December 26, 2025

Why We Built Data Palaces on Hot Sand

We built castles in the cloud. On sand. In the desert.

Remember when everyone lost their minds building offices with glass walls and beanbags, only to realise noise carries and no one likes agile stand-ups next to the ping pong table? That’s kind of what's happened with data centres.

A recent study has dropped a spicy truth bomb: most of the world’s data centres are built in completely the wrong climates. I’m talking about concrete boxes packed with humming servers… dropped right into hot, humid, high-cost regions.

We’ve unintentionally designed a cloud empire that drinks power, sweats coolant, and buckles under peak loads — especially now AI and cloud computing are turning every millisecond into a monetisable asset.

So let’s get into it: how we ended up here, what it means for tech leaders juggling growth and green commitments, and how smarter infrastructure thinking is about to become a recruiting edge too.

The heat is (literally) on

Let’s address the obvious: servers hate heat. They generate tons of it already. Combine that with a hot, humid external climate and you suddenly need massive cooling infrastructure, round-the-clock energy surges, and backup power redundancy like you’re prepping for the sequel to Mad Max.

Yet, per the TechRadar report, only 13% of global data centres are located in what would be considered suitable temperature and humidity zones. The rest? Sauna Season 365.

And this isn’t some abstract infrastructure quirk. As cloud workloads skyrocket, generative AI demands GPU-heavy server farms, and consumers expect everything-now digital access—cooling inefficiency becomes a business threat.

Here’s why that should spook you:

• Cost spike: Cooling accounts for 30%-50% of a data centre’s power bill. That’s not peanuts—it’s roasted cashews gold-plated in liquid helium.
• Reliability risk: Overheated centres throttle performance or force emergency failovers. Good luck explaining that outage to your investors.
• ESG exposure: For companies claiming “green” creds, running hot servers in hotter countries looks downright medieval.

Why did we build them there in the first place?

This is where it gets fun — or painful, depending on how many data centre contracts you’ve signed in southern England.

The placement of data centres has, until recently, been driven by things like proximity to corporate HQs, fibre/backbone access, cheap labour, or power prices—not climate suitability.

In the 2000s and 2010s, that made sense. Enterprise IT was centralised, lift-and-shift was hot, and nobody expected AI to hoover bandwidth and GPU like a Dyson on creatine.

But now? The software got smart, the workloads got heavier, and the infrastructure stayed stubborn.

The result?

Scale-up CTOs are trying to optimise for real-time collaboration, AI model hosting, and data sovereignty, while legacy infrastructure still acts like it was designed during the dial-up days. Something’s gotta give.

So what happens next? A cold (and smart) awakening

The good news: people are finally waking up. We’re seeing Nordic countries, Canada, Scotland — cold, renewable-rich regions — becoming magnets for new data builds.

In fact, some Governments are throwing tax breaks at “eco-efficient” data centres like confetti at a fintech launch party.

Hyperscalers like Microsoft and Google are experimenting with sub-sea centres (literally sunk into the ocean) or repurposing old mines. Why? Because they don’t want to be the ones trending on DownDetector during the World Cup Final.

And smart COOs and CTOs are asking new questions like:

• Can we optimise latency by distributing workloads across multiple cooler-region data nodes?
• Are we factoring in total energy lifecycle — not just access cost — when we pick providers?
• Does our infrastructure help or hinder our ESG hiring narrative?

Because here’s the kicker: infrastructure strategy now impacts talent strategy.

Cooling heads will win the hiring race

Tech talent is resourceful, values-led, and not interested in working on infrastructure set to explode every summer.

I’ve seen star candidates ask about cloud regions during interviews. “Are you building on green infra, or do I have to defend a diesel backup generator in Arizona?” It’s partly ESG concern, partly “I don’t want my work crashing at 37°C.”

For tech companies scaling fast — especially in fintech or greentech — the choices you make now about infrastructure could define your reputation tomorrow.

You can’t out-hire a data centre disaster.

If you want talent who care, you need to show you care about where the work happens — not just who does it.

• Rethink vendors: Don’t just chase uptime. Prioritise sustainable regions
• Add infra to EVP: Build green infra and shout about it. Techies notice
• Collaborate early: Get IT, People and Ops at the same table — location matters

Final thought — time to move the castle

We’re entering an age where data infrastructure is not just backend plumbing — it’s brand, it’s carbon footprint, it’s employer value proposition.

The smart companies are already routing workloads away from sweaty server farms and into smarter, cooler homes.

If you’re still clinging to data castles on hot sand, look around. That cloud bubble? It’s heating up fast. And not all towers will survive the melt.

Thinking about your infrastructure stack — or your hiring strategy — and want to chat to someone who’s seen it all go sideways (and right)? Give me a nudge. Always happy to talk warm servers and cool minds.

Back to news