How AI data centres cut annual water consumption by 50%

2026-05-22 · Source: AI Magazine · Field: Technology & Digital — Cloud Computing & IT Infrastructure, Artificial Intelligence & Machine Learning · Depth: Intermediate, short

Summary

Schneider Electric asserts that AI data centers can cut annual water consumption by up to 50% by adopting advanced liquid cooling systems instead of traditional air chillers. Tuan Hoang, Head of Cooling Technology and Product Development, emphasizes that "zero water is needed to cool AI data centers" and that water consumption is a "geographical choice." Modelled scenarios for facilities in Dallas and Paris demonstrated significant reductions: Dallas saw a 48% drop from 382,000 to 197,000 cubic meters per year, while Paris achieved a 53% reduction from 108,000 to 51,000 cubic meters. The company's Uniflair XCA line, featuring air-cooled chillers with up to 2.4MW capacity and a closed-loop design, exemplifies a water-free heat rejection strategy. Rich Whitmore, CEO of Motivair by Schneider Electric, states that liquid cooling is now essential for AI racks approaching 400kW.

Key takeaway

For AI Architects and MLOps Engineers planning or upgrading data centers for high-density AI workloads, you should re-evaluate liquid cooling strategies. Adopting closed-loop liquid cooling systems, such as Schneider Electric's Uniflair XCA line, allows you to meet the thermal demands of AI racks approaching 400kW while significantly reducing or eliminating annual water consumption. This approach can improve site selection flexibility and enhance your facility's environmental sustainability profile.

Key insights

Liquid cooling for AI data centers can eliminate water consumption, addressing environmental concerns for high-density compute.

Principles

• Liquid cooling is essential for AI racks approaching 400kW.
• Data center water consumption is a design choice, not a liquid cooling requirement.
• Closed-loop liquid cooling systems prevent onsite water use.

Method

Transition from traditional air cooling to advanced liquid architectures, specifically employing closed-loop systems that radiate heat externally using factory-sealed cooling fluid.

In practice

• Deploy closed-loop liquid cooling systems like Uniflair XCA.
• Model water savings for liquid cooling transitions in specific geographies.

Topics

• AI Data Centers
• Liquid Cooling
• Water Conservation
• Schneider Electric
• Thermal Management
• Closed-Loop Systems

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Architect, MLOps Engineer, IT Professional

Related on AIssential

• How some data center operators are tackling their water use problems — Data center water use is a complex, localized issue with inherent trade-offs between water consumption and energy efficiency.
• AI Data Centers Are Wasting Heat Cooling Chips. I Built a System That Feeds a Greenhouse Instead. — Data center waste heat is a valuable, predictable resource for secondary revenue streams, not an unavoidable externality.
• Microsoft and Chevron plan one of the largest gas-powered data center projects in US — Microsoft and Chevron are building a massive gas-powered data center, highlighting the energy demands of AI despite sustainability pledges.
• The Data Centers Have Arrived at the Edge of the Arctic Circle — Former industrial sites are being repurposed for AI-driven data infrastructure at the Arctic Circle's edge.
• Emilio Tenuta: Advancing Sustainability Through Growth — Embedding sustainability into business strategy is crucial to decouple growth from resource consumption, particularly for data centers.
• Nvidia says its AI data center design runs hotter to use a lot less water — Nvidia's Rubin data center design achieves near-zero water usage by employing high-temperature liquid cooling and direct chip heat capture.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Magazine.