I’m no data centre engineer.
I’m not a hardware architect.
I did recently complete a diploma in business sustainability — which probably makes me a slightly dangerous amateur.
But there’s a question that increasingly gnaws at my ageing leg.
Not:
“Should we refresh?”
But:
When does the total lifecycle delta justify the embodied carbon and capital reset?
Because that feels like the grown-up sustainability question.
The Comfortable Narrative
We’ve absorbed a simple storyline:
Old = inefficient
Legacy = waste
Evergreen = responsible
Stay current.
Stay supported.
Stay modern.
It sounds virtuous.
But infrastructure doesn’t respond to slogans.
The Mainframe Reality
Let’s talk about mainframe.
Not as mythology. Not as nostalgia.
Just as engineering.
Many large estates — especially modern enterprise mainframes — are:
• Highly utilised
• Densely engineered
• Extremely resilient
• Secure by design
• Efficient per transaction
They run the boring, critical stuff:
Payments.
Pensions.
Insurance claims.
Tax systems.
They don’t trend.
They just work.
If you’re running:
• An efficiently governed estate
• With limited MIPS growth
• With some workloads gradually peeling off to cloud
• And no regulatory or security cliff looming
Then what exactly are you solving by refreshing every three years instead of five?
The Carbon Question Nobody Enjoys
New hardware is not carbon neutral.
There’s embodied carbon in:
• Raw materials
• Manufacturing
• Supply chains
• Logistics
• Decommissioning
Yes — newer generations improve performance per watt.
Yes — consolidation improves density.
Yes — encryption and security capabilities advance.
But the question is not whether improvement exists.
The question is whether the improvement is material enough to justify:
• The embodied carbon
• The capital reset
• The operational risk of change
If the estate is stable and efficient, shortening the refresh cycle may simply bring forward emissions rather than reduce them.
Neatness is not sustainability.
The Subscription Layer We Don’t Talk About
There’s another factor creeping into this conversation.
Across enterprise software, support models increasingly favour:
• Staying on current versions
• Shorter support windows
• Subscription-based licensing
• Commercial incentives aligned to “latest release”
That isn’t sinister. It’s just the direction of travel in software economics.
But it does subtly influence refresh decisions.
If software support, pricing models, or licensing terms increasingly nudge organisations toward earlier upgrades, then the refresh cycle can become commercially driven rather than sustainability driven.
Which raises a fair question:
Are we refreshing because the infrastructure delta is material — or because the commercial model makes delay uncomfortable?
That distinction matters.
Cloud Makes It Even More Interesting
If applications are moving to cloud:
• MIPS growth may flatten
• Capacity pressure may reduce
• Consolidation opportunities may increase
Refreshing too early risks designing for yesterday’s demand.
A refresh aligned to a five-year horizon — after consolidation, after cloud migration stabilises — may actually reduce total lifecycle impact.
But that requires discipline.
And a bit of courage.
This Isn’t Anti-Mainframe. Quite the Opposite.
For clarity: I think modern mainframe technology is brilliant engineering.
I spent 15 years at IBM, mostly as an account manager and services deal lead. I wasn’t building processors — but I saw from near and far how seriously these systems are designed.
They are not museum pieces.
They are industrial-grade infrastructure.
This isn’t a swipe at vendors.
It’s a plea for better questions.
The Grown-Up Question
Instead of asking:
“What’s our standard refresh cycle?”
Maybe we should be asking:
“At what point does the lifecycle delta justify the embodied carbon and capital reset?”
If the efficiency leap is structural — refresh.
If the security uplift is transformative — refresh.
If consolidation materially reduces footprint — refresh.
But if the delta is marginal?
Sweat the asset.
Run it well.
Govern it tightly.
That might be the greener move.
A Slightly Dangerous Amateur’s View
As someone with:
• 35 years in enterprise services
• A diploma in business sustainability
• And just enough knowledge to be mildly troublesome
I’ve seen organisations spend fortunes modernising things that weren’t actually broken.
Sustainability is not about chasing “new”.
It’s about measuring impact honestly.
And sometimes the most responsible decision isn’t the one that looks freshest in the board pack.
It’s the one that quietly asks:
Are we refreshing because it truly reduces lifecycle impact —
or because it feels safer to keep up?