Model Lifecycle
Inference
The step where a trained AI model is actually used to generate an answer — as opposed to training, when it learns.
Training happens once, or periodically for updates; inference happens every single time someone sends a prompt. Inference is what runs when you chat with an assistant, generate an image, or call an AI API inside your own app.
Most of the industry's day-to-day GPU demand — and a growing share of AI capital spending — is now inference rather than training, since usage scales with the number of people and products running on AI, not the number of new models being built.