저자: 미상 · 외국인
요약 보기
엔비디아 CEO 발언 인용: 컨센서스 대비 실적 상향(100~200%)에도 주가는 20%만 상승, 12~14배 PER. 젠슨 황은 GW당 기회 400억 달러 이상, AI 인프라 확장을 강조. 내레이터는 RTXSpark 노트북(2,399달러~), 메르세데스 레벨4 자율주행 파트너십을 소개하고, 동시에 상승쐐기 붕괴·버블지수 급등 등 조정 신호를 언급하며 결론은 유보적.
다듬은 글
This company's consensus forecast has been revised upward by 100-200% . Yeah. That's amazing. In my more than 20 years of doing this job, it's rare to see a consensus forecast deviate by such a large margin . Yet the stock price has only risen by 20% . Well, we literally doubled this company's earnings forecast and free cash flow forecast .
It depends on who you're looking at, but right now it's trading at 13 to 14 times the profit . So, what do you think the market is saying about Nvidia, or about its confidence in market expectations ? The reason this is happening is that, normally , if the forecast were revised upwards this much , the stock price should have risen much more significantly .
Well, first of all, the fact that a large company like ours is gaining market share speaks volumes about resilience . yes. And now, regarding the scale and scope of our competitive advantage. I agree. So, is that advantage increasing or decreasing? Is the moat expanding or contracting? It is increasing.
It is increasing . It is increasing. Your competitive moat is growing with scale. Let me give you an example. Let's conduct a thought experiment right now. Let's say you need to hand someone a $100 billion check in order to purchase the next generation of products . Yeah. If you want to get in line for Vera RubinUltra , your order time is approaching.
That's amazing. Without seeing the actual item. Yeah. Right ? But they are confident . Now, let's say you have to bet $100 billion on someone . One. How do you know if something is good ? I agree. Since it was created by Nvidia, I know it's better than VeraRubin . I agree. And VeraRubin is fantastic .
I agree. And the reason everyone trusted Vera Rubin was because Grace Hopper was fantastic . Yeah. And in the world of computing, the workloads these computers perform are extremely complex. Measuring only one benchmark is insufficient. I agree. That's 50 billion dollars. Yeah. And the number of technologies that need to be combined to achieve that performance is enormous.
So the question is, who can you place an order for $50 billion or $100 billion without seeing the actual product? Yeah. Literally, it's just us. Yeah. Previously, even issuing a purchase order for $1 billion was a huge undertaking. But if you issue a $100 million purchase order and have the system right in front of you , you can benchmark the chip .
Right ? How do you benchmark these systems? It's a miracle just to be able to build it. Right ? So, if someone wants to evaluate it , I'd be happy to support them , but it would take several years to set up and renovate a massive data center . So, we can't even benchmark these things anymore . That's interesting.
The exact same thing happened at TSMC . Yeah. The reason we entrust 100% of our operations to TSMC is because they are the only company in the world capable of creating plans five years in advance . That's right . Even without seeing it in person , you can tell that this transistor is truly excellent .
Because it's better than that, and better than this . Yeah. Furthermore, TSMC has the production capacity for me, knows how to work with my business team , and has the agility to handle all my ups and downs , such as the fact that I moved from gray to blue to green . that's right . That's exactly right.
Yeah. It needs to be large enough to help with the transition from gray to blue and then to green . That's exactly right. No other company in the world can help me . That's right. Furthermore, it is necessary to have all kinds of process technologies, from silicon photonics to logic, memory, and push-pull.
The packaging process, that is, the sheer volume of work we do, is very dynamic. How confident are you that in five years, we will have made the right decisions, and that thanks to our investments and everything we do, the world will be a much better place, as we describe? Sitting here today, are you more confident now than you were 12 months ago ?
I'm 99.99%. However, to serve and show respect to good friends , we don't use 0 or 100. Well, I'm 99.99% certain . Well, many of us are deeply concerned with the outcome and dedicate our strength, intelligence, and resources to responsibly and optimistically getting there . Yeah. So, there are really a lot of people doing the right thing .
Now, as you know, in many cases there are people who speak authoritatively about these things, drawing information from various sources , but they fail to notice the many people, industries, brilliant researchers, and engineers doing amazing things outside their sphere of influence and unrelated to them .
Just because they didn't say it doesn't mean that many intelligent people aren't doing intelligent things. And, as you know, I always say this within the company . Even without my leadership or strategy, the company is doing smart things on its own . Smart people do smart things. While organizing and collaborating can lead to greater impact , truly brilliant people often do smart things on their own .
So, if someone tells me they're worried about me , it's a good thing for them to be worried . And we should express our concerns . We should express our concerns . However, we should not try to predict the end of the world . Because they underestimate the state of human beings and underestimate human capabilities .
And perhaps due to arrogance or haughtiness , they believe that nothing can be accomplished unless they do it themselves. that's right . that's right. Nothing can be accomplished unless we do it ourselves. that's right . In fact, whether I say so or not , this company has achieved something remarkable .
that's right. And I accept both praise and criticism, but in both cases, I am fair. Especially when it comes to praise. And the reason for that is that it creates a situation where you are surrounded by wonderful people and everyone desires the same results . that's right . Good result. A better version of that.
And I think most people would agree . that's right . That is precisely why we are saying we must take responsibility and accountability for ourselves . that's right . However, we should express optimism about the future. surely. And I am 99.99% certain that I will get there surrounded by amazing people from all over the world .
Yes, I agree. And I believe you are a national treasure . Nvidia's stock is being called a bubble this weekend . In the same few days , Nvidia embedded its chips into Windows laptops , delved deeper into the next Mercedes you might be riding in, and its CEO told the world that every gigawatt of AI capacity is worth more than $40 billion to the company .
