Ask HN: What is the evidence for a stock market bubble in AI?
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?
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?
Discussion Highlights (6 comments)
vga1
Number go up
rvz
> 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.
pestatije
valuation vs earnings...nothing special about this one
cyanydeez
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
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
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.