저자: JimCramer및게스트Mike,VinceLaruso · 외국인
요약 보기
CoreWeave가 Nvidia Vera Rubin 플랫폼 첫 고객(Cognition)을 확보했다고 발표, 데이터센터 규제가 수요를 줄이지 않고 입지만 옮긴다는 주장. 금리 상승에도 마진이 금리보다 빨리 늘어 수주가 증가한다는 설명. Vince Laruso는 AI 투자가 GPU 병목에서 소프트웨어 매도, 이후 인프라 병목 중심으로 진화 중이며 메모리 사이클이 길 수 있다고 봄. 어도비·인튜이트는 AI 대체 위험으로 우려, 사이버보안은 수요 증가 기대. Nvidia는 Morgan Stanley 상향(목표가 $300)과 FY2028 매출 성장률 70% 가이던스로 신고가 경신, 자사주 매입 여력 약 $235B 언급.
다듬은 글
This could continue the demand for hyperscalers for AI computing, but of course there are also the controversial points we just mentioned. This morning, CoreWeave announced its first client on the Nvidia Vera Rubin platform, the current iteration that Nvidia has. So join us now, for the first time on CNBC—CoreWeave co-founder and CEO Mike, who is not new to our show.
Hi Mike, you weren't at the conference yesterday, you weren't with the president. Tell us where we are now and what you're doing about building data centers, about the objections people might have, and, frankly, about this initiative with Vera Rubin. Good. So, first of all, thank you for inviting me.
I always appreciate the opportunity to perform on your show. We are currently in San Francisco for our Fully Connected conference. This is very exciting. It's fantastic to be here. We were expecting 2,000 people. And 5,000 people came and joined us to talk about how to build on the CoreWeave cloud. This is really very interesting.
Vera Rubin. OK. This is an incredible moment to introduce the latest technologies . We have attracted our first client. This is the first customer in the world to launch the Vera Rubin architecture into production. Cognition is our wonderful client. Great company. They will be able to use this technology to really grow their company, and I think you will see this technology spread across the cloud as we bring more and more customers onto our infrastructure.
Okay, Michael, let's talk about Nvidia's involvement. From what I'm reading, Nvidia should have a CPU partner, but Nvidia is the leader. Who do you work with to implement Vera Rubin for a new client? Yes. Look, the thing is, Vera Rubin is an interesting technology because it combines a GPU, as everyone knows, and also a CPU from Nvidia, and that's how they combine it.
We have a huge number of processors outside of the Vera Rubin server that we ship, but integrating these two pieces of hardware into one server is incredibly powerful and will really provide significant returns on the investment that companies make because these two components are so well synchronized within the same server.
Is this a processor from ARM Holdings? it's an Nvidia processor that they're shipping, so it's a really exciting product for us. Yes. Good. So, since we communicate quite often, I want to know: you had a great presidential meeting yesterday with all these people, and everything is going according to plan, but you have worked with different governments over time, including some democratic governments.
Where are we at now in terms of the backlash from states saying, “Oh, we can’t build this,” and what does that mean for you? Is there a decrease in the number of orders? There was absolutely no decrease in orders. I want to make it clear that the regulations being implemented for data centers will only move the infrastructure to another location.
They will not reduce demand in any way. The market needs more infrastructure because there are companies that build it and profit from its creation, and it's just a question of where to put it, nothing more. Michael, is the betting situation becoming problematic? Listen, you are absolutely right. Rates are rising, making everything more expensive.
Actually, that's why the Fed raises rates—to make everything more expensive, to slow down the economy, to beat inflation, right? We all know this. The important thing to take away from this discussion is that the margins we get are growing faster than the rates. So the overall profitability of the infrastructure we are launching is higher than it was a year ago.
Therefore, you cannot consider rates in isolation from everything else. We need to look at the economics around AI comprehensively, and comprehensively we get more orders because demand is growing and the profitability of each new element of infrastructure is increasing. This morning's data: The " Magic Seven," software companies like Adobe and Intuit, some big outsiders, and generative AI stocks like HPE, are all performing well today .
Vince Laruso— CEO of Cloud Capital Partners . He's here to...You know, Vince, when I look at the charts, at the markets in general lately . I'm having a little trouble, maybe you can help me, to formulate a clear picture of what's happening. What does this basket of stocks showing growth today signal to you?
Yes, hello, Kelly. I think there is an evolution of AI stock trading happening right now. You know, I think the first stage of this was the shortage of GPUs , the "bottleneck", and wanting to have everything related to GPUs: semiconductors and memory. And on the other hand, you wanted to play the bearish game in everything related to software, because of the fear that AI would replace it.
And I think as this investment thesis develops, more nuances emerge. You want to look specifically at the AI infrastructure: where do we have “ bottlenecks” today? We must recognize that some aspects of this process are cyclical. This may be one of the longest cycles we've ever seen, perhaps with new peaks in the memory and chip sectors.
But ultimately, what we're thinking about here at Cloud Capital today, and analyzing these steps: where do we see investment- attractive companies in "bottlenecks" that will have a return on invested capital in the future, not just during this cycle? Let me get to that, but I like where you started.
You said that the first stage is a game of upgrading the GPU and downgrading the software. What is the second stage, or are we already in the third? What kind of deal are market players making now ? Because I think I see something like long positions on Mag 7 and agent AI. And I'm not sure, I can't even describe it, what would you call it?
Yes, I think we're approaching this even more subtly. That is, "Mag 7" now consists of companies with different business strategies, different models, different—I mean Meta, as an example of an evolving company. You would historically think of it as a social company , generating significant revenues from WhatsApp, Instagram, and Facebook, but it seems to us and the industry that they are turning into a “railroad,” that is, the infrastructure itself.
