저자: TomLee · 외국인
요약 보기
미국 재정적자 확대와 장기금리 상승(10년물 4.71%)이 구조적 문제지만, 채권시장은 이를 반영 중이라 주식시장은 괜찮다는 입장. 유가는 80달러 수준이 정상이며 경제는 양호하다고 봄. 엔비디아 투자 확대는 공급 생태계 지원 목적이며, 광통신부품·반도체장비·프로세서(CPU) 쪽에 사이클 초입 기회가 있다고 주장. 수요 과잉(버블) 우려에는 실수요 기반으로 반박.
다듬은 글
What is your reaction or opinion on this? I mean, it's pretty shocking that we have an economic boom and the deficit is growing. Um, so Rockets are expensive. Yes. Interest is expensive. Yes. Yes. Yes, it seems like there is a lot of outflow happening. I agree with those who believe this is a structural problem.
But the bond market is signaling today that it fully agrees with this . So I think as long as he agrees with that, the stock market is fine. So that little jump in 30-year bonds that we saw when the deficit suddenly became an issue last week during the sell-off, you don't think that was a really significant move or that it's going to last?
Yes, this is a symbolic rate, because, as you know, perhaps more companies depend on the spread to 10-year bonds . True. So, that's a more important bet to watch, but again, I don't think it makes sense to not be a financially stable country. So that's a problem. It's very hard to fathom, you know, what's strange is, this 10-year bond range has been so narrow for so long, and everything points to it being higher, but that's the craziest part.
He was below four. I take Rick's words from his mouth. We were below 4% on the 10-year just before the Iran war, and oil prices skyrocketed. But now the price of oil has fallen, and 10-year bonds are at 4.71 . You could argue that oil should be more expensive today and all that, but it's not. Yes. Dryden, a brief final thought?
Well, the short answer is : 80% oil is normal. At just above 80, we're fine. The economy is doing well, and when the price drops to 70, we get $3 gas again . So I think it will be resolved sooner or later. Listen, economics always trumps politics. I'm not sure about that. And sooner or later we will find ourselves in a situation where we have a better economic Dryden, I want to believe you.
I hope I can believe you. Tom, thanks so much for joining us , guys. A lot of the talk around Nvidia is about cyclical funding. I think it's also interesting to talk about this in the context of what her portfolio is like, right ? After all, it is quite unusual for a large public company to invest in other public companies, including some of its clients.
Um, should people follow Nvidia's portfolio and invest in it on their own? Well, definitely Julie. I think what's happening is this : Nvidia is trying to support the ecosystem, and they're investing in two ways. First, they try to support their suppliers, so wherever they see any bottleneck or something that could be very important for their expansion, they invest there.
So from that perspective, they usually turn out to be very good opportunities. And on the other hand, they're trying to support the customers or the people who buy their chips, and that's a little more debatable, right? So, there are companies here that may not be able to get financing on their own , and with Nvidia's support, they will be able to get better credit terms.
Therefore, I believe that if you plan to follow Nvidia's actions, the signals are much more significant from the suppliers than from the customers. By the way, at least some of Nvidia's investments coincide with what you invest in your ETF. I have actually reproduced the top 10 assets of your portfolio if they are correctly reflected on our Alpha Space platform.
Many of them are companies that make optical components, but you also have companies like ARM. You have Coherent, which we already mentioned. We have Lumentum, which also just reported. In addition, if we look at the results of many of them since the beginning of the year, even despite the recent volatility, the indicators, as we see, are quite good.
The white line here is Coherent, which is showing truly outstanding results this year. So how do you see suppliers being the companies worth being with? It's obvious that you're betting on suppliers. Tell us about this thesis , because the " picks and shovels" strategy has been popular for quite some time .
What stage of this cycle are we at now? Well, Julie, what happens with the "pick and shovel" strategy is that the bottlenecks move from one place to another. For example, at the beginning of the year, a big bottleneck was memory. Looking ahead to next year, we believe that optics will have the same dynamics as memory, meaning it is just as cyclical and there is a shortage of supply.
