The AI race is no longer happening in apps. It’s happening in infrastructure.

Over the next three years, the world is expected to build one of the largest expansions of digital infrastructure in human history - hyperscale AI data centres designed to power everything from autonomous systems to scientific discovery.

And behind every AI breakthrough sits something surprisingly physical - Water.

The scale of AI is forcing the industry to rethink how water infrastructure should be engineered from the ground up.

The Hidden Layer of the AI Boom

Large AI models run on enormous clusters of GPUs packed inside hyperscale data centres operating 24/7. These facilities generate extraordinary amounts of heat, especially as compute density rises with next-generation AI hardware.

Cooling becomes mission-critical.

Historically, many hyperscale facilities relied on evaporative cooling systems - an efficient method, but one that consumes large amounts of water over time. Some estimates suggest a single hyperscale data centre can use up to 1.5 million litres of water per day depending on climate and workload.

How do we scale intelligence without scaling resource consumption in the same way?

Closed-Loop Systems Will Be the New Standard

The old model of “consume and replace” is being redesigned into something far more intelligent: closed-loop cooling systems.

Instead of continuously evaporating water, closed-loop systems recycle and recirculate coolant internally. Once the system is filled, the same water continues cycling through the infrastructure with minimal loss.

This matters because hyperscale AI facilities are becoming denser, hotter, and more powerful every year. Traditional cooling approaches simply won’t scale efficiently forever.

Microsoft recently announced that all new data centres designed from August 2024 onward would implement zero-water evaporation cooling systems using closed-loop liquid cooling technology. The company estimates this could save more than 125 million litres of water annually per data centre.

Why Water Efficiency Will Become a Competitive Advantage

Water-efficient infrastructure isn’t simply about regulation or environmental pressure. It’s becoming a strategic advantage.

Why?

Because the AI economy will ultimately reward efficiency at scale.

• lower long-term operational risk,

• easier approval in water-sensitive regions,

• improved community relationships,

• reduced infrastructure strain,

• and future-proofed against rising resource constraints.

In other words:

The most sustainable systems may also become the most economically resilient systems.

This is especially important as AI workloads continue scaling exponentially. Researchers estimate global data centre electricity demand could more than double by 2030, with AI workloads responsible for a disproportionate share of the growth.

Cooling efficiency is no longer a side conversation.

It’s core infrastructure strategy.

Companies like TerraX are building toward this future now - developing closed-loop sustainable water networks designed for the next generation of AI and hyperscale infrastructure.