저자: DanIves · 외국인
요약 보기
엔비디아는 AI 혁명의 핵심이며 데이터센터 매출이 분기당 57B달러까지 갈 수 있다고 주장. 실적 가이던스는 보수적이나 시장은 분기마다 강한 실적을 원하고 있음. 블랙웰 칩 과열·공급 병목은 노이즈로 보며 수요가 공급을 크게 초과(15:1)한다고 봄. 소버린 수요와 하이퍼스케일러 capex(약 3400억 달러)로 AI 시장은 초기 단계이며 4~5조 달러 규모까지 갈 수 있다고 강조. 다만 6분기 후 성장 둔화 가능성과 출구 타이밍의 어려움도 언급.
다듬은 글
this is just a beginning of the AI Revolution. $4 trillion is on its way, sure, and in the long term you may very well be right. But are you not concerned about this plus or minus 2% the company is reporting on revenue, simply because the market wants the company to beat, and beat well, every quarter?
Yeah, if you looked up "sandbag" in the dictionary, you'd see this guidance. I think Street is going to look through it. I think there's a stock that will probably be up even on the conference call, because if you look at everything they did here in terms of the data center, with Hopper, we could be talking $5, $7, $8 billion per quarter. Many thought it could just be a few billion. So as this all plays out, and you start to look at numbers over the next 12 to 18 months, the godfather of AI, Jensen, and Nvidia continue to be on that Mount Rushmore. It shows it's starting with Nvidia, and now the second, third, and fourth derivatives are playing out. That's why it's 9:30, 10 p.m. in this AI party that goes to 4:00 a.m.
Analyst Summary: Wedbush on Nvidia
Wedbush Securities highlights Nvidia's strong performance, positioning the company as a key player in the AI revolution with significant growth potential. He praises Nvidia's revenue growth and guidance, emphasizing its market leadership in AI and its expanding role across various tech sectors. He also acknowledges some production challenges, such as Blackwell chip overheating and supply bottlenecks, but believes Nvidia can manage them. He sees the company's dominance in AI as a key driver of its success and maintains strong market confidence in its future. Overall, he emphasizes Nvidia's importance in the US tech sector and its role in shaping the global economy, noting that while growth may slow in the future, Nvidia remains at the heart of the AI boom.
As the market digests this news, Dan, we're seeing a little pairing of those declines you were looking for. A mic-drop quarter, could you call it that?
I think you print this press release off, put it in the loo, and frame it. Because look, what you're really seeing here is a stock that's going to be green tomorrow. What they're putting up in terms of Hopper, in terms of overall growth, is exactly what the bulls want, and he'll continue to deliver on the promise. Next quarter, that could be a "4" in front of it in terms of ultimate revenue. We talked in the green room before the show about how important this one stock is to a constellation of trades that make the stock market go higher or lower. It's hard to think back over literally centuries and find a single company that had so much sway over the global economy or the global financial markets. I don't think there's one that hasn't had some backing, per se, from a government, even something like the British East India Company in 1990, not as big as this, relatively speaking. So it is incredibly important, not just for people trading tick by tick right after the earnings and the call, but also when you take a step back and think about the macro picture. Honestly, this is the heart of US exceptionalism, when we talk about why the US has outperformed for decades or more. This is it: it's tech dominance, it's the weight of tech in our equity market, and it's the fact that we have companies, the Mag 7, that over the next 12 months will invest half a trillion dollars in capex and R&D. Put that in context: the CHIPS and Science Act was $52 billion, so this is many multiples of that. It's the scale of our dominance, and Nvidia is the cornerstone of it. Whether they disappointed or not, to me it's the fact that you have this structural advantage.
Let me make one point about that, which I think is really interesting. That's the estimate for half a trillion dollars in capex and R&D over the next... what did you say, next year? Okay, so if it comes in at $450 billion, just think about what happens to these stocks. They're going to go down a lot. It wouldn't incrementally take, because estimates are going up. The one thing I'll say is, if you go back to Microsoft, they're a 19% customer of Nvidia, and we talk about capex, and how much it went up. It didn't go up that much right after they reported. So we're not seeing the sort of step functions we saw, let's say, three quarters ago. They also talked about being capacity constrained, which is one of the reasons they didn't have higher revenues. Now, their cloud business is growing 31% year over year, which is fantastic. They're taking market share from AWS. I just feel like we're going to go from capacity constrained to capacity overbuilt, that sort of thing, in the not-so-distant future. So when you're beating these numbers by a hair and the stocks are still sticking around, it shows investors still want the exposure. But sooner or later I think it's going to be one of those things where you start seeing a couple of quarters of declines, not just from the picks and shovels but also from the hyperscalers. I think that's a ways off, though. The hard part of owning Nvidia is this: if everybody knows that six quarters from now they're going to have slower growth, when do you exit? You want to exit before everyone else decides to exit. That's the rub.
