Only 20% of European datacenters AI-ready as skills shortages threaten infrastructure expansion

Only 20 percent of datacenters across Europe and the Middle East are currently equipped to handle the intense demands of AI workloads, according to a new report from datacenter consultancy BCS, raising concerns about the region's ability to meet growing artificial intelligence infrastructure needs.

The findings, drawn from more than 3,000 industry respondents across 41 countries, paint a picture of an industry struggling to adapt existing facilities to the power-hungry requirements of modern AI processing. While the report projects AI-ready capacity could reach approximately 70 percent by 2030, demand may still outpace available infrastructure.

The AI infrastructure gap

The core issue lies in the fundamental mismatch between traditional datacenter design and AI requirements. Most operational sites were built for conventional enterprise or cloud workloads, but racks of power-hungry GPU servers demand significantly higher power densities, more robust cooling solutions, and greater operational resilience than these facilities were originally designed to provide.

"This means that while many sites may have sufficient overall capacity on paper, they lack the ability to deliver power at rack level or manage the sustained heat loads AI infrastructure generates," BCS explains. The 20 percent figure doesn't signal industry failure, the consultancy emphasizes, but rather reflects how high the bar for AI-ready infrastructure has become and how much of the existing estate was built for an earlier generation of demand.

Skills shortages emerge as critical bottleneck

Perhaps most concerning is the industry's struggle with human capital. According to BCS data, 95 percent of respondents expect the shortage of skilled professionals to worsen over the next twelve months. Half of developers report missing deadlines or client objectives due to skills shortages, while 53 percent indicate that supply chain issues will directly influence future site selection.

This skills crisis compounds other well-known datacenter construction challenges that increasingly occur simultaneously on the same projects. Across much of Europe, developers face a shortage of available land to locate facilities, a dearth of materials and heavy equipment, difficulties getting connected to the energy grid, and critically, a lack of skilled staff.

"These are no longer isolated issues, but are increasingly felt together," writes BCS chief James Hart in the report's introduction. "As a result, the industry is entering a new phase, where growth in the market is still expected, but the ability to deliver on that ambition is ever more constrained."

Market dynamics and regional variations

Despite the challenges, the report identifies potential growth areas. The UK market still has a potential pipeline for more than 10 GW of new capacity over the next ten years, while Germany, Austria, and Switzerland are tipped for more dramatic expansion this year. In a significant shift, London may even lose its place as the top European datacenter hub to Frankfurt by 2031, according to a separate report cited by BCS.

The market for cooling technology specifically designed for AI workloads is particularly hot, with BCS noting that "the market for gear that stops GPUs losing their cool is red hot as Trane gulps down LiquidStack."

Industry adaptation strategies

BCS suggests that firms best positioned to weather these challenges will be those able to manage collaboration across teams, disciplines, and partners that don't traditionally work closely together, and to navigate trade-offs under pressure. This requires better organization and the ability to coordinate complex, multi-stakeholder projects.

The report's findings come at a critical juncture for European technology infrastructure. With 93 percent of industry professionals polled expecting to see rising demand for datacenter capacity over the next twelve months, yet available supply forecast to shrink, the region faces a potential infrastructure bottleneck that could impact AI development and deployment across multiple sectors.

The situation underscores the need for coordinated policy responses, investment in technical education and training programs, and innovative approaches to datacenter design and construction that can meet the unique demands of AI workloads while addressing the practical constraints of European markets.