저자: DanIves/다수발언자 · 외국인
요약 보기
메타 AI·클라우드 확장 과제와 4% 하락 부담 논의. 자본지출(120~130억 달러) 대비 성과 증명 압박이 커짐. Matrix 측은 마이크론·메모리주는 신규 매수 자제, 빅테크(MSFT·AMZN·META·AAPL)는 장기 관점 매수 기회로 봄. 애플은 메모리 가격 상승 우려로 하락했으나 가격 매력 부각. AI 추론 수요 확대로 메모리·인프라 수요는 장기 지속 전망이나 사이클 논쟁은 남음. 엔비디아 목표 1000달러 언급.
다듬은 글
Add a business and you're just throwing a bunch of everything at the wall to see what sticks . After all, Zuckerberg probably also said this at the shareholder meeting, or at least opened the door for a cloud business. Is this something unrealistic or really serious? Okay, I think cloud technology is maybe a bit of an ambitious project, but it's acceptable to me .
How do you gradually move away from addiction to advertising? Advertising will, of course, remain their main source of income. But if you look at what they are doing, it will be a continuation of diversification. And I consider this a positive, as their user base is second to none . Mm. So as they monetize it, part of our optimism is about what I'm talking about: $900 and ultimately, it could be higher .
This is a key year for Zuckerberg and Meta to prove their mettle. This is a step in the right direction after a quarter when everything was focused on capital expenditures , which felt more like an excuse for not doing homework. Well, let me ask you about this, because while you were talking, I was looking at the stocks.
So, they have fallen by almost 4% since the beginning of the year. How much pressure do you think they feel within the company? And what pressure does Zuckerberg himself feel to release something that will hit the market and make investors come back, because apart from Microsoft, which also fell in a year, everything looks pretty good with the others.
Besides these two. The pressure is definitely increasing, because when you spend 120-130 billion on capital investments , now you need to show results, and I think the latest quarterly reporting was a wake-up call. Because when you compare it to Amazon and others, for Meta it's a kind of investor penalty area.
You have to start demonstrating it. Katz, President, Chief Investment Officer, Managing Partner of Matrix Asset Advisors. Good morning to you. I find it very interesting in your notes how you talk about the rotation, that the sell-off of the " magnificent seven" was excessive. But another group, like Micron and memory companies, may have been very vulnerable as they rose in price.
So it's a bit like a "ferris wheel." I mean, how should people act: should they buy tech stocks or not? It's confusing, is n't it? This is quite confusing. We puzzle over what is happening in the market every day, and not only in the technology sector, but also in other industries. When semiconductors go up, money is pulled out of healthcare and consumer staples .
Therefore, some stocks rise by seven, while others fall by five. We believe that the key to investor success from this point on is a long-term perspective: don't get hung up on daily trading, it's better to evaluate the market now and think about where it will be in a year, and prepare for it. You mentioned that we prefer hyperscalers, mega-cap stocks.
These actions made headlines every day . A year ago, everyone loved them. Now they are about a third cheaper. Many of them are sold at price/ earnings ratios of 18 to 22, which is not usually the case or hasn't been the case for a very long time. We think the long-term prospects for Microsoft, Amazon, Meta, and Apple are very good, and you can buy them at great prices right now.
So pay less attention to daily fluctuations: know that in the short term they may fall, but in the long term, this is a great opportunity to buy great businesses at attractive prices. What do you think about yesterday, when Apple posted its worst performance in a year, or something like that? Given the increase in MacBook and iPad prices, as well as speculation of possible further growth, Apple shares have indeed fallen sharply.
Do you see this as a buying opportunity? We think so. Although it is an impulse-driven market , it is like it is on steroids. When people want to get into stocks , they push them up, like Micron did yesterday, and when they get out—there is no bottom . We believe that in the long term, Apple will eventually become a strong player in artificial intelligence.
There will be a move towards edge computing, which means having to buy more phones to use AI more effectively, and we believe that Apple, whose stock has been performing well , just gave away 3 months of profits in one day. We would use this as an opportunity. Yesterday's news about price increases is nothing new and should not come as a surprise.