So, companies will have to rush to increase supply, but they won't be able to do it in time. Therefore, optics is one of those areas that we really like. Another area is capital equipment. As memory and processor capacities need to be increased, you will see semiconductor manufacturers benefit . Companies like Lam Research and KLA supply such equipment.
This is another area where the cycle is just beginning, as companies are just starting to invest, unlike GPUs, where we are already several years into the cycle. Okay, when you say they're investing in this, are we already seeing these types of companies, whether it's optics or semiconductor equipment, already building up capacity in anticipation of growth?
That is, they have already seen a surge in demand, but are they expecting it to continue? Certainly. In terms of optics, most companies plan to double their capacity this year and double it again next year. But judging by the demand, it could actually be 7-10 times larger. This is because in addition to their traditional business , such as supplying fiber optic cables, there is also a data center sector that is already growing.
Now optical components will go into the servers themselves for scaling, and this is a whole new market that didn't actually exist before. That's why demand is growing 7-fold, while supply barely manages to double every year. And yes, we see this everywhere. So, semiconductor equipment manufacturers will be the companies that will provide this doubling of capacity.
Of course, you are well aware of any questions or concerns about this cycle. There was an interesting article this morning by Bloomberg macro strategist Simon White, who, in particular, pointed to Nvidia's new financing structure: it is partnering with Wall Street giants to potentially inject $500 billion into the system to fund this computing power.
In his article, he drew parallels between this and the mortgage crisis, and in particular, AIG's financing schemes. And he says, "During the housing crisis, the rise of financialization artificially inflated demand, drowning out the market's signal that there was already too much supply overall. Home prices fell nationwide for the first time since the 1930s, an event that no model had accounted for." So, do you have any concerns in this cycle that because of the huge amount of money that is being poured into the system, if there is a change in demand from the end consumer, it will be masked?
Well , Julie, definitely . If demand falls, when we have such an influx of supply— and these industries are cyclical. That 's why tracking demand is so important. We are talking to companies in both the private and public sectors, trying to assess what demand actually exists. And there is also a way to track spot prices.
And now we see a completely opposite trend . People don't necessarily chase them like they do with housing. The difference is that people buy them as investments or for resale. People here are not reselling GPU power. They buy them for their own needs. So what we're seeing with pricing... You could argue that most homes are bought to live in, right?
Well, when you see big booms, they're usually fueled by investment activity, right? So people don't necessarily buy 10 houses to live in, right? This really has to be something different. Bubbles are more driven by investment activity, right? Same here: if you start seeing companies that are perhaps renting GPU capacity without having any real buyers, that would be a concern.
But now we see more buyers for these capacities who need these chips to work. What about the fact that GPUs already operating in data centers are in many cases not being used to their full capacity? Well, that's something that will change over time, and that's where we see a lot of opportunity today .
So companies that offer ways to better utilize GPUs and, by the way, CPUs, really benefit from this. So any company...Cloudflare is an example where they have a wide network of CPUs and help companies that want to build AI applications use those processors efficiently. So companies like that really benefit, and I think we'll see a significant increase in usage over time.
So, you talked about optics, you talked about semiconductor manufacturing equipment. Are there any other areas that you think...I mean, these areas seem to be pretty clear overall, right? So are there any other bottlenecks that you see ahead that maybe aren't as predictable yet? Well, I think optical components are relatively new, as is semiconductor manufacturing equipment.
Processors are also, in my opinion, not yet well studied. We see that this cycle has just changed thanks to AMD and ARM. That's why we own shares of ARM. I think that with the development of agent-based AI, AI agents will require significantly more processors per GPU than the data centers themselves.
That is, or for training models. So I think we're in for an exponential growth in demand for processors. So I would say that in the context of data centers, it is definitely worth paying attention to processors, optics, and semiconductor equipment. And in a broader sense, we are very inspired by space themes.
I think this will open up many more opportunities and new markets, whether it's defense, communications, or data centers.
Comments 0