Circular economy is one of those phrases that gets used a lot in sustainability discussions and very little in procurement ones. That is a problem worth solving, because the circular economy in data centres has more practical commercial substance than its framing usually conveys.
In data centre terms, circular means three things. First, designing equipment so it can be repaired, upgraded and reused rather than retired whole. Second, extending the working life of equipment already deployed, through component-level upgrades and refurbishment rather than wholesale replacement. Third, recovering the materials at end of life when the equipment really has reached the end of its useful service.
Each of these has a measurable economic and environmental case behind it. The honest reason none of them is yet standard practice is that the buyer-side incentives have not historically pointed that way.
What changed
The case for circularity in data centres started shifting when the underlying assumption about hardware generations began to fail. For most of the last three decades, the working belief in IT procurement was that each generation of server hardware roughly doubled in energy efficiency. That made refresh to new the obvious lowest-carbon path, and made the embodied carbon of replacement hardware effectively rounding-error in the procurement calculation.
That assumption no longer holds. Measured benchmarks across recent server generations show much smaller gains, often inconsistent, and frequently outweighed by the embodied carbon cost of the new machine. When the carbon spent in manufacture is larger than the carbon saved in operation, the carbon arithmetic of refresh to new inverts.
Why this matters at supply chain level
The other reason circularity is climbing the agenda is upstream. The EU has identified 30 critical raw materials in short or politically unstable supply; 23 of them are used in data centre IT hardware. Only about 10% of those materials are currently recovered at end of life in the sector. The supply chain risk is increasingly recognised by procurement functions that previously treated raw materials as someone else's problem.
A circular approach to data centre IT (component-level upgrades where they make sense, refurbishment where it does, structured material recovery at end of life) addresses both the carbon and the supply-chain exposure at the same time. It is not a sustainability nice to have. It is procurement risk management.
What Interact contributes
We have spent years measuring the performance of the hardware that this argument depends on, and contributing to the research and policy work that turns the argument into adoptable practice. The dataset behind our tool sits underneath the cost and carbon comparisons that make the case for refurbished options defensible against a vendor-led new-build alternative.
The point of the research is not to be right in principle. It is to be defensible in front of the procurement committee that has to sign off the cheque.