원문 보기
Let's ask Jeremy Siegel. He is a professor of finance at the Wharton School of Business and chief economist at WisdomTree. Nice to see you again, as always. So, the employment report turned out to be better than expected. We are currently in a historically difficult month. How do you see the situation today?
Well, I think this is a very good and non-inflationary employment report, because more people actually entered the labor market than were employed . The unemployment rate, if not rounded, actually rose slightly, which means it was a response of supply, not excess demand. Wages are completely under control.
You know, I think this is a great report. Now I know that the probability of a raise has increased a little bit , and, you know, I'll tell you: It's going to be a really important meeting for Kevin in September. And I would say that if the midterm elections weren't on the horizon, the Fed would raise rates.
But, my God, take Trump's tweet today; He's not just saying "leave them unchanged," he's saying " if you don't lower them, I'm going to start raising tariffs on all of these countries." This is just madness. So let's hope that does n't happen, but don't think he won't be under political pressure at the next two Fed meetings.
Oh, I don't think there's any doubt about it, but if he had raised the stakes, as you say, there's a chance that, as I said earlier this afternoon, the roof would have been blown off 1600 Pennsylvania Avenue, and you know that's what it would have been . But you don't really think they're going to raise the stakes, do you ?
The data tells me there is room for some interest rate hikes. If he postpones them until after the elections, that is, until December, will it be a tragedy for the economy? No. Of course, if the situation escalates, if inflation continues to rise, and Iran is a " dark horse," they may decide to step up their missile attacks by that time, which will push up oil prices and, as a result, bond yields.
The bond market could force Powell to act by saying, " Listen, I'm being told I have to take action to get inflation under control." By the way, I think a decisive move by the Fed could actually lower long-term rates by saying, "Hey, the credibility of the Fed is at stake, and we're going to act." But listen, it will be ...
He can try to convince Trump of this . Although I don't think Trump would agree with that logic. Well, good luck with that, right? Well, let's say this, Professor, for the sake of conversation, that they will raise rates in September. How will the market react to this ? I think the market will react negatively at first, but then positively because they will be happy that the Fed is taking steps to get rates under control.
Again , let's hope that the roof on Pennsylvania Avenue holds up and we don't get a significant tariff increase, assuming that it stays under control, because ultimately, you know, there is trust in the central bank to fight inflation. Now again , don't forget about next Tuesday, we have the Producer Price Index (PPI) and that very important Consumer Price Index (CPI) coming out.
We know that Waller drew a clear line as to whether it would be 0.2 or 0.3 percent. So I mean, there's still some very important data that's got to, you know, come out. But if you look at, you know, the futures markets, the long-term bonds, the two-year, and by the way, as you know, I look at the money supply—something that Kevin Warsh actually mentioned in his speech at Jackson Hole as an important factor.
Since the start of the Iran war, the M2 money supply has been growing at 10% per year, which is excessive. And to me, that actually means he needs to start [ raising] those rates. Very, very low compared to what we see in other companies. So , in these companies you still have security. You also have the recognition factor of these companies, and you have the valuations that Jim points to, which are pretty attractive almost across the board .
So the concentration is happening right here and you look at others like CAT. CAT, you know, fluctuates back and forth , but it's still far from its highs. Deere, Innova, far from their peaks, they are coming back, but they are actually much more volatile than these names from the "Magnificent 7." Well, because these are the very companies that are at the center of the data center debate.
And you would expect the graphs to look exactly like, you know, you could plot a data center rollback versus AI power companies like Vernova, for example, because you would see one going like this and the other going like this . So this has largely become part of history too. C3.ai, which I own shares in, is the most complex example to look at.
They had great news, and stocks, you know, are swinging around here. Yesterday was a good day , but they dropped to 184, why? Precisely because of what you are talking about. The path of AI is entering another important phase, and this time the focus is shifting to one question. How much more computing power will the next generation of AI actually need?
The final argument is that businesses are abandoning simple experiments with AI in favor of large-scale implementation. This transition can be extremely important , as agent-based AI doesn't just answer questions and stop. It can perform multiple actions, interact with the software, and continue working on the task.
This means that the need for computing power could increase dramatically. One estimate being discussed in the market suggests that generative AI could require thousands of times more computation than simply answering a search query. If this happens on a large scale, the implications for the AI infrastructure ecosystem could be enormous.
And that's why investors are closely watching companies in the chip, memory, energy, and data center infrastructure industries. Nvidia remains at the center of the debate. But memory companies like Micron and SanDisk are also becoming increasingly important. The argument is simple. More AI agents means more workload.
More workload means more logical inferences, and more inferences mean more calculations. Nvidia CEO Jensen Huang also spoke out against extreme predictions of an AI-driven apocalypse. Huang argued that the idea that AI will destroy the world by 2030 is not realistic. He says that spreading fear around these scenarios is unnecessary and irresponsible.
His argument is also that Nvidia has a vested interest in ensuring that AI is implemented safely and productively. If the technology cannot safely create value, the entire AI ecosystem will ultimately suffer. But this debate has created another uncertainty for investors. This is regulation. Until recent warnings from some AI industry leaders, it was assumed that the US would likely maintain a relatively soft approach to regulation.
This assumption is now being questioned. Some AI companies are asking for more government intervention, while also proposing ways to self-regulate. The important question is whether this will lead to significant restrictions or simply create additional oversight over an industry that continues to grow rapidly.
So far, the infrastructure demand story remains strong , with one investment firm specializing in AI reporting a reduction in its stake in the sector at the start of the year. Then they increased it again , and infrastructure trade resumed. Their overall share of AI reportedly rose from around 13% to the mid-tens of percent in June, before returning to around 13% by mid-August.
The firm's argument is that its AI models were able to look beyond short-term headlines and focus on underlying demand. And this underlying demand remains significant. The expected shipments of Nvidia chips are another part of this story. Management is pointing to a potential doubling of chip shipments along with very strong expected revenue growth.
The overall picture is that customers continue to actively spend money on AI . And there may still be a significant amount of catching up work ahead in infrastructure.
Comments 0