There is no question that AI is a bubble, but that doesn't mean its going to pop necessarily, and if it does, that statement doesn't decide the timing. Most value investors are more skeptical than the average investor so I am sure many of us have been calling bs for years. We may be right but if we were to short 2 years ago we would have missed out on 50% of gains by trying to be the smartest in the room instead of waiting for a real data point.
I was listening to Prof G Markets the otherday, who, by the way tend to be very bearish, had Steve Eisman on. Steve Eisman is a very intelligent guy who shorted the market in 2008 with Micheal Burry. He was asked about weather he sees eye to eye with Burry's calls on the market and he said this. "In 2008 we had a real data point that we could point to every month (monthly mortgage securitization data)." You could point to this data and visibly see how bad this bubble was getting.
We simply don't have that data point, or atleast it hasn't been publicly found for AI bubble. The two biggest weak points in my opinion are neocloud companies (for obvious reasons), and private credit funding these highly depreciating GPUs. However private credits books are not public, and further more, spvs make it harder to see anything going on with the real time health of this financing.
I am currently watching neocloud credit spreads as a data point but certaintly isn't enough, any thoughts?