원문 보기
Everybody has been asking me to ask you, are we going to have an AI bubble? The question every investor is afraid to ask, asked directly to Jensen Huang's face. And what he said next is exactly why Nvidia could skyrocket. We uh worked together to get this company started and off the ground, and just got an incredible customer with Elon.
Uh could you imagine a startup company approximately zero billion dollars in revenues now going to build a data center for Elon. There are real, concrete arguments for why Nvidia moves before the end of this month, and I'm going to walk through every single one of them right now. Starting with the most important one.
Argument one, Jensen Huang made these announcements hours before Nvidia's own earnings call, not after, not in a press release, live, on stage, in front of the world. That is not something you do unless you are extremely confident in what the numbers are going to show. By the way, the Nvidia earnings call is later today.
Think about what [clears throat] that moment actually means. Jensen is sitting on a stage in Riyadh, Saudi Arabia, next to Elon Musk announcing billion-dollar partnerships, cracking jokes, and casually dropping the fact that his own earnings call is happening later that same day. CEOs don't do that when they're nervous about their numbers.
They go quiet. They stay cautious. They let investor relations handle the communication. Jensen did the opposite. He went on stage in front of the world's cameras, made major commercial announcements, and telegraphed confidence so clearly that anyone paying attention should have picked it up immediately.
When a CEO is that relaxed, that public, that willing to make news the morning of his own earnings call, the results are good. That's argument one, and it sets the tone for everything that follows. Argument two is about the scale of what was actually announced in Riyadh, and most people covering this story are massively underestimating it.
The Kingdom of Saudi Arabia signed a full artificial intelligence strategic partnership with the United States the day before this event. Jensen was there. Elon was there. And the deals that followed weren't ceremonial. They were commercial. 500 megawatts is gigantic. Uh this company is off the charts uh right away.
500 megawatts is not a number that most people have context for. So, here's the context. A single megawatt of data center capacity costs roughly two to three million dollars to build out. 500 megawatts puts that deal alone in the range of 1 billion dollars minimum. And that's just phase one, starting at 50 megawatts and scaling from there.
Amazon Web Services came in right behind it, starting at 100 megawatts with an ambition to reach a full gigawatt. Jensen confirmed Nvidia is at the center of both. The chips, the software, the infrastructure, all of it running on Nvidia hardware. And this isn't Saudi Arabia being the only buyer. This is Saudi Arabia being the template.
When one nation commits this aggressively to artificial intelligence infrastructure, every other government watching starts doing the math on their own position. The ripple effect from Riyadh extends well beyond this single event. These aren't future promises. These are contracts being signed right now, which means revenue flowing directly into Nvidia's books in the quarters ahead.
That's argument two. But here's where it gets interesting because argument three is not about the deals Jensen announced. It's about something much deeper. And there's a part of this story that almost nobody is talking about. I'll get to that in a moment. First, you need to understand the three layers of demand that Jensen laid out because without that context, the rest of this doesn't land the way it should.
Most people think about artificial intelligence demand as one thing. Companies buying chips to train and run large language models, chatbots, image generators, coding assistants. That's the layer everyone sees. But, Jensen described two entire layers of demand happening underneath that, completely separate from artificial intelligence as most people understand it.
The first invisible layer is the death of Moore's Law. For 50 years, the technology industry ran on a simple assumption. Traditional computer processors would double in power every 2 years, automatically, without anyone having to do anything differently. That free performance upgrade is over. The physics ran out.
And the entire computing industry has had to find a new way to keep up with the demand for more processing power. The answer was accelerated computing. Specialized chips designed to handle specific types of heavy computation, far more efficiently than general-purpose processors. That's what Nvidia builds, and that transition is not coming.
It already happened. CPUs were 90% of the world's supercomputers, top 500 supercomputers, 6 years ago. This year, less than 15%. The most powerful computers in the world have already flipped almost entirely away from traditional processors. That's not a prediction. That's a fact sitting in the data right now.
And Nvidia captured the majority of that transition. The second invisible layer is recommendation systems. Every time you scroll a social media feed, every time a website shows you an ad, every time a streaming platform suggests what to watch next, a recommender system is running underneath it. These systems power most of the internet's revenue, and they are quietly migrating from traditional processors to graphics processing units right now.
That's an entirely separate wave of demand for Nvidia hardware that has nothing to do with chatbots or language models. Now, here's why all of this matters for the bubble question, the one that started this video. The single biggest fear hanging over Nvidia right now is that the artificial intelligence spending boom is a bubble.
That companies are throwing hundreds of billions of dollars at infrastructure that won't generate real returns, that the stock is built on hype. Jensen's three-layer breakdown directly addresses that fear. And his answer is more powerful than most investors realize. It's really important when you look at what's happening around the world and go back to first principles of what's happening in computer science and computing.
There are three things that are that's happening. The first thing is that we all know that Moore's law has run its course. That's argument three. The demand for Nvidia's hardware isn't being driven by one trend. It's being driven by three separate structural shifts happening simultaneously. The end of Moore's law forcing the entire computing industry to switch to accelerated chips, the migration of recommendation systems that power internet advertising onto graphics processing units, and then on top of both of those, the agentic artificial intelligence layer that everyone is watching.
Three independent demand drivers all pointing in the same direction, all requiring the same hardware. If you take that into consideration, you'll come to the conclusion that in fact what is left over to fuel that revolutionary agentic AI is not only substantially less than you thought and all of it justified.
That last part is the key. The compute available to power the artificial intelligence applications everyone is focused on is actually smaller than people assume. Because so much of it is already being consumed by the two infrastructure shifts happening underneath. Which means the demand for Nvidia's chips isn't inflated.
If anything, it's being undercounted. That's the opposite of a bubble. Now, here's the part almost nobody is talking about. And this is the reason the long-term case for Nvidia is bigger than even the bulls are pricing in right now. And perhaps in the four or five-year time frame, the lowest cost way to do AI compute will be with solar-powered AI satellites.
Jensen is saying that the cheapest way to run artificial intelligence computing will not be on the ground. It will be in space. The reason is straightforward. Cooling a data center on Earth is enormously expensive. Every rack of chips generates massive heat that has to be managed with complex water cooling systems.
In space, you radiate the heat away passively. You have continuous solar power with no night cycle and no need for batteries. The solar panels are cheaper because they don't need glass or protective framing. The economics are, according to Jensen, overwhelmingly better than building on the ground. That's not a distant science fiction scenario.
That's a five-year prediction from the person currently building the chips that will power it. And when that market opens, the total addressable market for Nvidia doesn't grow incrementally. It expands by an order of magnitude. And then, there's Elon's contribution to this conversation, which reframes everything else being said on that stage.
My prediction is that work will be optional. Optional? Optional. Um so, We'll take that. Elon is saying that the combination of artificial intelligence and humanoid robotics will generate enough economic output that human labor becomes a choice rather than a necessity. Whether you believe that timeline or not, it tells you something crucial about the scale of the infrastructure being built right now.
Comments 0