Data Centres vs. the AI Power Monster: A Race Against Consumption

The rapid proliferation of AI tools, spearheaded by innovations like OpenAI’s ChatGPT and Google Gemini, has catapulted artificial intelligence from a background enabler into mainstream consciousness. Global leaders and tech titans alike now convene, acknowledging both the immense economic potential and the significant geopolitical risks inherent in this transformative technology. Whether AI ultimately serves as humanity's greatest ally or a formidable challenge, its development trajectory remains unequivocally upward.

What is known, however, is that relentless AI development hinges on robust digital infrastructure and a consistently reliable energy supply. And let there be no ambiguity: AI has a dirty secret – it’s insatiable thirst for energy. Put plainly, the computational demands of operating AI are exceptionally energy-intensive for data centres.

Consider the findings of a recent survey of US and European IT buyers and companies adopting AI, where a staggering 88 per cent reported a dramatic surge in computing power needs, with nearly half (47 per cent) experiencing a doubling of demand. Further underscoring this trend, another study projects that the AI industry's total annual energy consumption could rival that of the Netherlands by 2027, potentially reaching 85 to 134 TWh. Currently, data centres globally account for 3 per cent of electricity demand, a figure Vertiv predicts will escalate in tandem with AI's increasing prevalence and its deeper integration into products and services.

This rise in energy consumption inevitably translates to rising operational costs for data centre operators. In a UK business landscape already grappling with volatile energy prices and financial pressures, data centres face a critical juncture. They must strategically navigate the AI revolution, transforming potential risks into tangible opportunities.

Finding Profitability in Net Zero

For data centres seeking to expand capacity and capitalise on the AI boom, the pursuit of Net Zero offers a compelling synergy between profitability, energy efficiency, and responsible consumption. According to Mckinsey, embracing clean energy sources, actively reducing electricity use, and diligently minimising carbon emissions (both direct and indirect) are now pivotal key performance indicators for businesses committed to Net Zero.  

With the UK's legally binding commitment to achieve Net Zero by 2050 and mounting pressure from policymakers for UK businesses to align with these national objectives, energy-intensive sectors, including data centre operators, stand to gain a significant first-mover advantage by proactively implementing Net Zero strategies for data centres.

This strategic adoption involves a holistic approach, encompassing cost optimisation in data centres, emissions reduction, the integration of renewable energy sources for data centres, and the implementation of robust energy efficiency measures.

Crucially, these elements must be underpinned by rigorous monitoring and transparent reporting practices, woven together into a comprehensive data centre sustainability framework. Such a proactive stance not only shields data centres from the volatility of energy markets and provides resilience against economic fluctuations but also significantly bolsters their sustainability credentials and enhances their competitive standing in the B2B energy sector.

The Power of Energy Measurement

Accurate energy measurement in data centres forms the bedrock of any effective energy management and decarbonisation strategy. It establishes the essential baseline against which progress can be objectively benchmarked. Traditionally, data centre temperature set points have ranged between 18 and 27°C, complemented by specific humidity parameters.

In the UK, Power Utilisation Effectiveness (PUE) serves as the primary metric for gauging data centre energy consumption, calculated as the ratio of total facility energy to the energy powering ICT systems. UK data centres typically operate with a PUE between 1.5 and 1.8, compared to an EU average of 1.6. A PUE of 2.0 or higher signifies that more energy is being expended on supporting infrastructure than on the critical ICT equipment itself, often indicating excessive energy use in cooling. Therefore, PUE serves as a crucial indicator of data centre operational efficiency.

To achieve meaningful progress in data centre energy management and decarbonisation, operators must possess a granular understanding of their IT estate's energy consumption. Transitioning to renewable energy sources, whether through on-site renewable energy generation, dedicated renewable energy supply contracts, or Power Purchase Agreements (PPAs) for data centres, is paramount to achieving decarbonisation targets. PPAs, in particular, are gaining traction among data centres, offering transparency in renewable energy investments and enhanced long-term price certainty. Solutions can be tailored to meet the specific demands of both single and multi-site operations.

Incremental Gains, Big Efficiency

Due to the volume of energy consumed, surprisingly simple measures can often yield significant energy efficiency gains in data centres. Many established data centres rely on legacy cooling systems that may be inherently energy-inefficient. A clear understanding of your current PUE can provide invaluable insights for making informed decisions about upgrading cooling system performance.

Similarly, real-time active monitoring of energy consumption across all data centre sites enables the swift identification and resolution of anomalies. This monitoring can be further enhanced by the integration of occupancy sensors and other smart technologies that dynamically adjust energy usage based on real-time requirements, such as automatically switching off lights in unoccupied areas and optimising heating levels.

To see this approach in action, we invite you to contact us for a personalised demo. What may appear as small, incremental changes can collectively lead to significant cost savings—resources that can be strategically reinvested into high-efficiency equipment and advanced technologies, driving long-term performance and sustainability across your data centre operations.

Final Thoughts

Energy costs associated with feeding the AI algorithm are set to continue to grow and there is no universal panacea for optimising energy efficiency in data centres. Each data centre will benefit from engaging specialist expertise to identify tailored short-term and long-term strategies for achieving stringent energy control. For many in the sector, the immediate priorities are clear: to decarbonise data centre operations and demonstrably mitigate environmental impact. In this evolving landscape, strategic collaboration with a commercial energy specialist like Big Energy Group can be a decisive factor in achieving sustainable and cost-effective operations.

About Big Energy Group

Big Energy Group is a privately held, British-owned energy brokerage with an established track record of helping clients successfully navigate the energy market. The company has offices in Harrogate and the Tees Valley and serves more than 400 businesses across the UK. For more information, please visit bigenergygroup.co.uk.

‍