POWER
the system and pushed genuine projects back considerably. Reforms announced in early 2026 aim to address this by prioritising projects with strategic economic value, including AI-related infrastructure, and clearing out applications with no realistic prospect of delivery.
These reforms are necessary but they are not the whole answer. Transmission and distribution networks need significant investment to handle the loads that are coming. High-voltage direct current connections, grid-scale battery storage and intelligent demand management systems all have a role to play. The scale of current demand means the data centre industry now has to be an active participant in shaping how grid infrastructure develops, rather than simply consuming whatever capacity exists.
Renewable energy and the PPA model
The power challenge has an environmental dimension that cannot be separated from the engineering one. Data centres are large, predictable, long-duration energy consumers, which makes them well suited to power purchase agreements with renewable generators. Leading operators now source significant proportions of their power directly from offshore wind and solar through long- term PPAs, providing the revenue certainty that enables new renewable projects to secure financing.
This is one of the more underappreciated ways in which data centre growth contributes to the energy transition. By acting as anchor off takers for renewable capacity, operators are helping bring new generation online that would otherwise struggle to attract investment. The operational challenge is matching consumption with generation in real time. Battery storage at the facility level, combined with intelligent workload scheduling, is increasingly being used to smooth the mismatch between variable renewable supply and operational demand.
Efficiency as an engineering discipline
Power usage effectiveness (PUE), the ratio of total facility power to IT load power, has been the industry’s primary efficiency metric for years. Best-in-class facilities now operate at PUE figures approaching 1.2, meaning that for every unit of power consumed by IT equipment, only 0.2 units are lost to cooling, lighting and ancillary systems.
The engineering driving this improvement spans several domains. Cooling system design has changed fundamentally, with hot
aisle and cold aisle containment, direct liquid cooling and rear-door heat exchangers all contributing to reduced thermal overhead. Free cooling, which uses external air or closed loop water to cool facilities without mechanical refrigeration, is viable for a greater portion of the year in the UK climate than in warmer markets, giving domestic operators a genuine efficiency advantage. AI workloads are pushing the envelope further. The graphics processing unit (GPU) clusters used for large-scale training and inference generate heat at densities that air cooling struggles to handle effectively. Liquid cooling, both direct-to-chip and immersion- based, is moving from niche application to mainstream requirement as rack densities climb above 50kW and in some cases well beyond 100kW. Facilities being designed today need to accommodate cooling infrastructure that has only recently become available at scale.
On-site generation and resilience Resilience requirements add another layer of
complexity to the power engineering picture. Data centres are expected to maintain availability at levels that leave no room for grid interruptions. As a cleaner backup generation technology, hydrogen fuel cells are attracting significant interest although the infrastructure for reliable hydrogen supply at the volumes required by large facilities is still maturing. Battery systems capable of bridging the gap between a grid outage and generator start-up are now standard, and in some configurations are large enough to handle short outages entirely without generator involvement.
What the industry needs from policymakers
The UK has genuine advantages in the power dimension of this challenge. Its renewable energy capacity is growing, its climate supports efficient cooling for more of the year than most comparable markets and its regulatory framework provides a degree of long-term stability that investors value. What the sector needs is for policymakers to treat energy access for strategic digital infrastructure with real urgency. Grid connection reform needs to deliver meaningful reductions in wait times. Industrial energy pricing, which currently sits among the highest in Europe, needs to be addressed if the UK is to remain competitive as a destination for data centre investment. And the planning framework needs to move at a pace that reflects the speed at which digital infrastructure demand is growing. Power is the central engineering and commercial challenge of the next decade for this industry. The operators, policymakers and grid managers who treat it as such will determine where the industry grows and where it does not.
JUNE 2026 | ELECTRONICS FOR ENGINEERS
11
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50
