저자: SanjayMehrotra/JimCramer진행 · 외국인
요약 보기
마이크론 CEO는 역대급 분기 실적(매출 1,330억달러, 전년비 250%↑, 4분기 매출총이익률 87%, 다음 분기 86.25% 전망)을 강조하고, 직원 보너스 비용이 마진에 반영되지만 경쟁력 유지를 위한 것이라 설명. 고객 장기공급계약이 2031년까지 연장되며 수요 가시성이 높다고 주장. 후반부는 엔비디아 밸류에이션 매력(PER 14~15배 수준 언급), GPU 내용연수 연장(구형 A100도 활용 가능), 수요가 공급을 크게 상회하는 AI 인프라 상황과 반도체 업종 순환매를 논함.
다듬은 글
You guys are probably wondering why stocks are falling. Maybe this is a wrong impression . Data center revenue has increased 11-fold, but that's not the main story. We'll find out what went wrong last night, because I think calling it a mistake is a big mistake. We are joined by Micron Chairman and CEO Sanjay Mehrotra in an exclusive interview with CNBC.
First of all, I want to congratulate you, Sanjay, and thank you for coming to the show. Thanks Jim, it's great to be on your show again. Oh, thank you. So, what I want to find out is: the gross profit here is simply incredible. Anyone who has followed your company for many years knows this—Andy Grove, perhaps the greatest semiconductor executive of all time, would say that's impossible.
Your numbers are impossible. But Wall Street seems to have remained unsatisfied. The more I dig into it, the more I understand: since 1994, the shortage only occurred when too many companies started producing chips. It wasn't a shortage. It was your decision to pay employees more money. This is, in fact, a share of the profit.
Can you explain to people the difference between the so- called drop in profits due to the decision to pay people a little more and the influx of new chips, which usually destroys gross profit? Jim, we delivered an exceptional quarter and an exceptional year with revenue of $133 billion , up more than 250% year-over-year, with a gross margin of 87% in the fourth quarter.
And we gave a forecast of 86.25% gross margin, about the same for our first quarter, the current quarter. So, the Micron team really delivered an exceptional quarter. And of course, you know, along with our record quarter, we rewarded our team with record bonuses. This impacted operating expenses in our fiscal fourth quarter .
As the fourth quarter production employee bonuses are rolled into the first quarter cost, this impacts gross margin, but we are forecasting a very good margin of 86.25%. I am very pleased and proud to have been able to reward our employees in line with our record results. It is important for us to be competitive .
It is important to attract, retain and motivate our talent to maintain our current momentum. Jim, memory is a key factor in the age of AI or superintelligence. This memory is what makes intelligence possible. This improves intelligence, and Micron has a comprehensive portfolio of technologies, products, manufacturing capabilities, and we continue to capitalize on rapidly growing opportunities.
So, Sanjay, I visited you on August 20th in Boise, you had 26 strategic customer deals. Since then you have been 16 . You just said 26. Where did you get the 10 new ones from on August 20? Jim, not only do we have more customers looking to enter into long-term supply agreements with us, but customers with whom we already have strategic arrangements are also increasing their orders.
In fact, we cannot fully meet the demand of our customers. Customers are extending these agreements even now until 2031. We have implemented both extensions and new agreements with a horizon until 2031 . This gives us tremendous visibility, certainty in demand, and it helps us plan investments in production capacity with confidence .
And, of course, this provides a guarantee of supply for our customers, allowing them to grow their businesses with the same confidence. So this just reflects the importance of memory in the age of artificial intelligence. Memory is critical for scaling AI to create value for end users and, of course, deliver powerful experiences, as memory enhances the value of computing platforms.
Professor at Tulane University and author of a book about Elon Musk, among many others. Walter, it's nice to see you today. Glad to be back, Meg. You are an incredible researcher of history. You probably observe this by looking back at what happened in the past. Absolutely, and we've never had a technology that seemed so dangerous, perhaps since the atomic bomb .
But every time these things come up , new groups are created, including industry associations, as we saw earlier this week when Trump brought them together. But ultimately, we also develop regulatory regimes for them. I have a series coming out soon called " Builders"—it's a podcast series. Every one of these people is there, and everyone thinks it could be dangerous.
They all talk about P (doom), meaning the probability that the technology will lead to the complete end of the world. They say maybe it's 10-15%. I asked, then why are you doing these things? Why don't you get together and regulate this? And, of course, competition. The competition between them makes this impossible.
Competition with China makes this impossible. I said this as a joke yesterday, but wait, I want to be the first to destroy the planet. Yes, that's right. I mean, what is your goal here? One of the things that was interesting. I look at Mustafa Suleiman, whom I am interviewing for " Creators". He asks the basic questions: what is the purpose of creating this?
