(The AI Stack)
Energy, chips, infrastructure, models, applications: the whole AI industry in five layers, in about half an hour.
A free, visual course on the five-layer cake, NVIDIA CEO Jensen Huang's framework for the AI industry: energy, chips, infrastructure (AI factories), models and applications. See how a single chatbot answer depends on all five, why power and chips limit how fast AI can grow, and how demand at the top ripples down through the whole stack.
Every big AI headline, from power-plant deals and chip shortages to trillion-dollar data-center plans and new models, is news about one layer of the cake. Know the layers and you can tell what a story actually means.
Every answer is made from electricity, on the spot. Why power is the layer that limits all the others.
Billions of multiplications per word, all at once. Why AI chips are built for parallel math and fast memory.
Land, power, cooling and networking that turn tens of thousands of chips into one machine: an AI factory.
What most people think AI is. Not just chatbots: models for proteins, chemistry, weather and robots, open and closed.
The top layer is where the money is made, and its demand pulls on every layer underneath.
The slowest layer sets the pace. Bottlenecks, jobs, the bubble debate, and how to use the cake to read AI news.