So, is Nvidia having problems , or is it just getting started ? Let's see what actually happened . Let's start with something that should make bears uneasy . Jensen Huang stated, "We are expanding our market share because we are providing more to global capital investment ." On the surface, this is just the CEO talking about his book, but the numbers behind it tell a story.
Analyst Shea Ballware pointed out that Nvidia's revenue opportunities from AI capacity per gigawatt have more than doubled to over $40 billion . Why is that important? For many years, in bearish scenarios, custom ASIC chips from companies like Google and Amazon have eroded Nvidia's GPU business and have actually expanded their market share .
However, let's take a look at what Nvidia supports . They're no longer just selling GPUs . They sell CPUs, networks, and the entire AI system that encompasses them . In other words, even if rivals gain market share in chips, Nvidia can earn more revenue from each data center that is built. For every new gigawatt, the bill will increase.
This represents a transformation from a chip company to an infrastructure company , and it is at the heart of Nvidia's claim that it can continue to grow even as competition intensifies. Now, let's move on to the parts that most people actually interact with. Nvidia has officially entered the Windows laptop market .
On October 16, Nvidia announced that RTXSpark laptops, which integrate the GPU and CPU onto a single chip, will be released by Asus, Dell, HP, Lenovo, Microsoft, and MSI. This means that the six major PC companies will be releasing their products from the very first day of launch . What's interesting is the chip.
Combining a Blackwell RTX GPU with a Grace CPU, and sharing up to 128GB of integrated memory , AI performance reaches 1 petaflop. This memory value is the real point of interest . Most laptops come equipped with 16GB of memory . With 128GB of memory directly available to the GPU , you can run large-scale language models on the machine itself, rather than renting time from the cloud .
Nvidia is also marketing this as a machine for AI agents. We have released a 74 billion parameter model called Nematron3.5V, built for tool calls and multi-step tasks. The idea is a personal AI assistant that runs on a laptop and doesn't require a server connection. Price becomes the issue. The headlines at the time of the announcement stated that these laptops would be sold starting at $2,399 .
NVIDIA has offered two starting prices depending on the configuration . This is not a low-priced machine. Early reports on HP's list showed some models priced at nearly $3,000 , but these were unconfirmed, pre-release lists. This is a premium product for developers, creators, and anyone who wants serious AI power in a thin and light laptop.
Wall Street took notice of this chip when it was first announced . Following this news, Intel's stock price fell by 6% and AMD's by 5% , but NVIDIA's rose by 4% and Microsoft's by 3% . NVIDIA's challenge in the home territories of Intel and AMD is a big deal. The unresolved issues are whether software compatibility and battery life will be maintained in the real world .
Because that's what determines whether a WindowsonARM machine succeeds or fails . The third is a car. Mercedes-Benz and NVIDIA announced on October 16th that they are expanding their strategic partnership to build next-generation driver assistance and autonomous driving systems. The contents of the stack are as follows: Mercedes provides the MBOS operating system and safety engineering .
NVIDIA provides computing, driving software, and safety layers . NVIDIA's Hyperion processes automotive computing and sensors. Alpha Mayo is driving software with AI inference , and Halos is a safety system that provides redundancy and guardrails. The target is future Level 4 vehicles, including robot limousines.
Level 4 means that the car can operate autonomously under defined conditions. Mercedes executives have called NVIDIA their most important strategic partner in autonomous driving . Let's face reality here . You have that right. The release does not specify the release date, vehicle , or market. This is a development roadmap , not a car that will be available for order next month .
This partnership itself is not new, as the two companies first collaborated in 2020 . What's new is the expansion and the focus on Level 4. Nevertheless, it's important for the same reasons as the discussion about gigawatts. Nvidia doesn't just sell chips to automakers . From training in data centers to in-car computers, they're selling the whole brain .
Now we get to the unpleasant part. While all of this was happening, Benzinga published a headline on October 18th stating, "Nvidia enters correction phase after rising wedge breakdown , AI stock bubble index surges ." Simply put , an ascending wedge is a chart pattern where the price continues to rise, but the range of fluctuation narrows, which traders see as a sign that the buying momentum is weakening .
When prices fall from their lowest point , technical traders interpret this as a warning. A correction means that the stock price has fallen significantly from its recent high . It's important to note here that chart patterns are not predictions . This explains how prices have moved up to now , not how companies will move in the future .
The rising wedges broke down and recovered many times, but there were many that did not. Furthermore, the bubble index is an indicator of sentiment , not a judgment . However, this tension is worth enduring. The price movements indicate that investors are feeling nervous . According to business news, Nvidia is expanding its business to every level, from laptops and automobiles to data centers .
These two things could be true at the same time . The market often becomes afraid of a story long before it turns out to be untrue. Here's my perspective. The bullish scenario is that Nvidia is turning every gigawatt, every laptop, and every car into a bigger revenue stream . A bearish scenario is one in which expectations outweigh reality, and a perfect stock price doesn't need bad news, only slightly imperfect news to cause a decline .
Neither scenario necessarily means that Nvidia is a bad company . The issues are price, patience, and how much this expansion will actually translate into profits. Therefore, please pay attention to the following three points . Will RTXSpark laptops sell for their premium price ? Will the partnership with Mercedes actually create a timeline for Level 4 autonomous driving ?
And as custom chips gain market share, will the figures per gigawatt continue to rise? If you'd like me to continue tracking all of this , please hit the subscribe button and let me know in the comments whether this is a healthy adjustment or the beginning of something bigger .
Comments 0