Demis Hassabis emphasizes GPU/TPU inference compute is becoming dominant capex vs. training compute; all three dimensions (pre-training, post-training, inference) scaling simultaneously with inference potentially becoming the larger portion of required compute

“There's the amount of compute you have for training...But there's also because now AI systems are in products and being used by billions of people around the world, you need a ton of influence compute now. Um and then on top of that, there's the thinking systems...where they get smarter the longer amount of inference time you give them at test time. So, all of those things need a lot of compute...the training side actually is is only just one part of that. It may even become the smaller part of of what's needed in the overall compute that that's required.”