The Less Water Data Centers Use, The More Power They Consume. Here’s Why

An independent exercise, not a statement by the video creator or verification of every claim in the video.

Can on-site water savings increase the combined site-and-power total?

For a hypothetical facility, hold the period and useful IT workload constant. Assume the established method consumes 100 m³ on site and uses 100 MWh of total electricity, while an alternative consumes 20 m³ on site and uses 150 MWh. With a power-generation water-consumption factor of 2 m³/MWh for both, combined totals are 100 + 100 × 2 = 300 m³ and 20 + 150 × 2 = 320 m³. On-site savings of 80 m³ are outweighed by an extra 100 m³ in power generation, increasing the sum by 20 m³. With the same factor instead set to 0.5 m³/MWh, the totals are 150 m³ and 95 m³, so the sum falls too. Break-even occurs at (100 − 20) ÷ (150 − 100) = 1.6 m³/MWh; on-site savings alone do not determine the overall direction.

Check for yourself

1. Is the water figure withdrawal or consumption, and does its boundary include only the site or power generation too?
2. Are workload and period matched, using a water-consumption factor appropriate to the actual power mix and season?

Limits of application

This hypothetical sum covers on-site and operational power-generation consumption, not actual cooling performance, factors or generation mix. Equipment manufacture, construction, water treatment and transmission losses are excluded; a common factor and no double-counting are assumed. Adding water volumes from different sources does not imply equal scarcity or ecological impact.