저자: JimCramer · 외국인
요약 보기
AI 인프라 투자는 계속 가속 중이며 수요는 여전히 공급을 앞선다는 주장. 반도체 급락은 펀더멘털 훼손이 아니라 레버리지 청산에 따른 포지션 리셋이라고 진단. 메모리(마이크론·SK하이닉스)는 장기 고객계약과 증설 지연으로 공급 제약이 지속될 것이라 봄. 애플은 AI 인프라 대신 생태계 활용 전략이라 평가, 엔비디아는 데이터센터 핵심 지위 유지. 테슬라는 AI 기업으로 평가되지만 FCF 압박이 단기 우려.
다듬은 글
I was sitting next to Steve Ballmer at our college's 35th anniversary reunion , when he was still the CEO of Microsoft, and he pulled out some gadget that made me giggle. It was Windows mobile. Does anyone else remember this? I doubt it. So, in its core business, Apple is invulnerable, and they have never spent much on AI.
I want you to think about that as you watch Apple stock drop seven dollars today. Did anyone switch to Samsung during this AI tragicomedy? Someone said, “Enough . I'm fed up with this. I'm tired of the 17 Pro Max. " Bring me Lenovo." Perhaps experts would like Apple to spend $200 billion, about 70% on data center infrastructure and energy, perhaps give the rest to Google's TPUs, or buy expensive GPUs from Nvidia.
As someone who owns Apple stock in their trust, I love that they let Alphabet spend billions on AI and then make Alphabet pay them a nominal amount to be in the iPhone. Uh, the pros have a term for this Apple and Alphabet deal . Do you know what it's called? A good deal. Oh, I also want to say—I know people think I've been unfair to Nvidia lately.
They say I criticized them for not buying back more than $80 billion in shares, considering they're not that cheap when you look at next year's earnings estimates. You know, it's actually more of a friendly nudge, folks. There have always been companies that claim to value Nvidia highly but want to make their own chips.
You know what I'll say? Go ahead, make Nvidia's day better. But if the customer wants Nvidia, what will you do? What do you tell them? “No, no, you see, we don’t have Nvidia. You know, we have these—the worst "delicacy ". I mean, they'll go somewhere else. Nvidia is the heart of the data center, they actually invented the data center.
They have no real competitors from China. The fact is that these machines, and these, by the way, are machines stuffed with hardware and software, are the envy of the whole world. I...I try, Lisa Su from AMD is close, but there's really nothing like Nvidia. And that matters . In my opinion, instead of building each AI function independently, as we've seen in a number of things.
It increasingly seems like they are focused on customer experience, which is where they excel. Also, the recent news about Gemini integration into Siri, uh , I think that's very interesting stuff. And when I look at John Turnus, who will be coming in a month from tomorrow, an engineer with a lot of experience, I don't think it's a coincidence that the board of directors chose an engineer to lead the next chapter of Apple.
Because I think this next chapter will be defined by AI chips, hardware, supply chain execution, as well as software innovation. Okay, finally, we have to talk about Tesla. Negative free cash flow, declining operating margin, people didn't like the report we got from Tesla this quarter. How do you view Tesla?
And what kind of company is this, Tesla? Is Tesla still—do we still say “automaker, ” or has that train already left and we’re talking about AI? And next year, when we have this conversation, will Tesla still exist, or will it be a supercompany? Yes, great questions , Barley. When I think of Tesla, I think of AI vision.
So what they delivered last week was an improvement in revenue, but Wall Street really focused on AI spending. Elon says, "Hey, we're going to invest more than, I think, 25 billion this year in AI infrastructure, and also robo-taxis, Optimus." But investors don't question this vision. What they question is the timing.
So, investments are putting pressure on free cash flow, and shareholders are asking, "Well, when is all this going to start delivering tangible results?" Again, this is ROI ( return on investment). Tesla, to answer your question, I think is being valued more and more as an AI company, not as an automaker.
