저자: DanIves외GabelliFunds포트폴리오매니저JohnBelton · 외국인
요약 보기
AI 인프라 투자가 실질 수익으로 이어지고 있어 반도체(특히 엔비디아) 펀더멘털이 견조하다는 주장. 엔비디아 실적(8/28) 기대감, 공급 제약 지속과 Vera Rubin 출시 임박 강조. 테슬라는 AI 투자 성과 부족으로 우려가 있으나 Ives는 장기 AI 혁명의 초입으로 봄. AMD도 에이전트 AI 수요로 수혜 기대.
다듬은 글
the thing is, investors are simply terrified of their spending on AI. Our next guest says: don't worry. Joining us is Dan Ives from Yorkville, Ives&Co. I mean, that's almost your motto. Don't worry, be happy. But in this case, let's talk about Tesla first. I just read: this company's market capitalization is greater than that of the next 36 largest automakers combined.
Isn't it time for her to satisfy her appetites a little ? this is on Power Lunch. Thank you. Please. Look, the reality is, they're not just a car company. When you look at it from an investor's perspective, it's more about disruptive AI technologies. And I think when you look at Tesla, the problem remains , as it does at Alphabet, in the capital investment, let's say, at Tesla—in physical AI, when it comes to Autopilot or Optimus.
But you don't see results in the near term, and this continues to be a test of strength. fell by more than 14%. Does this mean that investors simply don't believe in it? I think patience is running out, isn't it? Because from an investor's perspective, if you show capital expenditure figures and back them up with results, you will be given credibility.
And I would say that Alphabet is a different story compared to the growth rates they are showing in cloud technologies. I think for Tesla this is not something that will happen tomorrow. You need to increase the production of robotaxi. You need to show a success story with physical AI. Look , that's all Musk does.
And I think there's a view that with SpaceX and, of course, Tesla— Musk's two darlings—it's unclear where this will ultimately lead. I continue to believe that we are still only at the beginning of the AI revolution, and this is just the beginning of these stories. We'll get back to AI later, but to expand on the topic: does Dan Ives of Yorkville Ives, by the way, think Tesla and SpaceX will eventually merge?
I think there is over an 80% chance that by the end of next year, SpaceX will eventually absorb Tesla. It's that you are more optimistic. The spotlight is on John Belton, portfolio manager at Gabelli Funds. Nice to see you. What a vote of confidence from Elon Musk to say that Nvidia's Vera Rubin architecture and their processors are the best, and that's why Elon Musk's SpaceX wants to work exclusively with Nvidia.
What do you think about this? Because it certainly helped boost Nvidia's weekly performance by 10.5% this week. Yes, thank you, Nicole. I mean, besides being a key vote of confidence for Nvidia, I think it has clear implications for financial performance as well. Another thing SpaceX talked about this week is plans to add 4 to 8 gigawatts of computing power next year.
This effectively means that they will scale their AI infrastructure at the same pace as the largest hyperscalers. So the AI infrastructure market from SpaceX will see real money coming in next year. And it seems that a significant part of these funds will go to Nvidia. Of course. We'll talk about AMD a little later, but for now let's stick with Nvidia.
We are also awaiting earnings reports. We've already heard a lot from hyperscalers regarding capital expenditures, and Nvidia will report in a few weeks, on August 28th. What are your thoughts on Nvidia's upcoming news and earnings report? Will you pay attention to any details there? Tell me more . This is going to be another great quarter.
What have we learned from the hyperscaler reports in this cycle? We learned, first, that every hyperscaler is still limited in supply, the demand is too great to meet. Second, hyperscalers' capital expenditure forecasts continue to be revised upwards . We also saw several encouraging demand signals during the quarter, for example, H100 rental prices continue to grow quite strongly.
Token consumption continues to grow rapidly. So all of these factors point to a very strong demand environment for Nvidia. This should be another strong quarter, and beyond the numbers, we expect to hear the first signs regarding the Vera Rubin product launch , which will begin next quarter, in October.
The company's stock has also been somewhat of an underdog when you look at the semiconductor group, although as you noted, valuations here are historically low. Can you say now that many, looking at the P/E ratio and valuation, find it more attractive than at any time in a long time? Would you agree with that ?
Yes, I think given the fundamental profile, the stable demand that they're seeing, the growth that they're still showing, combined with the low valuation, it still looks like a good risk-reward ratio, especially if this AI infrastructure investment cycle continues beyond 2027, which it now looks like it will.
When you mentioned AMD in your notes, we also heard from AMD that they also had an AI conference. What are your thoughts when you compare these two companies? AMD is often considered something of a little brother to Nvidia. Is AMD in good shape, in your opinion? Yes, I think they are two separate companies with different strengths .
AMD is clearly benefiting from the surge in demand for processors from agent AI. This market seems to be really growing rapidly. AMD has been increasing its share of this market for a long time. It looks like this will continue. I think the outlook they've talked about for the coming years for their server processor business is very strong, and I think that's the basis of what's driving AMD's earnings.
Additionally , AMD is trying to enter the market as a second commercial GPU vendor to compete more directly with Nvidia. They are also releasing a new series of products later this year. I think by definition they are going to The AI sector just got a significant vote of confidence, and this time the argument is n't just about how much companies are spending.
And whether these expenses bring real profits. Because this was one of the biggest questions hanging over the world of AI. How do you justify the hundreds of billions of dollars spent on GPUs for data centers, networks, and memory? And the latest results from hyperscalers are starting to provide an answer.
Amazon Web Services grew by 37% and Microsoft Azure by 43%, with even faster growth expected next quarter. And Google Cloud's business showed an extraordinary growth rate of 82%. But income growth is only half the story. Amazon and Google also received over 15% additional operating margin on this growth.
This changed the course of the conversation. We are no longer talking about companies spending huge amounts of money with no visible path to profitability. We see businesses spending heavily as customers demand more computing power, while the infrastructure itself becomes increasingly monetized. And that is why the development of AI may continue for a very long time.
And the most important constraint right now is not necessarily demand. These are our capacities. There simply aren't enough computing resources to satisfy everyone. And when supply remains constant and demand continues to grow, the companies providing this infrastructure gain enormous pricing power.
And this is where they become particularly interesting. Stocks had one of their strongest weeks in a long time. Growth of over 10%, but the more important story is not just in the price momentum. And the fact that Nvidia is starting to attract attention again at a time when investors are actively looking for a new source of momentum.
After a month of lagging in the semiconductor sector, Nvidia is starting to look not like yesterday's winner , but like a potential market leader. Fundamentals still play a crucial role, remains high, and Nvidia's new architecture is nearing release. And customers are still dealing with supply constraints.
And this is a rather unusual combination. Typically, when a company becomes so dominant , investors immediately start asking when growth will slow down . But the current problem seems positive. How quickly can the industry semiconductor trading is risk-free, not at all. One of the biggest risks is competition.
When margins become exponentially attractive, competitors eventually emerge . This is capitalism doing what it does best. High profitability attracts capital. Capital for capacity creation. Capacity ultimately puts pressure on pricing, and this is a traditional cycle for semiconductors. But the debate today is whether demand for AI is growing fast enough to absorb this new capacity, which is why Micron is particularly exciting.
Memory , but AI has created a new source of demand that was not present during the previous cycle. High- bandwidth memory is becoming critically competitive.
Comments 0