저자: JimCramer · 외국인
요약 보기
영상은 Nvidia·AI 투자 심리를 다룬다. Oracle 콜 이후 Nvidia 주가 2% 하락이 언급되며, 발언자는 Nvidia 칩이 중국 외 매출을 낼 수 있을지에 대한 우려에 반박하고 차세대 Rubin 칩을 근거로 든다. 이어 OpenAI(Sam Altman) 측 발언을 인용해 Gemini 3 경쟁 대응(code red), 추론(inference) 단계 수익화 전망, TPU 등 대체 칩 위협에 대한 Nvidia 입장을 소개한다. 구체적 가격 목표나 방향 예측 수치는 제시되지 않는다.
다듬은 글
It is pretty interesting Jensen's comments on Joe Rogan. Uh Sam Alman's going to do Jimmy Fallon on Monday, right? I mean, this is where we're getting to different areas of coverage. Sam Alman ought to focus on raising money so that we can take the existential off the table. That's what he needs to do.
He maybe needs to go to Microsoft. Maybe he needs to, I don't know, come public and be able to demonstrate that there's there's a there there. But but Jimmy Fallon is not going to raise the money that he needs. It's just not. I mean, maybe maybe Jimmy Kimmel. I mean, let's let's stop being stupid. I didn't go to college to be stupid.
Stupid. The session begins with a 2% decline in Nvidia shares, still rising roughly 34% year-over-year. Mr. Kramer, where are you? I know. I am the owner of it. Don't exchange it. Oh no, a new one has arrived. I understand you're resisting the Head and Shoulders chart. Jim, since $2 ago, I have battled the Head and Shoulders chart.
What will I do then? People do, in my opinion, believe that there is a lot of open oracle stuff and that Oracle encourages you to bring your own. That was taken as a stunner somehow. And then we hear this very, "Would you mind clarifying what you mean?" Oh, when they said, "Listen, you can bring your own processor." At one point during the Oracle call, folks responded, "Well, that's really sort of a step toward being Nvidia agnostic in a sense.
In other words, you can bring your own chips or TPUs if you'd like. However, nobody thinks Nvidia will actually make any money outside of China. I don't agree with that. You'll have to wait till Nvidia's frequent sell-offs are over. This is a really serious however I am not afraid because there is a chip around the corner that will transform everything you were worried about in terms of their reasoning and capabilities.
This is the be a rubin and will end if it's different catch will suddenly release something at the very least frogi can do that however it seems to me that you are focusing on AMD's present performance rather than Nvidia's upcoming release they'll be prepared Jensen said maybe that's going to be cut and we're going to have a lot better chip that will make us think wait a second that 10% that's wrong additionally if you were to question what Disney is doing It's too in the details.
However, I believe you would be designing the hell out of things with Disney if they could get the back rub. It's interesting. Yes. Yes. Do you believe that customers who wish to create Disney animations on Sora will have a genuine need? No. I believe that there is a genuine market for cartoons products, entertainment, and um slow motion videos.
By the way, this is fantastic for meta. Although we don't discuss it much, the notion is that you can get better and more creative more rapidly. You've publicly stated, I think, that the annual revenue rate is about $20 billion. Does an offer like this make you consider increasing your revenue and/or computing needs?
Yes, I guess. We anticipate that this will increase revenue and provide our users with enjoyable services. We anticipate that numerous other factors will result in more revenue. Well, but we need additional computing power to deliver on that. Therefore, we shall keep expanding compute. Over the course of the following year, we'll attempt to significantly increase that 1.9 G and we'll continue from there.
Yes, naturally, we have discussed Oracle extensively this morning and a lot of the investment view on that firm revolves around your own and your capacity to fulfill the $300 billion commitment you made to purchase compute from them. Why do you feel so confident? Because people are undoubtedly doubting your capacity to fulfill that responsibility over the course of the next 5 years.
We are sure that we will be able to significantly ramp income because of the value that people see in this technology and consequently their willingness to pay already today. To say nothing of the models that we have coming in the future months and years growing this quickly at this scale is definitely unusual.
But that's what our current data shows. the model's capacity to perform economically significant work in ever larger chunks. With our upcoming model, I believe you'll begin to notice this. And with the next and subsequent models, you'll see it more. We are certain that we can keep increasing revenue in order to reach this computing ramp.
