How Much Water Does Your AI Really Drink?

For years, the environmental cost of artificial intelligence has been a bit like a ghost in the machine. We knew it was there, but it was almost impossible to see or measure. We’ve all been so focused on the potential of AI that we haven’t spent much time thinking about its physical appetite for energy and resources.

Until now.

Last week, French AI company Mistral AI did something pretty remarkable. In a move that champions the transparency we desperately need, they published the first-ever comprehensive lifecycle analysis (LCA) for their Large Language Models.

They didn't just count the electricity used for training. They measured everything, from the manufacturing of the servers to the water consumed for cooling, giving us a proper, grown-up look at the real-world footprint of this technology.

And for HR and Talent Acquisition leaders, this isn't just a curious fact. It’s a new, critical dimension of AI governance.

The Numbers Are In (And They're Thirsty)

So, what did the study find? Let's put the data into perspective.

To train their Mistral Large 2 model, the total environmental cost included:

• Carbon Emissions: 20,400 tonnes of CO₂ equivalent.

• Water Consumption: A staggering 281,000 cubic metres of water. To make that number easier to grasp, that’s enough water to fill 112 Olympic-sized swimming pools.

The study also looked at the impact of a single query. A 400-token response from their AI assistant (a typical short answer) consumes about 45ml of water. It doesn't sound like much, but think about the millions of queries happening every single day across every organisation. It all adds up.

This data confirms what many suspected: these powerful models require a significant amount of resources, not just to build, but to run.

Why This Matters for HR: Your Vendor Checklist Just Got Longer

As leaders responsible for people, technology, and increasingly, workplace culture, this information lands squarely in our remit. It directly impacts three key areas:

1. AI Governance & Risk: Your responsibility for deploying AI safely and legally now has a green dimension. Choosing a vendor is no longer just about performance, cost, and data security. You are now, by extension, accountable for the environmental impact of the tools you bring into your organisation. This is a new pillar of "Responsible AI" that must be part of your governance playbook.

2. Recruitment Technology & Procurement: When you issue an RFP for a new AI-powered HR tool, are you asking about its environmental footprint? Mistral's report highlights a crucial learning: a model that is ten times bigger can have an impact that is an order of magnitude larger. Using an overly powerful model for a simple task is like using a sledgehammer to crack a nut - it's incredibly wasteful. Your vendor scorecard needs a new section for environmental efficiency.

3. AI Workforce Readiness: The study points out that user behaviour can make a significant difference. Inefficient prompting and using AI for the wrong tasks doesn't just lead to poor outputs; it actively wastes energy and water. This makes AI literacy more important than ever. It's not just about getting better results; it's about using these powerful tools responsibly.

Three Pragmatic Steps You Can Take Today

This is about making smarter, more defensible decisions. Here’s how you can start:

• Update Your Vendor Scorecard: Start asking potential AI vendors the tough questions. "Can you provide a Lifecycle Analysis (LCA) for your models?" or "What steps are you taking to improve the efficiency of your models?" Their answers (or lack thereof) will tell you a lot about their maturity and commitment to transparency.

• Champion Efficient AI Use: The need to upskill your workforce in how to use AI effectively now has a sustainability argument. Assessing AI competency is now an environmental issue. Building skills in your teams to write clear prompts and choose the right tool for the job reduces wasted computing cycles. (This is a core principle behind the skills measurement we do with genAssess).

• Integrate AI into your ESG Goals: Your organisation's Environmental, Social, and Governance (ESG) strategy needs to account for its technology stack. As an HR leader, you have a crucial voice in ensuring the company's AI strategy aligns with its public commitments to sustainability.

This study from Mistral AI is a welcome, and frankly overdue, step towards maturity for the entire AI industry. It moves the conversation from abstract hype to concrete accountability. For us, it’s a powerful reminder that our role is to empower the human side of this transformation, making decisions that are not only effective but also ethical and responsible.

While you’re here: Public AI Masterclasses

Something I don't normally do...

My three-hour AI workshops are usually reserved for private corporate teams. But after continuous requests, I'm opening up my calendar for a limited number of public sessions this August for the very first time.

These aren't one-hour overview webinars. They are comprehensive, capability-building sessions designed for individual HR and Talent Acquisition professionals. To ensure a high-quality, interactive experience, seats are strictly limited to just 20 per workshop.

Choose the track that's right for you:

Track 1: For Recruiters & TA Professionals: The AI-Powered Recruiter Workshop This is a practical deep dive into the "how". We'll move beyond basic prompting to build the skills you need to work faster, smarter, and more strategically. You will leave having mastered the PRIME framework in a hands-on session. (Dates: 5th & 26th August)

Track 2: For HR Directors & People Leaders: The AI Readiness Workshop This is a strategic session focused on the "why" and "what". We'll cover your role as the "Ethical Guardian", build "Guardrails" for your organisation, and develop a responsible AI roadmap. (Dates: 7th & 28th August)

If you're ready to move beyond the hype and build real, practical AI skills, this is your chance. Places are offered on an approval basis to ensure the right mix of professionals.

Question: Are you currently considering environmental impact when you evaluate new AI vendors for your HR or TA function?

Reply and let me know…

• Yes, it's already on our checklist.

• No, but this has made me think we should.

• We're not evaluating any AI vendors right now.

All the best,

Martyn

H.A.I.R. (AI in HR) 

Putting the AI in HR. Safely.