6 ms·

Ask HN: What is the evidence for a stock market bubble in AI?

11 points by roschdal 26d ago

Should the circular cash flows between the top AI companies, Nvidia, Anthropic, Openai, Google and Meta, be banned because of the systemic risks to the global economy?

vga1 26d ago
Number go up
rvz 26d ago
> What is the evidence for a stock market bubble in AI? Nvidia becoming a bank and having to save / bailout failing AI and semiconductor companies. There is a reason why Jensen does not want a ban on China for selling his GPUs there: China will build their own GPUs or AI accelerators for cheap regardless, but they will get there faster if there was an NVIDIA GPU ban. They are already there on open weight models and Jensen knows that it is only a matter of time until China catches up with GPUs or other AI accelerators.
- locknitpicker 26d ago
  > There is a reason why Jensen does not want a ban on China for selling his GPUs there: You don't need a conspiracy or a bubble theory to understand why a multinational company doesn't want to arbitrarily exclude itself from a market with a population of 1.4B.
  - rvz 24d ago
    The reason was already known. Jensen publicly said that if any US government done a GPU export ban in China, they will just build an alternative GPU themselves. [0] Their point was to slow down China in 2023, but Jensen already knew that doesn't work. The government might as well have just done a ban on India as well, but obviously only China. [0] https://archive.is/MAZYk https://archive.is/MAZYk
pestatije 26d ago
valuation vs earnings...nothing special about this one
just60sec 26d ago
[dead]
cyanydeez 25d ago
Circular financing is the big one. Enron and the dotcom bubble were both doing it, wether it was actual fraud or fraud adjacent, that's how recent bubbles pop.
sajithdilshan 25d ago
I don’t think all the companies you’ve mentioned are in a bubble territory. Even without AI some of these companies have a good source of revenue. However some are over valued. As an example Google is a company with a lot of products and diversified revenue streams. Last year the revenue was around 350 Billion and the ratio of price to earnings was around 17. But if you take Nvidia, their main product is GPUs and last year the revenue was around 190 Billions, but the price to earnings ratio is 33. Both companies have around 100 Billion net revenue. However, let’s say the demand for AI is dropped or disrupted, Nvdia would be vastly overvalued because they sell only one product and now the demand dropped for it and won’t have the same revenue moving forward. But Google would survive because their revenue is diversified. We don’t know any official revenue to expenditures on Anthropic or OpenAI. They have to publish them when going public and then only we would have a good idea how profitable or growth opportunity AI actually have and that would be a deciding point stock market
sloaken 25d ago
I am a big P/E (price vs earnings) fan. Value companies typically around 10. Growth companies usually exceed 20. IMHO anything over 40 is too risky. Although this would prevent me from being an early investor in many successful companies. It also keeps me away from the larger number of failures. <edit> For what it is worth, they have trailing P/E (history) vs foreword P/E which is expected. As of today NVIDIA Trailing P/E is 33.24 and forward is 24.75.
Jlagreen 23d ago
Everyone is looking what Big Tech is doing with their cashflow without understanding what is really going on. What is the TAM of AI and therefore the future AI compute demand? Answer this question and you might understand what is going on.

11 points by roschdal 26d ago

Should the circular cash flows between the top AI companies, Nvidia, Anthropic, Openai, Google and Meta, be banned because of the systemic risks to the global economy?

vga1 26d ago

Number go up

rvz 26d ago

> What is the evidence for a stock market bubble in AI? Nvidia becoming a bank and having to save / bailout failing AI and semiconductor companies. There is a reason why Jensen does not want a ban on China for selling his GPUs there: China will build their own GPUs or AI accelerators for cheap regardless, but they will get there faster if there was an NVIDIA GPU ban. They are already there on open weight models and Jensen knows that it is only a matter of time until China catches up with GPUs or other AI accelerators.

locknitpicker 26d ago
> There is a reason why Jensen does not want a ban on China for selling his GPUs there: You don't need a conspiracy or a bubble theory to understand why a multinational company doesn't want to arbitrarily exclude itself from a market with a population of 1.4B.
- rvz 24d ago
  The reason was already known. Jensen publicly said that if any US government done a GPU export ban in China, they will just build an alternative GPU themselves. [0] Their point was to slow down China in 2023, but Jensen already knew that doesn't work. The government might as well have just done a ban on India as well, but obviously only China. [0] https://archive.is/MAZYk https://archive.is/MAZYk

pestatije 26d ago

valuation vs earnings...nothing special about this one

just60sec 26d ago

[dead]

cyanydeez 25d ago

Circular financing is the big one. Enron and the dotcom bubble were both doing it, wether it was actual fraud or fraud adjacent, that's how recent bubbles pop.

sajithdilshan 25d ago

I don’t think all the companies you’ve mentioned are in a bubble territory. Even without AI some of these companies have a good source of revenue. However some are over valued. As an example Google is a company with a lot of products and diversified revenue streams. Last year the revenue was around 350 Billion and the ratio of price to earnings was around 17. But if you take Nvidia, their main product is GPUs and last year the revenue was around 190 Billions, but the price to earnings ratio is 33. Both companies have around 100 Billion net revenue. However, let’s say the demand for AI is dropped or disrupted, Nvdia would be vastly overvalued because they sell only one product and now the demand dropped for it and won’t have the same revenue moving forward. But Google would survive because their revenue is diversified. We don’t know any official revenue to expenditures on Anthropic or OpenAI. They have to publish them when going public and then only we would have a good idea how profitable or growth opportunity AI actually have and that would be a deciding point stock market

sloaken 25d ago

I am a big P/E (price vs earnings) fan. Value companies typically around 10. Growth companies usually exceed 20. IMHO anything over 40 is too risky. Although this would prevent me from being an early investor in many successful companies. It also keeps me away from the larger number of failures. <edit> For what it is worth, they have trailing P/E (history) vs foreword P/E which is expected. As of today NVIDIA Trailing P/E is 33.24 and forward is 24.75.

Jlagreen 23d ago

Everyone is looking what Big Tech is doing with their cashflow without understanding what is really going on. What is the TAM of AI and therefore the future AI compute demand? Answer this question and you might understand what is going on.