Europe risks falling further behind the US and China in the development of AI infrastructure, according to a new study from Roland Berger.

The report predicts that Europe’s share of global installed data centre capacity will fall from 13% today to around 10% by 2030 as investment in next-generation AI facilities accelerates elsewhere.

That decline could have consequences beyond the data centre industry. With access to computing capacity becoming increasingly important to the development of AI, Europe’s ability to attract new infrastructure investment could have a direct impact on its wider industrial and technological ambitions.

The US has already benefited from faster grid connections, comparatively lower energy costs and a regulatory environment that is seen as more supportive of large-scale data centre development. Europe, meanwhile, continues to face many of the same infrastructure barriers that have delayed energy and industrial projects across the continent.

According to Roland Berger, those constraints are structural rather than temporary, meaning Europe will require significant reforms to permitting, grid access and energy policy if it hopes to keep pace with global investment.

Europe’s infrastructure disadvantage

The rapid growth of AI is placing new pressure on electricity networks, with next-generation data centres requiring significantly more power than many traditional facilities.

That creates a particular challenge for European markets where developers are already facing long waits for grid connections, while high electricity prices can make operating power-intensive facilities less attractive.

This has already borne out in the UK, with OpenAI backing out of plans for Stargate UK, a follow-up to its massive Stargate project that is partially built in the US already. The company blamed the UK’s high energy costs as a key reason behind the cancellation, but it also dropped similar plans for Stargate Norway. 

While energy costs and the time to connect have been huge burdens for data centre operators, OpenAI also cited regulation as another key reason behind its decision to drop Stargate UK. 

Despite the UK and Europe looking to keep up with the data centre industry in terms of regulation, with strict controls on environmental protections and energy efficiency requirements, it remains slow. Additionally, a complicated approvals process, with data centres in the UK having to go through long-winded planning applications often decided by people who don’t have much experience with the industry, can also have an impact on investment. 

Michael Knott, Partner at Roland Berger, noted, “AI is now an infrastructure race – and Europe is not keeping pace. The US is accelerating rapidly, supported by faster grid connections, lower energy costs and a more enabling regulatory environment. Europe’s constraints are structural, not cyclical. Without decisive reform, the continent will struggle to match the scale of investment and capacity growth needed to remain competitive in the AI era.

“Every year of delay widens the gap and raises the cost of catching up.”

The warning will sound familiar to those working across the sector. Access to power has become one of the biggest obstacles facing new developments, with grid capacity increasingly determining where facilities can be built and how quickly they can become operational.

Simply attracting more data centre investment will not solve that issue. Governments will also need to support new generation, reinforce transmission and distribution networks and create a clearer route for projects to connect to the grid.

Could Europe’s suppliers benefit?

Despite the expected decline in Europe’s share of installed capacity, Roland Berger believes the continent still has an opportunity to benefit from the global data centre expansion.

Europe has an established network of companies specialising in areas such as power electronics, electrical distribution, advanced cooling and on-site power generation. Those technologies are becoming increasingly valuable as operators attempt to accommodate higher-density AI workloads without overwhelming local electricity networks.

That means European companies could capture a significant portion of the investment flowing into AI infrastructure, even where the data centres themselves are ultimately built in the US, China or other markets. After all, big names such as Legrand, ABB, Schneider Electric, Eaton and Rittal are all based in Europe. 

Siyi Hao, Principal at Roland Berger, added, “This is now a pivotal moment for Europe. The continent’s ability to participate meaningfully in this growth will depend on whether policymakers and industry leaders can act decisively: removing structural barriers, unlocking grid access, and enabling the continent’s supplier ecosystem to compete on a global stage. Europe still has a window of opportunity, but it is narrowing fast.”

Europe may not be able to close the infrastructure gap overnight, especially given the time needed to build new power generation and reinforce electricity networks. However, failing to address those barriers risks leaving the continent increasingly reliant on AI infrastructure developed and operated elsewhere.

The supplier ecosystem gives Europe an opportunity to retain some influence over the market, but without faster planning decisions, improved grid access and a more competitive energy environment, its share of the infrastructure underpinning the AI economy looks set to continue shrinking.