저자: EdYardeni · 외국인
요약 보기
AI 투자 과잉 우려 속 토큰 가격 하락과 경쟁 심화로 승자·패자가 갈릴 것이며 창조적 파괴가 가속된다는 주장. 케빈 워시 연준이 인플레이션 2% 달성을 위해 금리 인상 가능성이 있다고 봄(미 국채 10년물 약 4.48%). 금은 2000달러 돌파 후 5500달러까지 올랐다가 현재 4000달러대, 연말 목표 5000달러로 하향, 장기적으로 10000달러 보유 입장 유지. 주식시장 장기 상승(연말 10000 전망)과 금 리밸런싱 수요 언급.
다듬은 글
Let's talk to Ed Yardini, president of Yardini Research , about the show. Ed, first a comment. Do you think the CEO of Palantis said what people say in closed rooms but haven't had the courage to say out loud about tokens , or maybe this is an extreme view, tokens don't create any value because this huge capital investment in AI continues to, you know, continue.
Uh, let's look at what's happening with software names and even these hyperscalers from big tech companies. Yes, I think it's clear that there 's a lot of confusion about how AI will impact many businesses. There is a lot of confusion about whether all that money spent will achieve a certain rate of return.
Uh, I-I think this is going to be an ongoing problem, and, uh, some people are going to find reasons to be outraged by what's going on. But I think that markets will eventually solve these, uh , problems. We are already seeing token prices fall, and this is partly due to the competition between open AI and anthropic.
Part of this is because companies have some pretty interesting conversations about how much it all costs them. The fact that agent AI just runs on its own and tells itself what to do, that it just keeps these models running 24/7, just caches, you know, like a cash register constantly ringing up, charging for the work of these models.
And I think that's the problem that companies are facing right now. Even Microsoft, uh, has abandoned Anthropic and seems to be thinking about using a Chinese model, which is much cheaper. M, have you heard of this company? I looked, I looked, I managed to see how half of a company called Edged ET Ced was founded by two young guys in the US, and they claim that their new AI chip is well -funded, by the way.
They just announced that they've received huge funding from some of the biggest names, and their new AI chip is about 10 times more efficient. Uh, and oh, that's true. You know, it's creative destruction at speed. This is something extraordinary. Creative destruction is what happens in the capitalist system, especially in the technology sector.
The technology sector is literally devouring its offspring. As soon as they release one version of new software or hardware, they are already working on creating another that will make it obsolete. We see this with Nvidia. Nvidia will sell you a very expensive chip that requires a lot of power, and they are already working on a solution to the problem of these things requiring too much power.
IBM just released a chip that, um, looks very interesting in terms of being much more powerful, um, and yet, um, requires less power. You know, we dealt with someone from Black Rock yesterday and asked him the same thing . Of course, there's all this new hardware, meaning, well, you know, new chips, which, and it's incredible, isn't it?
Nvidia spends $20 billion a year on research and development, and you have these new young, young, well, I think 24-year-olds who at least claim, with a lot of funding, that they, they launched this chip that's 10 times faster than, you know, regular GPUs. But, but anyway, leaving aside the hardware, there's certainly a hardware component, but there's also, well, I mean, just the way AI is used, right?
Routing... meaning not every query you type into LLM needs to be sent to an AI cloud data center , like a hyperscale data center. So maybe the rooting also starts, and it's done on your computer, no need to send it to the cloud in some data center. Are these risks, Ed, and do you think they will be talked about more and more because, ultimately, this is a technology that is evolving as we speak.
Yes. I don't think I would call it a risk. I would describe it as the nature of the beast. Artificial intelligence is a potentially terrifying program that can do a lot of good, but also a lot of bad. And I think we're just seeing that the speed of technological innovation, you know, can just be mind-boggling.
And, uh, these hyperscalers who spend all this money on modern chips might regret it in a year or two , because if they had just waited six to twelve months, they could have gotten a better chip that's actually cheaper and uses less power. So, uh , there will be a lot of winners and losers. It's just impossible to say for sure.
