원문 보기
The first message: we have created a supergod, and venture capitalists are like lemmings. They run around and say, "I need to be on the right side of the supergod." And then they switch to the other two: “This is a real weapon. Let's shut down the world, and why don't you regulate everything?" They just go here and there.
We had this game. The problem is the scale of the trillion-dollar market cap, and this game has a ripple effect that affects everyone in the industry. That's why, by the way, I think Alex Karp deserves a medal. This is an incredible person who had the courage to come here. Everyone watched this clip.
All these S&P 500 CEOs were calling me after that clip and saying, "Help me interpret this." And I'll tell you, he was on the right side of history, because he shows you that, wait a second. It's like you need a much more predictable way of doing business. Because if you, you know, are exposed to the volatility of private company funding cycles or want private companies to meet valuations up to trillions of dollars, you're taking on a level of risk that you probably didn't know you had.
That's a fair criticism, I think. Hmm, we need to pause for a second. One last question related to Alex Karp's topic. In fact, he says that for Palanteer you need some kind of intermediate layer. I believe so too. Hmm, partly for your protection. I mean, he was arguing with LLM in a way, after all. The thing is, LLM does a very good job of a very superficial form of privacy called zero- data-keeping (ZDR).
That's the term they use. And they say you need to turn this on and everything will be great and everything will be fine. But if you look at the fine print, let's say Joe enters the secret formula for Kentucky Fried Chicken into the model and says, "Help me optimize it ." This part is called edr. But if you look at the little thing, for example, if he just clicks the " Like" button, is there any guarantee that this information is not somehow stored, that he liked such a thing?
And the honest answer is technically no, because we don't know how to do it. Another interesting thing I would like to point out to you is that when Alex came out, I think he said something extremely legitimate. I think this applies to companies. This also applies to countries. So at some point we should talk about this.
And Anthropic never once said, you know what, he's completely wrong . Here are the exact technical details, so there's nothing here. And I thought to myself, why don't you be completely specific right after this clip started going viral, right? And I think it's funny that we're here 10 days later , and not a word has been said except for Alex's version, and the reason is because we all know that's the true version.
It actually came out with a pretty strong report, but there seem to be some questions about the outlook right now. So this is the focus of morning trading. Alex Kofioffe joins us for a more in-depth look . Good morning, Alex. I mean, you and I analyzed these numbers today. At first glance, they look pretty strong.
We saw an initial spike before actually going negative. There seems to be some concern about this slowdown in orders. Just tell us what exactly they said here , because I heard the CEO essentially say that it's smart to start the year, uh, on a more conservative note. Yes, Cisco had a strong quarter, Seme .
Revenue and profit results beat expectations significantly, but investors seem focused on perhaps more modest AI prospects and some margin pressure, uh, in early trading, as you mentioned, the stock was down almost 7 %. For the first four quarters, Cisco earned $122 per share, and they did it on revenue of $17.25 billion .
Both of these figures exceeded consensus forecasts. Revenue from products, networks, security and collaboration exceeded expectations, but product orders also increased 35% year-over-year . The company also reported four billion orders for hyperscale AI-powered devices this quarter and said demand remains strong across the business.
Looking ahead, some of Cisco's fiscal 2027 forecasts exceeded expectations. They expected revenue of $72.2 billion to $73.4 billion and adjusted earnings per share of between $505 and $511 per share. Their first quarter forecast or current quarter forecast also exceeded expectations . However, some analysts have pointed to lower expected gross product margins as a key issue as AI-related hardware impacts profitability.
While the company continues to benefit from demand for AI-powered networking technologies , investors were expecting perhaps a slightly larger contribution from AI and appear to be questioning how sustainable the current growth acceleration will be . Uh, despite that, several firms, including Wells Fargo, Key Bank and Morgan Stanley, raised their price targets after the report, citing confidence in Cisco's networking cycle, strong demand trends and continued AI-related opportunities through fiscal 2027.
However, this is not enough, Cisco shares fell by 7%. OK. Yes, the bar is high when it comes to technology, but overall, I mean everything is holding up. As I mentioned, the CEO says they are starting the year cautiously. So let's take a look. I really appreciate the details, Alex, and this reaction.
Let's now exchange names with Joe Tay, portfolio manager of the Rational Equity Armor Fund. Good morning, Joe. Uh, how would you approach this, for example? Yes, good morning, Sam. Thank you for inviting me. I think I would generally agree with Alex. I think that, uh, I personally think that this looks like a very strong quarter, very strong results for the company.
And for me, just looking at the AI trade overall, this quarter confirms, uh, the direction of the trade. Uh-huh , strong growth. There is a demand for network technologies. Um, I think there's been a history of healthy skepticism in the field of artificial intelligence, looking at potential liabilities and predictions for the future, and most people have ignored a lot of that.
Cisco says we have many such RPOS. We are going to make a lot of money. I think most people are just skeptical about it, and, uh, showing it now, with this little digression. Also, the margin of the AI boom is creating a new memory supercycle. The biggest mistake that innovators can make now is to look at Micron as if it's the same memory company it was five or ten years ago.
Business is still cyclical, but the demand driving this angle is completely different. AI is taking memory from a background component to one of the most important concepts in the entire computing system, and Migrant is investing billions to make sure it is positioned for it, changing the bet to $10 billion .
Magnan is building a big research operation out of the guys with a planned investment of about $10 billion. The goal is ambitious. Migrant wants to bring together its own researchers with universities, startups, customers and technology partners to develop the next generation of memory technology. The focus is not just on producing more chips.
It's about solving the problem created by , involving new materials and technologies for beginners, memory architectures, and ways to integrate memory with the computer. In other words, Migrant is not just preparing for the next product cycle. He is preparing for the next era of computers. For a decade, investors have typically thought of computing in terms of processors, with CPUs processing instructions and GPUs handling an increasing amount of workload.
Memory was in the background, and artificial intelligence changed this relationship. Modern artificial intelligence models generate vast amounts of information, and this information must be moved through the system, stored, retrieved, and processed. As the model gets larger and impacts the workload, the system needs more bandwidth, power, speed, and efficiency.
This is why high-speed memory has become such an important part of the AI infrastructure. The GPU can perform calculations, but without enough memory, the processing power cannot work efficiently, and this creates a bottleneck . The memory partition becomes larger. Here's the important part. AI is not only increasing the number of GPUs in data centers, but also increasing the amount of memory required for those GPUs, and the next generation of AI is likely to increase this requirement even further.
Think about what will happen when AI goes beyond chatbots and cloud applications. You will get ambitious vehicles, automated vehicles, and humanoid robots. You will get industrial automation. You will get an intelligent machine that operates continuously in the physical world. Each of these systems must process and store vast amounts of information , potentially creating another wave of demand for memory.
Why do migrants want to control technology? Migrants have been through a brutal cycle of memory before. The industry has historically suffered from a familiar pattern. Demand is growing and manufacturers are adding capacity. Supply is catching up, and prices are falling. Margins disappear , and then companies cut production.
The cycle eventually begins again. But today Micro is making a completely different argument. The company believes that AI creates a stretched increase in memory requirements, which could make this cycle sustainably longer. The clear investment in research suggests that management does not view this as a temporary blip.
Comments 0