Bull thesis for $NVDA and the cloud providers in a nutshell: Though token prices are declining, task based cost of intelligence is still up 60% over the past year. This is because token prices are down by just 15% since January while token consumption doubled. This means we need to add a lot more capacity to efficiently serve the excess demand so task based cost of intelligence can actually start going down. We simply need more chips and data centers.
View on X ↗