2024 Undergraduate Research Showcase

Project Natick with Expanded AI Integration

Document Type

Student Presentation

Presentation Date


Faculty Sponsor

Dr. Sondra Miller


Underwater data centers dramatically lower the energy cost of data centers, but they are not enough on their own. Artificial intelligence (AI) can be used to further reduce costs through predictive resource allocation and workload management. The goal of this text is to inform readers about the benefits of using AI to optimize efficiency in underwater data centers. This entails analyzing current AI management practices’ ability to operate efficiently within underwater data center settings. We will examine existing algorithms for workload management, resource allocation, and effective cooling and power management. The assessment of these AI management systems will involve simulating underwater data center environments and evaluating their performance across various workload conditions. This research holds significance as it can contribute to reducing energy consumption in large underwater data centers. We aim to study established AI methods, validate their applicability in underwater environments, and refine them to accommodate unforeseen environmental factors. Additionally, we are dedicated to developing comprehensive guidelines and best practices for the deployment of AI-powered algorithms in underwater data centers, which will include conducting cost-benefit analyses to quantify their advantages.

This document is currently not available here.