Designing Datacenter Power Delivery Hierarchies for the AI Era
Grant Wilkins, Fiodar Kazhamiaka, Alok Gautam Kumbhare, Chaojie Zhang, Ricardo Bianchini

TL;DR
This paper presents a framework for designing efficient datacenter power delivery hierarchies tailored for the AI era, accounting for evolving workloads, hardware, and power density challenges.
Contribution
It introduces a comprehensive evaluation framework that models deployment dynamics and operational factors, enabling better long-term power delivery design for AI datacenters.
Findings
Multi-resource stranding significantly impacts capacity and cost.
Rising AI system density alters deployment and performance metrics.
Effective planning should focus on deployable capacity over time, not just installed power.
Abstract
Demand for AI accelerators is rapidly increasing rack power density, with projections approaching 1MW per deployment by 2027. This poses a major challenge for datacenter power delivery designers. As power densities increase, a datacenter designed for a different target density may strand power, i.e., may be unable to use all the power that its delivery hierarchy has provisioned. Designs must remain efficient over long datacenter lifetimes and multiple hardware generations. Power utilization is particularly important as grid power capacity is a scarce resource in the AI era. Designing an efficient power delivery hierarchy for the long run is difficult because rack placement feasibility, workload impact, and cost depend jointly on electrical topology, deployment granularity, placement policy, power oversubscription, and workload mix. Moreover, each of these factors evolve over time,…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
