저자: TomLee · 외국인
요약 보기
Mag 7의 설비투자 전환과 AI 인프라 확대로 자본집약도가 상승하고, 과거 광통신·무선 사이클처럼 주식은 사이클 중반에 시장수익률 수준으로 정체될 수 있음. 장기적으로는 타워 같은 수익 구조나 AI 시대의 '피자'(인력에 필요한 에너지·소비재) 등 승자를 찾아야 하며, 오프라인 경험의 가치가 높아질 것이라는 견해. 구체적 자산에 대한 1~3개월 방향 예측은 없음.
다듬은 글
I haven't I saw somebody make a comment about this and I thought it was interesting. Good one to talk to you about is we're now starting to see the Mag 7 actually doing capital investment, right, which they've never done. They just bought their own shares back. Is that a negative or positive in their performance?
Economic performance is probably good because capital investment was something we'd been missing for two decades. but no share buybacks or less share buybacks or does it still generate even more free cash flow that they end up just doing both? Yeah. Well, I think it's highlighting two trends. You know, one is um they didn't do R capex, but they did R&D, which until recently was a P&L expense, not a capitalized expense.
and and so they're supplanting those dollars for capital dollars that become balance sheet assets. In a world of labor shortage, that means you have less opex because that's wages, but you'll have more balance sheet intensity because now you have a computer that you're operating. So I think in a world of labor shortage, balance sheet intensity will go up.
So it makes sense that these companies are big spenders of capex now. And but that does create another wave of innovation because cooling and power supply and uninterrupt power and all these become strategic advantages which will create another wave of technology innovation. So I think that there's a lot of opportunity and waves of opportunity being created because of that change that you've just highlighted.
Yeah, I think you're right because it's going to be a game of efficiencies because the the cost of these models is going down fast. Um, and the compute is going down, but therefore they can increase output. So, it's just this tremendous race to get more efficient at every single component part of that flywheel from energy to chips to compute.
Yes. And we've even seen it in in the financial industry. You know, the the financial industry used to trade labor intensive, so had a lot of traders and people. And then they replaced all the traders with computers. But someone naively might have said, well then financial companies capex will go down because they already have computers trading for them.
No, they capex intensity and tech intensity of the banks has actually only gone up since then. So that arms race without even AI is consuming more dollars and that's of course I think going to be the case with with AI itself. Putting your stock picking hat on, how do you pick through what is going to win in the next few years because yeah sure we can you know sure Microsoft will do fine etc etc but but how do we make money out of this?
And I it's not very clear yet with AI outside of Nvidia and a few other stocks how to really make money out of this out of robotics. A lot of these kind of exponential age technologies pretty tricky because not much around. There's pretty slim pickings because there's been no IPO market for bloody forever.
I've got two simple analogies to share that I think we should be looking for in AI. So we just published a study looking at the wireless fiber boom and when the stocks peaked relative to capital spending and many people so let's say it was a 20-year capex cycle for fiber and wireless um the stocks relative to the broad market peaked eight years into the cycle.
So in other words like halfway into the cycle the market will suddenly say enough capital has flowed into the space that they're market performers. However in that post like now in retrospect the 25 year look back one industry in wireless pulled away from everything. Two actually two did it was the first one is the tower industry.
It has produced better returns than handset chips, carriers, spectrum owners, the towers and it's as you know it's a boring business that grows in a unit with demand. Um the second of course was Apple came along but in ' 07 so 13 years into the industry and after the industry's stocks all peaked Apple came and and essentially gobbled up the economics of the ecosystem.
So, I think that in AI it'll be it'll pay to look for the tower equivalent and the Apple that Apple the next iOS that'll come out on top of everything. But I think the second thing that's going to be important is that you you just described it. AI is infinite um workers with PhD qualifications today.
Today, don't forget in two years time, yeah, it'll be super intelligent, right? super intelligence. We It's something we've never had before. Correct. And yet these super intelligent entities have to eat pizza. And what I mean by that is like the best performing stock of the last 25 years is Domino's pizza because everybody has to eat.
So if you could find the pizza of the AI worker, so that's not the infrastructure, it's what you have to feed the worker. That's another money maker. And I don't know what that is going to be. So I've got two thoughts on that. Obviously, one is the energy flywheel, right? If you're feeding AI, it's going to be probably something to do with the energy mix and how that's delivered and what.
And I don't know again because the energy grid is going to get reconstructed and changed anyway. The other one that I've come up with is we are so digital now. We're going to become more digital. We don't our roles in whatever a workforce is or the economy is going to change but tomorrow will be more digital than today.
That is a fact and that is not going to stop for the rest of our lifetimes. So it's going to cause an equal and opposite reaction. Somewhere within this tangible experiences become premium experiences. And you know, if you look at the rise of LVMH, there's something in that story which is equivalent to your pizza, but the luxury pizza.
Somewhere within this is going to be an equal and opposite story that's going to be immensely valuable to us as humans. The thing that is the the most nonfgeable, you know, that's right. That's exactly right. Yeah. Yeah. That's a good point because you're right. We're we're going to be polluting and corroding so many human experiences, especially like with android robots and fake friendships and, you know, things that get really polluted.
Well, look, I look Tom, we do this now. We're so used to doing this on Zoom. Go back 10 years. You and I would have had to have met an event or whatever. And those experiences matter because they actually form true relationships and stuff. And you don't do it this way otherwise. Yeah. In fact, I don't know if you know this.
There's a really someone did a an experiment and you should just check this out. Whenever people shake hands, okay, and the research study this, you 100% of the time end up smelling your hands because you're you're trying to detect the pherommones of someone else, but you don't realize you're doing it.
And they said that it's h it 100. So, it's still instinctual that we need to meet in person. Yeah. Because pheromone exchange is what we do and we don't realize it cuz we see dogs do that and cats do that. We don't realize that we do, but we do. Tom, I know you need to get out of here, but listen, fabulous conversation.
I think there's a lot more we can talk about because we didn't really cover much of the future and where it's going, but I just think it's also very important to hear the past because people don't know why you make decisions that you make or why you have the understanding that you have until they understand where you came from.
So, thank you for sharing that as well and uh I look forward to getting together again at some point soon. Yeah, that'd be great. And probably in person next time, right? Absolutely, my friend. Okay, take care. Take care.
댓글 0