0:00

But I took a look at the financials of Anthropic, and it's a nightmare.

0:07

Now, don't take this as investment advice, of course, but if you look at how much Anthropic is going to have to pay to all of these, uh, service providers and infrastructure providers, including companies like Broadcom, AMD, Microsoft, Google, and many others,

0:28

if you do the math on that, and there's some variation depending on the circumstances, but over the next few years, like to 2031, Anthropic is going to have to come up with 800 billion dollars just to meet its signed obligations,

0:46

or, or nearly 800 billion. It's just under.

0:50

And that's already, you know, on paper. And yet, the company is burning cash like you can't believe.

0:58

Its cash flow is, you know, very small. Its revenues are a tiny fraction of 800 billion, you know, not even 20 billion.

1:09

And I don't see any possible way that Anthropic can earn what it needs to earn in order to just service the obligations it's already committed to.

1:22

I did the math on that, and came out to a number of about...

1:27

It really needs to earn about 100 billion a year, give or take, in order to be profitable.

1:35

And that includes everything, you know, operating expenses and all the leases and additional hardware investments and so on.

1:42

It needs about 100 billion a year in revenue.

1:45

Again, right now, it's only at a fraction of that.

1:48

And it's also clear that this needed revenue, of course, cannot come from end users with 100 dollar a month subscriptions or whatever it costs, or even 200 a month.

2:00

I mean, this isn't about serving end users.

2:04

The, the individual end user customers or even the individual corporate customers, these...

2:11

Their fees will never represent the level of revenue that Anthropic needs to justify its ongoing investments and its expansion, especially in infrastructure, even, uh, a massive number of TPUs leased in order to do model training.

2:29

I mean, it's, it's massive. And, and by the way, to Anthropic's credit, their models are really outstanding.

2:36

Opus 5.5 is largely regarded by many people, including those I trust, as the top coding model in the world.

2:46

Now, I'm not using it because I've, I've gone to everything local.

2:52

I'm using DeepSeek for coding right now, and I'm quite happy with DeepSeek.

2:56

Opus 5.5 is a little better than DeepSeek.

2:59

But then again, it has a lot of refusals in it.

3:02

So you're trying to do certain things, and it will say, "No, I'm not willing to do that."

3:06

I mean, even just regular security type of code or trying to troubleshoot, uh, security breaches or things like that, Opus will refuse.

3:15

And that's one of the problems with using cloud-based models, is you don't have control, you could be cut off at any time, et cetera.

3:22

And that's why more and more people, including companies, are moving to open source AI that they can run locally, or they can run even uncensored or obliterated models that don't refuse things.

3:36

So regardless of what you think about the capability of the models, there are still disadvantages to using them in the cloud, and that's why local AI will always have these advantages.

3:48

But if you think about it, the only way Anthropic could actually continue to exist, you know, without going bankrupt, well, there are really two ways.

3:55

One is they have to keep borrowing money and just keep rolling it over, almost like a Ponzi, keep bringing money in, which of course depends on the cost of borrowed money remaining relatively low, which it isn't.

4:12

Uh, interest rates for AI-related companies are rising dramatically, and some of the debt of some companies like CoreWeave has been put at about, uh, well, 16% yield in order to entice buyers to carry that debt.

4:31

16%. That's, I mean, that's getting crazy, right?

4:36

So AI companies aren't going to be able to borrow at 6 or 7 or even 8% any longer.

4:43

And so that becomes sort of a self-defeating thing if you have to keep borrowing in order to stay alive, but the borrowing becomes more and more expensive, and you have to roll over old debt into new debt at much higher rates,

4:56

then, you know, the end is near. Bright Videos.

End of transcript · 729 words

iPhone app Translate videos on your iPhone Paste a TikTok, Instagram or X link, add subtitles in your language and save the video to Photos.

Made with TranslateMom

TranslateMom transcribes and translates videos from X, TikTok, YouTube and Instagram. Paste a link and you get a page like this one, in your language.

Transcribed with TranslateMom. Paste any video link to get subtitles and a transcript in your language.

© 2026 TranslateMom · AI video translation and subtitles Transcripts Sources Pricing Legal