AI energy · Topic
Cooling and water use
Dense AI hardware generates heat that must be removed efficiently. Liquid cooling is now common for frontier AI, and some cooling approaches consume water, which raises local resource questions.
Key facts
- Direct-to-chip liquid cooling and immersion cooling are increasingly standard for dense AI racks.
- Cooling efficiency directly affects total energy use, often measured as power usage effectiveness (PUE).
- Some data center cooling consumes water, which can be a sensitive issue in water-stressed regions.
- Operators are investing in closed-loop and waterless cooling to reduce local impact.
Takeaway
Cooling is both an energy cost and, in some designs, a water cost; both shape where AI can be sited.