Tech giants and AI startups are all hungrily awaiting Nvidia's new Blackwell chip. We found out overnight that Japanese tech firm SoftBank will be the first to get their hands on them. Nvidia CEO Jensen Huang announced that at an AI summit in Japan. Let's turn to Dear Jaosa for what brings us to this point in that relationship. For today's tech check, Dear, who else but Masayoshi Son? Everything about this announcement was very on brand for Son, basically telling the world that he and SoftBank are going to be key players in the gen AI shift, or at least make a very big run for it. For Nvidia, though, it does raise a question: if demand for Blackwell is as insane as Jensen Huang has said, why put Son at the front of the line instead of OpenAI or Google, or a company at the forefront of LLM advancements? A source tells me it was simple: SoftBank's order was early, and it was big. It also comes just as the debate is heating up over whether AI advancements have peaked. A big deal with SoftBank could also let Huang and Nvidia diversify away from their large US customers toward new opportunities and sovereigns like Japan. Now, overnight in Tokyo, Huang praised Son as the only entrepreneur and innovator in the world who has selected or partnered with the tech winner in every generation, from Bill Gates to Jerry Yang to Jack Ma. He was even the largest shareholder of Nvidia at one point, leading to this moment.
Nvidia, yes. It's okay, we can cry, we can cry together. Has the time changed at all since you came out with that? It was 9:00 p.m., and now we're at 9:30? Got it. My other question for you: there's so much commentary on Blackwell and how production will ramp up into fiscal 2026. Product shipments will begin in the fourth quarter of fiscal 2025, which is the current quarter. Does Nvidia break out Blackwell revenue, or is this something you and investors always need to reverse-engineer out of the data center numbers?
Essentially, investors reverse-engineer it. But if you look at some of the wording, they're basically saying, don't worry about those heating issues. I think that's going to give more and more comfort. This is going to ramp significantly, and that continues to be the story. We've talked about from our Asia trip: demand to supply is 15 to 1, and there's only one game in town, and he's wearing a black leather jacket. His name is Jensen. That's why this is still the early days of what I believe we'll be talking about as a $4 trillion market.
You mentioned the overheating issues. These are Blackwell overheating issues, which might require a redesign of the servers housing the chips. The companies that make the servers, Dell, HPE, Super Micro, do they end up carrying the cost of having to redesign everything?
Yeah, but ultimately I think that's mostly noise. From all the checks we do, when you're at the top of the mountain, you're always going to have people taking shots at you. If you look at how Nvidia is playing out here, there are really no production issues as this all ramps. That's why, when you hear from Jensen today, I think investors are going to realize this AI revolution is now about second, third, and fourth derivatives. Look at what's going to play out across software, across the rest of the semis, across cybersecurity. If you're a tech bull, this combined with the last few weeks in terms of Trump, Musk, and everything else, this is what I think is going to be a setup for a Santa Claus rally.
So we're seeing Nvidia say that Blackwell is going to be supply-constrained for several quarters, but that there are some supply constraints now. Dan, tell us the truth: where have you seen those bottlenecks, or those potential places where supply will be constrained?
Everyone's waiting in line. They're at the velvet rope, and there's only one party, and that's Nvidia. The issue there is that you want those next chips. That's what everyone's after. It's the new oil, or gold.
Yeah, but that's the customer side. I'm talking about the supply side, Nvidia actually being able to deliver these.
It's ramping. Like any new production, any new chip, it's going to take time to ramp. But I'd say, compared to three months ago, expectations are different. So if you're sitting there right now, and Jensen's flying the plane, and you're in 3A watching Netflix feeling pretty good about pricing power, can Jensen Huang raise the prices of those Blackwell chips even more than he already has? What are you expecting on that front, and should people look for increased prices for those chips?
They definitely can, but right now it's all about market share. They're just going to continue to go further and further out. Even when you look at what the hyperscalers are spending, you have $340 billion of AI capex just in your sights. Then you think about sovereigns, look at private, look at everything else. It shows we are in an unparalleled, once-in-a-40-year tech trend. That's why Mount Rushmore has Jensen, along with Lisa Su and others.
You talked earlier about the velvet rope, and a lot of people waiting to buy these chips. The hyperscalers are at the front of the line, but if you bring sovereign buyers into the mix, where do they stand? Do they get to jump the line?
Sovereigns are in the middle of the line. They were saying, "Oh, my name's on the list," and they're like, "Look, you can't get in front of these people at the front." Big tech is first, sovereigns are still in the middle of the line, and they will get in. Then behind them is the rest of the Fortune 2000. That's why we're going to be talking about $4 trillion, $5 trillion next year or two. This tech bull market is still in the early days.
In conclusion, Dan Ives emphasizes Nvidia's dominant role in the AI industry, predicting continued growth despite challenges like supply constraints and Blackwell chip production issues. He believes Nvidia's strong position will drive the AI revolution and boost the broader tech market, with the potential for AI to reach trillions in value. Nvidia remains central to this transformation. Ives also highlights Nvidia's crucial influence on the global economy and US tech dominance. While some growth slowdown may occur, Nvidia's leadership in innovation will keep it a major player in the tech landscape. How will Nvidia's AI advancements affect other industries in the future, given supply challenges, and how should investors approach Nvidia's long-term prospects?
댓글 0