From Tokens to Energy Flexibility: Quantization-Enabled Demand Response for Data Centers with LLM Inference Workloads

Bojun Du, Xiaoyi Fan, Ershun Du +5

The rapid growth of large language model (LLM) inference is creating significant data-center loads that face increasing energy-management challenges under tightening grid condition…