The Hidden Energy of Building and Running Modern Computing

The Two Faces of Energy

Modern computing consumes energy in two main ways: the embodied energy that goes into manufacturing the hardware, and the operational energy that powers the devices and data centres while they are in use. These two components together determine the total carbon footprint of any digital service.

Embodied Energy in Hardware

Embodied energy refers to the sum of all energy inputs required to extract raw materials, manufacture components, assemble devices, and transport them to the point of use. According to the IEA’s Energy and Compute Test report, the manufacturing phase can account for up to 30% of a server’s lifetime energy use. This figure includes mining of rare‑earth elements, fabrication of silicon wafers, and the assembly of complex circuitry.

The life‑cycle assessment of a typical laptop shows that the energy spent on producing its processor, memory, and display can be comparable to the energy it will consume during its five‑year usage period. This insight is echoed by studies from the U.S. National Science Foundation, which highlight that supply‑chain emissions often dominate the early life stages of high‑performance computing equipment.

Operational Energy in Data Centres

Once hardware is deployed, the bulk of energy consumption shifts to the data centre environment. Cooling, networking, and power‑delivery systems can add 20–40% to the total energy draw of a single server. NASA’s infrastructure, which supports large‑scale scientific simulations, relies on advanced cooling techniques such as liquid immersion to reduce operational energy demands.

Data centres are increasingly designed around renewable electricity sources, but the overall efficiency—measured by the Power Usage Effectiveness (PUE) metric—varies widely. A PUE of 1.2 indicates that 20% of the energy is used for cooling and other ancillary services, while a PUE of 1.5 means 50% of the energy is not directly powering compute workloads.

The Supply Chain Footprint

Beyond the hardware itself, the supply chain contributes a significant portion of the embodied energy. Mining operations, especially in remote regions, can be highly energy‑intensive, and the transportation of raw materials across continents adds to the carbon budget. The NSF’s research on supply‑chain transparency suggests that companies that map their material flows can identify high‑impact stages and target them for improvement.

Mitigation Strategies

Reducing the total energy footprint involves both design‑time and operational measures:

1. Material substitution – Using recycled aluminium or biodegradable polymers for casings lowers embodied energy.
2. Modular hardware – Components that can be upgraded individually reduce the need to replace entire systems.
3. Efficient cooling – Deploying free‑air or liquid‑cooling techniques can lower the operational energy share.
4. Renewable integration – Aligning data‑centre power use with on‑site solar or wind generation cuts grid reliance.
5. Lifecycle management – Implementing take‑back and refurbishment programs ensures that hardware is reused rather than discarded.

Looking Ahead

The trajectory of digital growth means that both embodied and operational energy will continue to rise unless systemic changes are made. Emerging research from the IEA, NSF, and NASA points to a future where edge computing, more efficient silicon architectures, and circular supply chains converge to create a sustainable digital ecosystem. Understanding the distinct energy pathways of modern computing is the first step toward designing technologies that are both powerful and planet‑friendly.