저자: TomLee · 외국인
요약 보기
연준은 주택·임금 약세로 코어 인플레 압력이 약하다고 보며 금리인상 사이클 필요성이 낮다고 봄. AI 트레이드는 아직 유효하나 마진부채 축소와 실적 우려로 조정 중이며, 하이퍼스케일러 설비투자 삭감 우려에도 설비투자 사이클은 계속된다고 주장. 이익 모멘텀 피크 통과로 반도체·대형기술 리더십이 약화되는 '퀄리티 로테이션' 진행 중이라 진단하고, 조정이 거의 끝나가며 S&P500은 횡보 후 상승 가능성 언급. 엔비디아 관련 Hugging Face 인수(129억 달러) 영상도 포함되어 오픈소스 AI 생태계 확장 중요성 강조.
다듬은 글
And funds want to insure themselves against a binary event. That's why you have 30%, because someone needs to insure something. But I would say the probability is low. I mean, I wouldn't expect them to raise the stakes. What about any point in the future? You know, they 'll watch how the data develops .
So I don't think it's a statement addiction, but rather a reaction to the data. It seems to us that the core inflation story has really weakened because the housing sector is very weak, and we will see that tomorrow in the Case-Shiller report. And if wages are not putting pressure, then the only real source of core inflation we are seeing right now is tariffs and oil, and that is not what the Fed needs to start a rate hike cycle for .
I don't want to use that bad word " transient". They are potentially temporary , right? Yes, and oil is falling now, because certain negotiations are underway, and, by the way , it may rise again tomorrow. We have no idea what will happen. Depends on the negotiations. But the bottom line is that if and when some lasting peace comes, the price of oil will likely go down again.
The market seems to want to push it back down. This should influence the Fed's thinking . Yes, because then they just need to see this little bubble that they have to somehow overcome. I mean, in a sense, they could just reduce the balance sheet instead of changing the discount rate. Quantitative amplification?
Right. Yes, I think that's the best way to dose. They can say, "Hey, let's try to put a little pressure on growth, but not to the point of intentionally slowing the economy." Could this somehow upset the stock market a lot? You know, the stock market will eventually see the idea that "the Fed is reducing the balance sheet, which means they can lower rates." And then they will perceive the rate cut as something positive.
So I think this will actually pave the way for future rate cuts. So if the Fed isn't the most important factor for stock markets right now, then what is? Well, I think AI trading remains the most important story, and people, of course, have doubts about its durability. But you know, if you look back at the period from '94 to 2000, there were many moments when the history of the Internet and even stocks like Cisco raised doubts about their sustainability, and I think we're going through just such a moment of doubt right now.
But I think it's still in very good shape, and the second big problem is that the margin debt still needs to come down after such a high level of growth, as happened in Korea, where there was a kind of margin call, and I think that's why the stock is slowing down now. But, as for me, I believe that AI will continue to work actively until the end of the year.
Wonderful transition . This morning we had Steve Eisman on Squawk Box, a well-known investor. Here's what he said when asked the big question of the reporting season: what would happen if Big Tech started cutting capital spending on AI. Nvidia, I think when they reported last quarter, had 85% revenue growth.
So if hyperscalers cut costs, it would be 85%, and, you know, maybe that would be good for the long term, but I think the market would go down sharply after that kind of news. It's interesting to know your thoughts, your reaction to this. Uh -huh. And do you think that's even likely or worth discussing now, as opposed to what Google did—raised costs and still got penalized?
Yes. Well, on the one hand, I would say that there is some sense in Steve's logic , but the fact that so many people are talking about it is a sign that we are not at the peak, because people are questioning the length of the cycle. So, I think this is actually an optimistic signal . Secondly, do you know if this is likely?
I doubt there will be any cuts because these companies still have access to the bond market and the bond market is not denying them capital. So, as you know, CFOs raise money when they can, not when they need to. So I think the visibility of the costs will be very high. We were premature because people thought that there would likely be problems in the labor market and we could end up in a recession.
