Study Finds Europe at a Disadvantage in AI Development
July 27, 2026
Source: E & M powernews
The strategy consulting firm Roland Berger expects global data center capacity to nearly quadruple by 2035. Europe could lose ground because expansion is proceeding too slowly.
According to a study by the Munich-based strategy consulting firm Roland Berger, global installed data center capacity will nearly quadruple by 2035. This is primarily driven by the growing demand for infrastructure to support artificial intelligence (AI) applications. At the same time, the authors warn that Europe is likely to lose ground in international competition unless grid connections, permits, and energy supplies are expanded more rapidly.
The study, “The AI Datacenter Study – Part 1: The Global Build-out Race,” forecasts an increase in global installed data center capacity from 88,000 MW in 2025 to as much as 340,000 MW in 2035. The share of AI applications in total data center load is expected to rise from the current 20 percent to just under 60 percent during the same period.
The analysis is based on the Roland Berger Data Center Model, which, according to the company, examines trends in capacity, energy demand, and investment. In addition, Roland Berger surveyed 70 executives from the European data center industry.
International Competition Is Intensifying
According to the authors, the expansion of AI infrastructure is intensifying international competition. The study indicates that as early as 2025, the U.S. will have more data centers than Western Europe and Asia combined. For Europe, analysts expect a further decline in its share of global installed data center capacity from the current level of around 13 percent to about 10 percent by 2030.
The study cites long wait times for grid connections, high energy costs, and complex permitting procedures as the main causes. Connecting a new data center to the power grid can take up to seven years in Europe. In the U.S., grid connections are provided more quickly, or operators more frequently use their own energy supply solutions.
Electricity Too Expensive in Europe
Many market participants are also critical of the energy supply. According to the survey, 69 percent of industry representatives surveyed expect the availability of electricity to continue to deteriorate by 2035. At the same time, according to Roland Berger, electricity accounts for around 40 percent of a data center’s operating costs. Europe’s energy prices are higher than those in many competing regions.
Regulatory hurdles also play a role. According to the study, data center operators view lengthy approval processes and complex regulations in particular as obstacles to expanding new capacity. The study highlights the rising energy demands of modern data centers. According to Roland Berger, an AI data center with a capacity of 1,000 MW consumes as much electricity as a city with about one million households.
In-House Data Centers Are a Matter of Security
Edeltraud Leibrock, Global Managing Director at Roland Berger, therefore sees a need for further action. In-house data centers are a prerequisite for processing sensitive data in accordance with European law and for building a competitive AI ecosystem. To achieve this, Europe needs more robust power grids, faster planning and permitting processes, and competitive energy prices.
The strategy consulting firm also attaches great economic importance to this expansion. Partner Nikolaus Lehmann explains that AI and data centers are already making a noticeable contribution to economic growth in the U.S. So far, Europe has primarily benefited in specific supplier industries, but is increasingly becoming dependent on others. Without sufficient computing capacity, there is a risk of falling behind in this key technology.
According to the study’s authors, financial support programs alone are not sufficient. Rather, what is needed is forward-looking network planning and a technology-neutral regulatory framework that accelerates rather than delays investment.
The data center study is available online upon user registration.
Author: Susanne Harmsen