Why Data Centres Are Moving Closer to Power

Why Data Centres Are Moving Closer to Power (And What That Signals About AI’s Future)

After more than 20 years delivering infrastructure at scale, I can tell you this:

When location strategy changes

It’s usually not a technology decision.

It’s a constraint.

And right now, data centres are starting to move closer to power.

That’s not a trend.

That’s a signal.

What People Think

Most people assume data centres are built based on:

·         Connectivity

·         Market demand

·         Land availability

And historically, that’s been true.

You build close to users, close to networks, close to major hubs.

But that model is starting to shift.

What’s Actually Happening

AI has fundamentally changed the requirements of data centres.

We’re now seeing:

·         Higher-density compute environments

·         Continuous, non-stop workloads

·         Massive increases in power consumption

(And just to be clear—when I say “compute,” I mean processing power at scale: thousands of specialised processors running continuously to handle AI workloads.)

The result?

Power is no longer just a requirement.

It’s becoming the limiting factor.

Where It Breaks

This is where infrastructure reality takes over.

On paper, if you need more power

You upgrade the grid.

In reality:

·         Grid capacity is already constrained

·         Upgrades take years, not months

·         Regulatory approvals slow everything down

·         Transmission infrastructure is complex and expensive

I’ve seen similar patterns in telecommunications projects.

Demand increases.

Plans are made.

But the underlying infrastructure can’t keep up at the same speed.

The Shift: Follow the Power, Not the Market

What we’re now seeing is a reversal of traditional thinking.

Instead of:

Build where demand is

We’re moving toward:

Build where power is available

That means:

·         Regional or remote locations

·         Proximity to generation sources

·         Co-location with renewable energy projects

Because if you can’t get power..

You can’t operate.

Operator Insight: This Isn’t Optional

From an operator’s perspective, this isn’t a strategic preference.

It’s a constraint driven decision.

You don’t choose to move.

You’re forced to move.

Because:

·         Demand is accelerating

·         Infrastructure timelines are fixed

·         And systems can’t wait

I’ve worked on programs where the plan looked solid.

Until it hits a physical limitation or other factors unknown at time of planning.

Example: many are considering the price of fuel and availability a concerning factor outside of planning controls.

That’s exactly what’s happening here.

Moving closer to power creates new challenges:

1. Connectivity Trade-Offs

You may be further from fibre backbones or users.

Which increases:

·         Latency

·         Network complexity

·         Cost of interconnection

·         Capacity increases on existing infrastructure not historically built to handle large traffic.

2. Workforce Limitations

Remote locations often lack:

·         Skilled labour

·         Established delivery teams

·         Operational support structures

3. Infrastructure Coordination

You now need to align:

·         Energy

·         Network

·         Land

·         Logistics

All at once.

And alignment at scale is where most projects struggle.

The 5 Constraints Showing Up in Real Time

This is a clear example of the 5 Constraints of AI Infrastructure Delivery:

·         Power - driving location decisions

·         Fibre – major critical infrastructure projects in build stages from all carriers and network providers.

·         Land - becoming strategic

·         Workforce - harder to access

·         Capital - increasing due to complexity

These aren’t theoretical.

They’re already shaping decisions.

What This Means for Industry

We’re entering a phase where:

·         Energy access becomes a competitive advantage

·         Infrastructure planning becomes more integrated

·         Location strategy becomes constraint driven

And importantly:

Infrastructure teams are no longer support functions.

They’re becoming central to decision making.

What Happens Next

Expect to see:

·         More data centres built near energy hubs

·         Increased investment in transmission and interconnection

·         Closer alignment between energy providers and tech companies

And potentially:

A shift in global AI leadership based on infrastructure capability not just technology.

Final Thought

In infrastructure, location is never random.

It’s always driven by constraints.

Right now, the constraint is power.

And when that happens.

Everything else starts to move around it.

References & Further Reading

Footnote
This article is part of a series exploring the physical infrastructure behind the AI economy.

Disclaimer
The views expressed in this article are my own and are intended for general information and discussion purposes only. They do not represent the views of any employer, organisation, or client.

© 2026 Rodney Terry – Digital Backbone. All rights reserved.