Recirc IT's Server Offering
Refurbished Servers at Half the (CO2) Cost of new
Gone are the days when you had to upgrade every two years to increase your performance output. Now you can get the same output without having to change what you have.
Your 5-year-old servers aren’t the problem. Your refresh cycle is.
Many IT departments still operate on the assumption that servers need to be replaced every few years to keep up with performance. But here’s the truth: it’s not the age of the servers—it’s the way we treat them.
In reality, 5-year-old servers are still perfectly capable of handling modern workloads, especially when optimized through upgrades like additional memory, processor swaps, or proper reconfigurations. The real issue? Constantly refreshing hardware when it’s not necessary.
Consider this:
1. Old servers waste more energy when they’re left to run inefficiently, but with simple tweaks, they can perform just as well as newer models—without the carbon footprint of manufacturing new hardware.
2. Your refresh cycle creates unnecessary CO2 emissions:
3. Manufacturing new servers contributes to around 1333 kg of CO2 per unit. Every time you refresh hardware without a real performance need, you’re contributing to the problem.
4. Sustainability isn’t just about what’s new: It’s about maximizing the useful life of your existing assets. By keeping servers longer and optimizing them, you’re not just saving money—you’re cutting your emissions too.
It’s time to shift from a mindset of constant refresh to one of server longevity and optimization. Your 5-year-old servers could still be doing the job—they just need a little care.
How long has it been since you questioned your refresh cycle? Let’s talk.
What Are We Actually Talking About Here?
Moore's Law is Broken
Here’s the Proof:
For decades, the IT industry relied on Moore’s Law—the belief that processing power would double every two years while efficiency improved. This justified the relentless push for server refresh cycles every 3-5 years.
But the numbers don’t lie: Moore’s Law is no longer delivering the efficiency gains it once did. Here’s why:
1️⃣ Performance-per-Watt Has Stagnated
Historically, smaller transistors meant lower power consumption and higher efficiency. But since 14nm lithography, those efficiency improvements have flatlined.
Idle power consumption is rising. To maintain performance gains, newer processors use more cores, which increases baseline energy consumption—even when servers are underutilized.
Most servers run at only ~25% utilization, meaning much of this extra power is wasted.
2️⃣ Upgrading Isn’t Always Worth It
The dynamic range of servers (difference between full-load and idle energy consumption) is shrinking, making newer servers less energy-efficient in real-world scenarios.
Replacing a 3-year-old server with a new one often results in only marginal efficiency gains (under 15%)—not enough to justify the cost or environmental impact.
Optimizing older servers (through memory upgrades, CPU swaps, or configuration adjustments) can achieve comparable efficiency to new ones—without the embodied carbon cost of manufacturing a new unit.
3️⃣ The Real CO₂ Cost of New Servers
Manufacturing a single rack server generates 1,333 kg of CO2e—before it even powers on.
40% of data center servers are over 5 years old but consume 66% of total energy while providing only 7% of compute power—a clear sign of inefficiency. But replacing them with new servers isn't always the answer.
Refurbishing or remanufacturing hardware cuts costs and emissions while extending hardware lifespan in line with circular economy best practices.
4️⃣ Payback Periods Don’t Lie
The financial case for refreshing servers younger than 5 years is weak—the return on investment (ROI) for energy savings is often longer than the expected server lifespan.
In contrast, extending the lifecycle of existing hardware through refurbishment or targeted upgrades delivers efficiency gains with a faster payback period.
Bottom Line: Newer doesn’t always mean better.
Moore’s Law is breaking down, and the old refresh cycle model is failing IT leaders.
Instead of defaulting to new, rethink how you manage your IT estate.
A Real Life Case Study
Project | Financial – Decommissioning
(2000+ Servers)
Servers for analysis were due to be decommissioned and replaced over the next 12-18 months
• These could be replaced with Servers for analysis were due to be decommissioned and replaced over the next 12-18 months
• These could be replaced with 406 PowerEdge R7625 Servers at
the same average utilisation of 40%
• This would lead to a total estate reduction of 78% in Energy and
Carbon or saving of over 38 million kWh and 8 thousand Metric Tonnes of C02
• Saving, at today's energy prices, Servers for analysis were due to be decommissioned and replaced over the next 12-18 months
• These could be replaced with 406 PowerEdge R7625 Servers at
the same average utilisation of 40%
• This would lead to a total estate reduction of 78% in Energy and
Carbon or saving of over 38 million kWh and 8 thousand Metric Tonnes of Co2e
• Saving, at today's energy prices, £6.1 Million pounds
• Decommissioning strategy required to optimise benefit by removing lowest performing first and identify hosts with potential re-use within larger estate
• Decommissioning strategy required to optimise benefit by removing lowest performing first and identify hosts with potential re-use within larger estate PowerEdge R7625 Servers at
the same average utilisation of 40%
• This would lead to a total estate reduction of 78% in Energy and
Carbon or saving of over 38 million kWh and 8 thousand Metric
Tonnes of Co2e
• Saving, at today's energy prices, £6.1 Million pounds
• Decommissioning strategy required to optimise benefit by removing lowest performing first and identify hosts with potential re-use within larger estate
Get in touch
Send us a message and we'll get back to you as soon as we can.