Case Study: Managing Fluctuating Power Demand to Meet AI Workloads
3:40 pm - Tuesday 18th August 2026
- Examining how steep, fast AI load swings cause voltage and frequency deviations that trigger unnecessary transfers to backup systems
- Highlighting the importance of meeting ride through requirements to prevent cascading grid impacts
- Exploring rack-level and low-level technologies like embedded UPS to smooth spikes in load and improve power stability
- Evaluating strategies such as large-scale battery storage, capacitors and other power mitigation techniques to handle unpredictable AI workloads
