They begin at recycling.
At disposal.
At carbon offsets.
At ESG reports.
But the real sustainability decision usually happened years earlier, quietly, during procurement, design, software selection, or an upgrade meeting nobody thought twice about.
That’s the conversation I keep coming back to lately.
Not because it’s fashionable. Because I think we’ve accidentally normalised a system where perfectly functional things are constantly thrown away, not because they failed, but because the ecosystem around them stopped supporting them.
And the more industries I look at, the more I realise this isn’t just an IT problem.
It’s everywhere.
We’ve been sold the idea that newer automatically means better.
New platform.
New version.
New model.
New rollout.
New subscription.
New dashboard.
New “ecosystem.”
It sounds like progress.
Sometimes it is.
But underneath that story is a cost almost nobody measures properly.
Every replacement carries embedded carbon from manufacturing, transport, mining, assembly, packaging, logistics, deployment, training, and disposal.
Every upgrade retrains staff.
Every software change reorganises workflows.
Every unsupported platform forces otherwise healthy equipment into landfill long before its engineering life is over.
And the strange thing is, I don’t think this is usually malicious.
Planned obsolescence isn’t always some evil boardroom conspiracy.
Most of the time it’s just short-term rational decisions stacking on top of each other for decades.
A cheaper coating instead of a permanent material treatment.
A sealed component instead of a serviceable one.
Software licensing replacing ownership.
Support cycles becoming shorter than hardware life.
Subscriptions replacing tools that once worked independently.
Every individual decision makes sense in isolation.
The aggregate effect is enormous.
Here’s the question I think more organisations should ask:
“If we had to keep this system operational for ten years with no replacements, how differently would we design it?”
That one question changes everything.
Because suddenly you stop thinking about quarterly upgrade cycles and start thinking about survivability.
You notice which components were always designed to fail.
You notice the consumables disguised as infrastructure.
You notice the vendor lock-in hidden behind words like “lifecycle management.”
You notice how many systems you don’t truly own anymore.
And once you see it, you can’t unsee it.
The four costs always appear eventually:
• Upgrade cycles
• Parts replacement
• Subscription creep
• Vendor-controlled support timelines
What fascinates me most is the hidden human cost underneath all this.
When businesses replace systems, they talk about replacing hardware.
But they’re actually replacing accumulated human understanding.
That’s the expensive part.
Switching a fleet management platform doesn’t just replace software.
It retrains dispatchers, supervisors, maintenance teams, operators, reporting processes, workflows, habits, terminology, and troubleshooting knowledge.
Moving from diesel to electric equipment isn’t just changing drivetrains.
It changes maintenance philosophies, safety systems, charging infrastructure, training pathways, ventilation assumptions, operational behaviour, and incident response models.
Autonomous haulage didn’t just change trucks.
It changed the entire mine around the trucks.
The surrounding human system has to reorganise itself every time technology changes.
And almost nobody costs that properly.
The irony is that many of the “solutions” we were sold have already become the next problem.
Control systems built on Windows XP were replaced.
Then Windows 7 systems were replaced.
Then newer cyber requirements forced another replacement cycle — even when the actual industrial control hardware was still perfectly functional.
The PLCs could survive decades.
The software stack couldn’t.
Same with communications infrastructure.
Leaky feeder.
Then Wi-Fi.
Then LTE.
Then 5G.
Every generation marketed as the final answer.
Every generation obsolete within years.
The original infrastructure often kept working.
It was the upgrade path itself that aged out.
That’s the part I think we need to discuss more honestly.
Because sustainability isn’t just emissions.
It’s durability.
It’s maintainability.
It’s repairability.
It’s interoperability.
It’s whether a thing can survive outside a vendor roadmap.
And this becomes deeply uncomfortable when you look at waste streams honestly.
Australia generates enormous amounts of industrial waste that most people never see.
• Off-road tyres
• Conveyor belts
• Hydraulic hoses
• Industrial electronics
• Control systems
• Sensors
• Battery systems
Most of it is buried, shredded, or discarded long before the value is truly exhausted.
Meanwhile we’re layering newer technologies on top of old waste problems we still haven’t solved.
Electric equipment is important.
But battery recycling infrastructure is still immature.
Autonomous systems are powerful.
But electronic waste recovery pathways are weak.
Smart systems generate incredible operational data.
But most companies don’t even own their own historical data properly because it lives inside proprietary vendor ecosystems.
That one worries me more than people realise.
Many organisations don’t understand they are effectively renting access to their own operational history.
Switch vendors and years of data can become unusable.
That’s not a technology issue.
That’s sovereignty.
And then there’s the skills cliff coming.
The generation who knew how to keep ageing systems alive through ingenuity, fabrication, adaptation, and practical knowledge are retiring.
A lot of younger technicians are exceptionally skilled — but they’ve often grown up inside vendor-supported ecosystems where replacement is assumed before repair.
That changes how industries think.
It changes procurement.
It changes maintenance culture.
It changes engineering philosophy itself.
Which brings me back to something I keep saying lately:
The most sustainable device is often the one already sitting in front of you.
The most reliable carbon saving is usually the replacement that never happened.
Because once embedded emissions exist, they already happened.
The steel was already smelted.
The battery already manufactured.
The transport already completed.
The mining already done.
Keeping a working asset alive longer isn’t just financially sensible.
It’s often environmentally sensible too.
And I think we’re heading toward a cultural split.
One side will continue living inside permanent upgrade cycles because that’s what modern business defaults to.
The other side will begin asking harder questions:
Can this be repaired?
Can this be remanufactured?
Can this operate independently?
Can this survive without constant replacement?
Can we design around longevity instead of churn?
Those organisations are going to think very differently.
And honestly, they’re the ones I’m most interested in watching.
Because designing for longevity changes everything.
Who you buy from.
How you train.
What you measure.
What you value.
Even how you define progress.