저자: TomLee · 외국인
요약 보기
AI 버블 우려에 대해 Nvidia 등 AI 기업 밸류에이션은 1998년 Cisco 대비 훨씬 낮아 버블과 거리가 멀다고 주장. 기업들의 AI 지출 성과가 나오고 있고 AI 지출은 가속 중. 과거 버블(바이오, 주택, 인터넷)과 달리 현재는 과열 행동이 없다고 봄. 금리 환경과 실적 성장(S&P 이익 약 15% 성장 전망)을 근거로 주식이 여전히 저평가라는 입장.
다듬은 글
All right, Tom. I thought a great place to start the conversation is everyone is worried about the AI bubble. We know that AI is real technology. We know that companies are using it, but people still think that there's too much spending going on. There's this big bubble that's going to pop. The world's going to end.
Stock market's going to crash. What do you think? Uh I mean, if if we're measuring bubbles based on like risk reward, we're pretty far from it because in 1998, September, which is similar to today cuz Fed was on pause and then they started cutting again September '98. Cisco, which was the biggest company at the time, was trading at 56 times forward PE.
18 months later, it got to 210 times. Now, Cisco made a commodity router that had many competitors. Nvidia today is the singular most scarce company in AI. Uh they make a chip that everyone who is building a premier model needs. It's trading at 27 times forward earnings. Costco trades at 50 times forward earnings and Walmart's 34 times.
So, Nvidia is still cheaper than a membership club company. So, I I'd say that's why we're pretty far from a bubble. And when you bring that up to people who are skeptical or critical of kind of the AI companies, what is their response to that? I think people then say, "Well, there's a generalized stocks are expensive problem." And uh cuz I was just in Latin America and that's what a lot of the pension clients said.
But to me, the problem people have is that stocks are trying to discount the future and AI is a not only a labor replacement, so we can calculate, you know, job replacement and that's a use case. But then there's this agentic and super intelligence potential, which creates productivity. So, how do we discount that in the stock prices today?
I I actually think stocks are still cheap and I know Jensen was in New York recently, met with some analysts and I think a lot of the takeaways people had is that in the AI world, companies, enterprises that are spending are starting to see payoff on the AI spend, which means spending is is picking up.
I actually think AI spending is actually accelerating. Well, it always is funny to me how people are yelling bubble, but the more people who yell bubble, it seems like the less likely we're in one. Yeah. Then, how do you have a bubble if every time you hear an AI CEO open their mouth, they're like, "We can't keep up with demand." Yeah.
And then on top of that, the earnings reports are coming out and there's companies that are trillion plus dollar market caps that are like, "We're growing faster than we thought we were going to grow. We're driving more profitability with less employees, which means that they're becoming more efficient, more productive, hence they should be more valuable, right?"
Yeah. Yeah, and you you make a key point cuz I've experienced three bubbles in my three decades doing research. The biotech bubble was the most recent. The one before that was the housing bubble. And the one previous that was internet bubble. And three of three times, not only were people denying the existence of the bubble because they could explain it away, but we saw behaviors that were indicative of a bubble.
In the biotech bubble, hedge funds, health care hedge funds were funding university projects so they could get in front of the molecule. I mean, if that that was like bubble behavior to me. Uh and as you know in housing, people were doing ludicrous and absurd things. Uh that's not happening in AI. In fact, everyone's so measured in AI today that it doesn't feel like a bubble.
There could be a bubble somewhere else in in the market, but it wouldn't be. So, you mentioned valuation multiples and one of the things that a lot of the older crowd who's been around for a while, their kind of brains break and smoke comes out of their ears when I say this, but I actually think maybe you can't use historical valuations the same way that you can today.
And my logic has always been the companies are more valuable, but also there's like a monetary premium in stocks now that maybe previously didn't exist. And it's hard to quantify, so I don't know, you know, what what are like comparisons, but the same way you have like adjust inflation adjusted prices, there almost feels like there needs to be like a monetary premium adjusted share, you know, multiple or something.
Uh yeah. And also people uh people are are willfully overlooking that valuations are actually they're they're dynamic. You know, from 1940 to 1967, the tenure went from 2% to 6. And everyone should say PE should have fallen in that 27-year period. The PE went from eight times to 27 times. So, in other words, as nominal rates rose, earnings accelerated and so the PE went up.
And that was in a time when 70% of the earnings generated were cyclical. Today, most of the earnings is recurring revenue in the S&P, which is higher multiple. And in fact, the 10-year bond is at 4% at 25 PE. I mean, the bond market's expensive. So, yeah, so I I think you're correct. I mean, plus in the last five years, we had six black swans.
Uh we had COVID. We had the supply chain bubble with the fact we had an inflation bubble, the Fed's fastest hikes in history. We had the tariff terrors and US bombing Iraq's nuclear facilities. Six black swan events and the market grew earnings during that time. I would argue PE should be higher because we just battle tested the stock market.
So. Well, if you go back to 2020, I mean, I remember living through that and once we were on the other side of the recovery, I sat down with a group of friends and I said, "Is there anything that we could do that would be more disastrous for the US economy than like lock everyone in their home, shut down their businesses, velocity of money goes to zero?" And we're back to all-time highs, which then begs the question, Peter Lynch is famous for the like if you spend 13 minutes on economics, you spend 10 minutes too long.
Um I don't know. In the '70s and '80s, like maybe actually the stock fundamentals were more important than macro. Today, it almost kind of feels the opposite. Like, yes, you still need to pay attention to the company fundamentals, but like what the Fed does or what is happening in terms of like the momentum and and you know, kind of the sentiment online feels like it dominates the company fundamentals.
Yeah, that's true, but you know, in the in that six-year period since COVID, earnings delivery has been amazing. So, the substance has been there. Yeah, so like even this quarter, uh 85% of companies are beating results right now and it's still going to result in basically 15% earnings growth. I mean, so that's not late cycle earnings growth.
Late cycle would be single digits. So, companies are still finding ways to grow margins and earnings and there's plenty of multiple expansion. I mean, financials, business results look more like tech companies now, but they still trade like financials. So, I think there's a lot of groups that can can re-rate.
Comments 0