It's hard to gauge who's who, and, uh, the rules of the game are constantly changing. But listen, I use AI all the time, uh, throughout the day as a research assistant, and I find it very useful, very useful. It's kind of like a Google search for speed. Uh, but it's very useful. Right. Uh, you know, we're talking about AI, but in the near future, we're also listening to what Ken Wash said last night.
He said that inflation might be a little too high. Uh, you know, after most people perceived his first statement as somewhat belligerent, there was another, uh, you know, comment that inflation above 2% might be unacceptable. Looking at the US bond yield, it is now closer to 4.48. This time, there was little change in bond yields following these comments.
What do you think about them? Is this, uh, just a comment, or does it hint at some action in the near future? Yeah, this is all very interesting because Kevin Worth said he wanted to, uh, reduce the forecast. He doesn't want the markets to be polluted by what the Fed says. He wants the Fed to learn from the markets, and in fact he's like, uh, saying he's going to monitor the bond market.
Uh, it seems like the " bond controllers " are controlling how Vors conducts monetary policy, and I think you're absolutely right. He made it clear both at this press conference and today in Central Portugal that really the main objective of the Fed, his and the rest of the FOMC, is to achieve price stability, which means, and he defined very clearly, getting inflation down to 2%.
Uh, so despite all his talk about not making predictions about the future, every time someone asks him a question, he says, "Well, we'll send a task force to look into it ." But the reality is that they have already provided many predictions for the future. The FOMC statement was actually very belligerent, and the war was also very belligerent.
So either inflation will just miraculously drop to 2%. Er, or the Fed would actually have to raise interest rates for that to happen. Who would have thought, right? Just a few months ago, we thought Kevin Walsh would be extremely questionable when he took office. He will come out and lower rates under pressure from, well, you know, the President of the United States.
Well, that's not happening. Now we are talking about a potential rate hike by the US Federal Reserve. But he also talked about, um, you know, Ed, the potentially profound impact of artificial intelligence on inflation, because he said it has implications for monetary policy, and we're already seeing how it's affecting the price of memory chips, and the fact that Apple has had to go further and raise prices quite dramatically.
On the one hand, we have commodity deflation, oil has fallen, but we have inflation caused by artificial intelligence. So do you want to comment on that , and quickly, if time runs out, do you hold the $10,000 an ounce gold mark? Yes. Good questions. Listen, I think that, um, obviously, um, Kevin Warsh's Fed surprised people with how quickly it became "hawkish." And I think that's because the reality is that the economy is doing pretty well.
The unemployment rate is still around 4.3%. Um, it's just inflation, so the Fed should be ashamed, and the worst thing is that it's ashamed that the Fed hasn't reduced inflation to 2% for over five years. About the gold, I was partly right. Um, when gold, uh, went above $2,000 an ounce a couple of years ago, I said, “You know what?
"I don't know gold very well, but I do know that if central banks are going to buy gold, uh, because they're worried that their reserves will be frozen by the United States in the event of a geopolitical crisis, then that should be an optimistic signal for gold." And that took us up to 5,500. Uh, and now we're down to 4,000 again.
4,000 is a very important level. I think we'll find support at that level, and, uh, yeah, I'm still going to hold on to 10,000, uh, for the rest of the decade. And I think a lot of that will be because the stock market is going to go up, and I predict that by the end of the decade the stock market will hit 10,000, and I think there will be a lot of rebalancing.
You know, when people make a lot of money in the stock market, they get a little nervous and they want to diversify their investments, and they invest in bonds, and they invest in gold, and I think gold, um, will still work. I've lowered my year-end target from 5,500 to 5,000. And I admit that, um, you know, gold has been a pretty sloppy trade since the war.
You would think that I thought everyone thought that war would cause gold prices to rise even more. But instead, some central banks, ahem, like Turkeykey, saw that their currency was under a lot of pressure and sold gold to intervene in their currency. So it's a complex market. Thank you , uh, and a great, uh, conversation covered so many topics about, uh, artificial intelligence, you know, commodities, uh, you know, what the Fed is probably going to do.
Thank you very much for joining. Thank you. To get more news and updates, you just need to follow CNBC TV18 on all our digital platforms.
댓글 0