AI is changing the power challenge facing data centres. Not only is it increasing overall electricity demand, but it's also creating faster and less predictable changes in load as GPU clusters ramp up and down. At the same time, constrained grid capacity, rising sustainability expectations and growing scrutiny of resilience are forcing operators to reconsider how their critical power infrastructure is designed.

As rack densities soar, data centres are being forced to rethink their cooling strategies. For those wishing to ride the AI wave, air cooling is no longer sufficient. That’s led many to investigate the possibility of switching to a liquid cooling setup — but what do you need to know before making the switch?

The IEA has recently predicted that electricity consumption from data centres, AI and the cryptocurrency sector could double by 2026. Demand for capacity is increasing exponentially – but the UK power grid may struggle to support the data centres needed to meet that demand due to escalating energy requirements from advanced technologies like AI, limited infrastructure capacity, and the need for substantial upgrades to handle increased and variable power loads efficiently.

AI is the technology that is promising to usher in a new era for human civilisation, transforming industries across the globe. Yet, for AI to succeed, it needs data centre capacity to keep up with the ever-growing demand, and that poses some critical challenges.

The digital age is undergoing rapid transformation, placing unprecedented demands on data centre design and operations. Recently designated as Critical National Infrastructure (CNI) by the UK Government, data centres are essential to supporting advanced technologies such as AI and HPC. However, this rapid growth brings complex challenges, from managing surging power demands and strained grid capacity to meeting ambitious sustainability goals and evolving regulatory standards.

Among several critical power industries, but particularly in the hyperscale data centre market, part or all of data hall power trains are manufactured off-site in a containerised modular structure. This presentation from WBPS considers several key factors to consider when considering this approach in an OFCI environment.

In addition to the unprecedented growth in internet usage and data processing since the beginning of the pandemic, data center clients have been managing even more complicated challenges involving global supply shortages and labor shortages.

Join Data Centre Review, ABB and Red Engineering as we discuss new strategies to succeed in today’s unique and challenging environment.