Liquid Cooling Technology in Data Centers: How It Supports AI Workloads
The short version: Equinix's slick explainer on direct-to-chip liquid cooling sells efficiency and sustainability, but skips the water, energy, and grid tradeoffs communities actually live with.
Our Take
Equinix's video, a clean corporate explainer on direct-to-chip liquid cooling, makes a real technical point: liquid cooling lets racks run denser and can cut the energy overhead of chasing heat with fans and chillers. That matters as AI training clusters push power draw per rack to levels air cooling was never built for. Credit where it's due — this is a genuine efficiency gain over legacy air-cooled halls.
But "optimizing energy use" is doing a lot of work in that pitch. Liquid cooling reduces cooling overhead, it doesn't shrink the underlying compute load driving the power bill, and many of these systems still lean on water for heat rejection somewhere in the loop — a detail glossy vendor content rarely lingers on. The efficiency story also conveniently sidesteps the questions that matter to the towns hosting these buildings: what utility rate case is paying for the substation upgrade, what water permit was issued and by whom, and whether ratepayers or taxpayers are subsidizing the very density this tech enables.
Better hardware inside the building doesn't answer what's owed to the community outside it. Check our facility map to see what's been approved near you, and use our take-action resources if you want to ask your utility or county board those questions directly.
This is GridWatch the USA’s original commentary. The video above is the work of Equinix, published on YouTube — full credit to the creator.