We know that their prices are increasing due to the cost of DRAM chips, and they are just partially passing that on to customers. So, we like Apple. A week ago, at prices of 310 or 320, we were less enthusiastic about it. At today's price, we believe this is a very good time to open a position. Yes, and today there was also news about Apple.
Uh, you know, I'll just add that, David, so people know. Apple may now be considering working with Chinese memory manufacturers to help with production. So, in a week, the shares fell by 7.5%, and this month, as of June 26, they fell by almost 12% . So, you see this as an opportunity. When you talked about Micron's takeoff, you said this group is more vulnerable.
You don't invest in such companies, even though they've rolled back a bit ? We own a number of semiconductor companies and have made excellent profits from them. We are very happy with this. We would n't invest new money in them. Our expectations for this group going forward are much more cautious than they were before, simply because they have achieved profitability in a year, two or five years in less than three to six months.
Therefore, we do not believe that now is the best time to invest new money in them . If you own a lot of these stocks, we would advise you to gradually reduce your holdings on days when the market is rising wildly. Again, you won't know that stocks will start falling six months before... The next phase of AI.
Although Nvidia may consider its AI infrastructure to be still in its infancy. Over the past few years, investors have watched AI evolve from an existing technology into one of the largest capital investments in modern history. But an important question remains: how long can this last? According to many semiconductor executives , Wall Street analysts, and industry experts— much longer than most investors currently expect.
AI goes beyond training models. The first wave of AI spending focused on training large language models. Companies competed to build larger clusters, more GPUs, larger data centers, and faster networks. Now the industry is entering the next phase: inference. Every AI response, every chatbot conversation , every programming assistant, every AI search query, and every enterprise AI application requires inference.
Unlike learning, inference happens every day, millions, eventually billions of times. This creates a constant demand for AI infrastructure. Every request to AI requires memory. Most people think of GPUs, but AI systems cannot function without memory. Each token generated by an AI model depends on the rapid movement of huge amounts of data.
As models become smarter, contextual windows become larger, reasoning becomes deeper, and multimodal capabilities become broader. Memory requirements increase along with computing power. This is in AI servers. Supply is still years behind demand. One reason why memory companies continue to report extraordinary results is simple.
Supply cannot grow fast enough. Expanding semiconductor production capacity is not the same as opening another warehouse. Companies must purchase land, build clean rooms, install advanced equipment, optimize product yield, qualify the product with customers, scale up production, and this entire process often takes several years.
Meanwhile, the demand for AI continues to grow. The industry is investing for the next decade. Today, capital expenditures are not planned for the next quarter. They are designed for the next decade . Billions of dollars are being poured into modern semiconductor factories, packaging facilities, energy infrastructure, networks, cooling systems, data centers, and power grids.
These investments only make sense if customers expect demand for AI to remain strong for many years to come. And so far, that's exactly what management teams continue to report . Nvidia remains the platform leader. Competition continues to grow. Specialized accelerators are being improved. Large cloud providers are developing their own chips.
But Nvidia still retains several important advantages. CUDA software, AI network computing systems, continuous software optimization, a developer ecosystem, and integrated hardware and software solutions. These benefits go far beyond simply creating faster chips. They create an entire computing platform.
AI spending is becoming global. The race in AI is no longer limited to Silicon Valley. Countries around the world are now striving to have regional AI infrastructure. Governments want in-house computing power, universities want supercomputers, healthcare systems want AI research clusters, manufacturers want AI factories, financial institutions want AI- powered risk analysis .
Demand is becoming increasingly global. This significantly expands the total available market. Why are there still discussions about cyclicality? Despite strong fundamentals, many investors remain cautious. Why? Because the history of semiconductors teaches such caution. Memory has historically gone through turbulent cycles.
Demand for chips regularly soared and then fell. Investors naturally ask whether AI will repeat the same scenario. Industry leaders are increasingly arguing that this cycle is fundamentally different, as AI infrastructure has become critical, not optional. Will history repeat itself remains one of the biggest debates in the market.
Comments 0