Among all the scrutiny surrounding AI's environmental impact, data centres frequently make headlines for their water consumption. A recent Guardian article highlighted concerns about the UK's largest proposed data centre in Northumberland, suggesting water use could reach 621 million litres annually when accounting for both onsite and offsite footprints.

That's a massive number. But what does it actually mean?

Putting 621 million litres in context

Here are the things you can do with 621 million litres of water

• Make 62,000 pairs of jeans
• Irrigate 25 standard UK golf courses
• Produce 1.9 million avocados

For Northumberland's 325,000 residents, that's enough for

• 1 new pair of jeans for 1 in every 5 people
• ~70% of all golf courses in the county
• 6 avocados per person yearly

Let's zoom out even further. This is how we use water globally:

• Agriculture: ~70% of all freshwater
• Industry: ~19%
• Municipal/Domestic use: ~11%
• Data centres: ~0.015%

Imagine if we put everything in a 10-litre bucket, it would look like this:

In terms of what the water is used for in data centres, electricity actually takes the biggest cut, accounting for 75% of the water footprints. For on-site usage, 80% is used for cooling.

Many others have covered the numbers in detail. If you want a deeper dive into data centre water consumption, Andy Masley's piece offers some insights.

What’s the point of talking about data centres’ water consumption?

Given data centres' negligible water footprint, is this conversation even meaningful?

Yes, there are three reasons.

First, it shines a spotlight on water risk. Compared to carbon footprints, we rarely talk about water. While carbon is the fundamental metric for climate change, most people and businesses will feel the impact via water.

Second, while their impact is minimal globally, local communities bear the real cost. Those 621 million litres equal the average yearly water use of more than 11,000 people. And water isn't always abundant in the UK. We need to move away from finger pointing to monitor and manage water resources more accurately and holistically.

Third, data centre water use is set to grow significantly as AI expands, making water expertise increasingly valuable. That 0.015% teardrop today could become much larger within a few years.

Job seekers’ prospects

Nonetheless, if you’re concerned about the water footprint of data centres, why not be part of the solution? Here are some pathways worth considering:

• Thermal Engineers - Design advanced cooling systems that prevent chips from overheating without relying on massive water towers.
• AI/ML Optimisation Engineers - Build digital twins of data centres that simulate and optimise cooling flows in real-time, reducing both energy and water consumption.
• Mechanical Technicians - Install, maintain, and repair sophisticated cooling systems. This role is particularly accessible for those with hands-on technical skills from other industries.
• Fluid Dynamics Engineers - Design liquid cooling flows and heat exchangers that maximize efficiency while minimizing water use.
• Water Stewardship Managers - Negotiate with local water utilities, develop watershed replenishment strategies, and ensure data centres balance their local impact.
• Water Infrastructure Manager - Oversee water systems and resource planning at scale, such as this opening at Amazon AWS.

These professionals can also tackle water issues in other industries. For example, AI/ML optimisation is a sought-after skill in precision agriculture and leak detection etc. Whether working on data centre cooling or precision agriculture, it’s solving the same fundamental problem: how to deliver the right amount of resources to the right place at the right time, using real-time data. Thermal and fluid dynamics engineers can move into industrial cooling for steel production, chemical processing, or food manufacturing. Water stewardship is fundamentally a multi-disciplinary position.

Final random bits

So, are data centres the "thirsty beasts" the headlines suggest? Not in the global scheme of things. But they are a massive engineering challenge. And for the right person, a career-defining opportunity.

Meanwhile, Northumberland is facing a paradox. The region is steeped in an industrial heritage of thermal systems, heavy machinery, and fluid dynamics - exactly the skills needed to solve the cooling puzzle. Yet, as the local demographic shifts, the talent crisis is looming. The proposed data centre and other water-heavy businesses may struggle to fill specialised roles locally. Or could it be the catalyst for Northumberland to reinvent itself?1

Footnotes:

1. Calculations are based on a water footprint of 10,000L/pair of jeans. Source.
2. Calculations based on data from the Golf Course 2030 Water Solutions report.
3. Calculations are based on a water footprint of 2,000L per kg of avocados.. Source.
4. Calculation based on data from Mytton, D. Data centre water consumption. npj Clean Water 4, 11 (2021).