AI Factory Case Study on Total Cost of Ownership

This white paper presents a case study on the total cost of ownership (TCO) for AI factories, driven by the increasing demand for generative artificial intelligence (GenAI). It outlines the necessity for rethinking data center design and location strategies to accommodate the projected power requirements, which are expected to reach between 171 to 219 GW globally by 2030. The document details how AI factories differ from traditional data centers, emphasizing the need for higher power density and liquid cooling solutions to manage the unique challenges posed by GenAI workloads. It discusses the benefits of utilizing stranded power sources and optimizing cooling strategies to achieve significant cost reductions. The analysis suggests that careful site selection and innovative design can lead to annual savings of $30-$50 million for a 100 MW data center, with potential long-term savings reaching up to $2.7 billion over thirty years. The paper concludes by highlighting the importance of adapting infrastructure to meet the evolving demands of AI technologies.