저자: JimCramer · 외국인
요약 보기
Cramer는 하락장에서 AI 데이터센터 관련 주식 중 승자주를 저점에 분할 매수하는 전략을 설명한다. 마이크론은 PER 12배 수준으로 저렴해 매수 대상으로 제시하고, 나머지 SanDisk, Western Digital, Seagate, Vertiv 등은 밸류에이션이 높아 추가 하락 시 관망하겠다고 말한다. 청취자 질문에 AT&T는 Starlink·Amazon Leo 경쟁 우려로 비관적이며, Salesforce는 FCF 9배 수준의 성숙 기업으로 성장주로 보지 않는다고 평가한다. 마지막으로 AI 추론 수요가 학습을 넘어설 것이며, 엔비디아는 피지컬 AI로 확장한다는 논리를 펼친다.
다듬은 글
I recognize that it always feels compelling to bet on the underdog or might you might call them bottom feeders than to go for the otherwise winning stocks that just so happen to get hit very hard today. But I want to buy winners on a momentary slide, not losers that may just be resting before the next decline.
So I go down the list of stocks in the S&P 500 that are down the most and see if there's anything that kind of strikes my fancy. Incredibly today's many many of today's markets worst performers were indeed linked to the data center. KLA-Tencor, a material company used by chipmakers is down 5%. We own this one for the Chopper Sauce.
It's had a parabolic move even after this pullback. I think it's too expensive for the moment. Comfort Systems is a heating, ventilation, air conditioning company with plenty of data center exposure, but it's a $1,854 stock that's nearly doubled year-to-date. I'm not messing with that one. Western Digital's down nearly 5%.
It's a memory stock that trades at 46 times earnings. That seems too rich to me. The fairly similar Micron sells for less than 12 times earnings and it seems compelling down 6% today. I've been waiting for Micron to come down. This may be the opportunity. BUY BUY BUY! NEXT there's Corning, another Chopper Trust company with a parabo- another parabolic move.
It's got to come in more than just down 7% the beat down's not enough to entice me to buy or at least buy more of it. SanDisk, that stock's up more than 3,000% over the past 12 months. Even up here it only trades at just 20 times this year's numbers, but that may be too expensive versus Micron. Got to let it come down more than today's 5% decline.
Seagate again too pricey nearly 50 times next year's earnings estimates. These are commodity companies. Vertiv had an amazing quarter, the best backlog of any of the data center stocks, but at 52 times earnings I got to say no thanks. Bottom this with Geron, which got blasted by a poor result on a metastatic melanoma cancer trial.
I along with many others thought that this trial would succeed. You can't buy a drug stock that falls that much on the first day of the decline. Most analysts seem to have Keto in this drug to recommend the stock. It's damaged. There could be some more downgrades. So, what do you do? As someone who's always looking for bargains, I think Micron's the only possibility that hurt in that whole list.
It's the hardware stock that can most easily turn itself around next move higher. I buy some right here and then wait for another say 2 to 3% decline to buy more. That's how I get started. Bottom line, I just showed you how to use a rotation to buy something not aggressively, but gingerly on the way down as we wait for all the variables I talk about, especially the endless back and forth about the war and about this trade of her moves and nuclear materials to play out.
We got to find out whether President Trump's inability to bend Iran to his will is ever going to end. Caleb in Texas, Caleb. Hey Jim, how you doing? Hope you're having a terrific Tuesday. Uh well, it's kind of more of a Monday feel to it, frankly. But, what's happening with you? Yeah, it's all right.
I mean, look, I time passes so quickly. Go. This is what I was thinking of right here what I wanted to talk to you about and looking at my notes, this is why I said Tuesday. So, last Tuesday, the FCC approved a purchase for this telecom company for them buying about 50 MHz of EchoStar's nationwide spectrum for roughly $23 billion to expand their 5G network.
Now, the project could require sustained capital and operating spends adding pressure already to the high debt load if returns are slow. But, the satellite coverage meaningful meaningfully could reduce dead zones and strengthen customer loyalty in rural and enterprising segments. And that company I'm talking about is AT&T.
Do you feel like it's a good time don't want to be in AT&T and I'll tell you I also am very concerned about rural. I'm very worried about Starlink. I think Starlink is I I also think Leo from Amazon. Both of those are going to be considerable competitors to the rural part of a company like AT&T. But thank you so much.
