AI Data Centers Meet Electrical Grids: A Review of Power Challenges and Coordinated Solutions
eess.SY, cs.SY
Submitted: 2026-10-06
Updated: 2026-10-06
Code: https://github.com/Azure/AzurePublicDataset
Project page: https://www.aeso.ca/grid/connecting-to-thegrid/large-load-projects
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Electricity Demand and Grid Impacts of AI Data Centers: Challenges and Prospects
- Enhancing Data Center Low-Voltage Ride-Through
- Data Center Model for Transient Stability Analysis of Power Systems
- Part-time Power Measurements: nvidia-smi's Lack of Attention
- Multi-Scale Datacenter Power Modulation
- Power Stabilization for AI Training Datacenters
- AI Load Dynamics--A Power Electronics Perspective
- DualScale: Energy-Efficient Disaggregated LLM Serving via Phase-Aware Placement and DVFS
- PALS: Power-Aware LLM Serving for Mixture-of-Experts Models
- Wide-Area Power System Oscillations from Large-Scale AI Workloads
- Lazarus: Resilient and Elastic Training of Mixture-of-Experts Models
- FFTrainer: Fast Failover in Large-Language Model Training with Almost-Free State Management
- The Unseen AI Disruptions for Power Grids: LLM-Induced Transients
- GAR: Carbon-Aware Routing for LLM Inference via Constrained Optimization
- MARLIN: Multi-Agent Game-Theoretic Reinforcement Learning for Sustainable LLM Inference in Cloud Datacenters
- Adaptive Multi-Objective Tiered Storage Configuration for KV Cache in LLM Service
- Architectural Implications of Agentic AI Workflows
- Sequential Operating Simulation of Solid State Transformer-Driven Next-Generation 800 VDC Data Center
- Datacenter Energy Optimized Power Profiles
- TokenPowerBench: Benchmarking the Power Consumption of LLM Inference
Related papers
- One Request, Multiple Experts: LLM Orchestrates Domain Specific Models via Adaptive Task Routing
- A Geometric Decision Procedure for STL Feasibility and Repair
- From Asymptotic to Designer-Assigned-Time Control: A Review of Stability Notions, Design Mechanisms, and Controller Architectures
- Battery Second Life: A Review of Experimental Studies
- Submodular Multi-Agent Policy Learning for Online Distributed Task Allocation in Open Multi-Agent Systems
- Policy-Level Recursive Self-Improvement for Embodied AI with a Criticality World Model