저자: JeremySiegel · 외국인
요약 보기
AI/반도체 랠리는 정점을 지났고 금리 상승이 주식시장 압박 요인. 엔비디아는 실적 발표 후 상승 전망(5년 내 두 배 가능), 마이크로소프트·알파벳은 강한 해자로 긍정적. 10년물 5%가 위험 신호이며 10~15% 조정은 매수 기회로 봄. 분산투자 강조, 일본 TOPIX 저평가 언급. 메모리 영상에서는 수요 대비 공급 부족이 지속되나 중국 CXMT 경쟁 우려가 주가 하락 요인이라고 설명.
다듬은 글
extraordinary amounts of debt, uh, to support these institutions. There are a number of large public companies that often look like private equity firms. And, of course, the construction of data centers themselves is, uh, often, uh, financed by debt. Well, I mean, and you gave the reason. I mean, they're so powerful and generate so much revenue for ISPs, hyperscalers, and everyone else.
And so, the rest of the market, you know, is not feeling very well. This, you know, was an AI or SI rally, if you want to call it that now. I mean, all that expansion that we saw in the first half of the year has probably peaked. Uh, what we call the " Leopoldian minimum" with the Aschenbrenner situational awareness fund has effectively disappeared.
We have not returned to the historical relative highs that AI and G7 stocks reached last October and November. But we've already made a lot of progress, and because they've provided the strength, that means they're keeping interest rates high . They provide profits, and that puts pressure on the residential sector and anything that's sensitive to interest rates, and anything that doesn't have 40-50% margins like hyperscalers, you know, but closer to 8, 9 , 10%—it's going to squeeze those sectors.
But, you know, they're just a part of the entire stock market. You are a fan of indexes. If you were buying an index now, would you buy a broad index or a very narrow one? And if so, which one exactly? Yes, I'm all for broad indexes. And despite the problems in Europe, you know, they really need to grow fast to provide you with a profit.
I mean, Japan has performed extremely well this year . And you can look at the Topix index, it has a PE ratio of 16. I know that stocks are low and they're not performing because I think they're being pressured by higher interest rates and higher oil prices. So they are at a relative disadvantage , but in the long run, I still think diversification will pay off for the investor and provide them with greater security than dominating one sector or another.
Let's go back to the air. To be honest, we don't see much of a reaction to the PCE report that came out this morning. I think we're in a waiting period right now while everyone waits for Nvidia, but you mentioned something in your notes that I want to start with. You say that the likelihood of a significant market correction over the next 12 months has increased.
So what potential triggers are you tracking for this? Well, obviously, the interest rate is certainly a big concern if we hit 5%. Everything will cost much more if 10-year bonds reach 5% —then we will have problems. Any, say, disappointments about earnings could also cause problems. But in reality, we are looking at a 10-15% drop scenario, and this is just a great buying opportunity.
We don't expect to enter a traditional bear market, Marcy, this time, because the earnings are so good. So you're saying you see the catalyst in earnings, potentially in disappointment with them. This leads me to ask what your expectations are for Nvidia today, as I can't say they've ever disappointed, but the response to them has been disappointing in many previous quarters.
So, what are your expectations not only from the company and its report or forecasts today, but also from the market reaction? I believe Nvidia has had a " sell on the news" over the past four reporting periods: the stock has fallen by incredible numbers. Therefore, the market has what I would call a fear of large numbers.
We've never had a five trillion dollar valuation before, Marley. So there is a lot of indecision here. But you know, Nvidia is reporting today , and the earnings will be amazing again. Their margins will remain unchanged, and they will provide us with updated information on the overall state of the AI ecosystem and its costs.
And, unfortunately or fortunately, it will be a stellar result. Will this be another " news sell-off" for the fifth time in a row? I don't think so. I think Nvidia stock will actually go up this time. You know, the options show a 5% increase one way or the other , but I see the stock going up. I think the price is quite reasonable, and of all the stocks I like, this one has the best chance of doubling in value over the next three to five years.
So, doubling in three to five years— how sustainable is Nvidia's "moat" when you consider competitors and even customers? I mean, we have concerns about Nvidia's customer concentration, but many of them are now actively developing their own alternatives. How sustainable will their “moat” be if this continues?
Yes, but they still use some Nvidia chips, those who don't design their own GPUs. And so, you know, I think if you look at their "moat," it's not just about the chips. This is CUDA, it's networking technology, software, and an ecosystem that is deeply integrated across all cloud providers and AI developers.
Further investments in AI and the next Rubin cycle will help the stock rise. Expectations are extremely high, Marley. So any slowdown in AI spending could hit the multiples of all these stocks pretty quickly, especially Nvidia. What other big tech companies do you like right now? Of course, Nvidia is in the spotlight because of its reporting, but is there anyone else you like in this sector or are you holding back on buying?