I think you've got a good thesis, but I trumped it with mine. I'm sorry. All right, you have to use the rotation to your advantage if you're going to just sit there and look at it and you got to think about what to do. Buy things think and what could be replaced, at least the growth could be replaced by AI.
Um Ittai, your target for Salesforce I guess nine times cash flow free cash flow for its enterprise value. Pretty significant discount, right? We're talking about 11% free cash flow yield. That suggests not a lot of long-term growth. Is there anything that Salesforce can do in your mind to kind of reorient itself and and capture some of that future growth that you don't see it getting now?
So, first, is it why? Why only nine times? Because it's not anymore a growth compounder. It's a mature, very stable company that will continue to be a cash cow, generate cash flow, and will grow below 10% a year. That's why it deserves about nine to 10 times free cash flow. Um what can it do? It needs to continue and invest in AI and try to find ways.
I don't have the answer. I can I can only observe it. Try to find ways to turn AI into growth and that means expansion. What we've seen, by the way, across the board is that companies try to expand elsewhere. Take for example, ServiceNow, just as an as an example. They not only they're expanding into AI, they also expanded into cybersecurity and they complemented the AI solution with cybersecurity.
I'm not saying that Salesforce need now to go and on a shopping spree and buy companies, but you have to grow into new areas in order not to stand still and not to turn into just a mature company, mature software company, rather than a growth compounder. Seems like Salesforce and ServiceNow are the opposite ends of the spectrum in this sort of SAS couples SAS apocalypse trade.
Hey guys, welcome back to my channel. I hope you all are doing well today. So without wasting time, let's start our video. And today next growth wave could be bigger than AI training. For the past few years, investors have focused on one thing, that is AI training. Building larger models, training larger foundational models, buying more GPUs, constructing bigger clusters.
That narrative created one of the greatest technology boom in the history. But something important is happening beneath the surface. The industry is shifting from the training to inferences, and inferences will ultimately become a much larger market. Every time an AI agent perform a task that require inferences.
Every time a user ask a question, every time a software write a code, every time an AI assistant schedule a meeting, inferences happen. And unlike training, inferences never stop. Training is a one-time event, and inferences is recurring. That is why the rapid adoption of AI AI is so important. Companies are beginning to deploy AI systems across their organization.
Developers are building AI-powered products. Consumers are interacting with AI tools daily. The result is a massive increase in compute demand. Not because new models are being trained, because existing models are being used. This didn't used to matter because for years a skeptic questioned whether AI infrastructure spending would ever generate a return.
Today we are beginning to see the answer. The leading AI companies are generating revenue at unprecedented rates. Enterprises adoption continue accelerating. AI coding tools are exploding in popularity, and AI assistants are becoming mainstream. And every one of those applications consume compute. That's why hyperscalers continue investing aggressively.
Despite concern about the values, and despite concern about the interest rate, and despite concern about the geopolitical tension, the largest technology companies in the world continue telling the investors the same thing. Demand exceeds the supply. The world remain compute constant. That phrase continue showing up across the earning calls because it's accurately describe the reality.
If there were excess capacity, then pricing would collapse. Instead, GPUs rental prices remain elevated. That's data center utilization remain high and new capacity is absorbed almost immediately. This is not the behavior of market facing weak demand, but the behavior of marketing market facing the scarcity.
This is where Nvidia's strategy become especially important. Nvidia isn't simply selling the chips. The company is building an entire ecosystem. CUDA networking software enterprises AI platform, digital twins, robotics, agent framework, physical AI. Every layer reinforces every other layers. The deeper AI penetrate the society, the stronger that ecosystem became.
And Nvidia is positioning itself for what's coming after today AI boom. That is physical AI, humanoid robots, automation systems, smart factories, intelligent machines. Jensen Huang has repeatedly advised that physical AI could become a multi-trillion dollar opportunity. If digital intelligence transform software, physical intelligence transform the real world.
Factories become automations and warehouses transformation and transportation manufacturing. Every one of those systems require compute. Every one of those systems require inferences and every one of those systems require AI infrastructure. Infrastructure. This is why Nvidia road map stretch year into the future.
Blackwall, Rubin, Rubin ultra fineman. Each generation is designed to support increase the complex AI workloads.
Comments 0