How can we make data centers more energy- and water-efficient and make AI more sustainable? To tackle this challenge, this team will design and build a miniature data-center cooling testbed that makes data center energy/water consumption as well as cooling challenges visible, measurable, and solvable. The core deliverable is a data center architecture built on a breadboard using modular components, including controllable heaters that reproduce realistic chip heat flux, as well as cooling systems across chip, rack, and facility levels. The modular structure enables rapid reconfiguration, comparison of cooling strategies, and long-term system expansion. A central goal of the project is to demonstrate that meaningful thermal and energy systems research can be conducted using low-cost, widely accessible sensors and equipment. The platform will integrate affordable sensors, microcontrollers, as well as compact pumps and fans. The team may develop different cooling architectures, such as air cooling, liquid cooling, and two-phase cooling. Performance will be benchmarked using standard metrics such as overall thermal resistance, power usage effectiveness (PUE), and water usage effectiveness (WUE). Over multiple semesters, the teams will iterate on mechanical design, instrumentation, controls, and system optimization, creating an extensible research platform that supports publishable results, undergraduate leadership, and scalable, affordable experimentation for sustainable computing.