“To operate a gigawatt cluster for a year, if you're talking about building the CapEx it's probably closer to, you pay for it upfront, so it's probably $50 or $60 billion. But then that's just one cluster for one part of training.”

“I actually think some of the more complicated ere, all right, so we're training this model and we know we want to train the model because this is key, but how much product demand do we think we're going to have?”

“There was a moment where a bunch of people were like maybe we should just be a little more conservative. And it's like, no, we're doing this. We're going to make it work.”

“If AI works overall, plenty of other people want our compute. So they're willing to pay if the cost is whatever, $15 billion a gigawatt, we could get a lot more for that. So I think that in order for it not to work out, it's not just that we have to fail, AI overall has to fail.”

“And then even if that happens, we still have all these other uses to productively use compute. So in the grand scheme of things, we were definitely building out this compute for inference ahead of knowing that the product was going to be ready.”

@finkd @alexandr_wang