Substack needs more articles like these. Too many articles read like sales pitches, rather than research projects.
Your conclusions are sound (I think anyways). Eventually supply will catch up with demand and then things will look very different. Sure that could be years away, but I think it's very reasonable to say that there are too many ifs there for now. I drew the same conclusions looking at 2g energy last year.
The power constraint is becoming a real part of the AI investment story. But a shortage alone doesn’t guarantee attractive returns — capital costs, equipment utilisation and the timing of new supply matter just as much.
Substack needs more articles like these. Too many articles read like sales pitches, rather than research projects.
Your conclusions are sound (I think anyways). Eventually supply will catch up with demand and then things will look very different. Sure that could be years away, but I think it's very reasonable to say that there are too many ifs there for now. I drew the same conclusions looking at 2g energy last year.
The power constraint is becoming a real part of the AI investment story. But a shortage alone doesn’t guarantee attractive returns — capital costs, equipment utilisation and the timing of new supply matter just as much.
I’m not sure about the real part of the power constraint. The answers are in the article :)