By Professor Aoife Foley, IEEE senior member and Chair in Net Zero Infrastructure, The University of Manchester who considers the challenges of sustainability while also encouraging innovation.

As the UK positions itself as a global tech superpower - including new commitments to expand the nation's data centre footprint - it can be easy to overlook the broader environmental consequences of this strategy.

According to the International Energy Agency (IEA), electricity demand from data centres is projected to more than double by 2030 to around 945 terawatt-hours (TWh), slightly more than the entire electricity consumption of Japan. These numbers highlight the urgent need to balance digital growth with sustainable energy practices.

The challenges of aligning sustainability and tech

Much of this surge in demand is being driven by AI workloads, which consume significantly more energy than traditional cloud computing tasks. While hyperscale operators are investing in renewable energy to counteract the effects, this alone is not enough.

However, energy demand from digital infrastructure goes beyond silicon and servers. We must also consider software, structure and the way businesses produce, store and process data.

The dark data epidemic

Modern enterprises continuously generate and store vast amounts of data, ranging from routine system activity to machine, sensor and user-generated input. Much of this data is unstructured, redundant or never accessed again. Commonly referred to as 'dark data', it's becoming increasingly prevalent and contributes to growing digital waste.

The proliferation of dark data not only reinforces the flawed belief that more data is always better, but also leads to significant energy waste. Storing and managing this idle data consumes considerable power and resources, without delivering meaningful value.

As such, infrastructure and operations leaders need to find meaningful ways to address the environmental impact of unnecessary data storage and prioritise renewable energy use. Some progress has been made through energy-efficient, hyperscale data centres, which can reduce electricity demand and emissions. But locally, these facilities place huge pressure on electricity grids and complicate the energy transition - particularly in smaller countries.

What can companies do?

While cooling technologies such as liquid immersion and direct-to-chip cooling can improve energy efficiency, they only address the symptoms and not the underlying inefficiencies in model design. Furthermore, with an increasing dependence on water for cooling, this adds strain to local resources and is often avoidable thanks to renewables.

We need to ask more relevant questions. Are organisations choosing the best model architectures for the task? Are they designing for efficiency or just scale? Are they optimising for performance or simply defaulting to brute force?

While this year's World Environment Day focused on ending plastic pollution, it also served as a reminder to address less visible forms of waste too.

Now is the time for leaders to strengthen data management policies. This includes identifying valuable data, eliminating dark or redundant content and working with providers that support distributed networks and dynamic load placement. As an example, shifting workloads to off-peak periods reduces demand, emissions and costs across the board.

Finally, the sector would greatly benefit from an internationally agreed approach to data archiving - to prevent unnecessary digital 'landfilling'. While change takes time, these efforts will ultimately reduce the energy, carbon, water and land footprint of today's data economy.

About the author

Professor Aoife Foley is a Senior IEEE Member and Chair in Net Zero Infrastructure at the School of Engineering in The University of Manchester. She is a leading voice in AI-driven energy systems, digital infrastructure, and the decarbonisation of power, heat and transport through techno-economic syste...