Cloud provider facing compute constraints limiting revenue growth

Bullish+0.01%

AI Summary

The author presents a bullish thesis on AI compute infrastructure, highlighting a severe supply/demand gap in GPUs and ASICs alongside power constraints that are limiting chip deployment and capping cloud revenue growth for major hyperscalers.

One of the topics discussed on the pod with @LoganJastremski, do we have enough chips or do we not? There are two arguments here: 1. We don't have enough power and this was emphasized by Satya Nadella in a podcast where he says they don't have warm shelves to plug chips into so they are basically being warehoused. 2. The users of compute (labs + hyperscalers) mention a compute constraint time after time, with companies like Amazon and Google referring to a higher cloud revenue if they had more compute. On top of that, CLSA Research claiming a S/D gap of 73% in ASICs and GPUs. It's also one of the reasons why Michael Dell claims he looks at neoclouds to survey demand of compute as they are able to get land, power, and shell. $NVDA $AMD $GOOGL $MSFT $AMZN $META $DELL

More Views About GOOGL