I left out the parts of the article where he looks at the numbers, including Michael Burry's numbers (Burry says it's a bubble). He doesn't provide a definite answer anyway.
As noted in the article, a big difference between the 1999 internet bubble and AI is that the companies at the center of AI are mostly huge and profitable, like Nvidia.
The internet bubble featured IPO's of dozens of companies with no path to positive earnings. Many barely had any revenue.
For the Polymarket "YES" bets on "AI bubble bursts" to pay off, certain specific things have to occur within a specific timeframe. They could be early and therefore they won't pay off even if all those things eventually happen.
Wow, wondering what the 2nd comment was. Seems like some stocks have gone up in a parabolic fashion. History says that will not hold up.
so I watch GPU rental prices
I rent time by the hour on NVIDIA HPC servers , and right now the spot prices are stupidly low. At current hourly rates I could rent an H100 continuously for two years and not pay off the estimated hardware, electricity, and infrastructure costs.
AI is a bubble. Every time a disruptive technology comes along there is usually an investment bubble. But that doesn’t mean the new technology disappears or goes away. It just means there was an overbalance in investment vs adjusted ROI.
Yes, its a huge bubble. Banks and other companies are making huge investments in AI and the increased productivity simply is not there.