저자: JimCramer · 외국인
요약 보기
엔비디아는 2025년까지 물량이 매진된 상태이며 AI 인프라 투자(하이퍼스케일러 4개사 12개월간 약 2270억 달러)가 유지될 것이라 주장. 블랙웰 칩 지연은 약 90일 정도 매출 지연일 뿐이며 시장점유율은 유지된다고 봄. AI 1.0(반도체·데이터네트워킹·서버 인프라)은 선호하고 AI 2.0 종목은 아직 검증 안 됐다고 회의적. 엔비디아가 10년간 100배 상승할 수 있다는 낙관적 견해 제시. 엘리엇 등의 AI 과대평가 비판에는 반대.
다듬은 글
that to the point where they mentioned that Nvidia is sold out. They're sold out for a long time. They're sold out probably through 2025. Yeah, listen, the spending plans we've heard, reaffirmed by the likes of Microsoft, Meta, Alphabet, or Amazon, have done nothing to dissuade anyone from thinking Nvidia is still going to be the key beneficiary. Now we have this big rotation going on in the marketplace. There's concern ultimately about the consumer, whether the Fed is now behind the curve. But none of that seems to go to the heart of the matter, which is that Nvidia is potentially going to be sold out for some time to come.
I mean, you listen to Amazon. Of course they have to build their own, but that's a defensive move, because they can't get enough chips. Building their own chips, designing their own chips—here's the way I look at Amazon. I think it's important, because the stock sold off so badly today, I would be a buyer. Amazon knows there's a lot of business, so they have to buy a lot of planes. They have to put up hangars. When you're buying a lot of planes, you don't have to put these people in the planes, so they buy the planes first. It costs a lot of money, and then you have to bet that they're right. I want to bet that the chassis is right. Or you could use a different analogy: they're buying a Stradivarius, and you can't just learn a violin like a Strad. You've got to learn it. It's a little harder to learn, a little harder to teach. I'm just trying to find an analogy so people understand it takes some time. They can't just flip a switch, and that's making people very upset.
And the fact that they mentioned the Olympics. I mean, people are like, are you kidding? The country is in a 2012 mode. They're watching the Olympics, they're not buying as much. What they are buying is because they're watching the Olympics—them, not me.
Well, after a week where we got Arm guidance and Western Digital guidance, and now Intel and now Microchip, you're still bullish on the space at large?
Uh, no. I'm bullish on companies that really are going to be the GPU or helping Nvidia, so it's a narrower set of bullishness. Look, when the conference call for Amazon happened, they absolutely know they have to spend money on AI.
Jim Kramer is enthusiastic about Nvidia's potential, predicting a dramatic increase in its value and suggesting a 100-fold return over the next decade. He describes this as a once-in-a-generation opportunity and urges viewers to stay tuned for more details. Kramer acknowledges recent updates from various tech companies and guidance from firms like Intel and Microchip, but emphasizes his continued optimism about Nvidia. He notes that despite the market's current rotation and concerns about consumer spending, Nvidia's dominance in the GPU market remains strong. He highlights that major players like Amazon are struggling to keep up with Nvidia's demand, which could mean Nvidia will be sold out through 2025.
Kramer also addresses recent criticisms and reports suggesting the AI technology hype might be overstated. He firmly disagrees with the notion that Nvidia is in a bubble or that AI technology is overhyped. Instead, he points to ongoing substantial investments by major tech companies in AI infrastructure as evidence that the technology is on a solid growth trajectory. He dismisses concerns about delays in Nvidia's chip production, asserting that Nvidia's technological lead and critical role in the AI supply chain will keep it ahead even if there are minor setbacks.
Finally, Kramer makes a distinction between what he terms AI 1.0 and AI 2.0. He remains confident in the infrastructure-building stocks involved in AI, such as semiconductor and data networking companies, predicting continued substantial capital expenditures in this sector. However, he expresses skepticism about AI 2.0 stocks, which claim to offer revolutionary products but have yet to prove their financial viability. For now, Kramer believes that investing in AI infrastructure stocks is sufficient for realizing the potential of the AI revolution without needing to rely on the more speculative AI 2.0 companies.
Would you give it a year? If it doesn't happen in a year, if the rhetoric doesn't change, would you start to get nervous?
I'd say if we sit here in January or February and we haven't seen it—we haven't seen the monetization from a software perspective, we haven't seen it from an actual outside view, just the Mag 7, to call it second, third, or fourth derivative—if Michael Dell comes and all of a sudden he's negative on AI, then of course we have to take a look at it. But right now it's actually more emboldened, because of what we see, not just from earnings but from our checks. I've never seen Bill McDermott negative on anything. Yeah, he has pivoted before. But Satya Nadella and Tim Cook are seeing something. I'll take them, despite many who are right now kind of throwing gasoline on the fire.
There's this report we've been seeing about the delay in the Blackwell chips and what that could potentially mean. Some of those customers, like Meta, Google, and Microsoft—how are you watching that story? Is this going to have an impact, do you think?
I think what happens, as it pertains to Nvidia, is they may have a delay of a Blackwell series chip, which could be as long as about 90 days. But in the meantime, they're so far ahead with their technology that no one is going to come into that vacuum and steal market share. So it's there for the taking. I think their revenues are simply going to be delayed for about a quarter or two. And because they're the most important cog in the AI supply chain, it's certainly going to impact the AI infrastructure building of some of those hyperscalers, as we call them in the States—the four big companies spending billions and billions on building their AI infrastructure.