We can't drive revenue growth without the computational ramp, of course, but we see many more reasons to be hopeful than to be negative. Naturally, we will remain accountable and make any required adjustments to our plans in the event that something unexpected occurs. However, considering the current state of the market, the overhang of what these models can do that the world hasn't yet figured out how to include into their workflows is also quite significant, isn't it?
However, I do wonder if it's more costly for you, specifically in relation to Sora and the conversation we're having. It takes more processing power and I'm assuming that you're losing more money on all types of videos than on a straightforward inquiry. Additionally, do you start charging a lot? Especially because you'll also be paying Disney a licensing fee.
For Sora Generations Beyond, we do charge, some of which are free, and I'm sure we'll keep improving the economic model there. However, users have not shown that they are unwilling to pay for creating videos that they enjoy. Actually, if anything, there has been a lot more of that than we had anticipated.
Yes, it is more costly, but consumers pay and anticipate paying far more than they do for a picture. And in terms of what you may charge, you observe price elasticity there. In the market, it appears to be it's the AI trade is now under a little more scrutiny after being quite bullish. There are questions even though I am aware that you are not public.
Regarding the income question, for example, could you briefly describe chat GPT trends after Google released Gemini 3 and EON outperformed many OpenAI models? Our KPS have been less affected by Gemini 3 than perhaps we we feared. We're quite happy about a new model that will be released very soon along with a few other enhancements in response.
We have experienced a competitive threat twice in the past and I think that when one occurs, you should concentrate on it and take prompt action. We refer to that as a code red. A period of 6 or 8 weeks of intense concentration after which you deal with whatever needs to be addressed, return to a position of clear leadership and go on with Gemini.
That's what we're doing for us. This is really common place. I anticipate that by January, we'll be out of it in a pretty strong position. Yes, because it's becoming a really fierce battle in terms of both technology and market share. That's kind of the point at hand. How do you think the competition for market share will unfold on both the consumer and business fronts?
From a customer perspective, I wouldn't want to switch jobs with any organization. It will undoubtedly be competitive, but I love where we stand in relation to everyone else. With these impending launches, I anticipate that distance to grow rather than shrink. And I believe chat GPT is the clear market share leader.
I anticipate an increase in our market share. The business side is still earlier, huh? We plan to enter, but it won't be simple because we have a lot of work ahead of us. The set of services we wish to provide has, I believe, finally become clear and the models are becoming robust enough that I believe businesses can truly change.
However, the corporations vying for it will be one of the major stories of 2026. Once more, I'm thrilled with our current position and believe that our consumer strengths will convert into enterprise success. And Sam, for those who argue, well, just based on the cost per token, per dollar, per watt, to use the words of one of your partners, Satia Nadellu, guys will never be able to outperform um Alphabet.
How do you react? we that is we anticipate defeating Alphabet. We can't wait for our next chip. What's coming up with your chip? When do you think you'll be deploying that in reality? I knew you were going to ask me as soon as I said it. We'll talk about it, but we're not quite ready to discuss that aspect yet.
Additionally, our goal is to have the lowest cost per token at the best level of intelligence in the business. In fact, you shared Larry Ellison's chip neutrality statement. you know, TPUs are suddenly going to pose a threat to GPUs and CPUs are still involved. Was Nvidia in a sense de facto thrown under the bus?
Yes, the word neutrality isn't one I particularly enjoy. Although Nvidia chips are incredibly costly, I believe there will always be a substantial preference for them when given the option. The word widen is what I would use. One see the market expanding which indicates that Nvidia the market leader is still in the lead and that other firms in the lead can join.
This relates to the monetization cycle which is mostly driven by inference which also creates opportunities for more stakeholders. In the end I would strongly consider it because the market will only continue to grow. So let's start making money off of AI. You are discussing this super cycle. Many people are not taking advantage of it.
What is the inference phase exactly? What does that mean to someone who is watching the show? Yes, therefore, there is a significant upfront cost associated with the training phase, which is where we have been for the past few years. Oracle just showed that to us last night. Because inference is so closely linked to utilization, it's the first time AI can be made profitable.
AI hasn't actually been made profitable yet. Inference is therefore somewhat synonymous with revenue because it's the point at which usage begins to flow in and these large tech businesses can finally begin to display exciting income.
Comments 0