can't produce enough of the needed memory fast enough . This is radically different from the traditional situation of excess supply.
Therefore, Micron is evaluated according to two completely different models. One says these profits are near the peak of the cycle. Another says that the demand for AI has fundamentally changed the economics of memory, and the market is currently trying to figure out which model deserves more attention.
And there's another reason why Micron has become so interesting. The company is not just generating higher profits, it is potentially building an increasingly strong balance sheet. And this is something that investors can miss by focusing only on the income statement . If profits remain extremely high for several years, companies can accumulate huge amounts of cash.
This cash can later be used for share repurchases, capacity expansion, debt reduction, and other capital allocation decisions.
Comments 0