NVIDIA's Rev Lebaredian breaks down Jensen Huang's 5-layer cake of AI: "Jensen aptly describes AI in layers. We call it the five-layer cake. The bottom layer of AI is energy. Everything starts from that. The maximum amount of AI you can have is dependent on how much energy you have" "Next layer above that is chips. This is the machine that converts the energy into tokens, into intelligence. Then you have infrastructure above that, which takes these chips and scales it out and makes it so we can really take it out to scale" "On top of that, we have the models. This is normally what we think of as AI. These are the GPTs and Claudes and Groks of the world" "The layer above that is arguably the most important, which is the applications. This is where we take all the AI, all this work we do below that, and we turn it into value that makes a difference in people's lives" @nvidia @JensenHuang