XWind: A Cross-site Router for Large Language Model Inference Serving at Renewable Energy Farms
The paper introduces XWind, a cross-site router designed for large language model inference at renewable energy farms. It addresses the growing demand for AI power and proposes a model called AI Greenferencing to optimize energy use from wind sources. The evaluation shows significant improvements in latency and efficiency compared to traditional methods.
- ▪AI power demand is increasing rapidly, while power grids struggle to meet this demand.
- ▪The proposed AI Greenferencing model aims to deploy modular AI compute at renewable energy sources, particularly wind.
- ▪XWind reduces end-to-end latency by up to 52% compared to the strongest competitor and by up to 98% over traditional baselines.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Distributed, Parallel, and Cluster Computing arXiv:2605.23348 (cs) [Submitted on 22 May 2026] Title:XWind: A Cross-site Router for Large Language Model Inference Serving at Renewable Energy Farms Authors:Tella Rajashekhar Reddy, Atharva Deshmukh, Liangcheng Yu, Chaojie Zhang, Mike Shepperd, Rohan Gandhi, Anjaly Parayil, Srinivasan Iyengar, Ajay Manchepalli, Debopam Bhattacherjee View a PDF of the paper titled XWind: A Cross-site Router for Large Language Model Inference Serving at Renewable Energy Farms, by Tella Rajashekhar Reddy and 9 other authors View PDF HTML (experimental) Abstract:AI power demand is growing at an unprecedented rate while power grids are often ailing and struggle to keep up.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
