How Far Have We Come in Sustainable Computing?
Sustainable technologies rarely emerge fully formed. More often, progress comes through a succession of improvements that make systems more efficient over time.
The first electric light bulbs transformed daily life, but by modern standards they were remarkably inefficient. Roughly 90% of the electricity powering those early filament bulbs was released as heat rather than visible light.
Over time, engineers introduced better materials, new lighting technologies and more efficient designs. Today, an LED bulb can produce the same level of illumination using only a fraction of the power.
Computing has followed a similar path
Digital technology consumes significant resources, but modern IT is far more efficient than it used to be. That progress is worth examining.
A Decade of Remarkable Efficiency Gains
According to the International Energy Agency, the growth of internet traffic and digital services during the 2010s was not matched by an equivalent increase in the energy required to support them.
Internet traffic in 2020 was approximately twenty-five times higher than it was in 2010.
The number of internet users worldwide more than doubled over the same period.
Data centres and transmission networks supported this rapid expansion without increasing their energy demand at the same rate.
At first glance, that seems counterintuitive. More traffic, more applications, more data and more connected devices should have translated into proportionally greater resource consumption.
Yet that is not what happened. While demand for digital services continued to increase, engineers, infrastructure teams and technology providers consistently found ways to make computing more efficient.
The result was a digital ecosystem capable of supporting vastly more activity than its predecessors while using resources more effectively.
Several developments worked together
The progress was not produced by one single innovation. It resulted from improvements across infrastructure, software, operations and electricity generation.
Data-centre design and cooling
Virtualization and workload consolidation
Cloud and hyperscale infrastructure
Better visibility into resource consumption
Cleaner sources of electricity
Data centres became more energy efficient
One of the most visible areas of progress was the data centre itself. Over that period, operators improved the way computing facilities were designed, cooled and managed.
Advances in server hardware enabled more computing power per watt, while improvements in cooling systems reduced the amount of energy required to keep equipment operating safely.
The industry also moved toward cloud and hyperscale data centres. Optimizations in facility design and operations helped reduce cooling and power-distribution overheads, allowing a greater proportion of electricity to be devoted to computing itself.
These developments meant that a greater share of the electricity entering a data centre could be used for computing rather than for supporting the facility around it.
Virtualization, cloud computing, and higher server utilization
For many years, organizations purchased physical servers for specific applications, often leaving much of their computing capacity unused. A server might spend large portions of its life operating far below its maximum capacity while still consuming power and requiring cooling.
Virtualization helped change that by allowing multiple workloads to run on the same physical server. Cloud computing accelerated the trend by pooling computing resources across large infrastructures and allocating capacity where it was needed most.
As a result, organizations were able to do more with fewer physical machines, increasing server utilization and making better use of existing resources.
These changes reduced waste, improved efficiency and supported growing digital demand without requiring a proportional increase in hardware. In many cases, progress came not from building more infrastructure, but from using existing infrastructure more effectively.
Measuring and eliminating digital waste
Not all efficiency gains came from larger infrastructure improvements. Some came from a better understanding of how digital services consumed resources.
As monitoring and observability practices matured, organizations gained new ways to identify inefficient processes, unnecessary data transfers, poorly optimized applications and other forms of digital waste.
Greater visibility made it easier to target improvements. Rather than adding infrastructure to solve every performance issue, teams could often achieve better results by optimizing existing systems, reducing complexity and eliminating inefficient behaviours.
This shift from estimation to measurement helped make sustainability a practical engineering concern rather than a purely theoretical objective.
Understand resource consumption
Collect reliable data about applications, infrastructure, transactions and digital services.
Locate areas of inefficiency
Detect waste, unnecessary complexity and behaviours that consume resources without delivering proportional value.
Track the effects of improvements
Assess whether optimization efforts reduce resource use while preserving performance and user experience.
Understand the environmental impact of your digital services
Modern observability and digital experience tools can help identify inefficient processes, quantify resource usage and track the effects of optimization efforts over time. Discover how Ekara Green supports a greener application portfolio.
Renewable energy and cleaner digital infrastructure
Efficiency was not the only area where progress occurred. Many technology providers and data-centre operators also invested in cleaner sources of electricity and more sustainable operating practices.
Renewable energy procurement became an increasingly important part of digital infrastructure strategies, helping to reduce the carbon intensity of the electricity used to power digital services.
These efforts did not eliminate the environmental impact of computing, but they complemented the efficiency gains achieved elsewhere.
Today’s chapter in sustainable IT: the AI challenge
The efficiency gains and growing use of renewable energy achieved over previous decades did not eliminate environmental concerns. Instead, they created a more sustainable foundation for digital growth. Artificial intelligence now presents the latest test of that model.
Electricity demand from data centres increased by 17% in 2025 alone, significantly outpacing overall growth in global electricity demand.
Global data-centre electricity consumption could double by 2030, with AI-focused facilities expected to experience even faster growth in power demand.
Recent studies show that data-centre electricity demand is rising rapidly as AI workloads expand. Industry research paints a similar picture, with increasingly powerful systems requiring more computing capacity, electricity and supporting infrastructure.
Yet it would be wrong to view this as a reversal of the progress already achieved. The same industry that improved data-centre efficiency, increased server utilization, optimized cooling and developed new ways to identify waste is now applying those lessons to a new generation of systems.
AI itself is increasingly being used to improve infrastructure operations, energy management and cooling efficiency. Researchers, infrastructure providers and technology vendors are also exploring innovations ranging from more efficient hardware and model architectures to new approaches for power generation and energy procurement.
Artificial intelligence has not ended the story of sustainable computing. It has opened a new chapter.
The question is no longer whether efficiency improvements matter; they clearly do. The challenge is now to apply those improvements quickly enough to keep pace with the current wave of digital and AI-driven demand.
How can computing efficiency improve quickly enough to keep pace with today’s wave of demand?