Datacentre demand is huge but power and skills hold things back, survey shows | Computer Weekly

"Those are the findings of The BCS Consultancy's recent Datacentre truths 2026 survey, which questioned 3,000 respondents in the datacentre space. The largest number - 25% of respondents - were colocation operators, with corporate occupiers (13%), developers and investors (12%), IT integrators and network operators (both 11%). Respondents accounted for 7.1 million metres 2 of datacentre space under their control in 41 countries."

"While all (100%) developers and investors questioned had increased their portfolios in the past six months, 93% of all questioned think 2026 will be characterised by supply constraints. Three-quarters (75%) reported higher labour and delivery costs, while 86% reported supply chain volatility. A massive 95% reported skills shortage affecting delivery of datacentres, with that likely to worsen further."

"BCS Consultancy COO Chris Coward said: "Demand will continue to outstrip supply, and capacity constraints are here to stay. How fast we overcome them depends on delivering new infrastructure and resources. But physical capacity isn't the only limit - skilled labour is a bottleneck. Without it, projects stall, timelines slip, and even viable investments can fail. Power may show where development is possible, but skilled people make it happen.""

Demand for datacentres across Europe exceeds delivery capacity because of power supply limits, skills shortages, supply‑chain volatility and planning constraints. Three thousand respondents in the datacentre sector represent 7.1 million m2 of space across 41 countries; 25% are colocation operators and other roles include corporate occupiers, developers, investors, IT integrators and network operators. All developers and investors recently expanded portfolios, yet 93% expect 2026 to be marked by supply constraints. High percentages report rising labour and delivery costs, widespread skills shortages disrupting projects, increased operating costs, staff pressure, missed deadlines and some lost orders. Most professionals consider current facilities unsuitable for AI deployments.

Read at ComputerWeekly.com