Because what she's asking investors to do now is to evaluate tomorrow's AI business. Hyperframe Research AI stack with me. Stephanie, I'm so glad you're with us. I just finished talking about Apple and Amazon. You know, there was a question we started the week with: should we be " light on capital expenditures"?
And the answer seems to be no, and you tell me, but Amazon has not been "light on capital expenditure." I mean, they actually increased capital spending from 200 billion to 220 billion, but they're delivering results in Amazon Web Services. So, capital expenditures are okay if you deliver results. Is that—is that the main conclusion?
Yes, this is definitely the main conclusion. This , you know, I feel like investors want to see evidence. They want to see evidence. They want to understand that this investment will bring ROI. If you look at, say, Meta, the reason Meta didn't get rewarded for its gigantic investment in AI is because it couldn't prove that those investments had a path to monetization.
So we see that companies that can justify their huge investments in AI, especially in infrastructure, are rewarded by investors, while others whose path to payback is less clear are not. So when you look at what's happening, you said that AI investments are now completely focused on monetization. People are tired of hearing about how much you invest.
It's like " show me the money" or, in your words, " show me the checks." When you think about it, and last week was Google, right? So, Google, Apple, Amazon, Microsoft, Meta. Is there anything else you would like to add about any of these specific companies based on the reports that have intrigued you one way or the other?
I think Apple is interesting because they are playing a different game with artificial intelligence. They are really investing more in AI as part of their ecosystem and improving the user experience, rather than trying to be an AI platform that invests in infrastructure. Let's see if this strategy pays off and how it works for them.
I think Google was interesting because they showed great results, but their investments couldn't offset the costs of a huge AI infrastructure. Nvidia's next move may surprise everyone. While traders have been asking the same question for the past few weeks: is the AI investment boom finally losing its steam ?
Shares of memory manufacturers have collapsed, and the shares of chip developers are constantly fluctuating. Even Nvidia, despite strong business performance, is struggling to sustain its growth. At first glance, it seemed that investors were preparing for the end of the AI cycle. But if you look deeper than the headlines, a completely different picture begins to emerge.
Companies are actually spending it's not slowing down. They are accelerating. Amazon recently increased its spending on AI infrastructure again. Management acknowledged that even after increasing capital investment plans, they still lacked the capacity to meet customer demand for the next few years.
Microsoft demonstrated another quarter of accelerating growth in cloud services, continuing its large-scale investment plans in AI. Alphabet remains committed to expanding its data centers. Meta continues to invest aggressively, despite concerns about rising costs. If demand for AI weakened, try to build faster.
This tells us something important. The demand is not going away. The offer is still barely keeping up with him. So why have semiconductor stocks become so volatile? was not a collapse of fundamentals. It was a reset of positions. For the past few years, traders have been investing in the same stocks using high leverage.
When sales began, forced liquidation quickly followed. Margin requirements triggered new selling, and momentum funds reduced exposure. And quality companies were dragged down along with everyone else, regardless of their individual performance. And this is not at all the same as a sudden weakening of business.
The story with underlying demand has hardly changed. Memory manufacturers remain a perfect example. Companies such as Micron and SK Hynix continue to discuss long-term customer agreements. As well as high prices and limited production capacity. Customers are still booking shipments years in advance as advanced memory remains essential for AI learning and inference.
Building new production facilities is not something that happens overnight. Creating advanced memory capabilities required huge investments, specialized equipment, and years of construction. Even if competitors announce expansion plans today, significant production may not begin until several years later.
Until then, the offer will remain limited. Another debate currently taking shape concerns funding. AI infrastructure requires a staggering amount of capital. As interest rates remain high, financing these projects becomes more expensive. Some investors fear that higher borrowing costs could negatively slow the development of AI.
This is a valid concern, but there is another side to this argument. The largest cloud providers are starting to show impressive returns on their AI investments . The cloud business continues to grow, and profit margins remain healthy . Cash generation improves, and these earnings give companies more flexibility to fund future expansion internally rather than relying entirely on external capital.
This is an important distinction. Instead of spending blindly, they are increasingly investing from a position of financial stability. Competition is another topic that investors continue to closely monitor .
Comments 0