AI, Sustainability and the New Data Centre Lifecycle: Turning Decommissioning into Value

As artificial intelligence accelerates global demand for computing power, the data center industry is undergoing its most profound transformation in decades. Facilities built for conventional enterprise workloads are being reconfigured or replaced to support high-density AI infrastructure, and with that shift comes a critical question: what happens to the hardware being retired?

For many operators, decommissioning projects are strategic rather than operational. The servers, switches and storage arrays leaving service still hold value, not only in their residual market price, but in the materials, energy, and embedded carbon they contain. Treating those assets responsibly is becoming both a commercial advantage and an environmental obligation.

The AI Upgrade Cycle and Its Consequences

The rise of AI and high-performance computing (HPC) has dramatically shortened refresh cycles. Equipment that once ran for six to eight years is now being replaced in three or four. According to DataCenter Dynamics, hyperscale operators will double GPU-capable capacity by 2026, with hyperscale capacity trupling by 2030,, driving one of the largest hardware turnovers in history.

This creates unprecedented challenges for lifecycle management. Older data-center servers - often powered by DDR4 memory and previous-generation Xeon or EPYC processors - are being phased out rapidly. Yet many of these systems remain serviceable for secondary markets, cloud resellers, or enterprise reuse programmes. Without structured asset disposition, that potential is lost, and unnecessary waste is created.

From Disposal to Circular Strategy

Forward-thinking operators are responding by embedding IT Asset Disposition (ITAD) into their sustainability strategies. Instead of treating decommissioning as a waste-management task, they view it as a circular process: recover, refurbish, resell, or recycle.

Each stage adds measurable value:

- Recover ensures every asset is securely collected, inventoried, and assessed for potential reuse or resale.

- Refurbish restores viable equipment to working condition, extending its life and reducing the need for new manufacturing.

- Resale generates direct financial return from retired assets and strengthens secondary-market availability.

- Recycle captures critical materials such as copper, aluminium, and rare earths, reducing the need for virgin extraction and landfill.

These practices also strengthen sustainability performance. Circular models directly reduce Scope 3 emissions, the indirect carbon footprint associated with supply chains, which now account for more than 70% of indirect emissions in most data-center operators’ reporting.

Manda’s Perspective

At Manda, we see the data-centre transition to AI workloads as an opportunity to redefine lifecycle value. Every decommissioned facility, rack or server is part of a broader sustainability equation. Our role is to help partners recover maximum financial and environmental benefit while maintaining the highest levels of compliance and data security. Whether a client is consolidating a regional site, migrating to cloud, or refreshing for GPU infrastructure, our process ensures no asset leaves value behind.

In an era where performance and responsibility must coexist, circular lifecycle management is fast becoming the new standard for the data-centre industry. The message is simple: the end of one generation of hardware marks the beginning of another, and with the right strategy, decommissioning can drive both sustainability and profitability.