So, I just think that's the advantage of fundamental stock selection at Cloud Capital—we can analyze individual companies in detail. I think it's harder to apply general approaches here, so we want to be in the infrastructure layer, and the markets seem to be saying that some software looks very promising, like cybersecurity.
Others may actually be like that . So, I think now is the time, maybe we're in the middle of the game, as they say, when you can act a little more tactically. I know you were skeptical, correct me if I'm wrong, about companies like Adobe and Intuit. They have become one of the underdogs even as the rest of the software sector has caught up with the market, these stocks are still 20% below their peaks, although they have rebounded a bit today.
I'm curious, again looking at each company individually, do they remain in the category of those that are losing their intermediary role for you? Hmm, are there any developments? Are they on the path to recovery? Do you have your own opinion on which software stocks, these and others, will rise now?
Yes. So I think the question becomes : as this functionality that you mentioned is rolled out, as we have Muse and other tools that are able to analyze the technology stack and the software stack , what will users be able to integrate into their AI ecosystem? Then you start giving up some of the services you already have.
So I think if you're—I talked earlier about a volume-based licensing company like Adobe—if your business model is based on a certain number of users paying a monthly subscription fee, and suddenly AI is able to provide the same services. I think such a company could be in danger. You know, if AI starts doing some of your tax work.
Um, but such integration doesn't mean that all software is now out of favor. I think the first move of the market was to sell off everything related to software. We currently own a number of cybersecurity companies in ETFs that I manage. And our view is that as we have more endpoints, agents, and identifiers.
You will need more cybersecurity, not less. Yes. And, you know, you can analyze it yourself . People can search for this information themselves. Many of the names in those baskets that you are more cautious about include some financial companies. As you say, you want to own the toll road, not issue loans against it.
This applies to Oracle, SoftBank, Blue Owl, etc. Um, you also have to be careful with the late stages of the consumer sector. I want to finish by asking you about Micron. How important is today's earnings report to what we think is happening with the entire AI sector? This will be very important. You know, I think memory remains a cyclical part of the market.
And as I mentioned earlier, this could be a huge cycle, with new peaks, and it could last longer than we've seen. But will the market be able to start reading the return on invested capital as demand, particularly for high- bandwidth memory, as strong as we expected? I think the trade will continue, but I think it's getting a little more complicated.
You know, some of the comments about demand and growth expectations are going to be hard to reconcile with the stock price movement , right? Because the stock has risen so much, I think the market will start to differentiate more, but we can get data from earnings results and extrapolate from that: where is the strength in this trade?
Is there still a return on investment? I think, Kelly, that the financiers of this will have some risks along the way. We won't be able to land completely smoothly. This is such a huge investment cycle. There will definitely be some glitches. But if we can work and find the main winners, which are the "toll roads" that collect fees for building this infrastructure, we think they will have very good business models.
And the memory data we get tonight will be informative about the duration and ultimately the height of this cycle. Before we get to today's story, if you want more in-depth stock market analysis, exclusive information, and members-only content, consider joining the channel. This will help support the channel and give you access to additional content dedicated to the market.
Now let's move on to Nvidia. Nvidia is back in the market spotlight. The stock just hit a new all-time high after Morgan Stanley reinstated Nvidia as a top pick among semiconductors, and the reason is pretty simple. Analysts are closely watching two things. Rising earnings estimates and the fact that investors are increasingly willing to pay for these earnings.
Morgan Stanley believes Nvidia's long-term revenue outlook may actually be stronger than management's current projections . Nvidia assumed approximately 17% annual revenue growth for fiscal year 2028. But Morgan Stanley believes actual demand could be closer to 100% growth, and the firm has set an “ outperform” rating on Nvidia and a $300 price target , and it comes at an important time for the company.
CEO Jensen Huang is also spending time in Washington with top AI leaders, working on a voluntary framework aimed at addressing risks associated with artificial intelligence. Meanwhile, Nvidia's sales have already grown by more than 21% this year, and now the stock is reaching new highs. So, the big question is: is this just another Nvidia rally, or is the market starting to realize something more?
Let's start with an optimistic scenario. One argument is that Nvidia has been growing profits for a long time, while its valuation has not grown at the same rate. The stock traded in a range of roughly $195 to $250 for a significant period of time, even as earnings continued to rise. The stock price had difficulty moving outside this range.
At that moment, something interesting was happening. Earnings have been rising, but the valuation multiple has been gradually declining, and now the market may finally be changing its view. One factor that is attracting attention is the huge share buyback program. The company has about $235 billion in available funds for repurchase.
This is significant financial power. Apple has demonstrated how powerful consistent share buybacks can become over time, and Nvidia could potentially leverage its own balance sheet in a similar way. Of course, it remains one of the most popular stocks on the market. So it will move in accordance with general market conditions.
But the underlying sentiment regarding these stocks appears to be changing. There's another reason investors are becoming more optimistic: the company is no longer just a GPU manufacturer, what can effectively be called a full stack for AI computing. You have a GPU, as well as the CUDA and Nvidia software ecosystem.
You have networking products like Spectrum-X, Ethernet, and InfiniBand. Nvidia is also moving deeper into the CPU and integrated computing space with its newer architecture . So, instead of selling a single component, Nvidia is increasingly providing the entire infrastructure needed to run AI. This is a much stronger competitive position, and the numbers show why investors are paying attention.
The data center business grew by more than 100% year- over-year, reaching approximately $8.9 billion in quarterly revenue during the period under review. This is a huge business in itself. But the more interesting part is the company's long-term outlook. Wall Street had expected growth closer to 45–50%, and the company issued a much more aggressive revenue forecast of 70% for fiscal 2028.
This changed the conversation because when a company this big tells investors that it can still grow at this rate, the market has to reconsider what its future earnings might be.
Comments 0