This capex cycle has the same hopes and dreams of all capex cycles: that big investment will allow us to produce more real world stuff at lower cost. That's how we get the pay off. You build bigger and better factories and you sell more widgets. But this AI capex cycle is not about expanding raw productive capacity to make more stuff, instead it's a bet that AI will provide enough unearned knowledge that we will effectively be GIFTED ways to produce more stuff at less cost, and then we will act on that knowledge, and that THAT will provide the pay off down the road. There's an extra link in the chain. If, for some reason, AI struggles to make us more productive, but in the interim it erodes the demand for human labor from companies trying to adapt, the payoff for all of this investment would be much more amorphous and harder to see.(𝐓𝐗𝐌𝐂)