Well, the obvious answer is to help humanity. If you're doing something that seems like it could harm humanity, it's probably worth saying: let's figure out how to regulate it. Yes, I didn't say about the planet. I think the planet will be fine. With the planet? Yeah, it's like a neutron bomb, right?
Exactly. Well, that's right . She only killed people while leaving buildings. Left everything else intact. You talked to all these people at "The Creators." I think you've talked to Sam Altman, Elon Musk, Demis Hassabis, there are a lot of names out there. Yes. Obviously. And Dario and Daniela together on my day.
And things they really agree on. Each of them said that there was a non-zero probability that this would lead to the destruction of humanity or part of it. They believe that this is a very real chance. I mean, there are two things to worry about. The first is that it can really free something up. I think "Hugging Bear" is for me.
I was at this show right when "Hugging Bear" happened. So I'm like, "Wow, that's Hugging Face, Hugging Face." This was our first " Hugging Bear". The thing is, I was an AI optimist. So I think that became a problem. This could lead to the fact that as we try to create these autonomous agents, they become more and more autonomous, doing things without our control, and then they will get out of our control.
The big difference in history between the personal computer revolution that started in the '70s and this one is that with Steve Jobs, Doug Engelbart, Alan Kay and others, the personal computer was supposed to connect you more closely to your computer. This should create agents who are disconnected and do things without you.
This is actually not a very good idea. What do you think? I was going to ask what you think their motives are. I argued that these people have been relatively consistent since I've known them on the issue of security. But there are other people, Jensen was sitting here earlier this week, who said that these security concerns are crazy.
Which is actually, you know, an engineering problem. Mark Zuckerberg will tell you that liability laws are already in place. So, you know, you don't have to worry about these things. Yeah, I mean, what we saw at the White House summit was a group, where David Sachs is the leader, who are very influential and say, " Look, let's not worry too much about this.
Let's not lose to China. We have the FTC, the Department of Justice , liability laws that can regulate this." But, to answer your question, yes, for 15 years. I mean, Elon Musk got into this game. Sam Altman entered this game. Dario entered this game because they feared that uncontrolled general artificial intelligence could be very dangerous.
So, this is not some kind of regulatory takeover attempt . This is a real concern. Let's not reject it. If you look at Nvidia today and compare it to some of the biggest names in tech, the valuation gets pretty interesting. Nvidia trades at around 14-15% annual earnings, while Apple is closer to 25.
Now add in bout 17% year-over-year. And the problem is not the lack of demand, but . So you , deeply rooted in the AI ecosystem, and it's still trading at valuations that some investors consider surprisingly low. For some analysts, this makes Nvidia a much more obvious option than slower-growing tech companies .
But there's another part of Nvidia's story that doesn't get enough attention. How long does AI infrastructure actually remain useful ? For some time, investors have been very quickly. A new architecture appears, companies buy it, and the previous generation depreciates. And within a few years, old equipment supposedly becomes worthless.
But this assumption may be too simple. One cloud infrastructure operator recently provided an important example with the Nvidia A100 GPU. The was contracted for additional workloads every AI workload requires the latest GPUs. The latest equipment can older accelerators can still be useful for physical AI batch computing, , the old equipment doesn't necessarily disappear.
It can simply switch to a different type of load. This means that the useful life of AI infrastructure may be much longer than some investors expected. And if a GPU purchased a few years ago continues to generate revenue for a few let's take a look at the deployment of Nvidia's latest infrastructure.
The first Blackwell-based initial deployments include an NVL72 configuration. But getting a GPU is only part of the process. To operate an AI cluster, GPUs, racks, networks, data center capacity, power, and software are required. Everything has to work as a single unit. You can't deliver 95% of the infrastructure and consider it ready.
The entire system must be fully functional. And this iy turn individual pieces of equipment into useful computing power for AI companies and enterprises. Now demand significantly exceeds supply . Some providers say that virtually every GPU . That's how unusual this AI cycle has become. Of course, the market reacts to shortages over time.
More data centers are being built , new power supply capacities are being connected . More GPUs are being produced , more cloud resources are being deployed. Eventually, supply and demand will converge . The question is not whether this imbalance will last forever. This is not true. The real question is how long demand can outpace supply.
And long-term forecasts suggest that this imbalance could persist for years. Meanwhile, capital flows within the semiconductor industry. AMD, Micron, SK Hynix, Marvell. Several stocks of semiconductor companies related to AI have started to show growth together. And this is important because market leadership is constantly changing. One group becomes overcrowded, the pace slows, and capital moves elsewhere .
Comments 0