SpaceX $40 Billion Debt Financing for NVIDIA GPU's - AI is Too Expensive

Eli the Computer Guy (0:00)

I want money, lots and lots of money, piled up to the sky. I wanna be Rick.

I guess that is the new anthem for the AI world. I mean, since AI is basically a 1980s sci-fi movie fever dream, we might as well start singing 1980s songs too. So yeah, I mean, this is kind of interesting. This is where it shows you that the amount of money all these AI company needs is absolutely and utterly insatiable.

The amount of pure cash that they are consuming is just wackadoodle, is the technical term, right? So we had SpaceX, SpaceX IPO'd a couple of months ago at this point in time. They're now back up. They're pretty high. They're at like $175 a share, which is something like a $2.3 trillion valuation. So basically, they have that going on. But even as they have that going on, they're still going out to get more and more debt financing because they need to buy more and more hardware for the AI boom, right? The AI boom that everybody keeps telling you about, the AI boom where you have been informed, this is changing your life every day.

Yeah, that is why they need more and more hardware, and they are just burning money to get it. If we look at this coming from Reuters, SpaceX seeks $40 billion financing to buy Nvidia chips. SpaceX is in talks with banks and asset managers to raise $40 billion for the purchase of Nvidia AI chips. The Elon Musk led company is in talks with asset manager Pimco for financing. SpaceX is looking to raise about $10 billion in bank loans and $30 billion in investment-grade debt for the chip order.

Apollo is expected to lead the SpaceX deal and help place the debt with a broad range of investors. The proposed funding underscores the enormous capital requirements of the AI boom as technology companies race to secure advanced processors and build off the computing infrastructure needed to support AI development. Morgan Stanley estimates that AI infrastructure will require $1.5 trillion in external financing by 2028, even as lenders and investors grow more cautious about funding the industry's expansion. Musk, who took SpaceX public in June in a record $86 billion IPO, said last month that XAI's Colossus II data center could more than double the number of Nvidia chips it uses by December. The company plans to use Nvidia hardware exclusively to build its data centers, Musk said. So basically, Musk is just building the hell out of those data centers. And one of the things I find interesting about this, too, is that they're going to be exclusively using Nvidia GPUs. Why that's kind of interesting, right? We've heard about the terafab. So I think they've broken ground on the terafab. The terafab is supposed to be like the largest chip fab in the world. It's going to be 100 million square feet of internal space or something ridiculous. So it's kind of interesting here that he's so gung-ho on Nvidia. But it might just be, you know, that is the current... that is what he currently can buy. Like he's not actually producing chips yet. So these are the chips that he can buy. And they're going forward with it. I don't... I don't know. There's just so...

I just don't get it.

They're just dumping so much money into this AI idiocy. Like, I think about it with, you know, with OpenAI. So OpenAI solved this math problem recently. So there's this math problem. It's like a hundred-year-old math problem. These mathematicians were using AI to help them solve the math problem. OpenAI heard about it apparently, and decided they were going to solve the math problem first. Anyways, that is what it is. The curious thing that came out was they spent something like 21, 22 million dollars on tokens. So, just for them to solve it... And so, this is a hundred-year-old math problem. Like, is it a real problem? Yes, it is in fact a real problem. Also, do people really care that much? Again, in the math community, it was important. For the rest of us, it doesn't really matter. Anyways, one of the interesting things there is that they burned over 20 million dollars to solve this problem, just to say that they could. Again, basically, no ROI. Even there was a prize for solving the problem. The prize for solving the problem was a million dollars.

So they were net minus at least 19 million dollars in order to solve this problem. And one of the things that I see with so much of what's going on with AI right now, is that companies are just burning money. And in order to burn money, you need GPUs, right? You need GPUs of the highest of quality to fully combust your cash. Right? I've talked about this before with the hugging face incident, right? You hear about the felonious attacks against hugging face that OpenAI committed. And you think, okay, so there's one agent, there's two agents, there's a few agents decide to go after hugging face. One of the most fascinating thing about that is apparently Percy NBC, as I've been able to read, it was 14,000 agents. 14,000 agents hammered the ever-loving shit at a hugging face for days or possibly weeks. And the thing is, with how agents run, they are just burning. It's a wild true loop of burning cash.