Well, I know Sam scratched this point. She raised this issue about the AI hype, being bullish on tech, which you are broadly speaking. But drilling down to specifics: if an activist fund manager with the credibility of Elliott Management says there is an AI overhype and that these stocks are overbought, do you think the pressure builds to the downside?
I think we need to distinguish between what I call AI 1.0 and AI 2.0. On 1.0, I continue to be very comfortable with the infrastructure-building names—typically semiconductor companies, data networking companies, and data server companies. Now, the four hyperscalers have promised—I guess nobody promises, but they've said recently in their conference calls—that over the next 12 months they are going to spend $227 billion combined on AI infrastructure. I don't think they are going to change those plans. They also say their capital expenditures in the year afterward are going to accelerate from that very high base. So I like all those stocks. Then you get to the AI 2.0 stocks, which I'm not sold on yet, so I don't own a single one. These are the companies that claim, after all the infrastructure spending, that they will have marvelous money-generating products featuring AI. I'll believe that when I see it. But for my thesis to be right, I only need the AI 1.0 stocks. I don't need the AI 2.0 stocks yet.
Yeah, I know, that was weird. Jeff has a note today saying clients are calling in asking about potential delays in Blackwell. And we did have the report yesterday about DOJ reaching out to rivals of Nvidia who have been complaining. Well, it's slower, it's slow, but they never really said it was going to ship in volume. So people who think it was going to ship in volume should go back and talk to Jensen. He never said it's going to ship in volume right now. Never. There is still this question about return on invested capital, whether there will be. By the way, the answers from the hyperscalers have been uniformly yes, and Jassy said the same thing. About Zuckerberg, remember? He's talking about—Jassy said it's going to get big fast, and it's largely all going to be built from the get-go in the cloud, which, as you know, allows the opportunity for those businesses to continue to grow. You've heard no letup whatsoever for the cloud. That's what we should focus on. But we're all focused on those who are making the capital allocation decisions.
Can I just say one thing? You can say many—I'm sorry if you were mad at me, I apologize. I'm not mad at—I got a little excited because it's a big day. But I regret that I did. The AI revolution story, I believe, is well intact, and all of our checks say this is not a time to panic, despite the white-knuckle moments we're seeing in the market.
TheStreet is reporting that Elliott Management told investors recently that Nvidia is in a bubble, the AI technology is overhyped, and the reality is not matching the hype. You totally disagree?
I couldn't disagree more. Relative to where we were before earnings and before conference calls—I think we even talked about it. Let's see what the numbers look like. Look at the Azure acceleration. Look at what Lisa Su talked about from AMD. We saw that Azure's acceleration guidance for the second half was what we thought, and that's where the disappointment would come. That really talks about the moment of demonetization being right here. My view is: the haters have continued to hate. The bears have been there for the last few years. It's easy right now to come out and yell "fire" into a crowded theater. In tech, we don't see it, because it's not just from the call—you see the value matching the hype. I actually think the street is underestimating what is going to be a trillion dollars of AI capex over the next three to four years across the board, and nothing we saw here changes that in any way.
None of them are backing off their capex plans. To your point, AWS actually accelerated in terms of revenue growth, up 19%. There is still a larger question as to whether you're going to ultimately get that return on investment over time. You obviously believe it will. But what do you see on the ground in terms of applications coming in the near term that will start to alleviate any of the fears some investors may have from the enterprise perspective? We'll hear it from Jensen and Nvidia in a few weeks.
The use cases and the number of enterprises on that AI path are up four or five times. These enterprises are on the path to use cases across healthcare, financials, insurance, and retail. As we've always talked about, it all comes down to the modernization. The spending is there. When you look at what Satya Nadella talked about, and what Sundar Pichai and Andy Jassy talked about, this revolution is happening. The modernization is now just starting. We're going to have an air pocket between the capex and when revenue comes. But to bet on this and to bet against it—to me, it's like betting against Cook and the 2.2 billion devices, in terms of where we ultimately see the actual installed base. And right now, an AI revolution coming to Apple is a tough bet.
Jim Kramer is highly optimistic about Nvidia's future, projecting a 100-fold return over the next decade and labeling it a once-in-a-generation opportunity. Despite recent market fluctuations and concerns about consumer spending, Kramer remains confident in Nvidia's dominance and the significant demand for its GPUs. He argues that major tech firms like Amazon are struggling to keep up with Nvidia supply, underscoring the company's critical role in the AI sector. Kramer dismisses concerns that AI technology is overhyped, insisting that substantial investments from tech giants validate the sector's growth potential and Nvidia's leading position. In contrast, he is skeptical of so-called AI 2.0 stocks that promise revolutionary products but have yet to prove their financial worth. He continues to endorse AI 1.0 infrastructure stocks, including semiconductor and data networking companies, due to their substantial and ongoing investments in AI. For now, he believes these infrastructure plays are where investors should focus rather than speculative AI 2.0 companies.
(Note: 원문 일부는 영문 원문을 그대로 두었습니다. 요청하신 한국어 편집본은 위 구성 중 영문 대목을 한글로 옮긴 형태로 다시 작성이 필요합니다.)
Comments 0