A cooling system can circulate liquid continuously without consuming an equivalent volume of fresh water, yet the infrastructure surrounding that system determines how much water a site ultimately needs. This relationship has become increasingly important as high-density computing changes how data centers capture, transport, and reject heat. Liquid cooling moves thermal energy away from processors through controlled fluid circuits, reducing dependence on the airflow strategies that traditionally supported server environments. However, removing heat from a processor represents only one stage of a much larger thermal process, because the energy must eventually reach an external environment or another useful application. The equipment responsible for that final transfer can rely on ambient air, evaporative cooling, hybrid arrangements, or combinations of these approaches. Consequently, the water implications of liquid-cooled computing depend less on the presence of liquid inside a server and more on the complete cooling architecture that supports its operation.