You said that we have already gone through a kind of cyclical recession. We are now in a state of cyclical recovery, and you still think so. Absolutely, and now we move on to the next stage. So, this week we turned to this, as in 2021, right? After the big recovery after COVID , the market around mid-2021, we had a peak of the review factors, right?
We've been talking about this for the last month or two. One of the reasons we were skeptical about semiconductors is that the peak rate of change for reviews has already been passed, leading to what we call quality rotation. So the expansion story is still working, but it's transforming , as you just talked about earnings, right?
We see Sherwin-Williams doing well, and then we see some of these technology stocks doing great numbers, but they get sold off. So, what's the deal here? It all depends on the peak rate of change. This does not mean that the capital expenditure cycle is over. It simply means that we are overly optimistic about the future-trading stocks.
They take a lot into account , just like they did a year ago when multipliers increased. Multipliers are now being lowered in anticipation of a slowdown next year. So this is exactly the adjustment we were expecting. Initially, you thought capital would flow out of the G7 and mega- cap stocks into economically sensitive sectors.
It has already happened. Well, early cycle. Remember, the economy goes through an early cycle, a mid-cycle, and a late cycle during a recovery. The early phase of this gradual recovery is already behind us. OK. By the way, this does not mean that all these stocks should fall to zero. It simply means they are no longer leaders.
Well, Sherwin Williams—is that quality? Maybe. I mean, we published a list, they're from many different industries. It may not be so sector specific. Now it's more company specific. And as you know, hyperscalers, the reason their stocks have lagged is that while their earnings have been good, their free cash flow generation is terrible.
And so the market punished them. I think a lot of that is behind us now, and they've moved on to semiconductors and some of the storage companies. This is also a fairly advanced stage. I mean, these corrections are quite serious. Look what happened in Korea last night. So, now that people, you know, agree with us on this, I'm probably more inclined to say that we're probably closer to completing this correction.
And, ironically, you know, the S&P 500 is the highest quality index in the world. Okay, let's be honest about this. But at the end of such corrections, you usually get, you know, an increase even in the index itself. So, I still think we're stagnating. I think it's because of the uncertainty of the Fed, because of the war and Hello everyone, welcome to our channel.
Before we break down this huge Nvidia deal, subscribe, because this channel is about the most important moves shaping AI technology and the stock market. Nvidia is officially buying Hugging Face for $12.93 billion . And at first glance, it might seem like Nvidia is just buying another AI company. But the real story is much bigger .
Nvidia gains direct access to one of the largest open source AI development communities in the world. And this could become incredibly important as the AI industry has split into two competing development models. So how did this deal even happen? Hugging Face CEO Clement Delang explained that talks began in the summer.
His team believed that open-source AI had reached a tipping point. The technology has become powerful enough to support much more serious applications. But Hugging Face also needed additional resources and scale to take this ecosystem to the next level. So they turned to Jensen Huang. Jensen Huang's response was essentially this: they should build it together.
But the relevance of open source AI did not arise from a single event. A notable example was when Hugging Face faced a cyberattack that used advanced AI capabilities. The team realized that some of the technology used against them came from a model that was not publicly available. This experience raised a serious issue.
If advanced AI is developed entirely behind closed doors, defenders may not have access to the same level of technology that sophisticated attackers use. This further emphasized the importance of open models. And now the AI industry seems to be moving in two very different directions. The first is a world where powerful companies build their own models, and everyone else simply accesses these systems through APIs.
You don't own intelligence , you rent access to it. The second is an open ecosystem where developers and organizations can access models, modify them , improve them, and build their own AI systems. Hugging Face wants the second model to be much more powerful, and Nvidia wants to accelerate that. , quite the opposite.
He openly admitted that closed AI services are extremely useful. They are ready to use. They are improving rapidly, and companies can use them immediately without spending a lot of money and time building their own systems. Wong even mentioned that Nvidia uses . So why would anyone need open models?
Because not every company can simply send its most valuable information to an external AI provider. Imagine a pharmaceutical company working with confidential research, a scientific organization with its own datasets, or a highly regulated business that needs full control over its technology. Or a company that aspires to have its own national artificial intelligence capabilities. In such cases, ownership and control become crucial.
Comments 0