Trends to watch: How shifts in data structure and volume demand new approaches to AI compute

Demand for AI compute is doubling every three months. Taking full advantage of data means using more of it, leading to larger, increasingly complex models with billions of hyperparameters requiring massive clusters of compute nodes, all while meeting ever-stricter latency and power requirements. Tingwei Huang explains how the way we compute AI has to be completely rethought so it can evolve to meet the promise of global business transformation.