New framework uses waste-to-energy to cool AI data centers, saving up to 264 MW

A new arXiv paper from Qi He and colleagues introduces a boundary-consistent screening framework for coupling waste-to-energy (WtE) plants with AI data center cooling. The core innovation is grade-matched cooling: treating waste heat as an energy service that can supply cooling directly, bypassing the need for electricity-driven mechanical chilling. The framework translates plant-side exportable heat into corridor-level planning metrics, accounting for thermal attenuation, absorption conversion, and parasitic electricity for delivery. This allows planners to screen whether WtE-coupled cooling is feasible over a given distance and scale.

Key numbers illustrate the potential: a regulated WtE plant processing 1500 t/day of municipal solid waste at 10 MJ/kg provides about 78.1 MWth of exportable heat. Over a 20 km corridor, this yields 53.0 MW of delivered cooling and 8.0 MWe of net avoided cooling electricity. For a large 1 GW IT campus at 70% utilization and a 5 km corridor, net grid relief ranges from 116.9 to 264.4 MW across scenarios. The framework identifies three regimes: fully corridor-feasible cooling, hybrid operation, and infrastructure-scale constraints. Required WtE plant footprints range from 3 to 148 plants depending on displacement targets. The approach is designed for screening and comparison, not detailed hydraulic design.