원문 보기
Does this mean you need to sell your data center-related stocks ? You will be like us in the trust fund. I think some of it is worth selling. You can't be that deeply invested in this topic because the price /earnings multiples won't increase. And if the data center supporters, whose ranks are thinning, lose in the upcoming elections, the people who sell these stocks now will look like geniuses after the election.
But, say, what about the money flowing into Procter & Gamble and Coca-Cola or United Health and financial institutions like J.P. Morgan and their clients? Will Apple, Mercedes, Western Digital, Seagate, Micron and others be able to keep their groups afloat? The answer is yes, because it's not just about these actions.
The list is as long as your arm. Almost everything in healthcare is suitable. Just like travel and vacation. People just want to get out of data centers, and these are convenient places to put capital. However, there is one saving grace, and it is quite significant. Currently, the cost of building a data center is increasing due to the fact that many companies are building them for speculative purposes, not just hyperscalers.
It's no coincidence, it's true that Amazon, Alphabet, Microsoft, and Meta stocks are growing rapidly, they're being bought up like crazy, just bought up like crazy. You see, they are the biggest beneficiaries because they can afford to offset the costs to local communities and build their own warehouses full of servers, while the speculative developers will be truly wiped out.
And they again have the opportunity to create this code of conduct. And if they weren't such fools, they would do it. If you get rid of the small speculative developers, you get rid of the fierce competition that has driven up the price of everything: from land to construction, labor , electricity, and, of course, the internal equipment of the data center.
Like Nvidia, which reports on Wednesday and has been falling for seven days in a row. I think it's mainly due to concerns about data center load. "The Magnificent Four"—here are the winners. Okay, what about the Fab Four? Yes, that's probably better. I think they'll continue to win because they were losing when people extrapolated the costs of building these data centers, right?
They're constantly thinking, "Oh, more capital expenditure, more capital expenditure. Sell, sell, sell." This political resistance could be a godsend for hyperscalers. Only a few of us think so, but the stocks are already moving. So, the money won't just flow to Johnson&Johnson. Have you seen this promotion?
They will go to the real hyperscalers, who are being criticized so much now. They can follow these new rules. Small players—no. A classic example of how government regulation destroys small businesses and supports big players. Here is the main conclusion. This resistance to data center construction could be a victory for big tech.
A true cynic would say they should support these new rules. Clear rules, a little money for each city, along with electricity subsidies and clean water commitments—they would be perfectly happy. This will help their profits and crush smaller players trying to compete. The situation is this: every note about Palo Alto Networks talks about great successes, but growth is starting to slow down a bit.
Growth rates are falling slightly, organic growth last quarter slowed to 14%, although everyone says the company is good. The stock has simply run too far ahead and its valuation is too high. But this is a quality company with strong recurring revenue and stable free cash flow, but as I said, every note talks about a premium valuation compared to the industry with growth normalized to 2027.
So people seem to be concerned about this. Most say stocks have had a great run, but they're just too expensive, Diana , Tom. Okay, it seems you support this analyst's view. Tom , what is your opinion? Yes, the analyst is generally raising target prices for all of these cybersecurity companies; but he downgraded Palo Alto , Okta and SentinelOne, although he raised the target prices for them, as well as for CrowdStrike, Fortinet and others in the sector, speaking of fair value .
You should start paying attention to the valuation of these companies. Kevin mentioned that their organic growth is slowing down, but if you look at their acquisition history over the last year—they bought Console, Port Key, Coy Security, CyberArk. So they achieve this through mergers and acquisitions, right?
And many of these companies are delivering immediate growth. They become profitable immediately after acquisition. So that's the advantage of their growth. You saw their numbers last quarter: revenue in the fourth fiscal quarter grew 34%. Their RPOs, or remaining contract obligations, exceeded $20 billion for the first time , ending the year at $ 21.2 billion, up 34% year-over- year.
So these numbers are starting to grow, and I think this whole talk about valuation is going to eventually become relevant, but at the same time, the amount of protection and cybersecurity that these companies are going to need as AI agents evolve is going to grow exponentially. We may not even know the total size of the market and exactly what type of cybersecurity firms will need in the future, and I think that's where the optimism in this area lies.
Okay, let's move on to examples of deals. Kevin, let's start with yours. What is your approach today? So, I looked at something, quite expectedly, from the bearish side, Diana and Tom. I looked at a two-week, or rather a three-week calendar spread. I have an expected movement on September 8th of around $23.
I have a move for October 2nd or, yes, October 9th for about $39. I chose something in the middle, around $30, and looked at a calendar put spread of 350 for October 9– September 25. And now it is trading around $5.25. That's where we opened it. It's probably trading ten cents higher because the stock has gone up.
He fell a little, so the day is in the plus. Nvidia is entering a new phase of the AI race, and specialized chips are becoming more common. Specialized defense systems are evolving, and Nvidia's answer is surprisingly simple. Surpass them. Jensen Huang made it clear that Nvidia already has the largest infrastructure base for AI.
Its platform is used in closed models, open models, clouds, and data centers. So instead of trying to eliminate all specialized chips, it's connecting those chips to its ecosystem. And this is exactly what we see in the example of D Matrix. This startup has developed a specialized data output processor designed for ultra-low latency AI workloads .
And now it is partnering with Nvidia to integrate its technology into the Nvidia platform via NVLink Fusion. Why is this important? Because Nvidia already has a huge installed base. Its equipment is deployed in clouds, countries, and data centers around the world. For a startup like D Matrix, rebuilding such infrastructure from scratch would take years.
Together with Nvidia, they can tap into an ecosystem that already exists. D Matrix has spent over 7 years developing memory-centric data output technologies. And time is very important. AI agents and coding tools have significantly increased the demand for rapid data extraction. When you interact with an AI agent, reaction speed matters: the faster the response, the more useful the system becomes.
And this is where specialized data output equipment can play an important role. D Matrix's upcoming Raptor XFuse takes a different approach to memory. Instead of relying on HBM, the company is developing a 3D- stacked DRAM architecture. This places the memory much closer to the compute unit to overcome one of the biggest problems.
Namely, the " wall of memory." Artificial intelligence systems constantly move huge amounts of data between memory and the processor. This creates serious challenges in terms of latency, power consumption, and cost. D Matrix believes that their architecture will help solve these problems. The company expects its first integrated product to hit the market in about a year.
The potential customer base is also quite broad. D Matrix says it is targeting hyperscalers, advanced AI labs, DNA analysis projects, cloud services, and high-frequency trading firms. So, it's not necessarily about replacing the Nvidia GPU. It's about adding specialized computations alongside them. And this distinction is important.
Nvidia seems to understand that AI won't be able to run on just one type of processor forever. Different workloads will require different architectures. Some will require Nvidia GPUs, while others may benefit from specialized chips. So Nvidia's strategy is probably to integrate such chips into its platform.
This turns a potential competitive threat into an opportunity for the ecosystem. But Nvidia's expansion into specialized AI chips is also drawing the attention of regulators. The Justice Department is reportedly investigating Nvidia's deal with Broadcom worth about $20 million.
Comments 0