After reading that Amazon's cloud capacity is largely sold out through 2027, and hearing his co-host's doubts about how much reserved GPU capacity is actually used, Ryan says it is notable that the cloud known for deep capacity now says it is constrained.
…being a poor use of capital supply, not demand, but it appears to be this constraint. Demand is described as coming from both AI labs and enterprises, which points to the capacity competition between frontier model builders and typical enterprise adopters worth discussing how this affects. Pricing.
…that on the back end. You know, Ryan here said, I need, you know, five GPUs, and Matt said he needed three, and I'm only using one. So Josh comes in and requests two. Are they actually available to the random on-demand user or not? I don't, I kind of wonder how they're reserving it on the back end.
Yeah. I mean, it's an odd thing. And it's like, if stuff's already being sold into 2028, if you come in and want reserved capacity, do you do you have to wait until 2028 to get that reserve? Like, will they even sell it to you? So it is sort of crazy. And it's Amazon. Like, Amazon's not never the cloud that has the capacity constraints. Like, I'm used to this from Azure and GCP, where it's just like, this isn't very cloudy, but when Amazon comes out and announces this, the AI hardware constraint is real.
Whatever happened to the Chinese model that was able to train their model on more generalized compute and older GPUs than requiring the newest ones.
It was a deep seek training exercise, I believe. Like it wasn't anything special, but it's like it's, yeah, it's achieving the similar level benchmarks with building that model on way different hardware. But according to OpenAI, though, that's a lot of distillation of the models of the GPT models…
Jonathan, come save us.