Well, you should be smart about what to buy in the technology sector today, because valuations are very high. So look for companies with a wide "moat." You know, Microsoft has a great "moat." Nvidia has a great "moat". Alphabet has a great "moat". Microsoft has one of the deepest enterprise ecosystems in the world thanks to Azure, Office, Security, GitHub, and Copilot.
Customers are already tied to this platform, which gives them huge opportunities to spread AI. In other words, they own the railroad , on which the AI is built , and the AI drives the train along these tracks. Alphabet is also a great investment option given its valuation and "moat." If you look at search, YouTube, Android, Cloud, Gemini—this is one of the best distribution advantages in the technology industry .
This is also the company I like the most in terms of evaluation regarding the quality of their business. Um, but that's where we are today. So, you like them, you find them the most attractive in terms of value. Is it because markets are underestimating Google's AI capabilities, or perhaps even overestimating the threat it poses to search?
Because it seems to me that every time they release a report , I'm talking more and more about how AI will impact search, and search seems to continue to perform well . I think Google is in trouble, and if I had $100,000, you know, I would put 50% in Nvidia, 30% in Microsoft, and the rest in Google.
But you're right, generative AI has the potential to disrupt the search economy at Google. And we have n't seen the full extent of it yet. We saw a slight decline last quarter, but I expect this trend to continue as people become more reliant on AI-powered search. I mean, we're already doing it. Everyone is already doing it.
And I know that you personally own the three companies mentioned. Your firm recommends them. How do you separate your investment conviction from the fact that you also have a financial interest in those assets ? Well, you know, that's good news because we don't go back on our word, right, Marley? I mean, we want to believe in what we recommend.
There is no ulterior motive here. Our investment dynamics will not affect the valuation of these companies. So, you want to own companies that have wide “moats” and high market valuations. They will have revenue growth regardless of what happens to AI spending, and they are attractively valued. And these companies, despite the fact that they are huge, and we are dealing with the fear of big numbers, are currently attractively valued.
So look at it this way. Nvidia gives us the greatest growth potential. Microsoft gives us the best quality, taking into account risk, and Alphabet gives us, let's say, the most compelling valuation. Before we begin, subscribe to this channel for news and analysis on Nvidia, AI, and the stock market.
Shares of memory manufacturers are falling. Investor confidence is weakening, and the market is starting to question one of the biggest assumptions behind the AI boom. But there is a strange twist. Demand for memory chips still far exceeds available supply. So why are Micron, SK Hynix, Sandisk, and other memory companies facing such a massive sell-off?
And this is part of a deeper problem. Is it possible that the market is overreacting to short-term fears? Let's figure it out. The DRAM ETF, which tracks companies in the memory industry, is headed for its third consecutive negative week. This would be its longest streak of weekly declines since the fund launched earlier this year.
It has large holdings in , among others, Micron, SK Hynix, Sandisk, and Western Digital. They were all included in a recent sale . But according to Dave Mazza, CEO of Roundhill Investments, something important is happening in this industry . Memory companies may be moving away from the business model that investors have known for decades.
To understand why this is important, we need to look at how memory manufacturers' stocks have performed in the past. Historically, the memory manufacturing business has been very cyclical. As demand grew, companies expanded production. As supply increased, prices often fell. Profits were declining and stock prices were under pressure.
Then the cycle started again. This “boom and bust” pattern made investors question the appropriateness of high valuations for memory companies: even with strong earnings, investors often believed that those earnings would eventually decline. But the boom in artificial intelligence is changing the situatio to process and use information.
This makes memory a critical part of the infrastructure for advanced AI computing. And demand is not limited to one type of memory. High-speed memory is essential for many advanced AI systems. Whereas regular DRAM is also needed for servers and data centers. As AI infrastructure expands, memory becomes a major bottleneck.
According to industry comments in this discussion , the available supply for 2026 and 2027 is already largely booked, and buyers are looking further afield—towards 2028. This is an important detail. If demand remains high and supply is limited, memory manufacturers could gain greater pricing power and better visibility into future revenues.
But there is a problem. A strong market does not function solely on the basis of business indicators. It also responds to interest rates, credit conditions, competition, and investor expectations. And now several of these forces are working against the shares of memory manufacturers. One concern is the possibility that China will become a stronger competitor.
The company at the center of this discussion is CXMT, a Chinese memory chip manufacturer. Recent reports have suggested that CXMT is researching a new technology that will eventually allow it to compete in a different segment of the memory market. And this news has raised concerns about future supply and prices.
But there is an important difference. CXMT has mostly focused on lower-margin memory products. The technology mentioned in a recent report will require further development, manufacturing capacity, and time before it can have a significant impact on a large scale. Massa argued that this potential threat may be more of a market fear than an immediate change in the industry's fundamentals.
In his opinion, a more favorable market environment might have perceived this news as a less significant event. Instead, investors are already worried about AI-related trading. This means that even an unconfirmed report of impending competition could trigger a much larger reaction. Time matters. Investors are looking at more than just what memory companies can produce today.
They are also trying to estimate what the market will look like in 2028, 2029, and 2030.
Comments 0